October 2013 VOL: XXXXVI, No.04 Commemorating 40 years of publication Printed by Karunaratne & Sons (Pvt) Ltd. From the Editor ... III SECTION I Factors Influencing the Service Life of Buildings by: Eng. (Prof.) W P S Dias Potential and Viability of Rice Husk Based Power Generation in Sri Lanka by: Eng. (Dr.) Asanka S Rodrigo and Eng. Shantha Perera Investigation on Efficiency of Repairing and Retrofitting Methods for Chloride induced Corrosion of Reinforced Concrete Structures by: Eng. B H J Pushpakumara, Eng. (Dr.) Sudhira De Silva and Eng. (Dr.) (Mrs.) G H M J Subashi De Silva Productivity in Construction – A Critical Review of Research by: Eng. (Dr.) D A R Dolage and Dr. Paul Chan Stream flow, Suspended Solids and Turbidity Characteristics of the Gin River, Sri Lanka by: Eng. (Mrs.) T N Wickramaarachchi, Eng. (Dr.) H. Ishidaira and Eng. (Dr.) T M N Wijayaratna Peak Electricity Demand Prediction Model for Sri Lanka Power System by: G V Buddhika De Silva and Eng. Lalith A Samaliarachchi SECTION II Floating Wetlands for Management of Algal Washout from Waste Stabilization Pond Effluent: Case study at Hikkaduwa Waste Stabilization Ponds Notes: ENGINEER, established in 1973, is a Quarterly Journal, published in the months of January, April, July & October of the year. All published articles have been refereed in anonymity by at least two subject specialists. Section I contains articles based on Engineering Research while Section II contains articles of Professional Interest. ENGINEER JOURNAL OF THE INSTITUTION OF ENGINEERS, SRI LANKA * Completed 40 Years of Publication * EDITORIAL BOARD Eng. Tilak De Silva -President(Chairman) Eng. W. Gamage - Chairman, Library and Publication Committee Eng. (Prof.) K. P. P. Pathirana - Editor “Transaction” Eng. (Prof.) T. M. Pallewatta - Editor “ENGINEER” Eng. (Dr.) U. P. Nawagamuwa - Editor “SLEN” Eng. (Prof.) (Mrs.) N. Rathnayaka Eng. (Dr.) D. A. R. Dolage Eng. (Miss.) Arundathi Wimalasuriya Eng. (Dr.) K. S. Wanniarachchi The Institution of Engineers, Sri Lanka 120/15, Wijerama Mawatha, Colombo - 00700 Sri Lanka. Telephone: 94-11-2698426, 2685490, 2699210 Fax: 94-11-2699202 E-mail:
[email protected] E-mail (Publications):
[email protected] Website: http://www.iesl.lk COVER PAGE Colombo Katunayaka Expressway (CKE) Constructed addressing a long felt need for expedient access from Katunayaka International Airport to the city of Colombo, CKE was declared open by H E the President of Sri Lanka on 2013 October 27. Construction commenced on 2009 August 18 on the 25.8 km four lane expressway with a segment from Kelani Bridge to Peliyagoda interchange having six lanes. With three interchanges and option to connect to Outer Circular Highway (OCH) at Kerawalapitiya this expressway reduces 90 minute travel time through route A-003 to 20 minutes. Courtesy of the Road Development Authority The statements made or opinions expressed in the “Engineer” do not necessarily reflect the views of the Council or a Committee of the Institution of Engineers Sri Lanka, unless expressly stated. October 2013 VOL: XXXXVI, No. 04 Commemorating 40 years of publication CONTENTS Vol.: XXXXVI, No. 04, October 2013 ISSN 1800-1122 III 1 9 19 31 43 53 63 by: Eng. (Mrs.) Sujatha Kalubowila, Dr. Mahesh Jayaweera, Eng. Chandrika M. Nanayakkara and Eng. Dhanesh N. De S. Gunatilleke The above Paper was placed First in the ‘Over 35 years of age’ Category at the Competition on “Water Resources Development and Future Challenges”- Role of Engineering meeting Future Challenges of Water Resources Development in Sri Lanka” 2012/2013 Sponsored by: International Water Management Institute (IWMI) II FROM THE EDITOR………….. Reducing the travel time of a 30 km route by as much as 75% is what has been achieved by the newly opened Colombo Katunayaka Expressway (CKE). After successfully defeating terrorism that had been plaguing the country for a long period, we are on the way to development with ambitious targets. In this scenario, the fact that transportation infrastructure becomes a prerequisite launch board, has not escaped our Engineers, political decision makers as well as funding agencies. It is a fact that hinder free expedient access from the first international airport to the commercial city of the country will facilitate corporate activities, industry and tourism. Emergence of better highways and expressways has been the highlight of transportation infrastructure in this country for the past few decades. However, in the development of any infrastructure in this era of fast depleting natural resources, very special considerations need to be granted to economy and sustainability in a global sense. When it comes to transportation, the primary consideration should be to minimize the cost per unit per unit distance, be it people or goods, without neglecting sustainability, environmental and social impacts. When above factors are given cognizance under a technical viewpoint, it would become very clear to you that Rail transport naturally comes to the lead. As a mass land transport mode with the lowest rolling resistance and predefined right of way, rail transport has the aptitude to keep the burden posed by travel and transport needs of an ever growing populace on our planet, at bay. This fact is amply endorsed by the so called ‘Developed’ countries of the world that have adopted rail as the primary land transport mode. It can not be denied that, our country has in fact, retracted pre independence rail track infrastructure and given undue prominence to roads. Though drastically late, we should at least now direct our energies at improving and expanding the railway infrastructure in this land. In support of this argument, suffice to say that, it could be the only way in a future devoid of abundant fossil fuels. Eng. (Prof.) T. M. Pallewatta, Int. PEng (SL), C. Eng, FIE(SL), FIAE(SL) Editor, ‘ENGINEER’, Journal of The Institution of Engineers. III SECTION I 1 ENGINEER Factors Influencing the Service Life of Buildings W.P.S. Dias Abstract: The service life of a building depends mainly on its chief structural materials and the environment it is placed in. This paper collates the evidence from condition surveys conducted on some buildings with ages of up to 125 years set in a humid tropical environment, and seeks to arrive at some generalizations. Load bearing masonry walls and timber floors had performed well, as had exposed steel sections that were well maintained. Buildings with such elements could be expected to last well beyond the ‘normal’ design life of 60 years. If a reinforced concrete building had been exposed to a chloride source, major repairs were required after just half this design life. Carbonation depth was found to broadly obey a correlation with the square root of time. However, it is shown that both depths of carbonation and surface chloride levels can vary considerably in different parts of the same building. These findings have direct implications for both construction (in the choice of materials) and inspection (with respect to sampling and use of multiple test methods). Keywords: Service life, durability, chlorides, carbonation, corrosion 1. Introduction 1.1. The Service Life of Buildings The service life of structures depend on a variety of factors, such as (i) their purpose; (ii) socio-economic considerations; (iii) materials of construction; (iv) surrounding environment; and (v) degree of maintenance (Dias, 2003). Very long service life of even up to 500 years or more would be desired for monumental buildings such as temples and churches. Public buildings such as town halls and parliament buildings could be expected to last for 100 to 200 years, whereas private structures such as offices and dwellings for perhaps 50 to 60 years. BS 7543 (1992), defines the ‘normal’ life of a building as 60 years. The new Eurocodes, e.g. BS EN 1992-1-1 (2008), assume this period to be a lower one of 50 years. Socio-economic considerations impinge on the above durations, some of which tend to reduce the lifespans of buildings, while others increase them. The changing needs of various owners, and indeed the changing face of the city or area in which the building is located, may cause a building to be obsolete even before it ceases to be serviceable. In the context of the above proneness to change, most investors or builders may not want to invest in a building with an excessive service life. On the other hand, owners sometimes try to use an existing building over and above its service life, because demolition and reconstruction may force them to comply with new planning regulations. Also, once a building exceeds a certain lifespan, the owner, or even other interested parties, may wish to prolong its life further, if it is considered a national heritage. The different materials of construction that are used in a building will give rise to different rates of deterioration. In general, steel and reinforced concrete will tend to deteriorate faster than masonry; and timber in internal environments. Heat and moisture are environmental factors that tend to accelerate deterioration. Where steel embedded concrete and structural steel are concerned, a chloride environment, inclusive of proximity to the coast, will significantly enhance corrosion. The factors affecting service life can vary, not only from building to building, but even within a given building. For example, (i) The quality of the substructure, superstructure and even roof structure in a building may vary if different subcontractors were responsible for them; (ii) the environment a building is subjected to will vary from external elements to internal elements and also from seaward side to landward side (if it is near the coast); and (iii) different building elements may receive different degrees of maintenance, depending on their accessibility and inspectability. 1.2. Changes in Construction Technology From a historical perspective, we can identify Eng. (Prof.) W.P.S. Dias, BScEng(Hons), PhD(Lond), DIC, CEng, MIStructE, FIE(Sri Lanka), Senior Professor of Civil Engineering, Department of Civil Engineering, University of Moratuwa, Sri Lanka. ENGINEER - Vol. XXXXVI, No. 04, pp. [1-7], 2013 ©The Institution of Engineers, Sri Lanka ENGINEER 1 ENGINEER 2 1945 as a point at which there was a worldwide shift in building materials usage – i.e. most structural steel used for beams and columns, and timber used for floors, was replaced with reinforced concrete. We can also identify a time around 1975 as a point at which there were a more subtle changes in the quality of building materials. There was a worldwide change in cement manufacturing processes, resulting in cements that developed greater strengths quicker (due to the increased percentage of tricalcium silicate in cements). This meant that a given strength of concrete (normally tested at an age of 28 days) could be achieved with less cement. However, concretes were now made not only with lower cement contents, but also with lower percentages of the ingredient in cement (i.e. dicalcium silicate) that contributed to longer term strength development. This resulted in a lowering of the durability properties of the concrete. There was also a worldwide scarcity of timber for construction, causing less durable species of timber to be used for construction. Although such species were chemically treated to improve durability, the efficiency of treatment may not always have been adequate. The passage of time (especially since 1975) has of course seen increasingly greater awareness of durability issues, and these have been reflected in codes of practice, especially in those for reinforced concrete. It can therefore be argued that the greater awareness of durability issues has compensated for some of the detrimental impacts on durability described above. In this same period, building durability has been improved through the availability of good quality waterproofing materials, performance enhancing admixtures for concrete and specialist repair materials (e.g. repair mortars). However, building longevity may have decreased due to the modern practice of hiding most of the structural elements behind ceilings, paneling and facades, thus making inspection (and hence the early arresting of deterioration) more difficult. 2. Objectives and Methodology The first objective of this paper is to analyse a set of condition evaluations carried out on buildings from the ages of 7 to 125 years, and to draw various lessons from that analysis regarding the factors that either increase or decrease the longevity of buildings. These evaluations have not been carried out in a random fashion, but rather in response to clients. At early ages, such evaluations are generally made only due to changes of building ownership. In mid life a need for evaluation often occurs due to unexpected deterioration. At older ages, clients request evaluations because of concern regarding the continued safety of their buildings. Many of these evaluations are based largely on visual inspection, but some of them are backed up by a reasonable degree of sampling for material properties, inclusive of durability indices (e.g. Dias, 1994; Dias and Jayanandana, 2003; Dias and Sivasubramaniam, 1989). The second objective of this paper is to analyse some data regarding the depths of carbonation and surface chloride levels of some reinforced concrete elements or structures, because carbonation and chloride ingress are the two main mechanisms that lead to the deterioration of such structures. Considerable focus is placed on variations of such indices within the structure itself. The carbonation depths have been obtained by spraying a 1% phenolphthalein solution onto freshly cut surfaces and noting the depth that remains colourless. The fresh cuts were made either by coring (where the core is used for a variety of other purposes such as strength testing) or by advancing a drill bit into the surface in increments of 5 mm, spraying the phenolphthalein solution into the hole at each increment and noting the depth at which the outflowing liquid is pink in colour. The carbonation depths in a structure at a given age will help us to estimate how much longer it will take for the carbonation front to reach the reinforcement (i.e. for the ‘incubation’ phase to be completed), after which the likelihood of corrosion increases significantly. If the front has already passed the level of the reinforcement, it signals the need to take stringent measures for ensuring that the reinforced concrete elements are waterproof. Procedures for determining the chloride profile within the concrete cover zone are well documented (e.g. see de Rooij and Polder, 2004). This can be done by using extracted cores, slicing them and obtaining the chloride contents at various depths. It is then possible to predict the time at which the chloride content at the level of the reinforcement will reach a certain threshold value for corrosion initiation by using the surface chloride concentration and the diffusion coefficient obtained from the ENGINEER 2 3 ENGINEER chloride profile. However, obtaining the diffusion coefficient in this way is difficult and tedious (de Rooij and Polder, 2004). As such, the above modeling can also be done by using values published in the literature (e.g. Bentz and Thomas, 2012). These depend on the water/cement ratio (which affects the initial diffusion coefficient) and type of cement (which affects the age dependent variation of that coefficient). Whatever method is used, the surface chloride content must be obtained from the structure being examined. Published literature also gives guidance regarding the rates of surface chloride build-up. However, such broad generalizations may not be accurate, and there could also be variations depending on the micro environment. The surface chloride contents for the case studies in this paper have been obtained by taking the surface layer of the element, whether a plaster coating (of around 10 mm thickness) or a surface slice of 5 mm from extracted cores, and determining either the total or water soluble chloride content in that layer by acid or water extraction respectively. 3. Case Studies of Deterioration Table 1 gives a few cases of buildings that were inspected for condition evaluation over a period ranging from 1988 to 2011, with the age of the building at inspection given in parentheses. They are listed in inverse order of year of construction, which also happens to correspond to increasing age when the inspection was made. The cases can be divided into three broad categories, separated by bold horizontal lines in Table 1. The 7 and 12 year old buildings, which do not show any visible deterioration, fall into the first category. In the next category are buildings of ages 25 to 30 years where distress of varying degree has occurred in reinforced concrete elements, due to chloride induced corrosion. The chloride source for both the Hotel Sunflower and Buddhist Girls’ School is sea spray. It should be noted that the much greater corrosion in the latter is due to poor construction. For the Puttalam Cement Works, the chloride source was the groundwater used during construction (Dias and Jayanandana, 2003), and for the Bandaranaike Wing of the Colombo General Hospital, contamination from the toilets (Dias and Sivasubramaniam, 1989). This suggests that serious repair work may become necessary after around 30 years if reinforced concrete elements of a building are exposed to a chloride source. In the last category are buildings that have survived for 65 to 125 years. It should be noted that the main structural elements are not of reinforced concrete in these buildings; note that these have been constructed prior to the year 1945, alluded to before. Two of the 100 year old buildings had masonry loadbearing walls and timber floors too - i.e. no steel at all. This combination is arguably the best combination for ensuring long service life. The rest of the buildings in this category are steel framed, with reinforced concrete floors or roofs. In almost all cases the reinforced concrete elements experienced significant corrosion, especially roofs and toilet areas. The Grand Hotel displayed only minimal deterioration in the slabs above the kitchens, which are moist and humid environments. In general hotel buildings tend to be well maintained, with defects attended to promptly. Steel columns and beams that were exposed (and hence easily painted) performed very well, as seen in the 125 year old Gaffoor Building (Figure 1), where the contrast with the reinforced concrete slab is stark. On the other hand, encasing the columns in concrete (Figure 2) and the beams in a mesh and plaster covering (Figure 3) resulted in undetected corrosion of the structural elements. Figure 1 – Exposed steel columns and beams in the 125 year old Gaffoor Building 4. Depths of Carbonation Apart from a chloride environment, referred to above, reinforced concrete in all environments will experience carbonation. When the carbonation front reaches the reinforcement, the chemical protection given to it by the concrete ENGINEER 3 ENGINEER 4 Table 1 – Analysis of Condition Evaluations Building Year & Age (yrs) Building Type Deterioration Comments Smart Shirts Factory, Katunayake 1991 (7) RC frame and slabs Not apparent Tourist Board Building, Colombo 4 1982 (12) RC frame and slabs Not apparent Buddhist Girls’ School, Mt. Lavinia 1977 (25) RC frame and slabs Columns and sunshades badly corroded Location close to coast; poor quality construction Hotel Sunflower, Negombo 1974 (25) RC frame and slabs Some corrosion Location close to coast Puttalam Cement Works 1970 (28) RC frame and slabs Some buildings badly corroded High chloride levels in groundwater used for concreting Bandaranaike Wing, Colombo General Hospital 1958 (30) RC frame and slabs Severe corrosion in toilet area slabs High chloride levels through operation Baur’s Tenemants, Grandpass Road 1936 (65) Steel frame; RC slabs and roof RC roof badly corroded; also open corridor and toilet slabs Poorly maintained Angoda Mental Hospital 1925 (72) Steel frame, RC slabs Toilet area slabs badly corroded Poorly maintained Grand Hotel, Nuwara Eliya 1911 (100) Masonry; timber & RC floors Some deterioration in slabs above kitchens Generally cool environment; well maintained. Central Point, Colombo Fort 1911 (100) Steel frame; RC slabs and roof Corrosion in internal cased columns and roof Institute of Aesthetic Studies, Colombo 7 1899 (100) Masonry; timber floor; RC roof All RC components corroded Gaffoor Building, Colombo Fort 1886 (125) Steel frame & roof; RC slabs Severe corrosion in slabs Figure 2 – Significant corrosion in concrete encased steel internal column in Central Point building Figure 3 – Corrosion in steel floor beam covered by wire mesh and plaster in Central Point building ENGINEER 4 5 ENGINEER will be lost, and corrosion will take place, especially in a moist or wet environment. The depth of carbonation is considered to be a function of the square root of time. Figure 4 gives a relationship for these two entities (the trend line is not forced through the origin) for some of the buildings in Table 1 together with a few others. Such a universal relationship will not strictly be possible for concretes of various qualities, and even moisture conditions. However, since most of the concretes are of around grade 20 quality, Figure 4 could be used to obtain approximately the likely carbonation depths in concretes of various ages. y = 4.924x R² = 0.850 0 10 20 30 40 50 60 70 80 0 2 4 6 8 10 12 14 D e p t h o f C a r b o n a t i o n ( m m ) Age ^ 0.5 (years ^ 0.5) Figure 4 – Carbonation depth as a function of the square root of time The slope of a logarithmic plot between depth of carbonation (d) and time (t) will give an estimate of the exponent ‘n’, in the relationship d = k(t) n , where k is a constant. Such a plot in Figure 5 gives the exponent as 0.69, somewhat higher than the usually adopted value of 0.5 (i.e. the square root of time law). This may be due to the possibility that the older concretes have a higher k value than the more recent ones (Richardson, 1988). A plot using only the first four data points in Figure 5 (corresponding to a maximum age of 30 years and years of construction from 1958 to 1991) gives the expected slope of 0.52. It should be noted that these four data points all fall below the regression line in Figure 4. Figure 5 is based on the average carbonation depths from the buildings surveyed. There can however be significant differences between various areas in a building. For example, Figure 6 shows the differences in depths of carbonation for toilet areas compared to corridor ones in the 30 year old Bandaranaike Wing of the Colombo General Hospital (Dias and Sivasubramaniam, 1989). The corridor areas were more carbonated because the concrete was dry, whereas the toilet areas were less carbonated because the concrete was wet, with 4 zero values too. However, there is greater scatter in the values for the toilets, inclusive of some high values – these regions would be very susceptible to corrosion if the carbonation depth exceeds the cover provided, because of the wet conditions. The phenomena of zero carbonation depths and widespread variation in such depths within a building have been reported before (Roy et al, 1996). Figure 5 – Determining the exponent ‘n’ in the expression d = kt n 0 5 10 15 20 25 30 35 40 1 2 3 4 5 6 7 8 D e p t h o f C a r b o n a t i o n ( m m ) Toilet areas Corridor areas Toilet average Corridor average Figure 6 – Carbonation depths at Bandaranaike Wing, Colombo General Hospital 5. Surface Chloride Levels Figure 7 gives the variation in the rate of surface chloride build up for different elements in 3 coastal hotels of varying ages. There is a clear difference between the rates for interior and exterior elements, with the former being significantly less than the latter and displaying much less scatter. It should be noted that these rates have been obtained from the water soluble average chloride contents in plasters of thickness around 10 mm. The published literature for total surface chloride build up rates on concrete surfaces can range from 0.004% to more than 0.1% by weight of concrete ENGINEER 5 ENGINEER 6 per year, depending on the proximity to the coast (Bentz and Thomas, 2012). 0.0000 0.0010 0.0020 0.0030 0.0040 0.0050 Interior slabs Exterior slabs Exterior beams Exterior columns C h l o r i d e b u i l d u p r a t e ( % / y e a r ) 21 yrs 28 yrs 34 yrs Figure 7 – Surface chloride build up rates at 3 coastal hotels of different ages Another example of within structure surface chloride level variations is given in Table 2 for the 100 year old Central Point building in Colombo Fort, a capital port city. These total (i.e. acid soluble) chloride levels were obtained from 5 mm thick slices cut from the bottom surfaces of 4 cores taken at each floor level. The very high levels in the fourth and fifth floors compared to the lower ones may be due to the shielding of the lower floors by surrounding buildings. It may also be due to sea spray contaminated rainwater entering the upper floor levels through observed leaks in the concrete roof. The evidence for the chloride level variation is confirmed by the degree of corrosion observed in the mesh reinforcement embedded in those cores. Figures 8(a), 8(b) and 8(c) give examples of the degree of corrosion defined as low (L), medium (M) and high (H). 6. Concluding Discussion - Implications for Practice The case studies in Table 1 indicate that if very long service life is required, consideration should be given to using construction materials such as masonry for walls, timber for floors and exposed (not encased) steel sections for columns and beams. Such materials, especially the steel sections, will however require continuous maintenance. Reinforced concrete, although very popular after around 1945, is good for normal life buildings (i.e. 50 to 60 years) that need little maintenance. If reinforced concrete is used for buildings in chloride environments, provision will have to be made for improved design (e.g. greater cover and better quality concrete); else the buildings may reach only around half their expected service life. Table 2 – Surface chloride content variations in Central Point building Floor Chloride content (w/w %) (range, average) Corrosion Level (H, M, L) Ground 0.02 - 0.04 (0.03) L, L, L, L First 0.03 – 0.05 (0.04) L, L, L, L Second 0.03 – 0.04 (0.038) L, L, L, M Third <0.01 – 0.07 (0.045) L, L, L, H Fourth 0.04 – 0.31 (0.150) L, L, M, H Fifth 0.07 – 0.50 (0.210) H, H, H, H When conducting forensic investigations on buildings, care should be taken to ensure that samples are taken from various exposure conditions, because carbonation depths and surface chloride levels can differ considerably from one part of a building to another. This will enable more nuanced strategies to be proposed for continued use. Greater confidence in the sampling can be obtained by combining and comparing different test results. For example, the surface chloride levels can be compared with the degree of corrosion in embedded reinforcement. Comparisons can also be made between the structure being investigated and the accumulated data from buildings in a (a) (b) (c) Figure 8 – Degrees of corrosion in embedded reinforcement: (a) Low; (b) Medium; (c) High ENGINEER 6 7 ENGINEER similar environment. For example, the estimate for the carbonation related incubation phase duration can be obtained both from the building concerned and also from an overall curve for buildings of varying ages References 1. Bentz, E. C. and Thomas, M. D. A. (2012). Life-365 Service life Prediction Model for Reinforced Concrete Exposed to Chlorides: Computer Program and user manual, v. 2.1. Life-365 Consortium II, January. 2. BS 7543: 1992. Guide to Durability of Buildings and Building Elements, Products and Components. British Standards Institution, Milton Keynes. 3. BSEN 1992: Part 1-1: 2008 Eurocode 2: Design of concrete structures: General rules and rules for buildings. British Standards Institution, Milton Keynes. 4. de Rooij, M. R. and Polder, R. B. (2004). What Diffusion Coefficient is used for Chloride Diffusion Modeling? In Advances in Concrete through Science and Engineering. RILEM International Symposium, March. 5. Dias, W. P. S. (1994). Structural Appraisal of Reinforced Concrete Buildings from In-Situ Material Properties - Some Issues and Insights. Transactions, Institution of Engineers, Sri Lanka, pp. 129-145. 6. Dias, W. P. S. (2003). Useful life of Buildings. University of Moratuwa, Moratuwa, June. http://www.slaasmb.org/USEFUL LIFE OF BUILDINGS.doc. Accessed 23/03/2012. 7. Dias, W. P. S. and Jayanandana, A. D. C. (2003). Condition Assessment of a Deteriorated Cement Works. ASCE Journal of Performance of Constructed Facilities. Vol. 17, No. 4, pp. 188-195. 8. Dias, W. P. S. and Sivasubramaniyam, S. (1989). Assessment of Floor Slabs in the Bandaranaike Wing of the Colombo General Hospital. Engineer (Sri Lanka). September, pp. 27-36. 9. Richardson, M. G. (1988). Carbonation of Reinforced Concrete: Its Causes and Management. Citis, London. 10. Roy, S. K., Northwood, D. O. and Poh, K. B. (1996). Effect of Plastering on the Carbonation of a 19 year old Reinforced Concrete Building. Journal of Construction and Building Materials, Vol. 10, No. 4, pp. 267-272. ENGINEER 7 1 ENGINEER Potential and Viability of Rice Husk Based Power Generation in Sri Lanka Asanka S. Rodrigo and Shantha Perera Abstract: Due to intense fuel dependency on energy production in the world, cost of energy is now heavily depends on the prices of fossil fuels. Most of the countries in the world are suffering due to this and Sri Lanka is no exception. It is in this context promotion of biomass, as a renewable source, is so vital to the country. Rice being the staple food of the country as well as the crop with highest land area under cultivation, rice husk (RH) generated in paddy processing was found to have a significant potential in power generation. This paper investigates the possibility of using rice husk as a viable source of power generation in Sri Lanka. It is clearly seen that there is a significant potential in the districts of Ampara, Polonnaruwa, Anuradhapura and Kurunegala for power generation using rice husk. It was found that 30% of excess RH can be exploited for power generation with an annual energy potential of 180 GWh. This potential can be exploited by (1) Commercial scale RH power plants, (2) Small scale power plants under net metering scheme and (3) Off grid RH power plants. Keywords: Rick husk, power generation, viability, gasification, combustion. 1. Introduction Sri Lanka’s energy sector has been ailing for the last two decades due to its excessive dependency on petroleum and lack of diversity in energy sources in the energy supply mix. In year 2011, the primary energy share in Sri Lanka was 46% from biomass, 42% from petroleum, 12% from Hydro and rest from the non-conventional renewables [1, 2]. Initially, almost entire electricity requirement of the country was met by hydropower whilst gradual increase in demand for electricity during the last decade of the 20 th century shifting electricity generation more towards thermal power. There are no proven fossil fuel resources in Sri Lanka. Therefore, there is a high emphasis on introducing non-conventional renewable sources to the electricity sector. Having realized this fact, the government of Sri Lanka set a target of achieving 10% of electricity demand from non-conventional renewable energy sources by 2015 as set out in the national energy policy of Sri Lanka [3]. It is expected to connect feasible renewables to national grid as distributed generators, which can also be used to optimize the network use [4]. Biomass has been identified as one of the most potential sources of renewable energy for power generation in Sri Lanka [3]. At present, biomass is confined to domestic cooking and to some industrial thermal applications. In Sri Lanka, rice being the staple food of the country as well as the crop with highest land area under cultivation, produces substantial quantity of rice husk (RH) as a waste product in paddy processing. Part of RH produced is used as a source of thermal energy in a few applications while the balance is burnt or dumped in the open air, causing a lot of environmental hazards. If this excess RH can be exploited for power generation that can be used to displace part of oil used in power generation and can add more security to energy supply. RH is converted to energy using different technologies such as direct combustion, co- firing, gasification, pyrolysis, and anaerobic digestion [5]. However, the two most proven and common technologies are the direct combustion and the gasification [5, 6, 7]. Most of today’s biomass power plants are direct-fired systems where the biomass fuel is burnt in a boiler to produce high-pressure steam, which is used to power a steam turbine driven power generator. Eng. (Dr.) Asanka S. Rodrigo, Ph.D., M.Sc.(Eng), B.Sc.Eng.(Moratuwa), MIEEE, MIEE, AMIE(Sri Lanka), Senior Lecturer, Department of Electrical Engineering,, University of Moratuwa. Eng. Shantha Perera, M.Sc., B.Sc.Eng.(Moratuwa), CEng, MIE(Sri Lanka), Chief Engineer, Ceylon Electricity Board. ENGINEER - Vol. XXXXVI, No. 04, pp. [9-17], 2013 ©The Institution of Engineers, Sri Lanka ENGINEER 9 ENGINEER 2 Gasification is another commonly used option that can be used to generate electricity using RH. This technology is now widely being used in rice growing countries for driving small scale power plants of the order of 10 kW- 100 kW, though the technology can be used for higher capacities of several Mega Watts [9, 10]. Therefore, RH is considered as one of the potential sources of power generation in Sri Lanka as it is renewable, sustainable and indigenous, as the rice is the staple food of the country as well as the crop with highest land area under cultivation. Furthermore, RH is considered as a waste material with negligible commercial value at the moment. Since it being a local fuel, it will provide financial benefits to the local community as well. Therefore, the main objectives of this study were to assess the potential of rice husk based power generation in Sri Lanka and to analyze the economics of harnessing this resource. 2. Methodology First, the paddy production data of the country over last twenty years were collected and analyzed to determine the potential of rice husk production at national level. From the initial investigations, highest RH generation districts were identified by analyzing paddy production capacity and milling capacity. As a case study, one of the highest potential districts was considered for detailed study. In order to collect data pertaining to availability and distribution of rice mills, their capacities and average rice production, type of rice being processed, a survey was conducted among the rice mills in the selected district. During this survey, data from 650 rice mills were collected and 338 mills were identified as the mills that have significant milling capacity. Using the data, the maximum size of the RH based power plant and the conversion technology to be used are identified. Finally, a financial analysis for the plant was performed to ensure the viability of the power generation using RH. Even though the case study was confined to a limited area of the country, the results obtained were used to assess the potential at the national level. 3. Paddy Cultivation in Sri Lanka The total land area under paddy cultivation in Sri Lanka is estimated to be about 870,000 hectares [8] at present and this is the highest land area occupied by any single crop accounting for almost 34% of total agricultural lands in the country [10]. As paddy is a wetland crop that needs a lot of water for its growth, it is cultivated seasonally in Sri Lanka so that it gets enough water from rainy seasons. The two seasons in the year that the paddy is cultivated are known as “Maha” and “Yala” which falls during North-East monsoon (November to February) and South-West monsoon (May to September) respectively. Figure 1 shows the seasonal variation of the paddy production in Sri Lanka. Figure 1- Paddy production in two seasons from 1952 to 2010 (Source: Paddy Statistics, Agriculture and Environment Statistics Division, Department of Census and Statistics-Sri Lanka [10]) Accordingly, it can be seen that production of Maha season is almost double that of Yala season. Since rice is the staple food of the country, demand for rice remains almost constant throughout the year. Therefore, paddy is stored and released to the market as rice at a constant rate. Rice production of the country has increased from about 2.2 million tons to about 4.5 million tons over a period of 20 years. This clearly shows that there is a continuous growth of paddy production in the country. In order to find the paddy production trend in Sri Lanka, paddy production data was fitted into a trend line as shown in Figure 2. Figure 2- Annual paddy production in thousand tons from 1980 to 2010 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 1 9 5 2 1 9 5 9 1 9 6 6 1 9 7 3 1 9 8 0 1 9 8 7 1 9 9 4 2 0 0 1 2 0 0 8 P a d d y P r o d u c t i o n ( 0 0 0 M t . ) Cultivation Year Maha Season Yala Season R² = 0.833 0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 1 9 7 5 1 9 8 0 1 9 8 5 1 9 9 0 1 9 9 5 2 0 0 0 2 0 0 5 2 0 1 0 2 0 1 5 2 0 2 0 P a d d y P r o d u c t i o n , P ( 0 0 0 M t . ) Cultivation Year, t ENGINEER 10 3 ENGINEER It is clear that there is an increasing trend in paddy production. After liberating the country from 30 year long civil war, most of the paddy lands which had been abandoned in the Northern and the Eastern provinces are now being added to paddy farming and this too will have a considerable impact on the paddy production in the future. 4. Availability of Rice Husk The outer cover of paddy grain is called as the RH which accounts for about 14%-27% of its weight [11]. The gross energy content of the RH can be determined using Husk to Paddy Ratio (HPR) and calorific value (CV) of RH. CV of RH has been reported by various research works to be in the range of 12.1-15.2 MJ/kg [7, 12]. Sri Lanka cultivates different variety of paddy. HPR values for most common paddy verities are given in Table 1. Since HPR varies between 18% and 23% for different varieties of paddy, an average value of 20% was taken for the future analysis. Similarly, average CV value of 13.6 MJ/kg is considered for the analysis. Table 1- Husk to Paddy Ratio values of common local paddy varieties Paddy Variety HPR H4 0.1972 BG 3-5 0.1816 Podiwee A8 0.2166 Pachchaperumal 0.2065 BW 78 0.2274 BG 400-1 0.2012 LD125 0.2300 BG33-2 0.1898 BW 170 0.2170 MI 329 0.2045 (Source: National Cleaner Production Centre, Sri Lanka [11]) Even though paddy is cultivated all over the country, paddy is not grown at same scale everywhere. It was found that Ampara, Polonnaruwa, Anuradhapura and Kurunegala districts show a greater potential due to their higher production compared to the rest of the areas in the country (see Figure 3). However, paddy production is not the only indicator that shows the real RH availability; instead the paddy processing/milling capacity of each district will be a better criterion. Table 2 shows the milling capacity of highest paddy production districts. Figure 3- Annual Average Paddy Production in Different Areas (Year 1999 to 2010) (Source: Paddy Statistics, Agriculture and Environment Statistics Division, Department of Census and Statistics-Sri Lanka [10]) Table 2- Milling Capacity of Higest Paddy Production Districts No. of Mills Milling Capacity (kg/Day) N o t f u n c t i o n i n g T o t a l < 1 0 0 0 1 0 0 0 < C ≤ 2 5 0 0 2 5 0 0 < C ≤ 5 0 0 0 5 0 0 0 < C ≤ 8 0 0 0 < 8 0 0 0 Ampara 9 4 8 11 55 3 90 Polonnaru wa 28 30 40 19 27 13 157 A’pura 16 18 27 26 20 6 113 Hambanto ta 2 2 3 11 28 2 48 Kurunegal a 2 10 15 6 4 0 37 By considering both production and milling data, it was found that Ampara, Polonnaruwa, Anuradhpura and Hambantota were identified as the districts with highest potential. Even though the paddy production is higher in Kurunegala district, milling capacity appears to be comparatively low where the harvest is taken away to another part of the country for milling. As there were practical limitations in collecting required data island wide, Polonnaruwa was selected for detailed case study, since it is the second highest districts in paddy production as well as a leading area of paddy milling. Ampara, the highest paddy production district, was not selected for the case study due to the lack of data collectors available in the district. 0 100 200 300 400 500 600 A M P A R A K U R U N E G A L A M A H A W E L I ' H ' B A T T I C A L O A B A D U L L A M O N A R A G A … M A T A L E R A T N A P U R A G A L L E K E G A L L E G A M P A H A V A V U N I Y A C O L O M B O J A F F N A A v e r a g e A n n u a l P a d d y P r o d u c t i o n i n ' 0 0 0 M T s ENGINEER 11 ENGINEER 4 During the survey, data from 650 rice mills were collected. Out of these mills, data from 338 mills which are considered as mills with significant milling capacity (higher than 2 tons/month) were analysed. Table 3 shows the details of selected mills. Table 3- Rice mills in Polonnaruwa categorized on capacity. Scale Capacity (tons/month) No. of Mills Output (Mt.) Small ≤ 10 191 10033 Medium >10 and ≤ 100 94 46359 Large > 100 53 243750 Total 338 300142 Even though, paddy mills are scattered all over the district, it was found that most of the large mills are around Kaduruwela and Minneriya whilst very small portion of mills lying in and around Medirigirya area. RH production in year 2010 was found as shown in Figure 4. Figure 4- Rice husk production [in Mt] in different areas of Polonnaruwa in 2010 RH in the district was found to be used in different applications. Among those applications, (a) paddy boiling and drying (55%), (b) brick kilns (16%), (c) poultry farms (5%) and (d) sales for other applications like tobacco curing, bakeries and as a fertilizer in agriculture (2%) are identified as the key applications and 22% of excess RH is left in the district after using RH for different other applications. 5. Power Generation Potential in Polonnarauwa District RH is used for most of the activities without any financial transaction. Considering the RH production of year 2010, minimum potential of the Polonnaruwa district can be calculated as follows; Total RH production in year 2010 - 109,480 Mt Excess RH left (22% of total RH production) - 24,086 Mt It was assumed that the total RH quantity left can be tapped for power generation, annual electricity generation potential can be calculated by taking CV as 13.6MJ/kg and Conversion efficiency as 15%. Then the annual electricity generation (Potential) - 13.65 GWh If about 60% of this potential can be tapped, the minimum capacity of the power plant that can convert this potential into electricity is about 1 MW at 90% plant factor. 5.1 Possible Conversion Technology. Even though there are several technologies available for biomass conversion, direct combustion and gasification are the two widely used technologies for RH conversion. Gasification is widely used for small-scale power plants because of its higher efficiency low cost and simple operation. But this technology is not hassle free due to producer gas produced in gasification contains lots of impurities like tar, ash particles and dust. Gas cleaning is therefore an essential requirement, when using gasification technology for power generation. Therefore, power plants driven by gasification technology require more maintenance. Compared to gasification, direct combustion is more efficient and easy to apply for cogeneration also. However, this method requires higher investments. Considering difficulties in maintenance and gas cleaning in gasification, direct combustion with steam cycle is selected as the preferred and used for financial analysis of this study. 6. Financial Analysis. Even though there is enough fuel in Polonaruwa to run a power plant of 1 MWe, this resource can only be tapped, provided such power plants are commercially viable. 6.1 Capital cost of the project Capital investment of this project includes costs associated with all the plant and equipment, lands, buildings, utility services, project management, design and consultancy, construction and installation works and approvals etc,. Project cost of number of biomass power plants (recently constructed) of the capacity ranging from 1 MW to 10 MW were collected and their capital investment were analyzed. This analysis revealed that capital cost per kW of most of these power plant falls within the range of 1000 US$ - 1500 ENGINEER 12 5 ENGINEER US$ [6]. Further, it was found that per kW capital cost of most of the power plants constructed in this region (in Asia using Asian equipment ) were closer to 1000 US$ while the same for other power plants in Europe and America remained closer to 1500 US$ margin or above. Therefore the capital cost of this plant was assessed on the basis of 1250 US$/kW which is the middle value of the above range. Accordingly, capital cost of this power plant was estimated to be LKR 143.75 million (Exchange rate was taken as 1 US$ = 115 LKR). 6.2 Fuel cost of the power plant In Polonnaruwa district, more than 70% of the RH produced was found to be used for different applications. From the balance, very small fraction is sold and the remainder is excess. Due to prevailing low demand, this excess RH has no commercial value and even taking away of this excess RH free of charge from mills is most welcome by millers as disposal is a problem to them. However, if a real demand arises, this situation can change and easily a firm price can be set for RH. Moreover, if RH is transported there will be an additional cost on top of RH price. As per the information received from the survey, it was found that cost of RH per kg at a site ranges from LKR 1.70 to LKR 3.00 depending on the distance of transportation. Considering RH price plus handling and transportation charges, per kg cost of RH was assumed to be LKR 3.00 with yearly escalation of 3% for this analysis. 6.3 Operation and maintenance (O&M) cost Different literature gives varying values for O&M cost of biomass power plants. From the records of power plants which are in operation, annual O&M cost found to be within the range of 3% to 7% of capital investment with 5 % yearly escalation [14]. Therefore annual O&M cost was taken as 5 % of capital cost with 5% escalation. 6.4 Debt /equity ratio of the capital investment Since Sri Lanka is new to biomass power generation field and success of this kind of projects is not yet proven, 70/30 typical debt equity ratio may not be achieved. Therefore, it was assumed that 60/40 debt-equity ratio is achievable for this kind of power plant. 6.5 Finance cost and loan repayment period It was found that the typical interest rate of project financing schemes of the local banking sector is around 12% per annum for a period of six years. 6.6 Tax and incentives Income tax of 28% is levied by the Inland Revenue for local investments and as per the new tax regime of 2012 certain tax concessions are offered to local investors in order to promote private investments within the country [15]. Under this, a period of tax holiday is granted depending on the type and the scale of the investment. The projects with investment between LKR 100-200 million fall into medium scale category where this quantum of investment is eligible for five-year tax holiday. However, renewable power projects are not listed under this category in Sri Lanka. Even though it is not clear, whether the renewable energy investments qualify for a tax holiday, depending on the quantum of investment it was assumed that five-year full tax holiday will be granted for this type of projects, because the country’s energy policy is to promote renewable power generations. 6.7 Returns of the project The direct return of this power plant will be the revenue generated from the electricity sale. The expected revenue from the power plant can be estimated using prevailing tariff published by the Public Utility Commission of Sri Lanka (PUCSL) under Standardized Power Purchase Agreements (SPPA) [16]. Under this scheme, a developer can choose one from the two tariff options available. Under option I, three tier tariff is declared while two component out of three in this tariff subject to yearly escalation. Option 2 is a flat tariff with no escalation at all allowable during the first 20 years. For the baseline case analysis, option 2 of the tariff, that is 17.71 LKR/kWh, was considered. Other than that, there can be indirect returns through clean development mechanism (CDM) benefits and ash sale. However, for this analysis these indirect returns have not been considered. Financial analysis for this conceptual power plant was performed considering different cash flows for a period of 20 years. Table 4 shows the summary of the analysis. Based on above analysis project IRR and equity IRR was found as 34.2% and 56.2% respectively for the baseline case while simple payback period comes to 2 years 8 months. Higher IRR values plus short payback period indicate that implementation of a 1 MW RH based power plant is financially viable in Polonnaruwa district. ENGINEER 13 ENGINEER 6 Table 4- Key Figures of Financial Analysis. Capacity of the Power Plant 1 MWe Plant factor 90% Expected energy export ( 90% of total) 7.0956 GWh/year Life span 20yrs Capital cost( @1250 US$/kW) LKR 143.75 million (@115 LKR/US$) Tariff ( Option 2 of SPPA) LKR 17.71 Expected annual revenue LKR 103.10 million Debt/Equity ratio 60/40 Depreciation period 20yrs Fuel consumption( @ 40 tons/day) 13,140 tons/year Fuel Cost( @ 3.00 LKR/kg) LKR 39.42 million Escalation of fuel cost/year 3% O & M cost ( 5% of capital) LKR 7.19 million Escalation of O&M cost/year 5% Cost of capital 12% Loan repayment period 6 years Tax/ tax holiday 28% / 5 years Discount rate 15 % Project internal rate of return (PIRR) 34.2% Equity internal rate of return (EIRR) 56.2% Simple payback period 2 years and 8 months However, at implementation stage the values taken can be changed, giving rise to different result. Hence, sensitivity of this project was checked as follows by varying the critical components in the cash flow. The behavior of Pooled Internal Rate of Return (PIRR) and Economic Internal Rate of Return (EIRR) under seven cases, were examined and summarized in Table 5. Table 5- Summary of sensitivity analysis Case Description PIRR EIRR 1 If tariff option I is selected in SPPA as against option II 32.0% 49.4% 2 If capital cost is based on 1500 US$/kW(Capital- 172.5MRs) 26.3% 39.1% 3 If the plant factor drops to 0.8 from 0.9 in the baseline case 31.5% 48.2% 4 If fuel price escalates 5% as against 3% in the baseline case 32.2% 53.7% 5 If O&M cost is 7% of capital as against 5% in the baseline case 31.6% 50.9% 6 If fuel price is 4 LKR/kg as against 3 LKR/kg in the baseline case 22.3% 32.1% 7 If no tax holiday is offered as against 5 yrs in the baseline case 27.4% 39.9% Results of the sensitivity analysis further shows that even under adverse conditions, the project is financially robust. 7. Surplus RH availability at country level It was recorded that 50% of the paddy has been found to be milled in commercial scale in bigger mills while the balance is milled by the farmers for themselves in small mills [11]. When these facts are summed up with the annual paddy production, surplus RH availability of the country was found as 30% of total RH production. This is the excess without any use which can be used to generate electricity with an annual energy potential of 180 GWh. 8. Harnessing RH Power Generation Potential in the Country Even though it was estimated that there is a potential of 180 GWh for power generation using RH, harnessing entire above potential may not be viable as a part of this RH comes from lesser known districts of paddy production in the country. Therefore, setting up of RH power plants in these districts has to be ENGINEER 14 7 ENGINEER ruled out due to shortage of fuel as well as collection difficulties. Hence, three different models in setting up of RH power plant were identified that can ensure maximum usage of RH for power generation. They are (1) Commercial scale RH power plants, (2) small scale power plants under Net Metering Scheme and (3) Off grid RH power plants. 8.1 Commercial scale RH power plants: Commercial scale grid connected RH power plants are the most welcome option when considering the current power situation in the country. In order to promote this kind of renewable energy power plants, the PUCSL has formulated an attractive tariff system with LKR 14.53 per kWh for biomass based power plants [16]. If RH power plants of this scale (order of several MW) are to be set up in the country, it is required to ensure that there are sufficient stocks of RH to drive such power plants throughout the year. To be more precise, there has to be considerable number of commercial scale paddy mills in the vicinity of a RH power plant of this scale to keep it running throughout the year. It was found that the Polonnaruwa district shows a potential of 1 MW power plant at viable scale. When looking at the paddy production in the country at district level and the current level of utilization of RH in the country, possibility of setting up RH power plants of the scale of several MW (exceeding 2 MW) seem to be far more remote. However, paddy production data of Ampara, Anuradhapura and Kurunegala also show that there is a substantial potential of setting up RH power plant of the capacity ranging from 1MW to 2 MW in these districts. But, the exact capacity can only be confirmed only after doing a survey on these districts to ensure availability of surplus RH and the viability of collection of the same. 8.2 Small scale power plants under Net Metering Scheme For commercial scale power generation mentioned above, RH is more feasible to be collected from commercial rice mills as RH is available in bulk at such mills with guaranteed supply over a long period. Hence, RH from other mills (medium and small scale) will be left for being exploited in small-scale power plants in the range of 30kW to 40kW. Results of this study done for Polonnaruwa district, show that the 80% of total mills in the district are medium and small scale. Even in other areas of paddy cultivation, this proportion may remain same due to limitation of large mills in the country. RH produced in these mills can be used to generate electricity for captive power requirements of these mills while any surplus electricity being exported to the national grid under the Net Metering Scheme introduced by the Ministry of Power and Energy. According to this scheme, any consumer with an electricity supply below the contract demand of 42 kVA is eligible to connect his renewable energy facility to national grid and thereby qualify to pay only for net energy consumption [3]. Therefore, mills which fall into the category less than 42 kVA electricity demand have a very high potential of developing RH power plant up to the installed capacity of 42kVA as fuel is available right there and of totally free of charge. Table 6 shows a financial analysis performed on 35 kW RH fired power plant operated under Net Metering Scheme by a rice miller. Table 6- Financial Analysis of 35 kW plant under Net Metering. Capacity of the Power Plant 35 kWe Plant factor 67% (16 hrs/day) Expected energy export ( 90% of total) 184.88 MWh/year Life span 10 yrs Capital cost( @1000 US$/kW) LKR 4.03 million (@115 LKR/US$) Tariff ( under Industrial category) LKR 12.08 (10.50 + 15% fuel surcharge) Expected annual revenue LKR 2.23 million Debt/Equity ratio 60/40 Depreciation period 20 yrs Fuel consumption( @ 1.11 tons/day) 365 tons/year Fuel Cost Free Escalation of fuel cost/year 5% O & M cost ( 10 % of capital) LKR 0.4 million Escalation of O&M cost/year 5% Cost of capital 12% Loan repayment period 6 years Tax/ tax holiday No tax Discount rate 15 % Project internal rate of return (PIRR) 52.07% Equity internal rate of return (EIRR) 89.69% Simple payback period 2 years and 1 months ENGINEER 15 ENGINEER 8 This shows that this option is very attractive to the small and medium scale rice millers. However, most of these rice millers are ordinary people with less technical knowledge. Therefore, it is necessary to promote this option among rice miller by providing necessary technology and financial support. 8.3 Off grid RH power plants At present, the electrification level of the country stands around 90% [2]. But, still there are some areas in the country with low electrification and most of these areas are remote places. Powering paddy growing areas without grid electricity such as remote villages in Ampara, Anuradhapura and Polonnaruwa by means of small off grid internal combustion engine driven power plants of the capacity of 30 kW-40 kW with RH gasifiers would be a viable option. 9. Discussion According to the data collected in this study, annual average paddy production of the country was found to be around 4 million metric tons. If assumed that entire paddy production is processed into rice, this much of paddy can produce about 800,000 metric tons of RH per annum with gross energy content of 10,880 pJ. But, in real terms harnessing this much of energy from RH is not at all possible due to low efficiencies associated with converting RH into energy. If it is assumed that conversion ratio of 20% is achievable, then total net energy potential of RH per annum comes to 1,632 pJ. Tapping this energy again in mass scale becomes challenging as the viability of using RH for energy production has other constrains such as collection and transportation difficulties as well as supply chain issues. According to the case study performed on Polonnaruwa district, about 22% of RH was found to be left for exploitation in Polonnaruwa district as that portion was surplus after using the balance for above mentioned applications. However, the excess RH level is not uniform all over the country due to variation in use of RH for different application from area to area. It was found that, about 50%-70% of RH produced in parboiled rice producing mills goes into paddy parboiling and drying within the same mills. Since paddy parboiling and drying is the number one application of RH, this has become the critical factor in deciding the surplus RH availability of an area or a district. It was estimated that about 30% of excess RH can be easily exploited for power generation in Sri Lanka with an annual energy potential of 180GWh. It was further revealed that the harnessing this resource for power generation is viable under three different option in Sri Lanka namely (1) Commercial scale RH power plants, (2) Small scale power plants under Net Metering Scheme and (3) Off grid RH power plants. However, it is understood that due to lack of technical knowledge and financial capability, still RH based power generation is at under exploited status. Therefore, the key players like ministry of power and energy, Sri Lanka Sustainable Energy Authority and Public Utilities Commission of Sri Lanka should take a fresh initiative to shed light on the issues that have hampered harnessing this resource for power generation and find lasting solution to such problems. Among such issues, prominence should be given to following. Most potential developers of this resource are rice millers themselves. But, their awareness on this kind of power generation is very poor. Hence, an awareness campaign on different aspects of RH power plants has to be launched among the prospective rice millers and other interested parties. Capital investment must be a critical factor that hampers such projects in the country. If financing is not made to such projects by the lending agencies, that will be a major stumbling block to the implementation of such projects. Therefore, the government should come forward to provide financial assistance to prospective RH power plant developers through state banks at least till such power plants are popularized in the country. Tax incentive for renewable power projects are now confined to large scale projects with capital exceeding Rupees 300 million [15]. For prospective developers of RH power plants, this ceiling would be too high to get eligible for the current tax concession. (Existing potential in the country is not high enough to put up RH power plants with capital more than Rupees 300 million). Therefore, the government should look at this seriously and need to include tax concession for projects of this magnitude too. The government should encourage more R&D works towards rice husk usage for power generation. ENGINEER 16 9 ENGINEER Usage of RH for non-energy application should be discouraged while introducing alternative sources for such applications. 10. Conclusion Based on the analysis presented in this paper the RH based power generation in Sri Lanka is viable with substantial financial return. Most potential developers of this resource are rice millers. But, this resource is not yet exploited due to their lack of awareness on this kind of power generation and lack of financial support. Key players like ministry of power and energy, Sri Lanka Sustainable Energy Authority and Public Utilities Commission of Sri Lanka should take initiative to harness this resource for power generation. References 1. Sri Lanka Sustainable Energy Authority, Sri Lanka Energy Balance 2010:An Analysis of Energy Sector Performance. 2. Ceylon Electricity Board, Statistical Digest 2010. 3. Ministry of Power and Energy Government of Sri Lanka, National Energy Policy and Strategies of Sri Lanka, 2010. 4. Rodrigo, A. S., Wijayatunga, P. D. C., “Pricing of Embedded Generation: Incorporation of Exernalities and Avoided Network Losses”, Energy Conversion and Management, Vol 48, No. 8, August 2007, pp 2332-2340. 5. Brown, G., Hawkes, A. D., Bauen, A., Leach, M A, “Biomass Applications”, Centre for Energy Policy and Technology, Imperial College, London UK. 6. Oliveira, M. O., Neto, J. M., Inocencio, M. C., Ando Junior, O. H., Bretas, Perrone, A. S. O. E., “Viability Study for Use of Rice Husk in Electricity Generation by Biomass”, Proceedings of International Conference on Renewable Energies and Power Quality (ICREPQ’12), Santiago de Compostela (Spain), 28th to 30th March,2012. 7. Balat, M., Balat, M., Kırtay, E., Balat, H., “Main Routes for the Thermo-Conversion of Biomass into Fuels and Chemicals: Part 2: Gasification Systems”, Energy Conversion and Management, Vol 50, No. 12, December 2009, pp. 3158–3168. 8. Dasappa, S., Paul, P. J., Mukunda, H. S., Rajan, N. K. S., Sridhar, G., Sridhar, H. V., “Biomass Gasification Technology – A Route to Meet Energy Needs”, Current Science- Special section: Application of S&T to rural areas, Vol 87, No. 7, October 2004, pp. 908-916. 9. Zhang, L., Xu, C., Champagne, P., “Overview of Recent Advances in Thermo-Chemical Conversion of Biomass”, Energy Conversion and Management, Vol 51, No. 5, May 2010, pp. 969- 982. 10. Department of Agriculture, Sri Lanka,. http://www.agridept.gov.lk/index.php/en/cro p-recommendations/808, Visited January 11, 2013. 11. National Cleaner Production Center, Sri Lanka, “Project on Converting Waste Agricultural Biomass to an Energy /Material Resource”; 2010. 12. Agriculture and Environment Statistics Division, Department of Census and Statistics-Sri Lanka, http://www.statistics.gov.lk/agriculture/Padd y%20Statistics/PaddyStats.htm, Visited November 25, 2012. 13. Kapur, T., Kadpal, T. C. and Garg, H. P., “Electricity Generation from Rice Husk in Indian Rice Mills: Potential and Financial Viability”, Biomass and Bioenergy, Vol 10, No. 5–6, 1996, pp. 393–403. 14. El-Kordya, M. N, Badra, M. A, Abeda, K. A, Ibrahimb, M. A, “Economical Evaluation of Electricity Generation Considering Externalities”, Renewable Energy, Vol 25, No. 2, February 2002, pp. 317–328. 15. Board of Investment, Investment with Tax Incentives (Section 17) www.investsrilanka.com, Visited January 20, 2013. 16. Public Utilities Commission of Sri Lanka, “Non- Conventional Renewable Energy Tariff Announcement, Purchase of Electricity to the National Grid under Standardized Power Purchase Agreements (SPPA)”, 2012. ENGINEER 17 1 ENGINEER Investigation on Efficiency of Repairing and Retrofitting Methods for Chloride induced Corrosion of Reinforced Concrete Structures B.H.J. Pushpakumara, Sudhira De Silva and G.H.M.J. Subashi De Silva Abstract: The corrosion of steel reinforcement bars is one of the major deterioration mechanisms of Reinforced Concrete (RC) structures. Once the corrosion signs appear on the concrete surface, it may be too late to prevent further corrosion. As a result, service life of RC structures would be reduced. Use of repairing and retrofitting methods at the appropriate time will contribute enormous saving of country budget, which is required for re-construction. This paper presents an experimental investigation of repairing and retrofitting methods for the RC structures corroded due to chloride attack. As repairing methods for delaminated areas of corroded RC structures, Fly Ash (FA) and Silica Fume (SF) mixed mortars were developed and their performances were evaluated. As retrofitting methods, Cathodic Protection (CP) and Electrochemical Chloride Extraction (ECE) were conducted. ECE method is similar to CP method except the anode was covered by Ca(OH)2 layer. RC beams with concrete of Grade 20 and reinforcement bars of 16 mm diameter were cast. Efficiency of both repairing and retrofitting methods was evaluated by measuring free and total chloride ion concentration and rust production. Efficiency of repairing and retrofitting methods was further evaluated by measuring resistivity and Rapid Chloride Permeability Test (RCPT) and current measurement, respectively. FA mixed mortar reduces the total chloride ion concentration near embedded steel reinforcement bars in concrete by 40% while SF mixed mortar reduces 25% compared to OPC mortar. FA and SF mixed mortars prevent the corrosion process by minimizing the diffusion of chloride ions into concrete. It was found that CP method removed around 49% of total Cl - while ECE method removed around 69% of total Cl - near the steel reinforcement of RC beams. Removing of total Cl - by ECE is grater (in 20%) than that of by CP method. Chloride contaminated concrete that is still sound can be retrofitted by using CP and ECE methods and the spalled and detached concrete can effectively be repaired by using FA mixed mortars. Keywords: Accelerated Corrosion Test Method (ACTM), chloride attack, corrosion, repairing mortars 1. Introduction Sri Lanka is an Island surrounded by the Indian Ocean and has large number of Reinforced Concrete (RC) structures near the coastal belt. These structures are suffered by chloride ion induced corrosion; therefore the service life of these RC structures may be reduced. When use of the structures is critical, the authorities have to use reconstruction measures. As a result, considerable portion of country’s budget are consumed for reconstruction of corroded RC structures. Figure 1 - Existing corroded bridge girders The use of effective repairing and retrofitting methods at right time helps to increase the service life of RC structures, as result there will be enormous saving of country’s budget. If a RC structure is cracked or spalled or delaminated (Figure 1), repairing of these damaged areas must be done using mortars before applying the retrofitting methods. Eng. B.H.J. Pushpakumara, B.Sc. Eng. (Hons) (Ruhuna), AMIE (Sri Lanka), Rresearch Student, Department of Civil and Environmental Engineering, Faculty of Engineering, University of Ruhuna, Sri Lanka Eng. (Dr.) Sudhira De Silva, PhD (Saitama), M.Eng (Saitama), PG. Dip (Strut.), B.Sc. Eng. (Hons) (Moratuwa), C.Eng., MIE(Sri Lanka), Member-JCI (Japan), Senior Lecturer, Department of Civil and Environmental Engineering, Faculty of Engineering, University of Ruhuna, Sri Lanka, Eng. Dr. (Mrs.) G.H.M.J. Subashi De Silva, PhD (Saitama), PG. Dip (Strut.), B.Sc. Eng. (Hons) (Moratuwa), C.Eng., MIE(Sri Lanka), Senior Lecturer, Department of Civil and Environmental Engineering, Faculty of Engineering, University of Ruhuna, Sri Lanka ENGINEER - Vol. XXXXVI, No. 04, pp. [19-30], 2013 ©The Institution of Engineers, Sri Lanka ENGINEER 19 ENGINEER 2 Most of repairing mortars minimize the diffusion of chloride ions to steel reinforcements by increasing the density of repair mortar, producing chemical reaction with chloride ions, applying water proof materials and applying hydrophobic agents with repair mortars (Jennifer et al. [8]). These repair mortars could normally consist of more than one type of cement (special cement, like ultra-fine alumina cement), additions (silica fume, slag or fly ash), aggregates (normal, light weight and special type fillers), admixtures (such as plasticizers), air-entrainers and accelerators, polymer additives and fine polymer fibers (Saraswathy and Song [11]). The Fly Ash (FA) and Silica Fume (SF) are normally used during concrete batching to increase the density and strength. Therefore, FA and SF would be effective admixtures to increase the density of repairing mortars. According to Detwiler et al. [3], the optimal replacement percentage of cement (by mass) in concrete by using SF is between 6% and 8%. The amount of FA between 25% and 60% of cement weight were normally used to produce sustainable, high performance concrete mixtures (Aggarwal et al. [1]). Cathodic Protection (CP) is one of most effective retrofitting method for the corroded RC structures. CP works by using current, to shift the potential of reinforcement steel in negative direction. If the potential is shifted far enough so that all of the steel reinforcements behave as cathode, in which corrosion will be terminated. An advantage of using CP method as a retrofitting method for RC structures is that only spalled and detached concrete need to be repaired. Chloride contaminated concrete that is still sound (i.e., no any visible external damages like cracking and spalling) can remain in place, because the CP system prevents further corrosion and in fact, reduces the concentration of chloride ions adjacent to the protected reinforcement bars. The Electrochemical Chloride ion Extraction (ECE) method, which is another retrofitting method, is similar to the CP method except that the anode in ECE method is covered by a chemical media. This chemical media absorbs Cl - which approach towards the anode area (i.e., concrete surface). ECE method is a new technology and it accelerates the efficiency of CP method by removing the Cl - effectively. The chemical media can be any of alkaline solutions which react with Cl - and produce chloride salt of particular cation (Virmani and Clemena [13]). Further, Ca(OH)2, which is an alkaline solution, would react with Cl - and produce CaCl2 salt. Formation of CaCl2 salt will prevent the present of free Cl - near anode, which helps to continue current path without restriction. Therefore, Ca(OH)2 would be an effective chemical media for ECE method. Use of repairing and retrofitting methods at right time, helps to increase service life of RC structures which save the financial expenditures of a country for reconstruction. Therefore, it is an urgent need to introduce suitable repairing and retrofitting methods for RC structures in Sri Lanka. The objectives of this study are; To identify the performances of Fly Ash and Silica Fume mixed mortars for corrosion To determine and compare the performances of Cathodic Protection and Electrochemical Chloride ion Extraction methods To identify the performances of Ca(OH)2 as a new chemical media for Electrochemical Chloride ion Extraction method 2. Methodology Fly Ash (FA) and Silica Fume (SF) mixed mortars were evaluated as mortars for repairing spalled and detached RC structures. Cathodic Protection (CP) and Electrochemical Chloride ion Extraction (ECE) were evaluated as retrofitting methods for chloride ion contaminated concrete which is still sound. 2.1 Repairing Methods Five RC beams were cast with concrete of Grade 20 and reinforcement bar diameter of 16 mm. Three steel stirrups with diameter of 6 mm were used in each steel reinforcement bar cage. Size of the RC beams is 380 mm x 80 mm x 130 mm and the clear cover is 10 mm (Figure 2). All the RC beams were immersed in water for 28 days of curing. After curing, RC beams were covered by using mortar layers. The mortars were prepared with the cement sand ratio of Figure 2 - Schematic diagram of cross section of RC test beam (Dimensions are in mm) Mortar layer Initial clear cover ENGINEER 20 3 ENGINEER 1:3. Surfaces of RC beams were chipped and 10 mm thick mortar layers were applied as the final clear cover was to be 20 mm (Figure 2). Cement in the mortar was partially replaced by adding FA and SF on weight basis. Two beams were treated with 50% of FA mixed mortar and other two beams were treated with 10% of SF mixed mortar. The remaining RC beam was used as the control beam with 10 mm thick OPC mortar layer. After applying mortars, RC beams were again immersed in water for 28 days of curing. After 28 days curing, the RC beams were subjected to Accelerated Corrosion Test Method (ACTM) until 0.2 mm corrosion crack width was attained. Then, free and total chloride ion concentration, rust production and resistivity of the RC beams were determined. In addition, chloride ion permeability through mortars was evaluated by conducting Rapid Chloride Permeability Test (RCPT). 2.2 Retrofitting Methods Eight RC beams having the sizes of 400 mm x 100 mm x 150 mm were cast using Grade 20 concrete. Four steel reinforcement bars with diameter of 16 mm were tied into a cage by using three steel stirrups, whose diameter is 6 mm. Two wires were set to each beam for the purpose of applying current to the beams during experiments. The cover depth was kept as 20 mm. All RC beams were immersed in a water tank for 28 days of curing. Corrosion in seven RC beams was accelerated by performing ACTM until 0.2 mm corrosion crack width was attained. The remained RC beam was not subjected to ACTM and it was labelled as “before performing ACTM beam” and used as control beam to identify effect of ACTM on corrosion. This control beam and one beam out of seven corroded beams by using ACTM were tested for free and total chloride ion concentration and rust production measurements in order to identify the effect of ACTM on corrosion of RC beams. Four RC beams (i.e., two beams for each retrofitting method) were retrofitted by using CP and ECE methods. Other two beams were used as control beams. Three of these beams (i.e., control beam and one beam from each retrofitting method) were tested for free and total chloride ion concentration and rust production measurements in order to determine effectiveness of each retrofitting method. The remaining three beams; (i.e., control beam and one beam from each retrofitting method) were again subjected to ACTM. Resistivity of these three RC beams was measured by using Resistivity meter (James – RM8000). 2.2.1 Cathodic Protection (CP) Method Copper meshes were connected to 400 mm x 150 mm surface of each RC beam as shown in Figure 3(a) and the mesh was covered by using wet sags for the purpose of contacting the mesh with concrete surface (Figure 3(b)). The current was supplied by using a Direct Current (DC) power supply unit; the positive terminal was connected to copper mesh (i.e., anode) and the negative terminal was connected to embedded steel reinforcements which would be the cathode. Since the higher current flow cause hydrogen ion generation, current should be maintained below 1 A (Jennifer et al. [8]). Hydrogen ion generation can cause the brittle failure of RC structures. Therefore, a current of 0.3 A was supplied initially. The current of 0.3 A could be supplied with the voltage of 11.5 V. This lead to which the current was supplied to the system with a constant voltage of 11.5 V and was monitored with time until a negligible current reading (current reading < 0.1 A) was observed. 2.2.2 Electrochemical Chloride ion Extraction (ECE) Method In this method, copper meshes were connected to RC beams as similar to that in CP method presented in Section 2.2.1. The meshes were covered by using hydrated lime (Ca(OH)2) layer of 5 mm thickness (Figure 4(a)). Similar to the CP method, electric system was prepared and the current was supplied (Figure 4(b)). In (a) (b) (a) (b) Figure 3 - CP method procedure (a) Copper mesh was fixed (b) Wet sags were laid on the beams Figure 4 - ECE method (a) Ca(OH)2 layer was laid on copper mesh (b) Experimental ENGINEER 21 ENGINEER 4 ECE method also, the current was monitored with time until a negligible reading (i.e., current reading < 0.1 A) was observed. 2.3 Laboratory Experiments 2.3.1 Accelerated Corrosion Test Method ACTM was based on the electrochemical polarization principle. The corrosion of embedded steel reinforcement bars were accelerated as specified in FDOT 2000 [4] and Sahmaran et al. [10]. The beams were immersed in a sea water container where the Cl - concentration of sea water was improved up to 5% by adding NaCl. The NaCl concentration of sea water sample was measured by using titration against AgNO3 and was observed as 34.8 g/l. The current for the system was supplied by using a DC power supply unit (Figure 5(a) and 5(b)). The positive terminal of the unit was connected to the embedded steel reinforcement bars and the negative terminal was connected to the copper bars which were placed under the RC beams. Then, a small current was supplied to the system with the constant voltage (5 V). Finally, the corrosion crack width was monitored with the time as shown in Figure 5(d), until 0.2 mm corrosion crack width was attained. 2.3.2 Free and Total Chloride ion Concentration Test samples were prepared by collecting the concrete powder at nine locations of each RC beam; three points of 400 mm x 150 mm surface: centre point and two points (100 mm away from the centre, at both sides). At each point, samples were collected at 20 mm, 40 mm and 60 mm depth. For each beam, the concrete powder samples, which were collected at the same depth, were mixed and separately sieved through 75 μm sieve pan. For each depth, two samples of 5 g were prepared to measure total and free chloride ion concentrations. One sample (5 g) was mixed with 50 ml of 0.05 mol/l nitric acid (HNO3) and stirred for 10 minutes using a magnetic stirrer to extract acid- soluble chlorides which were mostly equivalent to total chloride ions. Free chloride ion concentration was measured using the remaining 5 g sample. The sample was mixed with 50 ml distilled water and stirred for 10 minutes using a magnetic stirrer to extract water-soluble chlorides ions which were mostly equivalent to free chloride ions. Each solvent sample was filtered through a filter paper. Finally, all filtered samples were titrated against silver nitrate (AgNO3) and the free and total chloride ion concentrations were measured. 2.3.3 Rust Production Rust production of embedded steel reinforcement was measured by using two methods: rust weight and reduction in bar diameter. The steel bars were carefully removed from the beams and the bar surface was scrapped in order to collect rust. The weight of the collected rust was measured. The bar diameter was measured by using a vernier calliper and the reduction in bar diameter was determined. The average values of rust weights and reduction in bar diameter were calculated to determine the rust production of steel bars. 2.3.4 Resistivity Figure 6 - Procedure of measuring resistivity (a) Preparation of 8 mm depth hole (b) Preparation of additional hole at 50 mm apart (c) Use of conductive gel (d) Placing Resistivity meter (e) Measuring resistivity Resistivity meter (James RM-8000), which is capable in assessing the possible rate of corrosion in reinforcement bars with the electric (d) (e) (a) (b) (c) Figure 5 - (a) ACTM procedure (b) DC was supplied (c) Corroded beam (d) Measuring the crack width (b) (c) (d) (a) RC specimens 5 V DC supply unit 5% of Cl - solution Copper bars ENGINEER 22 5 ENGINEER resistivity measurement method, was used to measure the internal corrosion level. Two holes with the diameter of 6 mm and the depth of 8 mm, 50 mm apart (along the reinforcement bar), were drilled on the surface of beam (Figure 6(a)). The stainless steel template plate, which was given with the equipment, was used to identify the location of the additional holes (Figure 6(b)). The two holes were filled with dispenser conductive gel (Figure 6(c)). This was necessary to create proper contact between concrete and Resistivity meter probes, which, was connected to the holes as shown in Figure 6(d). Resistivity (in k Ohms cm) was measured and the possible corrosion rate (possible corrosion rate for resistivity: <5 = very high, 5- 10 = High, 10-20 = Moderate to low and >20 = Insufficient) of steel reinforcement bars was determined by referencing the resistivity levels provided with the Resistivity meter (Instruction manual [9]). 2.3.5 Rapid Chloride Permeability Test Rapid Chloride Permeability Test (RCPT) was performed for FA mixed and SF mixed mortars to evaluate permeability as specified in ASTM C1202 (Ganesan et al. [5]). Twenty one mortar cylinders (i.e., three cylinders for each mortar percentage and three OPC mortar cylinders) (Figure 7 (a)) having a diameter of 100 mm and a thickness of 50 mm were cast by using 1:3 cement sand mortar; cement in the mortar was partially replaced by FA and SF, in weight basis, as summarized in Table 1. Table1 - Admixture replacement level Type of mortar Admixture replacement percentages (%) Control (OPC) 0 Fly Ash (FA) 25, 50, 75 Silica Fume (SF) 5, 10, 15 After 90 days of curing, all the cylinders were set to the RCPT experimental setup (Figure 7(b)). Two copper electrodes were connected to terminals and the positive terminal was immersed in NaOH solution (with possessed normality of 0.3 N) reservoir while the negative terminal was immersed in 3% of NaCl solution reservoir. The cylinders were connected between two reservoirs as shown in Figure 7(b). A constant voltage of 60 V, DC was supplied to the system and current across the cylinders were recorded at 30 minute intervals over the period of 6 hours. The total Charge Passed (CP) through the specimen was computed by Simpson’s rule (Equation 1) as specified in the ASTM C1202 (Ganesan et al. [5]). CP = 9uu|I 0 + 2I 30 + 2I 60 + 2I t + ·+ 2I 330 + 2I 360 ] - - - -(1) where CP is total charge passed in coulombs, I0 is the initial current in Ampere (A) and It is current in Ampere (A) at time t, measured in minutes. 3. Results 3.1 Efficiency of Repairing Mortars Efficiency of the repairing mortars was investigated by free and total chloride ion concentration, rust production, resistivity and permeability and is presented in this section. 3.1.1 Free and Total Chloride ion Concentration Figures 8 and 9 show free and total Cl - concentration of RC beams, respectively. The Cl - concentration was determined at four different stages: before performing ACTM, after performing ACTM for OPC, FA mixed and SF mixed mortar laid RC beams. (a) (b) Figure 7 - RCPT setup (a) Repair mortar cylinders (b) RCPT experimental setup 0 0.4 0.8 1.2 1.6 20 mm 40 mm 60 mm C l - c o n c e n t r a t i o n ( k g / m 3 ) Depth (mm) Before performing ACTM After performing ACTM for control beam After performing ACTM for FA based beam After performing ACTM for SF based beam 0 0.4 0.8 1.2 1.6 20 mm 40 mm 60 mm C l - c o n c e n t r a t i o n ( k g / m 3 ) Depth (mm) Before performing ACTM After performing ACTM for control beam After performing ACTM for FA based beam After performing ACTM for SF based beam Figure 8 - Free chloride ion concentration Figure 9- Total chloride ion concentration ENGINEER 23 ENGINEER 6 It was observed that free and total Cl - concentrations decreased with increasing depth of the beam. This trend was common at all the four stages. For all depth levels (i.e., 20 mm, 40 mm and 60 mm), total Cl - concentration was higher than the free Cl - concentration. Generally, Cl - concentration is greater after performing ACTM compared to the concentration before performing ACTM. After performing ACTM, in repaired beams both free and total Cl - concentrations were less than those for the control beam. At 20 mm depth level, after performing ACTM, the total Cl - concentration is 1.514 kg/m 3 for OPC mortar. At the same depth, FA and SF mixed mortar applied beams which were subjected to ACTM simultaneously with the OPC mortar applied beam, have lower Cl - concentrations: 0.97 kg/m 3 for FA mixed mortar and 1.14 kg/m 3 for SF mixed mortar, respectively. 3.1.2 Rust Production Rust weights and reduction in bar diameter of repaired RC beams, which were tested after performing ACTM, are shown in Table 2. Table 2- Rust production Inspection Stage A v e r a g e w e i g h t o f r u s t ( W ) ( g ) A v e r a g e b a r d i a m e t e r ( Φ ) ( m m ) A v e r a g e r e d u c t i o n i n b a r d i a m e t e r ( Φ r ) ( m m ) Initial condition 0 16 0 A f t e r p e r f o r m i n g A C T M OPC mortar 7.8 13.57 2.43 FA mixed mortar 1.76 15.81 0.19 SF mixed mortar 3.64 14.75 1.25 It can be seen from Table 2 that, in the FA and SF mixed mortar applied beams, the average weight of rust and reduction in bar diameter (after performing ACTM) were lesser values compared to that of OPC mortar applied beams. The rust production (rust weight (W) and bar diameter reduction (Φr)) of beams repaired with FA mixed mortar (W = 1.76 g and Φr = 0.19 mm) was lower than that of the beam repaired with SF mixed mortar (W = 3.64 g and Φr = 1.25 mm). 3.1.3 Resistivity Resistivity of RC beams (OPC, and 50% of FA and 10% of SF mixed mortar applied) which were subjected to ACTM is shown in Table 3. Table 3- Resistivity of mortars Mortar condition Average resistivity (k Ohms cm) OPC 1.8 50% of FA 12.0 10% of SF 9.05 It has been clearly seen that resistivity in the OPC mortar applied beam was very much lower (i.e., 1.8 k ohms cm) compared to that of FA and SF mixed mortar applied beams (FA = 12.0 k Ohms cm and SF = 9.05 k Ohms cm). The resistivity of beams repaired by FA mixed mortar is slightly larger than beams repaired by SF mixed mortar. 3.1.4 Rapid Chloride Permeability Test Variation of current with time for the FA mixed and SF mixed mortar cylinders is shown in Figures 10 and 11, respectively. The current pass through the specimen decreases with increasing both FA and SF content in the mortars. This trend is clear from 25% to 75% of FA and from 5% to 15% of SF, although the trend is not clear when increasing FA content from 0% to 25% and SF content from 0% to 5% in the specimen. The calculated average charge passed (using Equation 1, 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0 3 0 6 0 9 0 1 2 0 1 5 0 1 8 0 2 1 0 2 4 0 2 7 0 3 0 0 3 3 0 3 6 0 C u r r e n t ( A ) Time (min) Ctrl 25% FA 50% FA 75% FA 0 0.1 0.2 0.3 0.4 0.5 0.6 0 3 0 6 0 9 0 1 2 0 1 5 0 1 8 0 2 1 0 2 4 0 2 7 0 3 0 0 3 3 0 3 6 0 C u r r e n t ( A ) Time (min) Ctrl 5% SF 10% SF 15% SF Figure 11- Current measurement of RCPT for SF mixed mortar Figure 10- Current measurement of RCPT for FA mixed mortar ENGINEER 24 7 ENGINEER Figures 10 and 11) across the mortar cylinders is shown in Figure 12. The charge passed values decrease with increasing of both FA and SF percentages. The average charge passed across the mortar cylinders cast by using mortar of having 75% of FA and 15% of SF are 3873 C and 4278 C, respectively. These values are lower compared to OPC mortar cylinder (7518 C). 3.2 Efficiency of Retrofitting Methods This section presents the efficiency of retrofitting methods (CP and ECE) that were evaluated by using chloride ion concentration, rust production and resistivity of RC beams and current passed through RC beams. 3.2.1 Free and Total Chloride ion Concentration Figures 13 and 14 show the free and total Cl - concentration, respectively, for the depth level of 20 mm, 40 mm and 60 mm at the four stages. . The free and total Cl - concentrations were greater after performing ACTM. Both free and total Cl - concentrations decreased with increasing depth. Comparing Figures 13 and 14, it can be clearly seen that for all depths, the total Cl - concentration was higher than the free Cl - concentration. For all depths, both Cl - concentrations were reduced after applying retrofitting methods. At 20 mm depth, the total Cl - concentration of control beam (non- retrofitted), after performing ACTM, is 1.64 kg/m 3 while it is 0.831 kg/m 3 for the beam retrofitted by the CP method and is 0.516 kg/m 3 for the beam retrofitted by the ECE method. At the same depth level, the free Cl - concentration of control beam (non-retrofitted), after performing ACTM, is 1.471 kg/m 3 while it is 0.346 kg/m 3 for the beam retrofitted by the CP method and is 0.25 kg/m 3 for the beam retrofitted by the ECE method. The free and total Cl - concentrations of beams with the condition of before performing ACTM are 0.282 kg/m 3 and 0.41 kg/m 3 , respectively, at 20 mm depth level. 3.2.2 Rust Production Table 4 shows the average weight of the rust production and reduction in bar diameter of embedded steel reinforcements. Table 4- Rust production Inspection Stage A v e r a g e w e i g h t o f r u s t p r o d u c t i o n ( W ) - ( g ) A v e r a g e b a r d i a m e t e r ( Φ ) - ( m m ) A v e r a g e r e d u c t i o n i n b a r d i a m e t e r ( Φ r ) - ( m m ) Initial condition 0 16 0 After performing ACTM 3.5 15.56 0.44 After retrofitting by CP method 4 15.48 0.52 After retrofitting by ECE method 3.75 15.53 0.47 Control specimen (without retrofitting) 4.28 15.41 0.59 The weight of the rust and the reduction in bar diameter were lower for the beams retrofitted with CP and ECE methods, compared with that of control beams. There was a negligible amount of rust production in retrofitted beams. The rust (rust weight (W) and reduction in bar diameter (Φr)) produced in beams treated by CP Figure 12- Charge passed average value 0 0.4 0.8 1.2 1.6 2 20 mm 40 mm 60 mm C l - c o n c e n t r a t i o n ( k g / m 3 ) Depth (mm) Before performing ACTM After performing ACTM After performing CP method After performing ECE method 0 0.4 0.8 1.2 1.6 2 20 mm 40 mm 60 mm C l - c o n c e n t r a t i o n ( k g / m 3 ) Depth (mm) Before performing ACTM After performing ACTM After performing CP method After performing ECE method 0 2000 4000 6000 8000 10000 C h a r g e p a s s e d ( C o u l u m b s ) Figure 13- Free chloride ion concentration Figure 14- Total chloride ion concentration ENGINEER 25 ENGINEER 8 method (W = 4 g and Φr = 0.52 mm) were greater than that in the beams retrofitted by ECE method (W = 3.75 g and Φr = 0.47 mm). As mentioned in the methodology, retrofitting methods were applied on the corroded beams (after performing ACTM). Therefore, rust is already produced on the bar surface. Comparing to after performing ACTM stage, the rust production and reduction in bar diameter of CP method (W = 0.5 g and Φr = 0.8 mm) and ECE method (W = 0.25 g and Φr = 0.03 mm) are lesser than that of control beam (W = 0.78 g and Φr = 0.15 mm). 3.2.3 Resistivity Table 5 shows comparison of resistivity of beams, retrofitted by using CP and ECE methods. Table 5- Comparison of resistivity Beam description Resistivity (k Ohms cm) Control beam 0.7 Beam retrofitted by CP 3.1 Beam retrofitted by ECE 4.5 From Table 5, it can be seen that the resistivity of the control beam was 0.7 k Ohms cm, which is significantly lower compared with the resistivity of other beams: beam retrofitted by CP method gives 3.1 k Ohms cm while beam retrofitted by ECE gives 4.5 k Ohms. The resistivity ofthe RC beam retrofitted by ECE method was slightly higher than that of the RC beam retrofitted by CP method. 3.2.4 Current For CP method, the initial current reading was 0.3 A for the first mesh, 0.23 A for the second mesh, and 0.16 A for the third mesh (the system was not continued with replacing third copper mesh after observing negligible current). Therefore the total current reduction was 0.14 A. The total time for the beam retrofitted by CP method was 31 days. For ECE method, the initial current reading was observed as 0.5 A. The initial current reading for the second mesh was 0.29 A, and for the third mesh it was 0.17 A (the system was not continued with third mesh). The total current reading reduction was determined as 0.33 A. The total time for beam retrofitted by ECE method was 23 days. 4. Discussion According to the JSCE (2001), the critical Cl - concentration is 0.3 ~ 0.6 kg/m 3 (Veerachai et al. [12]). Further, according to the JSCE (2001), when the Cl - concentration exceeds 1.2 ~ 2.4 kg/m 3 , the corrosion incorporating micro- cracks might be initiated (Veerachai et al. [12]). Simply, when the Cl - concentration of RC beams is higher than 0.6 kg/m 3 , the critical Cl - level was initiated (i.e., corrosion risk initiate) and when it was 1.2 kg/m 3 , the corrosion is initiated on embedded steel reinforcement. The threshold Cl - concentration for onset corrosion is considered as 1.2 kg/m 3 (Gunesekara et al. [6]). 4.1 Repairing Mortar Relatively high percentage of SF makes the concrete expensive and causes the difficulties in mixing and finishing. In the current study, 10% of SF and 50% of FA were used in producing mortars with high density and high strength, while reducing the difficulties in mixing and finishing. Further, 5%, 10% and 15% of SF and 25%, 50% and 75% of FA mixed mortars were used to identify optimum level of admixture percentages for reducing the Cl - diffusion, which was investigated by using RCPT. The most critical depth region related to corrosion was considered as less than 20 mm depth, because steel reinforcements were embedded to the depth of 20 mm. The total Cl - concentration, at 20 mm depth level of OPC mortar applied beam (1.514 kg/m 3 ) (Figure 9) is higher than threshold value of 1.2 kg/m 3 , hence the beam was corroded. The total Cl - concentrations of FA and SF mixed mortar applied beams (FA = 0.97 kg/m 3 and SF = 1.14 kg/m 3 ) were less than the threshold value of 1.2 kg/m 3 , hence the corrosion risk is minimized. The total Cl - reduction, at 20 mm depth level, in the beam repaired with FA mixed mortar was 0.544 kg/m 3 and in the beam repaired with SF mixed mortar was 0.374 kg/m 3 (Figure 9). FA mixed mortar reduces Cl - concentration near embedded steel reinforcement by 36%, while SF mixed mortar reduces 25% of Cl - concentration. Therefore, FA mixed mortar prevents the diffusion of Cl - effectively compared to SF mixed mortar. After the 30 days period of ACTM, the corrosion cracks (crack width > 0.2 mm) in OPC mortar applied beam were observed. Micro level corrosion cracks (crack width < 0.05 mm) were observed in the beams repaired with SF ENGINEER 26 9 ENGINEER mixed mortar and there was no surface corrosion crack in the beam repaired with FA mixed mortar. These results imply that the RC beam repaired by OPC mortar was totally corroded, and the beams repaired with SF mixed mortar was slightly corroded where the beam repaired with the FA mixed mortar seems to be free from the corrosion. The rust production (both rust weight and reduction in bar diameter) of OPC mortar applied beam is higher value compared to FA and SF mixed mortars applied beams (Table 2). Further, rust production of FA mixed mortar applied beam is negligible value (i.e., W = 1.76 g, Φr = 0.19 mm) compared to OPC mortar applied beam (i.e., W = 7.8 g, Φr = 2.43 mm). This implies that the OPC mortar applied RC beam was continued corrosion while FA mixed mortar prevents production of rust. According to Table 3, possible corrosion rate of OPC mortar (i.e., 1.8 k Ohms cm) applied beam is very high (resistivity < 5 k ohms cm) while that is high (resistivity between 5 to 10 k Ohms cm) in the RC beams repaired by using 10% of SF mixed mortar (i.e., 9.05 k Ohms cm) and moderate to lower (resistivity between 10 to 20 k ohms cm) in the RC beams repaired by 50% of FA based mortar (i.e., 12 k Ohms cm). These results imply that the internal corrosion level of the beams repaired by FA based mortar has significantly reduced. Charge passed through of 15% of SF and 75% of FA mixed mortar are lower values compared to that of OPC mortar (Figure 12). Therefore, chloride ion diffusion through these mortars (75% of FA and 15% of SF mixed mortars) is lower compared to OPC mortar. Compared to OPC mortar, 75% of FA mixed mortar reduces the diffusion of Cl - through mortars by 48.5% and 15% of SF mixed mortar reduces 43.1%. The uses of FA and SF with mortars reduce the diffusion of Cl - into concrete, which prevents the corrosion. The pozzolanic reaction in FA and SF converts the calcium hydroxide in to more of the calcium silicate hydroxide (CSH), thus increase the strength which leading to reduce permeability. Further, FA and SF have fine particles which might have contributed to minimize the voids in mortars. Therefore, the replacement of both FA and SF admixtures increases the density of mortars. As a result, FA and SF mixed mortars effectively minimize the diffusion of Cl - ions, moisture, oxygen and aggressive chemicals into concrete compared to OPC mortar. Considering the results of Cl - concentration, rust production, resistivity and charge passed values, the corrosion prevention performances of FA mixed mortar is a high value compared to SF mixed mortar. Jansen and Pratt [7] showed, by using microstructure studies, that the reaction product from the pozzolanic reaction of FA was CSH (calcium silicate hydrates) and CAH (calcium aluminate hydrates), where the latter can give rise to increased chloride binding capacity. With the increase of SF amount, the producing of CAH decreased which cause to decrease the binding capacity. Therefore, FA mixed mortar minimizes the diffusion of Cl - compare to SF mixed mortar. In addition, coal power generation has recently been established in Sri Lanka and one of by-products of the coal power plants is FA. Hence FA can be easily found in Sri Lanka with minimum cost. 4.2 Retrofitting Methods Before performing ACTM, both free and total Cl - concentrations were less than the threshold value of 1.2 kg/m 3 (Figures 13 and 14). After performing the ACTM, free and total Cl - concentrations of all test beams were greater than the threshold value of the Cl - concentration resulting to initiate corrosion. After retrofitting, the total Cl - concentrations in these beams reduced to a value less than the threshold value. This implies that the applying of retrofitting prevents or minimizes further corrosion of RC beams by reducing Cl - concentration near to embedded steel reinforcement bars. Table 7- Percentages (%) of chloride ion removed by retrofitting methods D e p t h ( m m ) Cl - concentrations as percentages (%) CP method ECE method Free Total Free Total 20 76.48 49.33 83 68.53 40 72.65 57.01 75.98 63.2 60 69.84 47.2 71.69 52.27 In the current experimental investigation, 20 mm depth level is the most critical level as the clear cover thickness provided is 20 mm. In this region, the ECE retrofitting technique removed 68.5% of total and 83% of free Cl - from the corroded beams (i.e., after performing ACTM) while the CP retrofitting technique removed 49.3% of total and 76.5% of free Cl - (Table 7). Clemena and Jackson [2], have carried out ECE method (Ca(OH)2 with H2O were used as ENGINEER 27 ENGINEER 10 chemical media) on a 28-year old bridge deck located in Arlington, Virginia and observed similar results for free Cl - ion concentration where ECE removes Cl - in concrete with about 75.8% of Cl - at the 6 to 19 mm depth, and 72.2% of Cl - at the 19 to 32 mm depth. In their study, ECE method was conducted using titanium mesh (i.e., as anode) which is better than copper mesh used in the current study. However, use of copper mesh for the anode is cost effective rather than use of titanium mesh. Compared to the current study (Cl - concentration of corroded RC beam was 1.64 kg/m 3 at 0 to 20 mm depth), the bridge deck was highly chloride contaminated (initial Cl - concentration was 4.97 kg/m 3 at 6 to 19 mm depth). Therefore, it is easier to remove free Cl - ions from bridge deck. Further, in the previous study (Clemena and Jackson [2]) ECE was conducted for bridge deck over a long period (58 days) while it was 23 days in the current study. However, removal of Cl - is less (75.8%) for bridge deck than that of the current study (83%) and final Cl - concentration (after performing ECE) of bridge deck was 1.2 kg/m 3 (which was equal to threshold value), while it was 0.516 kg/m 3 in the current study. It seems that conducting the ECE for bridge deck in-situ condition and for current study in the laboratory with less Cl - contamination condition would cause the slight variation in results for bridge deck and test beams. When the current was supplied to the CP setup, the steel reinforcement became cathode and the chloride ions are redistributed away from the steel bars and reached towards the copper anode (at concrete surface area). This might contribute to the reduction in chloride ion concentration near the steel reinforcements. In ECE retrofitting technique, the copper mesh is covered by using chemical media (calcium hydroxide) and the chloride ions which approached towards the copper mesh are removed by calcium hydroxide by producing CaCl2. This might attribute to increase the efficiency of removing chloride ions by the ECE method compared to that of the CP method. After retrofitting (CP and ECE methods), the free Cl - concentration of test beams decreases nearly down to the Cl - concentration of beams observed at 28 days curing (Figure 13). The retrofitting methods are based on the electrical system and the free Cl - contributes directly to conduct the current across concrete area. Therefore, admittedly, the free Cl - are redistributed from steel reinforcements and reached towards the concrete surface area. These free Cl - are removed by calcium hydroxide layer by producing CaCl2. On the other hand, free Cl - causes the corrosion of steel reinforcement. The bonded chloride ions may not contribute to the corrosion process. Possibility of removing free chloride ions by these retrofitting techniques is an advantage. The weight of rust production and reduction in bar diameter of retrofitted RC beams are lesser compared to control RC beam (Table 4). This implies that retrofitting methods contribute to prevent further corrosion. The resistivity of all RC beams (Control beams and retrofitted beams) is lower than 5 k Ohms cm (Table 5), implying the possible corrosion rate of all beams is very high. Before retrofitting (CP and ECE methods), RC beams were subjected to ACTM in order to make sure the RC beams were already corroded. After retrofitting, the beams were again subjected to ACTM. This allowed the Resistivity meter to measure the resistivity at both corrosion stages: before retrofitting and after retrofitting. The observed corrosion rate was significantly high value, possibly because the corrosion was accelerated by conducting the ACTM. Further, considering the measured values, control beam shows very lower value, means it would be highly corroded than other two beams. The free Cl - conduct the current through the concrete area, hence considering the current distribution through retrofitted RC beams, the free Cl - were redistributed away from the embedded steel reinforcement (cathode) area towards concrete surface area (copper mesh anode). Therefore, the reduction in current of ECE and CP methods implied the reduction in Cl - concentrations. The efficiency of removing chloride ion (based on the current reduction) for CP method was 46.7%, while it was 66% for ECE method. Time taken to remove Cl - by CP method is around 31 days while it is around 23 days for the ECE method indicating that the ECE method removes Cl - within shorter time period compared to the CP method. In both CP and ECE methods, the potential of embedded steel reinforcement were shifted to negative direction, making the steel reinforcement became cathode with time. When the steel reinforcement bars totally charge with negative ions, further continuation of current might help to produce hydroxyl ions (with the presence of water and oxygen) near the reinforcement bars, resulted in creating of alkaline conditions. This leads to re-passivate the steel reinforcements, produce iron oxide ENGINEER 28 11 ENGINEER (Fe2O3) film surrounding the steel reinforcement bars, prevent the diffusion of oxygen to steel bar area and hence prevent the further corrosion. In addition, these retrofitting methods help to redistribute the Cl - from embedded steel reinforcement towards the concrete surface. The Cl - which approached to concrete surface is removed by calcium hydroxide as producing calcium chloride (CaCl2). The use of calcium hydroxide as chemical media in ECE retrofitting method cause to increase the efficiency of the method, compared to CP method. 5. Conclusions As repairing methods, Fly Ash (FA) and Silica Fume (SF) mixed mortars were developed and Cathodic Protection (CP) and Electrochemical Chloride ion Extraction (ECE) were conducted as retrofitting methods for corroded RC structures. Efficiency of both repairing and retrofitting methods was evaluated by measuring free and total Cl - concentrations and rust production. Efficiency of the methods was further evaluated by conducting resistivity measurements and Rapid Chloride Permeability Test (RCPT) relevant to the repairing mortars and current measurement relevant to the retrofitting methods. The FA and SF mixed mortars minimize the corrosion of RC structures by preventing the diffusion of Cl - into concrete. It was found that the 50% of FA mixed mortar performs better compared to the 10% of SF mixed mortar. With the use of coal power generation, recently established in Sri Lanka, FA can be easily found with minimum cost, as one of by-product of the coal power plants is FA. The CP and ECE methods remove around 49% and 69% of total Cl - , respectively. These retrofitting methods prevent further corrosion of RC structures by removing Cl - near the steel reinforcements and producing alkaline conditions around the steel bars which helps to form iron oxide film around the steel bars. With ECE method, calcium hydroxide helps to increase the efficiency of the CP method process by nearly 20% with removing the Cl - approached towards copper mesh area. The spalled and delaminated areas of corroded RC structures can be repaired by using FA mixed mortars and the chloride contaminated concrete that is still sound can remain in place by using ECE method, effectively. The use of calcium hydroxide as chemical media with ECE method accelerates the performances of CP method. Experimental investigation conducted in this study confirmed that the use of both FA based repair mortar and ECE retrofitting more effectively help to prevent the chloride ions diffusion into concrete and enhancing the service life of structures. Acknowledgments The authors wish to express their special thanks to Research Grant of Transforming University of Ruhuna to International Status (TURIS-2011) for providing necessary funds and Faculty of Engineering, University of Ruhuna for providing technical assistance for carrying out the research work presented in this paper. Furthermore authors’ sincere thank goes to Prof. Hiroshi Mutsuyoshi, Department of Civil and Environmental Engineering, Saitama University, Japan for providing necessary experimental materials and technical guidance. References 1. Aggarwal V., Gupta S. M. and Sachdeva S.N., “Concrete Durability through High Volume Fly Ash Concrete (HVFC) - A Literature Review”, International Journal of Engineering Science and Technology, Vol. 2(9), 2010, pp. 4473-4477. 2. Clemena G. G. and Jackson D. R., “Pilot Applications of Electrochemical Chloride Extraction on Concrete Bridge Decks in Virginia”, Transportation Research Record, No. 1597, 1997, pp. 70-76 3. Detwiler R. J., Whiting D. A. and Lagergen E.S., “Statistical Approach to Ingress of Chloride Ions in Silica Fume Concrete for Bridge Decks”, ACI Materials Journal, Vol. 96, No. 6, Nov.-Dec. 1999, pp. 670-675. 4. Florida Department of Transportation (FDOT), “An Accelerated Laboratory Method for Corrosion of Reinforced Concrete using Impressed Current”, Manual of Florida Sampling and Testing Methods, Tallahassee, FL, September 2000, pp. 6. 5. Ganesan K., Rajagopal K. and Thangavel K., “Chloride Resisting Concrete Containing Rice Husk Ash and Bagasse Ash”, Indian Journal of Engineering & Materials Sciences, Vol.14, June 2007, pp.257-265. 6. Gunasekara M. P. C. M., Mutsuyoshi H. and Sumita A., “Renovate RC Structures with Newly Developed Mortar Considering Chloride Binding and Inverse Diffusion ENGINEER 29 ENGINEER 12 Phenomenon”, International Conference on Structural Engineering, Construction and Management (ICSECM), Kandy, Sri Lanka, December 2011. 7. Jansen H. U. and Pratt P. L., “The Binding of Chloride Ions by Pozzolanic Product in Fly Ash Cement Blends”, Advances in Cement Research, Vol. 2, No. 7, 1989, pp.121-129. 8. Jennifer L. K., David D. and Carl E. L., “Evaluation of Corrosion Protection Methods for Reinforced Concrete Highway Structures”, Structural Engineering and Engineering Materials, SM Report No.58, University of Kansas Center for Research, Inc., Lawrence, Kansas, May 2000. 9. NDT James Instruments Inc., “Instruction Manual - RM-8000 Resistivity Meter”, James Instruments Inc., Chicago, U.S.A. 10. Sahmaran M., Li V. C. and Andrade C., “Corrosion Resistance Performance of Steel- Reinforced Engineered Cementitious Composite Beams”, ACI Materials Journal, V. 105, No. 3, May-June 2008, pp. 243-250. 11. Saraswathy V. and Song H., “Evaluation of Cementitious Repair Mortars for Corrosion Resistance”, PortugaliaeElectrochimicaActa, 26/5, June 2008, pp. 417-432. 12. Veerachai L., Toshimitsu S., Yuichi T. and Masayasu O., “Estimation of Corrosion in Reinforced Concrete by Electrochemical Techniques and Acoustic Emission”, Journal of Advanced Concrete Technology, Vol. 3, No.1, February 2005, pp.137-147. 13. Virmani Y. P. and Clemena G. G., “Corrosion Protection-Concrete Bridges,” Report No. FHWA-RD-98-088, Federal Highway Administration, Washington, D.C., September 1998. ENGINEER 30 1 ENGINEER Productivity in Construction-A Critical Review of Research D. A. R. Dolage and P. Chan Abstract: The aim is to bring a fresh perspective to the construction productivity research agenda, which is congruent with the new demands in the construction industry and its ever changing nature. The articles which had ‘construction productivity’ as a keyword in the abstract, and were published in each of the three journals (JCEM, CEM and IJP) from the earliest year that the articles had been uploaded to the respective official website of each journal were identified. Out of 5862 articles searched, only 121 articles fulfilled the selection criteria, the titles of which were examined. The past decade has witnessed the continuation of the same relentless research interest in productivity studies. The findings revealed that, in the studies: five types of productivity have been examined; five data collection methods have been deployed; research objects can be classified under seven categories. The research objects in a high number of studies are devoted to ‘measurement of productivity’ and ‘examining the casual relationships with productivity’. The study ascertained that the main drawbacks of past productivity studies are the strong empirical inclination of methodologies adopted and the overwhelming positivist approach to examining productivity issues. The absence of follow-up studies to investigate the validity of productivity measurement techniques and the models and to test the claims made in productivity improvement studies, is a striking feature. Another impressive finding is the lack of scholarly attention to incorporate blue-collar worker perspective, employee involvement, and social dimension into productivity research. Key Words: Construction productivity, Critical review of Research, Worker integration, Interpretivism 1. Introduction In view of the substantial share of construction sector in the whole economy, in most countries its stakeholder attention is invariably focussed on improved resource utilisation or productivity (Kazar et al., 2008). The decline or the enhancement of its economic activity has a significant direct impact on the smooth functioning of the economy and the well-being, of any country. On the one hand, the competitive business environment in the present day has increasingly compelled construction organisations to focus on their core business activities to a greater extent (Sheng, 2002). On the other, the construction industry is under tremendous pressure to be efficient due to issues such as increasing cost of energy, labour, raw material and competition. The rational participants who perceive productivity as the ratio of output resources to input resources must strive to get higher outputs from lower inputs. In the construction industry higher productivity levels invariably lead to superior profitability, drawing the constant attention of the management. Owing to its significance to the profitability of construction projects, productivity is regarded as one of the relentlessly discussed topics in the construction industry by both the practitioners and the researchers. Nonetheless, several studies have shown that the construction industry has shown under-performance in comparison to other industries (Lee et al., Smith 2011). The nature of construction industry has been going through a transformation over the last three decades due to a number of reasons, the two main being the emergence of off-site production (OSP) of components and the increasing engagement of subcontractors by main contractors. The growing interest in sustainable construction, rapid technological improvements and increasing labour costs should provide opportunities for OSP to serve sustainable projects (Baba et al., 2008). According to Chan and Kaka (2007), past productivity studies bore a strong positivistic tradition, often maintaining the managerial perspective. In view of ever changing construction project environment the adequacy Eng. (Dr.) D. A. R. Dolage, CEng, FIE(Sri Lanka), BSc Eng. (Moratuwa), MSc (Reading), MA (Colombo), MBA (SJP), DBA (UniSA), Senior Lecturer, Department of Civil Engineering, The Open University of Sri Lanka. Dr. Paul Chan, BSc (Hons), PhD, ICIOB, PCAPL, FHEA, Lecturer, School of Mechanical, Aerospace and Civil Engineering, The University of Manchester, UK. ENGINEER - Vol. XXXXVI, No. 04, pp. [31-42], 2013 ©The Institution of Engineers, Sri Lanka ENGINEER 31 ENGINEER 2 of existing research paradigm is questionable. The scepticism of this inadequacy has been made justifiable further by the complexities in the modern construction project since it has become increasingly difficult to benchmark performance and record, manage and transfer information and knowledge effectively (Zhao at al., 2007). Nowadays, building projects are becoming much more complex and difficult (Chan et al., 2004). This invariably adds to concerns regarding productivity in construction projects. 2. Research gaps and objectives Construction productivity has been generating significant interest in both the construction industry itself and academia (Park, 2006; Ellis and Lee, 2006). Despite the many publications, a comprehensive critical review of ‘productivity research in construction’ with respect to levels of analysis, research objects and methodologies, has not appeared hitherto. The studies based on review of literature are gaining popularity, particularly, in order to critically review the existing body of knowledge with a view to examine plausibility of their findings and the relevance of emerging theories which are being hotly debated. However, critical reviews have been carried out in related areas such as ‘competitiveness research in construction’ (Flanagan et al, 2007) and ‘energy and building’ (Schweber and Leiringer, 2012). The aim is to bring a fresh perspective to the construction productivity research agenda which is congruent with the new demands in the construction industry and its transient nature. This study which is based on theoretical review of recent academic publications on ‘construction productivity’ aims to closely examine and recapitulate the body of knowledge. It is therefore considered pertinent to summarise the progression of research on productivity in construction through a comprehensive review, and to suggest new directions for further studies. The study has two distinct research objectives. First, to examine the perception of construction productivity, progression of research interest and the spectrum of research objects currently being investigated in the existing literature concerned with ‘construction productivity’. Second, to review these studies systematically with respect to each research object in order to assess the level of research and new directions for future studies. 3. Key theoretical concepts under investigation Productivity is about the efficient and effective use of all resources that go into the ‘production’ of an activity or a process. Resources invariably include labour, materials, equipment, space, energy, finance knowledge, information and time. In the literature, the term productivity is often used interchangeably with the term efficiency (Halligan, 1994). There are different measures of productivity and the choice between these depends either on the purpose of the productivity measurement and/or data availability. The productivity measures adopted by researchers are: total factor productivity, labour productivity, construction productivity, and process productivity. The adequacy of each measure needs to be investigated in the light of the changing nature of the industry and the new demands. Further, the validity of findings and claims made based on such measures need to be verified. The research interest in the application of concept of worker involvement, integration of perspectives of workers and managers on productivity in the construction industry is still at the nascent stage since Chan (2006, 2007) rekindled it to the research agenda. The possibility of widening the worker integration to encompass social relationship is also investigated. 4. Research Method The relevant research publications for the study were selected from potential journals according to reputation and impact ratings as in Schweber and Leiringer (2012). A preliminary survey, carried out adopting ‘Google Scholar’ using keyword ‘‘construction + productivity’’, revealed that in the first 50 articles, more than 90 per cent are from three Journals namely, Journal of Construction Engineering and Management (JCEM), Construction Management and Economics (CME) and International Journal of Project Management (IJPM). The first two journals are prominent ‘construction research’ journals while the latter is a ‘business and social science’ journal. The other ‘construction research’ journals which included relevant articles are: Structural Survey, Building Research and Information, Engineering Construction and Architectural Management. It was observed that the range of objects of research covered in the articles which appeared in the three series of journals amply covered those found in the former set of ENGINEER 32 articles. Hence, for this exercise, only the top three journals namely JCEM, CME and IJPM were considered. In all three journals, the authors carefully went through the titles of all the articles appearing in each issue of all the volumes looking for any articles which might be concerned with ‘construction productivity’. The abstracts of all the articles which had some relevance to ‘construction productivity’ were examined closely and the ones which had the keyword ‘productivity’ in the abstract were considered for the study. One of the objectives of the study was to find out the rate of publication of articles on ‘construction productivity’ in each journal over the years it has been in publication. The following information was extracted from the articles: i) The type of productivity being examined ii) The level of analysis at which the productivity is examined. iii) The data collection methods and methodological approaches adopted iv) The main research object under investigation The results obtained with respect to the above four aspects were graphically represented in terms of different types and corresponding percentage. 5. Research Findings 5.1 Trends in publication of research articles on construction productivity The articles which had ‘construction productivity’ as a keyword in the abstract and published in each of the three journals from the earliest year that the articles had been uploaded to the respective official website of each journal were identified. Out of 5862 articles searched, only 121 articles fulfilled the selection criteria, the titles of which were examined. While the highest attention to construction productivity research has been paid by the JCEM (66 out of 2390; 2.76 per cent), the attention of the CME (38 out of 1672; 2.27 per cent) is of the same order. The publication of the same in IJPM (17 out of 1800; 0.94 per cent), as can be expected, has been low. Figure 1 depicts the trend of publications of these articles, as a percentage of the total number of articles published in the given year for all three journals. The articles published since 1980 having been considered, Figure 1 shows that all three journals have published articles on productivity without a continuous upward or downward trend. Since 1990 until 2012 the publication of articles follows a similar trend regarding all three journals. This indicates that research interest in productivity still holds an important position in the construction management research domain. This provides a valid justification for the study undertaken. - ENGINEER 33 5.2 Types of productivity and levels of analysis In the articles reviewed, the researchers have been interested in investigating various types of productivity, namely TFP, partial productivities such as technology and labour, process and construction. The distribution, in the descending order of application is as follows; labour (57 per cent), construction (22 per cent), process (13 per cent), TFP (6 per cent), technology (2 per cent). Since construction is a labour intensive industry, it is understandable that the highest percentage of articles are devoted to the labour productivity. Construction labour productivity has become a buzz word and one of the most frequently researched topics (Jarkas and Camille, 2012). A high number of articles (41 articles) are devoted to investigate the labour productivity of construction industry. However, evidently, in these articles, little or no scholarly attention has been paid to the changing nature of construction workforce due to growing dependence on subcontracting, OSP and increase in self employment. The second highest percentage of research effort is expended to investigate ‘construction productivity’ which connotes site or industry productivity, in general. In almost all the studies which engaged questionnaire based surveys to elicit information, the respondents’ perception on productivity is based on their understanding on ‘construction productivity’. In some articles, although the objective of the article is to measure the construction productivity what has really been measured is either process productivity or labour productivity (eg. Thomas et, al., 1986; Portas and AbouRizk, 1997, Artdti, 1985; Herbsman and Ellis, 1990; Goodrum et al., 2009; Goodrum and Haas, 2002). Hence, a striking feature among some of these articles is the omission to provide a clear definition of productivity. The articles that mainly studied the productivity of operations and equipment have adopted the process productivity. It is observed that successive studies have been conducted to study productivity aspects of the same operation (eg. 4 on piling, 2 on crane, 3 on structural design, 4 on construction machinery). Although the construction industry constitutes a large number of operations and equipment that warrant studies that closely examine their productivity aspects, current research agenda has evidently failed to deal with most of them. The reason for the low number of articles focusing on TFP is because of the difficulty to accurately determine and measure all the input resources utilised to achieve the output. Only two studies have been devoted to investigate technology productivity, one dealing with changes in material technology (Goodrum et al., 2009) and the other dealing with changes in equipment technology (Goodrum and Haas, 2002). Figure 2 graphically represents the distribution of the five types of productivity examined in these articles. Table 2 displays the levels of analysis, types of productivity measures adopted in these articles, along with the areas that had been under investigation in these studies. - ENGINEER 34 5 ENGINEER Table 2 - Breakdown of Level of analysis, type and focus of productivity Level of Analysis Type of productivity Focus Industry Labour Construction industry (41) Construction Construction industry (17) Partial Construction industry (2) Total Construction industry (8) Sector Labour Housing, industrial construction, high-rise building Construction Highways (2), water and wastewater treatment, low-cost housing Operation and Equipment Labour Four operations, masonry wall systems, mechanical and sheet metal (2), Material and technology Construction Masonry projects (2), Process Piling (4), carne (2), structural design (3), Group of operations, structural concrete, concrete batching plant, bridge, asphalt operations, trenchless excavation, drywalls, Construction equipment (2), earth moving (2) Trade/Contractor Labour Group of Tasks, formwork (2), concrete work (5), beam fixing, bricklayer, mechanical work, fabricator, electrical and mechanical contractors, labour intensive contractors, on site and offsite contractors, steel drafting and fabrication (2), rebar fixing (3), masonry (2) Process Key building tasks The different levels of analysis performed in these articles can be grouped under four main categories. They are in the descending order of application; industry (57 per cent), Operations/Equipment (20 per cent), Trade/Contractor (16 per cent) and sector (7 per cent). A large majority of studies which have explored different types of productivity industry-wide are based on questionnaire surveys. Since the responses to the questionnaires are based on accumulated experience gained over the years being in the industry, they reflect the industry-wide perception. Further, the availability of panel data on information required to compute different types of productivity has encouraged the proliferation of a large number of studies examining the construction industry, holistically. Productivity being a phenomenon associated with social dimension, the validity of findings made in such studies may have room for error. Alternatively, if these studies had been focussed on different industry sectors instead of the industry as a whole, the validity of findings could have been improved. Only a minute percentage of articles having been dedicated to examine productivity at industry sector indicate the lack of rigor in the present research agenda. Of the studies 57 per cent considering the industry holistically, compared to 7 per cent of studies doing the same on the industry sector is grossly inadequate. A substantive number of productivity studies on associated with various operations and equipment are shown in Table 2. In a similar manner, a good number of articles have looked into the productivity of some trades and tasks of specific contractors. However, as shown in Table 2, a large majority of these trades are from a limited range (eg concreting of slabs, beams and columns, formwork and structural steel fixing). Mostly, these are related to concrete work or to tasks for which data available in abundance or to tasks which are easily measurable. A likely explanation for the situation is ‘methodolatory’ where researchers tend to conduct research, particularly positivists, to suit to the methods and data availability. Figure 3 displays the percentage distribution of levels of analysis adopted in these articles. ENGINEER 35 ENGINEER 6 5.3 Data collection methods and methodologies As one of the steps of the study, data collection methods and research approaches or methodologies deployed in the articles were examined. Unsurprisingly, most articles adopted more than one data collection method to varying degrees in their investigations; for example, every article had data coming in from literature review. Therefore, in order to keep the analysis simple, the data collection method that had the predominant bearing on their findings was picked for this study. Figure 4 shows graphically how productivity researchers have adopted the five different types of data collection methods. The highest percentage (35 per cent) of studies have adopted data maintained over a period at site or project offices concerning relatively stable, established, repetitive and simple construction activities. On some occasion, the analysis is based on a single data set (case study) and on the others the same was based on different data sets. However, the researchers seldom refer to possible differences among contractors regarding data collection and recording, which concern, if not addressed, could invalidate the findings. In 23 per cent of studies, the type of method adopted to collect data is the questionnaire survey. A great majority of studies are devoted to identify and rank factors influencing productivity. Scholars have not been adequately concerned about differences in defining factors on which perceptions are sought and the significance of providing productivity definitions to respondents. Construction industry data bases and industry panel data are used in 19 per cent of studies. While in 11 per cent of studies field measurements are used, in 12 per cent of studies literature surveys are adopted. This also shows a high percentage (91 per cent) of articles use site data, questionnaire surveys, field measurements, site observations and data bases evidencing the acute empiricist inclination in the contemporary productivity studies. The empirical studies favour positivist research approach and therefore all these articles are based on positivism. Positivism was defined as a scientific methodology that aims to reach the laws of human behaviour and social life (Tekin and Kitaman, 2013). The positivist approach pays scant attention to the complexity of social construction and interactions in its attempt to explain them through a one- dimensional linear cause-and-effect relationship. It is heartening to note that only a small percentage (12 out of 136, 09 per cent) of studies are based on pure review of literature, an approach which has some characteristics of interpretivist research. Interpretivism is a school of thought that builds on the meaning of social interactions. Further, most of the studies which adopted questionnaire surveys elicit perceptions of respondents on productivity related questions. However, the accuracy of - ENGINEER 36 7 ENGINEER 5.4 Types of research objects Each article fundamentally has been devoted to examine one type of productivity discussed earlier, at a particular level of analysis, with respect to a particular predominant research object. The different research objects investigated in the articles can be clustered into six different generic groups namely trends, measurement, means and assessment of improvement, identification and ranking of influencing factors, nature and causal relationships. The number of articles, percentages and focuses are presented in Table 3, under each generic group of research object. The research object of the largest proportion of articles (49 out of 136; 36 per cent) are basically concerned with productivity measurement although the stated research objects had been; assessment (25 out of 49), measurement (14 out of 49) and modelling (10 out of 49). In order to simplify the analysis and comparison, .it was decided to group them all under productivity measurement since all these are associated with measurement. A striking observation is the growing adoption of Artificial Neural Network (ANN), a technique adopted when variables involved are too complex for the human brain to handle, to model construction productivity (eg. Portas and AbouRizk, 1997; Sonmez1 and Rowings, 1998; Ok and Sinha, 2006; Zayed at al., 2005). Some other studies had adopted other advanced statistical models (eg. Zayed and. Halpin, 2005) and software programmes (eg. Everett and Slocum, 1993). The interest in the use of advanced techniques to estimate productivity can be ascribed to complexities in measurement. However, the concern still remains whether the intense research effort to measure productivity using dominant positivist approach is still valid when social and human facet of productivity have been ignored. Almost an equal number of articles (48 out of 136; 35 per cent) investigate the causal relationship of productivity with different attributes, namely, Technology (5), Operation (4), Design (4), Motivation (6), Benchmarking (6), Buildability (4), Physical environment (3), Labour deployment (6), Change orders (4), Management practices/actions (8);the number of articles are depicted in parentheses. Technology- The articles examine how changes in material and equipment technology cause improvements in productivity. Also some examine how new technologies such as findings in these studies is questionable particularly because productivity being a complex phenomenon, respondents could have different perceptions. The most conspicuous methodological feature of productivity studies concerns the involvement of positivist versus interpretivist approaches. Despite the adoption of advanced statistical techniques (eg. correlation, multivariate, principal component analysis and structural equation modelling), it can be argued that the rigid theory testing methodology of the positivist research paradigm cannot address the specific research context needed to understand a complex phenomenon like productivity. - ENGINEER 37 ENGINEER 8 advance machine guidance systems improve process productivity. Operation-The articles are mainly dealing with the assessment of productivity of standard operations like concreting, fabrication and joining. Design- The research objects of articles broadly falls in to two areas; impact on construction productivity due to changes in the design while construction in progress; impact on construction productivity due to improvement in the design process. Motivation-The impact of operative motivation and training on productivity and the impact of demotivating factors on productivity are dealt with in these articles. Benchmarking-The articles are devoted to explain how benchmarking could be effectively used to improve productivity. The articles are focusing on benchmarking across countries and across contractors. Buildability- The influence of buildability on productivity of operations and quantification of the influence in its relationship with productivity have been dealt with in these studies. Physical environment-The causal relationships of inhibitors such as climatic effects, rainfall and thermal effects with the productivity are quantified in these articles. Labour deployment-The impacts of scheduled overtime, overmanning, shift work on productivity are assessed in these articles. Change orders-The contribution of change orders or construction changes to labour productivity are investigated in these articles. Management practices/actions-The importance of management practices and actions such as BPR, quality assurance and participatory management on productivity are assessed in these articles. The purpose of a substantial number of articles (16 out of 136; 12 per cent) is to identify and rank the factors influencing productivity. This type of studies has been carried out in different countries: USA (Herbsman and Ellis, 1990; Dai et al. 2009), Singapore (Lim and Alum, 1995), Thailand (Makulsawatudom et al., 2004) and Indonesia (Kaming et al.1998) to name a few. It was observed that there is incongruence among the factors identified and variance in the classification of factors. Further, the utility of these studies is limited because these factors are country specific and has less relevance to the construction industry of another country. The research object of a reasonable number of articles (14 out of 136; 10 per cent) is to explore means and assessment of productivity improvement. These studies are separately looking into two aspects of productivity; namely areas where productivity improvement is possible (eg. Arditi,1985; Arditi and Mochtar,1996) and productivity improvement techniques (Cottrell, 2006; Chan and Kaka, 2007; Han et al.2008, Minchin et al.2008). It is striking to note that follow up studies have not been conducted to empirically verify the validity of claims and promises made in productivity improvement studies. Understandably, only a small number of studies (5 out of 136; 4 per cent) explore the productivity trends and these studies are focusing on TFP or labour productivity of the construction industry. Four of these studies have been carried out in the USA (Allmon, 2000; Arditi and Mochtar, 2000; Teicholz, 2001 Rojas and Aramvareekul, 2001) and of which three are looking into the labour productivity trends. Chau (1993) has conducted a study on estimating industry-level productivity trends in the building industry of Hong Kong. Again, only a small percentage of studies (4 out of 136; 3 per cent) look into the nature of construction on productivity and theories of productivity. Figure 5 graphically illustrates the distribution of research objects. ENGINEER 38 Table 3 - Distribution of Research objects Groups of Research Objects Focus Article Number Percentage Productivity trends Labour productivity and total factor productivity of different countries 5 4 Measurement Assessment (25) Sub contracting, scheduled overtime, change orders, scheduled overtime 49 36 Measurement (14) Modelling (10) Means and assessment of Productivity Improvement Design process, Measures to be taken, motivation, Productivity improvement officer, design integrated process/planning, quality circles, 14 10 Identification and ranking of factors Influencing the industry 16 12 Nature and Theories Variability of productivity, implications of nature, field studies, baseline theory and practice 4 3 Causal relationships Technology (5) 48 35 Operation (4) Design (4) Motivation (6) Benchmarking (4) Buildability (4) Physical environment (3) Labour Deployment (6) Change orders (4) Management practices/actions (8) 9 ENGINEER 6. Implication of findings on theory, practice and policy An impressive finding of the study is the overwhelming engagement of the positivist research in construction productivity research. The positivist approach ignores the complexity of social matters and interactions, and productivity being a concept associated with humans, this research approach alone is not adequate to understand the exact context of construction productivity. The construction scholars should adopt a research paradigm like interpretivist approach which is understandably new to construction productivity research domain. According to the review of articles, the scholars have investigated the construction productivity in a managerial perspective while paying scant attention to worker perspective. Construction professionals are subjected to a plethora of occupational demands that can have a negative effect on their psychological wellbeing (Love et al., 2010). These demands can have an adverse - ENGINEER 39 ENGINEER 10 influence on individual and organizational productivity. The studies need to be undertaken to investigate new areas like the worker happiness, nature of self and social supports, work stress and mental health among construction professionals directly focussed on worker wellbeing. These aspects which are directly associated with workers’ social construction can only be addressed through multidisciplinary and inductive research. Therefore, in this context, interpretivism is a much better choice than positivism. The findings suggest that emphasis on research objects such as productivity measurement and ranking of factors influencing productivity is overwhelming and the neglect of much pertinent research objects like worker involvement and social dimension is noticeable. The research effort should be diverted to investigate less debated areas like employee involvement, social dimension to productivity and integration of blue collar and white collar worker perspectives on productivity. The managers should be made aware of the necessity to look at the productivity improvement holistically considering the perspectives of blue collar workers too. The industry professionals should be concerned about the social dimension and get the workers involved in productivity issues and see that they are content and happy. The findings also suggest that productivity improvement studies should be followed up to empirically determine the veracity of these studies. The findings of the study will be useful to administrators of research grants to make more funds available to prospective studies on less researched areas identified above and to prospective studies that adopt new approaches. The industry professionals should be concerned about the social dimension and get the workers involved in productivity issues and see that they are content and happy. With the little evidence shown in the study, the percentage of ‘construction productivity’ research is higher in construction journals than in the business and social science journals. The scholarly interest in publication of productivity research in business and social science journals need to be examined more closely. The interest among the scholars who wish to publish in this domain should be promoted in order to bring in the multidisciplinary approach to construction productivity research. The articles reviewed refer to two productivity measures namely construction productivity (abstract measure) and process productivity (physical measure), in addition to the conventional ones such as TFP and partial productivities such as labour and technologies. The drawbacks of using ‘construction productivity’ which is not a well defined productivity measure and the necessity of defining productivity measure to lessen the subjectivity of findings were highlighted. The process productivity was formally recognised and defined. The construction scholars need to take the above concerns into consideration in future studies. They should also divert their attention to less researched types of partial productivities. 7. Conclusion This paper has sought to review the published research on ‘construction productivity’. It ascertained that the main drawbacks of past productivity studies are the strong empiricist inclination of methodologies adopted and the overwhelming positivist approach to examining productivity issues. A large number studies have explored research objects such as measurement, ranking of influencing factors and improvement techniques related to productivity. The absence of follow up studies to investigate the validity of productivity measurement techniques and the models and to test the claims made in productivity improvement studies, is a striking feature. Another impressive finding is the lack of scholarly attention to incorporate blue collar worker perspective, employee involvement, and social dimension into productivity research. Hence, in future, scholarly attention should be drawn to these areas so that new studies can be taken up. The encouragement to adopt interpretivist methodology instead of positivist approach for research studies is a necessity to bring insights into existing knowledge on construction productivity. It proves the existence of a vacuum of interpretivist research on productivity and cohabitation of approaches, interpretivism and positivism, should be encouraged. Further, this will encourage new audience of researchers adept at researching social dimension to investigate construction productivity. This provides opportunities to construction researchers to publish in business and social science journals. ENGINEER 40 11 ENGINEER The articles reviewed refer to two productivity measures namely construction productivity (abstract measure) and process productivity (physical measure) and in this article they have been formally recognised and defined. References 1. Abdelhamid, T. S., and Everett, J. G., (1999) “Time Series Analysis for Construction Productivity”, Experiments”, Journal of Construction Engineering and Management, Vol. 125, No.2, pp. 87-95. 2. Allmon, E., Haas, C. T., Borcherding, J. D .and Goodrum P. M. (2000) , “U.S. Construction Labor Productivity Trends, 1970–1998”, Journal of Construction Engineering and Management, Vol. 126, No. 2, pp.97-104. 3. Arditi, D.(1985), ”Construction Productivity Improvement”, Journal of Construction Engineering and Management, Vol.3, pp.1-14. 4. Arditi, D. and Mochtar, K. (1996), “Productivity Improvement In The Indonesian Construction Industry”, Construction Management and Economics, Vol 14, pp. 13-24. 5. Arditi, D. and Mochtar, K. (2000), “Trends in productivity improvement in the US construction industry”, Construction Management Economics”, Vol. 18, pp.15-27. 6. Chan, P. W. and Kaka, A. (2007), “Productivity Improvements: Understand the Workforce Perceptions of Productivity First”, Personnel Review, Vol. 36 No. 4, pp. 564-584. 7. Chau, K. W. (1993), “Estimating industry-level productivity trends in the building industry from building cost and price data”, Construction Management and Economics Vol. 11, pp. 370-383. 8. Cottrell, D. S. (2006), “Contractor Process Improvement for Enhancing Construction Productivity”, Journal of Construction Engineering and Management, Vol. 132, No.2, pp. 189-196. 9. Crawford, P. and Bernard Vogl, B. (2006), “Measuring Productivity in The Construction Industry”, Building Research & Information, 34(3), 208–219. 10. David W. Halligan,D. W., Demsetz, L. A., Brown, J. D., and Pace, C. B. (1994), “Action- Response Model And Loss of Productivity in Construction”, Journal of Construction Engineering and Management, Vol. 120, No.1, pp. 47-64. 11. Ellis Jr, R. D., and Lee, S. H. (2006), “Measuring Project Level Productivity on Transportation Projects”, Journal of Construction Engineering and Management, Vol. 132, No.3, pp. 314-320. 12. Everett, J. G. and Alexander H. Slocum, A. H. (1993), “Cranium: Device for Improving Crane Productivity and Safety”, Journal of Construction Engineering and Management, Vol. 119, No.1, pp. 23-39. 13. Goodrum, P. M and Haas. C. T. (2002), “Partial Factor Productivity and Equipment Technology Change at Activity Level in U.S. Construction Industry” Journal of Construction Engineering and Management, Vol.128, No.6, pp. 463-472. 14. Goodrum; P. M, Zhai, D. and Yasin, M. F.(2009), Relationship between Changes in Material Technology and Construction Productivity, Journal of Construction Engineering and Management, Vol. 135, No.4, pp. 278-287. 15. Goodrum, P. M., Haas, C. T., Caldas, C., Zhai;D., Yeiser, J. and Homm, D. (2011), “Model to Predict the Impact of a Technology on Construction Productivity”, Journal of Construction Engineering and Management, Vol. 137, No.9, pp. 678-688. 16. Herbsman, Z. and Ellis, R. (1990), “Research of factors influencing construction productivity”, Construction Management Economics”, Vol. 8, pp.49-61. 17. Jarkas, A. M. and Bitar, C. G (2012), “Factors Affecting Construction Labor Productivity in Kuwait”, Journal of Construction Engineering and Management, Vol. 138, No.7, pp. 811-820. 18. Jason Portas, J., J.and AbouRizk, S. (1997), “Neural Network Model for Estimating Construction Productivity”, Journal of Construction Engineering and Management, Vol. 123, pp.399-410. 19. Jiukun Dai1, J., Paul M. Goodrum, P. M., and William F. Maloney, W. F., (2009),” Construction Craft Workers’ Perceptions of the Factors Affecting Their Productivity”, Journal of Construction Engineering and Management, Vol. 135, No. 3, pp. 217-226. 20. Kaming, P. F., Holt, G. D., Kometa, S. T., Olomolaiye, P.O. (1998), “Severity Diagnosis of Productivity Problems--A Reliability Analysis”, International Journal of Project Management, Vol. 16, No. 2, pp. 107-113. 21. Langford, D. A., H. El-Tigani1, H. And Marosszeky, M. (2000), “Does Quality Assurance Deliver Higher Productivity?”, Construction Management and Economics, Vol.18. pp. 775–782. 22. Lim, E. C. and Alum, J.(1995), “Construction Productivity: Issues Encountered by Contractors ENGINEER 41 ENGINEER 12 in Singapore”, International Journal of Project Management, Vol. 13, No. 1, pp. 51-58. 23. Makulsawatudo, A, Sinthawanarong, K, and Emley, M.(2004), “Critical Factors Influencing Construction Productivity in Thailand”, The Journal of KMITNB, Vol.14, No.3 24. Naoum, S. and Hackman, J. (1996) “Do site Managers and the Head Office Perceive Productivity Factors Differently? “Engineering, Construction and Architectural Management” Vol.1, 2, pp. 147-160. 25. Ok, S. C. and Sunil K. Sinha, S. K. (2006), “Construction Equipment Productivity Estimation using Artificial Neural Network Model”, Construction Management Economics”, Vol. 24, pp. 1029–1044. 26. Park H. S. (2006), “Conceptual Framework of Construction Productivity Estimation”, KSCE Journal of Civil Engineering, Vol. 10, No. 5. pp. 311- 317. 27. Peter E. D. Love, P. E. D, Edwards, D. J and Irani, Z. (2010), “Work Stress, Support, and Mental Health in Construction”, Journal of Construction Engineering and Management, Vol. 136, No.6, pp. 650-658. 28. Rojas, E. M and Aramvareekul, P., (2003), “Is Construction Labor Productivity Really Declining?, Journal of Construction Engineering and Management, Vol.129: No 1. pp. 41-46. 29. Sonmez, R. and Rowings, J. E. (1998), “Construction Labor Productivity Modeling With Neural Networks”, Journal of Construction Engineering and Management, Vol. 124, No. 6, pp.498-504. 30. Tarek M. Zayed, T. M. and Halpin, D. W., (2005), “Pile Construction Productivity Assessment”, Construction Management Economics”, Vol. 131, pp. 705-714. 31. Teicholz, P. (2001), “U.S. Construction Labor Productivity Trends, 1970–1998”, Journal of Construction Engineering and Management, Vol. 112, pp.245-258. 32. Thomas, H. R., Mathews, C. T. and. Ward, J. G (2006), “Learning Curve Models of Construction Productivity”, Journal of Construction Engineering and Management, Vol. 112, pp.245-258. 33. Zayed,T. M., Halpin, D. W. and Basha I. M. (2005), “Productivity and Delays Assessment For Concrete Batch Plant-Truck Mixer Operations”, Construction Management and Economics, Vol.23. pp. 839–850. 34. Zayed, T. M. and Halpin, D. W. (2005), “Productivity and Cost Regression Models for Pile Construction”, Journal of Construction Engineering and Management, Vol. 131, No.7, pp. 779-789. ENGINEER 42 g 1 ENGINEER Streamflow, Suspended Solids, and Turbidity Characteristics of the Gin River, Sri Lanka T. N. Wickramaarachchi, H. Ishidaira and T. M. N. Wijayaratna Abstract: Human induced impacts on the river systems result in decrease in water quality, which is generally reflected by an increase of particulate matter in rivers. Turbidity and suspended solids are part of physical and aesthetic parameters and good indicators of other pollutants that are carried as sediment in suspension. Study objectives were to define the relation between turbidity and total suspended solid (TSS) concentration in Gin river at Baddegama (6°11'23" N, 80°11'53" E) in developing an estimation technique for TSS load, and to reveal how turbidity and TSS load vary with the streamflow. Linear regression model developed between turbidity and TSS concentration showed strong positive correlation (R 2 = 0.98). Results strongly suggest turbidity is a suitable monitoring parameter for TSS, where TSS evaluation is crucial when logistical and financial constraints make TSS sampling impractical. Mean daily TSS loads in the Gin river at Baddegama during 2000-2009 were modeled in the study using load-discharge rating curve for estimating constituent loads in rivers. Relatively strong relationship (R 2 = 0.85) was observed between the rating curve estimated and observed TSS loads. Estimated TSS loads were having substantial temporal variation and generally peaked in May and October, coinciding with the high flows. Turbidity which ranged between 2.3 NTU (Nephelometric Turbidity Units) and 195 NTU significantly exceeded the maximum permissible limits of the water quality standards set for the potable water as well the inland waters of Sri Lanka. Since there was no specific water quality standards developed for TSS in Sri Lanka to compare with the present values, TSS concentrations were compared with the permissible total solid levels. TSS concentrations which ranged between 2.4 mg/l and 204 mg/l were well below the maximum permissible total solid level cited in the Sri Lanka standards for potable water. Understanding on this turbidity and TSS characteristics in Gin river flow might be useful for water managers and planners to adjust operations accordingly at water treatment plants. Keywords: Linear regression, Load-discharge rating curve, Streamflow, Suspended solid, Turbidity 1. Introduction Increasing population and industrialization have resulted in a variety of impacts on the river systems, while increasing the demand for higher quality water. Increased sediment and nutrient loading in rivers adversely affect the river water quality. Water quality and river flows are correlated and this correlation varies spatially as well as temporally. Turbidity is an optical property of a liquid that causes light rays to scatter and absorb rather than transmit in straight lines through the sample. Turbid water results from the presence of suspended and dissolved matter such as clay, silt, finely divided organic matter, plankton, other microscopic organisms, organic acids, and dyes [1]. Suspended sediment and turbidity are thought to increase with increasing stream size, stream order, and drainage area because of the accumulation of sediment and nutrients from the watershed and stream banks [2]. The level of suspended solids in rivers changes rapidly and unpredictably with changing water depths and velocities related to anthropogenic causes or natural hydrologic events making the quantification of suspended solids critical. To capture these rapid changes in suspended solids, sampling must be conducted at a high temporal frequency that is usually impractical and expensive. A more practical method is to monitor a surrogate, some parameter that is closely related to the concentration of Eng. (Mrs) T. N. Wickramaarachchi, B.Sc. Eng(Hons) (Moratuwa), MPhil (Moratuwa), MJSCE(Japan), AMIE(Sri Lanka), Senior Lecturer, Department of Civil & Env. Engineering, University of Ruhuna. (Currently reading for PhD) Eng. (Dr) H. Ishidaira, B.Sc. Eng(Nagaoka ), M.Eng (Nagaoka), D.Eng (Nagaoka), Associate Professor, Interdis. Graduate School of Medicine and Engineering, University of Yamanashi, Japan. Eng. (Dr) T. M. N. Wijayaratna, B.Sc. Eng(Hons) (Moratuwa), M.Eng (AIT), D.Eng (Yokohama), C.Eng, MIE(Sri Lanka), Senior Lecturer, Department of Civil & Env. Eng., University of Ruhuna. ENGINEER - Vol. XXXXVI, No. 04, pp. [43-51], 2013 ©The Institution of Engineers, Sri Lanka ENGINEER 43 ENGINEER 2 suspended solid which can be continuously monitored [3, 4]. Turbidity measurements are theoretically well correlated to suspended solid concentration because turbidity represents a measure of water clarity that is directly influenced by suspended solids. As such turbidity based estimation models typically are effective tools for generating suspended solid concentration data [5].The use of turbidity as a surrogate for suspended solid concentration has become more common and shown by several river basin studies around the world [6, 7, 8, 9, 10]. Present study intends to assess the quality of water in Gin river because of its significance as the primary drinking water source for the Galle city. Gin river caters approximately 90 % of the drinking water to Galle, the capital city in Southern Sri Lanka. Human activities in the Gin river basin have increased substantially in the past few decades including alterations in landuse pattern; subsequent to the expansion of homesteads, cultivated area in the basin has been decreased by 10% between early 80s and late 90s [11]. These activities have been causing significant impacts on the quality and clarity of water in the river creating a need to monitor the quality of water. In spite of these activities, there remains a lack of comprehensive water quality analysis for the basin. Turbidity is a significant indicator of the quality of water. The correlation between TSS and turbidity is well documented for number of sites and turbidity has been utilized as less expensive and easiest to measure method enabling turbidity to serve as a surrogate for suspended solid. This study assessed how turbidity, as related to TSS, vary with Gin river flow at Baddegama. LOADEST, a load-discharge rating curve for estimating constituent loads in rivers is used in this study to develop regression model and estimate TSS load [12]. LOADEST is widely used to estimate constituent loads in rivers [13, 14, 15, 16, 17, 18]. Results of the study could be useful in monitoring turbidity levels to meet water quality standards, to prevent adverse effects on aquatic life, and to enhance aesthetic and recreational values. Moreover understanding on turbidity and TSS characteristics in river water assists in estimating real-time TSS based on automated turbidity record system. 2. Study Area Gin river originates from the Gongala mountains in Deniyaya having an elevation of over 1300 m and flows to the Indian Ocean at Gintota in Galle district. Gin catchment is about 932 km 2 and the catchment area at Baddegama river gauging station is 780 km 2 . The catchment is located approximately between longitudes 80°08" E and 80°40" E, and latitudes 6°04" N and 6°30" N and includes Galle (83% of the basin area), Matara (9% of the basin area), Rathnapura (7% of the basin area), and Kalutara (1% of the basin area) administrative districts. The catchment’s land is primarily used for human settlements, agriculture, natural forests and plantation forests [11]. Rainfall pattern in the catchment is of bi-modal, falling between May and September (Southwest monsoon, which is the major rainfall season), and again between November and February (Northeast monsoon) followed by the inter- monsoon rains during the remaining months of the year. Rainfall varies with altitude with mean annual rainfall above 3500 mm in the upper reaches to less than 2500 mm in the lower reaches of the catchment. Gin river annually discharges about 1268 million cubic meters (MCM) of water to sea [19]. It is the primary drinking water source, from which pipe-borne water is supplied to Galle city, the capital of Southern Sri Lanka, averaging 0.7 MCM per month. Gin river, its catchment location, and Baddegama river gauging station are shown in Figure 1. Figure 1- Gin river, its catchment location, and Baddegama river gauging station. Baddegama Gauging Station Sri Lanka Gin river’s Catchment Location Gin River ENGINEER 44 3 ENGINEER 3. Data and Analysis 3.1 Turbidity and TSS Correlation Turbidity is influenced by color, temperature, and shape of the suspended particles. Turbidity-TSS relationships have been reported, on site by site basis, and the reliability varies due to water color and suspended particle composition [20]. A direct correlation between turbidity and suspended solid concentration has been documented in many studies conducted around the world. Moreover the magnitude of turbidity in streams, lakes, and estuaries is often proportional to suspended solid concentration and the turbidity-suspended solid concentration relation has been quantified through linear regression analysis in number of studies [7, 8, 21, 22, 23, 24]. In addition, as one of the least expensive and easiest to measure methods, turbidity has been utilized to determine TSS concentration in the present study. Turbidity was measured on more than 100 samples collected over three months period at Baddegama (6°11'23" N, 80°11'53" E) and TSS concentration of the same samples were analyzed as per standard procedures [25]. This sampling effort over wide range of turbidity values, provided data for the development of regression equation to estimate TSS concentration from the turbidity. As a means of quality assurance, approximately 20% of the total number of samples were collected as duplicate samples and analyzed to determine the precision of the results. Almost Figure 2 - Linear regression of turbidity and TSS concentration including 95% prediction intervals (number of samples = 103). all duplicate TSS concentrations were within 5% of the corresponding sample TSS concentrations. Relationship between turbidity and TSS concentration was assessed using the regression analysis with 95% prediction interval. Linear regression equation was developed using the least squares method. A strong (R 2 = 0.98) highly significant (p < 0.0001) relationship existed between turbidity and TSS concentration (Figure 2). Figure 2 shows the regression best fit line bounded by 95% prediction intervals. Data fairly fit around the regression line and the model provides reasonably accurate prediction; for Gin river’s long term mean turbidity value of 25.2 NTU, the measured TSS concentration is 26 mg/l, and the model yields a TSS concentration of 26.4 mg/l with a 95% confidence interval from 25.7 mg/l to 27.0 mg/l. Regression models developed between turbidity and TSS concentration in most of the previous studies have shown similarity to the present study, based on the fact that the regression line passing through the origin. Slope coefficients in the regression models developed previously have been ranging between 0.9 – 1.3, producing comparable results to the present study. Variability of the slope coefficient from one study to another is attributed to the characteristics of suspended sediments and their transport processes in the catchment; high turbidity subsequent to fine suspended particles but low TSS concentration results in lower slope coefficient and vise versa. These acknowledge similar nature of the regression models that best describe turbidity TSS relationship though depend upon catchment land use and the associated hydrologic responses in the catchment. 3.2 TSS Load Model Development Turbidity data of 116 water quality samples at Baddegama (6°11'23" N, 80°11'53" E) tested at monthly frequency between February 2000 and November 2009 were obtained from the National Water Supply and Drainage Board (Southern), Sri Lanka. Sampling location of Baddegama is located near the streamflow gauging station maintained by the Department of Irrigation, Sri Lanka. Daily streamflow data at the Baddegama gauging station for the same period were collected from the Department of Irrigation. Water quality and streamflow data collected on a particular day are representative Standard error of estimate = 28.6 Fit line; Y=1.0457 X Coeff. off determination (R 2 ) = 0.98 ENGINEER 45 ENGINEER 4 of average concentration and average flow on that day, respectively. Using the linear regression equation developed, the corresponding TSS concentrations were calculated for the 116 turbidity measurements. The calculated TSS concentrations were used to estimate daily and monthly suspended solid loads. Load-discharge rating curve [12] for estimating constituent loads in rivers incorporated into the LOADEST, a computer programme developed by the United States Geological Survey (USGS) was used in this study to develop multiple regression model and estimate TSS loads over the period 2000 - 2009. LOADEST uses time series streamflow data and constituent concentrations to develop and calibrate regression model that describes constituent loads in terms of various functions of streamflow and time. The calibrated model is then used to estimate constituent loads using daily streamflow observations over specified time period. The calibration and estimation procedures within LOADEST are based on three statistical methods; Adjusted Maximum Likelihood Estimation (AMLE), Maximum Likelihood Estimation (MLE), and Least Absolute Deviation (LAD). AMLE and MLE methods are appropriate when the calibration model residuals (errors) are normally distributed and AMLE is the method of choice when the calibration data set contains censored data (constituent concentrations that are less than the laboratory detection limit). LAD is an alternative method to maximum likelihood estimation when the residuals are not normally distributed. For the special case where the calibration data set is uncensored, the AMLE method converges to MLE, resulting in a minimum variance unbiased estimate of constituent loads [26]. Since all the data sets used in this study were uncensored, model coefficients for the AMLE and MLE methods were identical and the AMLE was used to determine the model coefficients and estimate the log load. Two statistics, the Akaike Information Criterion (AIC) and the Schwarz Posterior Probability Criterion, were computed for the calibrated model [12, 27]. In this study, a regression model was developed and calibrated in estimating the TSS loads. TSS concentrations were used in conjunction with corresponding observed streamflow data to develop and calibrate the regression model using AMLE. The model with the lowest value of the AIC was then selected for use in load estimation. The regression model developed for TSS using LOADEST is shown in Table 1. AMLE results are contingent upon the assumption that model residuals are normally distributed. Once the model formulation and calibration were done, AMLE residuals were examined to see whether this assumption was valid. Checks for normality included construction of a normal probability plot, a plot of model residuals versus their Z-scores, which should yield a normal probability plot [12, 29].The linearity of the plot suggested that the residuals follow a normal distribution (Figure 3). This linearity was supported by the Probability Plot Correlation Coefficient (PPCC) of 0.97. Coefficient of determination (R 2 ) of the regression model for TSS load which represents fraction of the variance explained by regression is shown in Table 1. The relatively high R 2 value indicated that the model successfully simulated the variability in constituent loads. Table 1 - Regression model developed for TSS using LOADEST and the coefficient of determination (R 2 ). Regression model Model Coefficients (a) R 2 a0 a1 a2 a3 a4 a5 a6 Ln(L) = a0 + a1 LnQ + a2 LnQ 2 + a3 Sin(2 π T) + a4 Cos(2 π T) + a5 T + a6 T 2 10.88 (0.1) 1.69 (0.07) -0.08 (0.07) 0.03 (0.08) 0.30 (0.08) -0.02 (0.02) 0.02 (0.01) 0.85 Notes: (a) Standard deviation of the model coefficients are shown in parentheses. L is the constituent load; Q is the streamflow; R 2 is the coefficient of determination for the regression model. Relationships are considered to be significant at p < 0.05. LnQ = Ln(streamflow) - center of Ln(streamflow); T = decimal time - center of decimal time. Explanatory variables were centered to eliminate the colinearity [12, 28]. ENGINEER 46 5 ENGINEER Figure 3 - Normal probability plot for the model residuals of the AMLE regression. Notes: Model Residual [ln(L) - ln(ˆL)] is the difference between the observed and estimated values of log load L is in kilograms per day. * PPCC 4. Results & Discussion 4.1 Influence of Stream Flow on Turbidity Turbidity is an indicator of the amount of sediment and related constituents transported by a stream. According to Christensen et al. [30], turbidity and streamflow are related because streamflow can affect suspension of the sediment and related constituents. But the conditions between turbidity and streamflow affect this relation; Paustian & Beschta [31] have revealed that, first storm flow occurring after a dry period, results in higher turbidity than from subsequent larger flows due to an initial flush of suspended sediment. In Gin river, turbidity varies directly with the streamflow (Figure 4). A plot of turbidity vs. streamflow exceedance 1 2 4 8 16 32 64 128 256 4 16 64 256 T u r b i d i t y ( N T U ) Stream Flow (m 3 /s) Figure 4 - Variation in turbidity with the streamflow (2000-2009). probability shows that higher turbidity values are in the range of intermediate and higher flows (Figure 5). During all the flow regimes, turbidity levels indicated exceedance of the water quality standards set for the potable water as well the inland waters of Sri Lanka. In 92% of the samples, turbidity level exceeded 5 NTU, the maximum permissible limit for Class 1 Waters; drinking water with simple treatment, in the proposed ambient water quality standard for inland waters of Sri Lanka [32]. 8 NTU, the maximum permissible limit for Sri Lanka standards for potable water [33] has been exceeded by 70% of the samples. Sediments from catchment runoff subsequent to higher rainfall, eroding stream banks, and sand mining activities might be attributable to the high levels of turbidity. 4.2 TSS Load Estimation TSS concentration values derived from the turbidity based linear regression equation were used as the model input concentration data to develop TSS load-discharge multiple regression model using the LOADEST program. TSS mean load estimates were developed on daily and monthly basis from January 2000 to December 2009. 1.00 10.00 100.00 1000.00 0 10 20 30 40 50 60 70 80 90 100 T u r b i d i t y ( N T U ) Percentage of time Streamflow was equaled or exceeded Highest flows Lowest flows High flows Low flows Intermediate flows 8.00 Figure 5 - Turbidity vs. streamflow exceedance probability (2000-2009). Maximum permissible limit (8 NTU) for the potable water ENGINEER 47 ENGINEER 6 Model input TSS loads were plotted against the estimated TSS loads by the LOADEST to evaluate the fitness of the regression model developed (Figure 6). 1:1 line represents similar values for both model input and model estimated. Points were moderately scattered near the 1:1 line over a wide range of loads. More points clustered slightly below the 1:1 line indicated slight underestimation of the model predicted loads. 4.3 Variation in TSS with Streamflow It is important in any suspended sediment estimation study to collect samples over several years and over a wide range of streamflows to more accurately depict the sediment flux under all flow regimes [22]. Figure 7 illustrates the Figure 6 - Relationship between model input TSS loads and model estimated TSS loads. Note: ** R 2 for the regression model. Figure 7 - Flow duration curve and concurrent turbidity sampling. relation of the streamflow, at which the turbidity samples were collected during February 2000 – November 2009, to flow duration, based on the streamflow data at Baddegama station. Distribution of sample collection over such a wide range in the streamflow hydrograph, particularly at the high flows, indicates that the TSS concentration data would characterize TSS load information for most flow and turbidity conditions. TSS concentrations were compared with the permissible total solid levels, since there was no specific water quality standards developed for TSS in Sri Lanka. TSS concentrations which ranged between 2.4 mg/l and 204 mg/l were well below the maximum permissible total solid level (2000 mg/l) cited in the Sri Lanka standards for potable water [33]. Concentration of TSS in rivers increases as a function of flow. TSS concentrations have been shown to be strongly correlated with the streamflow with most of the sediment load transported during peak flow events [34, 35, 36]. The relationship between streamflow and TSS load is well established for Gin river (Figure 8). High TSS loads occur in May and October with the corresponding peak flows in the Gin river. This peak constituent load occurrence in Gin river in coinciding with the high flows is further supported by Wickramaarachchi et al. [18]. When considering annual TSS loads in Gin river, there was a Figure 8 - Mean monthly TSS load estimate and mean monthly streamflow of Gin River at Baddegama during 2000-2009. ENGINEER 48 7 ENGINEER decreasing trend (0.02%) in annual TSS loads and the trend was statistically significant at the 95% confidence level (p < 0.05). Figure 9 shows the model input and model estimated average daily TSS loads during 2000–2009. The estimated TSS loads closely resemble the model input TSS loads, particularly at the moderate flows, although some estimated loads during extreme flow conditions have some deviations. On the descending limbs of the hydrograph, model input loads are slightly overestimated by the model. As Clark [37] pointed out, the reason might be that the samples collected during the ascending limb and near the peak of the hydrograph typically contained higher concentrations of solids than did samples collected during the descending limb at the same discharge. This is because material that had accumulated in the stream prior to storm runoff becomes mobile as stream velocities rise and the concentrations measured later at the same discharge were low because the stream had been flushed of accumulated solids. However, the model does not account for this and as a result, changes in load resulting from rapid changes in streamflow may not be modeled accurately. 5. Conclusions Correlation established between the turbidity and TSS concentration in Gin river at Baddegama can be used to compute TSS concentrations beyond the period of record used in model development with proper sample collection and analysis. Moreover the understanding on turbidity and TSS characteristics in river water might be useful in exploring the potential to establish automated turbidity recording system that makes real-time sampling decisions to facilitate TSS estimation in Gin river at Baddegama. Due to the overall fitness reflected by the load- discharge regression model developed for TSS load estimation, it could be utilized in inferring the TSS loads from the flow data, during un- sampled periods. Results of the study could be used to better understand the fluctuation of turbidity and TSS under changing flow regimes and to assess quality of water in Gin river relative to the water quality standards in practice. This prevailing water quality conditions in Gin river provide necessary information for water managers and planners to adjust water treatment strategies accordingly. Acknowledgements Authors gratefully acknowledge National Water Supply and Drainage Board (Southern), Sri Lanka for providing the water quality data of Gin river and facilitating the water quality testing. Sincere appreciation is extended to University of Yamanashi, Japan and JSPS (Japan Society for Promotion of Science) for the technical and financial support for the study. Figure 9 - Relation of model input to model estimated TSS loads and corresponding streamflow during 2000-2009. ENGINEER 49 ENGINEER 8 References 1. ASTM International, D4410, Terminology for fluvial sediment, Annual book of standards, Water and Environmental Technology: West Conshohocken, Pennsylvania, 2007, 7 p. 2. Thorp, J. H., Thomas, M. C., & Delong, M. D., “The Riverine Ecosystem Synthesis: Biocomplexity in River Networks Across Space and Time”, River Res. Appl., Vol. 22, No. 2, 2006, pp. 123–147. 3. Leopold, L. B., Maddock, T., The hydraulic geometry of stream channels and some physiographic implications, U.S. Geological Survey Professional Paper 252, 1953, 57 p. 4. Susfalk, R. B., Fitzgerald, B., & Knust, A. M., Characterization of Turbidity and Total Suspended Solids in the Upper Carson River, Nevada, Desert Research Institute (DRI), DHS Publication No. 41242, 2008. 5. Jastram, J. D., Moyer, D. L., & Hyer, K. E., A Comparison of Turbidity-based and Streamflow- Based Estimates of Suspended-Sediment Concentrations in three Chesapeake Bay tributaries, U.S. Geological Survey Scientific Investigations Report 2009–5165, 2009, 37 p. 6. Gippel, C. J., “The Use of Turbidity Instruments to Measure Stream Water Suspended Sediments Concentration”, Department of Geography and Oceanography, University College, Australian Defense Force Academy. Monograph Series No. 4, 1989. 7. Lewis, J., “Turbidity-Controlled Suspended Sediment Sampling for Runoff-event Load Estimation”, Water Resour. Res., Vol. 32, No. 7, 1996, pp. 2299-2310. 8. Christensen, V. G., Xiaodong, J., & Ziegler, A. C., Regression analysis and real-time water-quality monitoring to estimate constituent concentrations, loads, and yields in the Little Arkansas River, south- central Kansas, 1995-99, U.S. Geological Survey Water-Resources Investigations, 2000. 9. Smolen, K. D., “Trout Creek Stream Restoration and Wildlife Enhancement Project: Water Quality Monitoring”, M.S. Thesis, Graduate Program of Hydrologic Sciences, University of Nevada, Reno, 2004, 134 p. 10. Daphne, L. H. X., Utomo, H. D., & Kenneth, L.Z.H., “Correlation between Turbidity and Total Suspended Solids in Singapore Rivers”, Journal of Water Sustainability, Vol. 1, No. 3, 2011, pp. 313–322. 11. Wickramaarachchi, T. N., Ishidaira, H., & Wijayaratna, T. M. N., “Projecting Land use Transitions in the Gin Catchment, Sri Lanka”, Res. J. Environ. Earth Sci., Maxwell publishers, in press (Accepted for publication June 2013). 12. Runkel, R. L., Crawford, C. G., & Cohn, T. A., Load Estimator (LOADEST): A FORTRAN program for estimating constituent loads in streams and rivers, U.S. Geological Survey Techniques and Methods, book 4, chap. A5, 2004, 69 p. 13. Hooper, R. P., Aulenbach, B. T., & Kelly, V. J., “The National Stream Quality Accounting Network: A Flux-Based Approach to Monitoring the Water Quality of Large Rivers”, Hydrol. Process., Vol. 15, No. 7, 2001, pp. 1089–1106. 14. Aulenbach B. T., Annual Dissolved Nitrate Plus Nitrite and Total Phosphorus Loads for the Susquehanna, St. Lawrence, Mississippi–Atchafalaya and Columbia River Basins for 1968 to 2004, Open File Report 2006-1087, US Geological Survey, Reston, VA, 2006. 15. Morrison, J., Colombo, M. J., Surface-water quality and nutrient loads in the Nepaug Reservoir watershed, northwestern Connecticut, 1999-2001, U.S. Geological Survey Scientific Investigations Report 2006-5272, 2006, 36p. 16. Foster, K., Kenney, T. A., Dissolved-solids load in Henrys Fork upstream from the confluence with Antelope Wash, Wyoming, water years 1970–2009, U.S. Geological Survey Scientific Investigations Report 2010–5048, 2010, 16p. 17. Stenback, G. A., Crumpton, W. G., Schilling, K. E., & Helmers, M. J., “Rating Curve Estimation of Nutrient Loads in Iowa Rivers”, J. Hydrol., Vol. 396, No. 1-2, 2011, pp. 158–169. 18. Wickramaarachchi, T. N., Ishidaira, H., & Wijayaratna, T. M. N., “Variation of Constituent Loads and Concentrations with the Flow in Gin River, Sri Lanka”, J. Natl. Sci. Found. Sri Lanka, in press (Accepted for publication June 2013). 19. National Atlas, 2 nd ed., Survey Department of Sri Lanka, 2011. 20. Packman, J. J., Comings, K. J., & Booth, D. B., “Using Turbidity to Determine Total Suspended Solids in Urbanizing Streams in the Puget Lowlands”, Confronting Uncertainty: Managing Change in Water Resources and the Environment, Canadian Water Resources Association annual meeting, Vancouver, BC, 27–29 October 1999, 1999, pp. 158–165. 21. Gilvear, D. J., Petts, G.E., “Turbidity and Suspended Solids Variations Downstream of a Regulating Reservoir”, Earth Surf. Process. Landf., Vol. 10, No. 4, 1985, pp. 363–373. 22. Uhrich, M. A., Bragg, H. M., Monitoring instream Turbidity to Estimate Continuous Suspended- Sediment loads and yields and clay-water volumes in ENGINEER 50 9 ENGINEER the Upper North Santiam River Basin, Oregon, 1998–2000, U.S. Geological Survey Water- Resources Investigations Report 03–4098, 2003, 43 p. 23. Rasmussen, P. P., Gray, J. R., Glysson, G. D., & Ziegler, A.C., Guidelines and procedures for computing time-series suspended-sediment concentrations and loads from in-stream turbidity- sensor and streamflow data, U.S. Geological Survey Techniques and Methods, book 3, chap. C4, 2009, 52 p. 24. Ellison, C. A., Kiesling, R. L., & Fallon, J. D., “Correlating Streamflow, Turbidity and Suspended Sediment Concentration in Minnesota’s Wild Rice river”, 2nd Joint Federal Interagency Conference, Las Vegas, NV, June 27 - July 1, 2010. 25. Standard Methods for the Examination of Water and Wastewater, 21 st ed., Eds: Eaton, A. D., Clesceri L. S., Rice E. W. & Greenberg A. E., Washington D.C., published jointly by the American Public Health Association, the American Water Works Association and the Water Environment Federation, 2005. 26. Cohn, T. A., Gilroy, E. J., & Baier, W. G., “Estimating Fluvial Transport of Trace Constituents using a Regression Model with Data Subject to Censoring”, Proceedings of the Joint Statistical Meeting, Boston, 9-13 August, 1992, pp. 142–151. 27. Judge, G. G., Hill, R. C., Griffiths, W. E., Lutkepohl, H., & Lee, T. C., Introduction to the theory and practice of econometrics, 2 nd ed.,John Wiley, New York, 1988, pp. 1024. 28. Cohn, T. A., Caulder, D. L., Gilroy, E. J., Zynjuk, L. D., & Summers, R. M, “The Validity of a Simple Statistical Model for Estimating Fluvial Constituent Loads - An Empirical Study Involving Nutrient Loads Entering Chesapeake Bay”, Water Resour. Res., Vol. 28, No. 09, 1992, pp. 2353–2363. 29. Helsel, D. R., Hirsch, R. M., Statistical methods in water resources, U.S. Geological Survey Techniques of Water-Resources Investigations, book 4, chapt. A3, 2002. 30. Christensen, V. G., Ziegler A. C., & Jian X., “Continuous Turbidity Monitoring and Regression Analysis to Estimate Total Suspended Solids and Fecal Coliform Bacteria Loads in Real Time”, Proceedings of the Seventh Federal Interagency Sedimentation Conference, March 25–29, 2001, Reno, NV. Subcommittee on Sedimentation, vol 1, 2001, pp III-94 to III-101. 31. Paustian, S. J., Beschta, R. L., “The Suspended Sediment Regime of an Oregon Coast Range Stream”, Water Resour. Bull., Vol. 15, 1979, pp. 144–154. 32. Proposed ambient water quality standards for inland waters of Sri Lanka, Central Environmental Authority, Sri Lanka, 2001. 33. Sri Lanka Standards for potable water – SLS 614, Sri Lanka Standards Institute, 1983. 34. Leopold, L. B., Wolman M. G., & Miller, J. P., Fluvial processes in geomorphology, San Francisco, CA, Freeman, 1964. 35. Allan, J. D., Stream ecology: Structure and function of running waters. London: Chapman and Hall, 1995. 36. Doyle, M. W., Stanley, E. H., Strayer, D. L., Jacobson, R. B., & Schmidt, J. C., “Effective Discharge Analysis of Ecological Processes in Streams”, Water Resour. Res., Vol. 41(W11411), 2011, 16 p. 37. Clark G. M., Occurrence and transport of cadmium, lead, and zinc in the Spokane River Basin, Idaho and Washington, water years 1999-2001, U.S. Geological Survey Water-Resources Investigations Report 02-4183, 2003, pp. 37. ENGINEER 51 Peak Electricity Demand Prediction Model for Sri Lanka Power System G.V.Buddhika De Silva and Lalith A. Samaliarachchi Abstract: Accurate prediction of daily peak electricity demand is a requirement for service reliability, system stability and operating performance of a power system in the field of electrical engineering. This has now become a very important factor for Sri Lanka power system, since the available power plants are to be dispatched in an economical and reliable manner especially during the peak demand period of the chronological load profile. Therefore the prediction of next day peak electricity demand to an acceptable accuracy is useful for the system control centre (SCC) of the Ceylon Electricity Board (CEB). However, presently the unit commitment to meet the next day peak electricity demand is being mostly done by the system control engineers based on their past experience in the field of operation with respect to the day, period and other factors. This research paper carefully identifies sensitive elements which affect the daily peak demand of Sri Lanka power system and develop two forecasting models, namely linear statistical “Multiple Regression” and feed forward “Artificial Neural Network”. Both models were developed and fine-tuned using recorded peak demands of Sri Lanka power system from year 2008 to 2011 taken from the SCC of CEB and tested for the calendar year 2012 and also for the first few months of 2013. Artificial Neural Network model was found to be the best fit model for the prediction of daily peak demand of Sri Lanka power system with the lowest Mean Absolute Percentage Error (MAPE). Keywords: Artificial Neural Networks, Multiple Regression Analysis, peak demand 1. Introduction 1.1 Objective This paper discusses step by step development of a best fit forecasting model for the prediction of next day peak electricity demand of Sri Lanka power system. Such a prediction model will not only help the System Control Engineers to prepare optimum allocation of generation, unit commitment and hydro/thermal coordination but also for the preparation of short term planned outages for scheduled maintenance of power plants eventually contributing to the system to maintain its stability and reliability. 1.2 Background One of the major and also a difficult task of a system control engineer is to prepare a plan showing the next day’s expected peak electricity demand and its associated optimum generation dispatch. Presently, this is being done based on past experience with respect to various factors by system control engineers and it involves lots of risks. Therefore the prediction of next day peak electricity demand to certain accuracy would help the system control centre to ride through the critical period without facing many difficulties. 1.3 Features It has been noticed that the demand for electricity in Sri Lanka grows steadily and continuously specially after the 30 year long civil war. Social life patterns and day to day activities of Sri Lanka exhibit two peak demand periods in chronological daily load profile [6] as shown in figure 1. One is in the day time and the other during night time. Night peak demand is the highest and control engineers are very much concerned about. Being a diverse country, home to many religions and ethnicities, peak electricity demand in Sri Lanka is mostly affected by official/non official holidays and also by religious and cultural events. Sri Lankan government offers 25 formal annual holidays and could be divided into two main groups such as P.B.M [Public, Bank and Mercantile] and PB [Public and Bank]. Peak electricity demand is observed to be different on a monthly Buddhist holiday called “Poya day”. Sinhalese and Tamil Cultural/Festival seasons, Eng. G.V.Buddhika De Silva, B. Tech. Eng. (Hons), Open University of Sri Lanka (OUSL), AMIE (Sri Lanka) Eng. Lalith A. Samaliarachchi, B. Sc. Eng. (Hons) (Moratuwa), M. Eng.(AIT), MIET(UK), MIE(Sri Lanka), C.Eng Senior Lecturer, Department of Electrical & Computer Engineering, OUSL. ENGINEER - Vol. XXXXVI, No. 04, pp. [53-60], 2013 ©The Institution of Engineers, Sri Lanka ENGINEER 53 ENGINEER 2 Figure 1 - Chronological Daily Load Profile identified as “New Year Season” during April and Buddhist religious activities during “Wesak” week in May show a unique but unsteady peak demand pattern and needed to be treated separately from the regular days. Peak demand value variation owing to the temperature and humidity changes and the impact owing to night time cricket matches such as ODIs & T20s during peak demand hours are noted and also needed to be considered. 1.4 Forecasting Two mathematical forecasting techniques are considered in this study and they are: Multiple Regression analysis (MRA) Artificial Neural Networks (ANN) Basically MRA is a linear statistical model traditionally used for forecasting and in here it is used to identify the relationships between the peak demand and its dependent variables. Statistical significance of each dependent variable has to be considered here. The aim of using ANN is to model any non linear complex relationships, if exists, between the variables that cannot be identified with traditional linear models. ANN has to be trained first and checked its ability for generalization [lowest output errors for the inputs which have not seen before]. Since most of the forecasting techniques are lagged behind their theoretical formulation and verification in the field of operation, study compares and identifies the best fit model for the prediction of peak electricity demand. 2. Literature Survey It has been observed that the wide variety of forecasting techniques are being used for the prediction of short term electricity demand, such as Similar Day approach, Time series methods (Autoregressive Moving Average ARMA), Autoregressive Integrated Moving Average (ARIMA), autoregressive integrated moving average with exogenous variables (ARIMAX), Multiple Regression Analysis, Artificial Neural Networks, Expert Systems, Fuzzy logic, Support vector machines [4] and Exponential smoothing methods [9]. ANN has been used and tested for the hourly demand prediction model developed for the Western part of Saudi Arabia [7]. Nigerian Electrical Power System [8] and South Sulewesi’s (Sulewesi Island–Indonesia) Power System have used multiple regression analysis [5] for the development of a demand prediction model. However, it was noted that very few publications have apparently verified the goodness of fitness of the developed models against actual values. 3. Methodology 3.1 Multiple Regression Analysis In here, the dependent variable peak demand (P) can be expressed as a function of many independent variables P = a1 + a2 × Xa + a3 × Xb + a4 × Xc + a5 × Xd + a6 × Xe + Error SPSS 16.0 (Statistical Package for the Social Sciences) software package is used to do MRA throughout the study. Embedded F & T tests in SPSS compares the sample data with the main distribution by making hypothesis and explore whether the regression results are owing to the random effect. The four basic assumptions to be fulfilled before the correctness of the results of MRA is verified are Linearity, Independence of residuals, Normality of residuals and Homoscedasticity [1] & [12]. An observation that is numerically distant from the rest of the data called “outlier” can lead regression estimates to be inaccurate and limit the ability to understand data. In this study outliers were identified by the scatter plot between residuals Vs predicted values. Common thresholds are |2.0| or |3.0| times the standard deviations [11] from error mean. Day time peak Night time peak ENGINEER 54 To have an efficient regression results, those outliers have to be removed from the data and analysis should be redone until all the standardized residuals are within the acceptable limits. The block diagram for the algorithm showing how the MRA model is developed and fine- tuned is given in Appendix A1, figure A1. network is a feed forward ANN, signals travel only in one direction. Output of one layer will be the input for the following layer. In this paper, two layer ANN architecture i.e. with one hidden layer and one out put layer were used. Figure 4 exhibits a two layer feed forward ANN structure. 3.2 Artificial Neural Network Neural networks are adaptive statistical models based on an analogy of the structure of human brain. It can be fine-tuned to estimate the parameters of the population utilizing limited number of exemplars (one or few) at a time. ANN maps an input space in to the desired output space by adjusting the connecting weights and biases (called learning). There are two methods of learning; “Supervised learning and unsupervised learning”. Supervised learning method is used in this paper analogue to the student guided by the teacher as shown in figure 2. Figure 2- Supervised Learning Method A single neuron model shown in figure 3 has multiple inputs (X) and a single output (Y). Each input is modified and multiplied by a weight (W). Weighted inputs are combined at a Summing point. Output will be determined with reference to the activation function (f) and threshold value (θ). Figure 3 - Single Neuron Model Neurons are organized to make up the neural network structures and the most concerned structure is the multilayer network. When the Figure 4 – Two Layer Feed Forward ANN For the preparation ANN model, two ANN architectures were treated and developed using three activation functions “Tan-sigmoid”, “log- sigmoid” and “pure linear”[2, 3 &10]. Deciding the number of hidden neurons in layers is an important part while developing the overall ANN architecture. Two traditional approaches were used in this study to find out the number of neurons in hidden layers and they are: Try different configurations, see what works best Rule of thumb methods. Latter provides a starting point to consider. 3.3 Data Pre-Processing While “Statistical or Z- Score Normalization” technique is used to normalize the inputs and targets to improve the training process, principal component analysis is used to reduce the dimensions of the inputs which have high correlations between their components[3 &10]. 3.4 Training of Artificial Neural Network Training of the ANN architecture needs to be done by using four different training methods [10] and could be stated as: trainbr - Automated Regularization trainlm- Reduced Memory Levenberg- Marquardt traingdx- Variable Learning Rate Back propagation trainrp - Resilient Back propagation ENGINEER 55 ENGINEER 4 The block diagram for the algorithm showing how the ANN model is prepared and fine-tuned is given in Appendix A2, figure A1 5.2 Results of ANN The best ANN architecture model giving lowest MAPE for the daily peak demand prediction of year 2012 is given in table 2. 4 Data Input Observed peak demand data and its associated variables identified in 1.3 from calendar year 2008 to 2011 are carefully sorted out, treated and analysed for both forecasting models. Table 1 – Input variables Numerical indexes are given to represent the data input vector as depicted in table 1 to get the output for forecasted day peak. Daily Temperature and Relative humidity values were taken as the mean hourly values recorded between 4.40 p.m to 8.40 p.m at Katunayaka [13]. 5. Results 5.1 Results of MRA Multiple regression Analysis is carried out for both linear and exponential models. Exponential model of MRA is shown to be the best fit and seven equations were totally derived to represent the final forecast, one per each day. The Results of MRA model applied to the calendar year 2012 is shown in Appendix A1, figure A2. A significant deviation between the actual [14] and the forecast peak value has been observed for 64 out of 365 days owing to the unprecedented power cuts and other inherent features such as peak power saving strategies imposed by CEB from time to time. However, results of MRA revealed a Mean Absolute Percentage Error (MAPE) of 4.72% for 301 out of 365 consistent data. Table2–The Best Neural Network Architecture The results of ANN model is shown in Appendix A2, figure A2. The deviation between the actual and the forecasted peak cannot be overemphasized here again owing to the reasons pointed out in 4.1. However, results of ANN revealed a MAPE of 2.50% for same number of days considered in year 2012. 6. Conclusions Two forecasting models are developed and presented in this paper for the prediction of next day peak electricity demand of Sri Lanka power system. Peak demand of Sri Lanka is mostly affected by features such as day of the week, temperature, Relative Humidity, PB or a PBM and full moon day. Special features such as whether the day is falling on a New Year week or a Wesak week has also shown some effect on the system night peak demand. Unprecedented events like ODIs and T20s has also shown some effect on the peak demand and is interestingly noted. The above features are carefully treated and analysed using widely used load prediction models such as traditional MRA statistical model and an alternative model which does not rely on human experience i.e. ANN. The Mean Absolute Percentage Error while predicting peak demand for the calendar year 2012 using MRA is found to be 4.72%where as using ANN is 2.50%. Since ANN is giving the lowest MAPE, it can be concluded that the Artificial Neural network model is the best fit model that can be used to forecast the Daily peak demands of Sri Lanka. Since the process is an adaptive method of forecasting procedure, it has to be updated at the end of each year. Such an updated exercise carried out Independent Variables Data Input Year 2000- 0...., 2012-12 Week 1-52 Day Mon.=1...Sun.=7 Temperature Forecast temp. Relative Humidity Forecast RH P.B. holiday 0...1.. P.B.M holiday 0...1.. Poya day 0...1.. New year week 0...1.. Wesak Week 0...1.. ODI or T20 0...1.. Description ANN structure No of hidden layers 1 No of Hidden neurons 17 Activation function of hidden layer Tan– Sigmoid Activation function of out put layer Pure linear No of neurons in the output layer 1 Training function traingdx ENGINEER 56 for the first four months of the year 2013 is shown in figure 5. Developed ANN model for the year 2013 gives the Mean Absolute Percentage Error of 2.08% for 110 out of 120 consistent data in the period of 01/01/2013 to 30/04/2013. 3. Eugene A. Feinberg & Dora Genethliou , “Load forecasting”, Applied Mathematics for Restructured 4. Eugene A. Feinberg & Dora Genethliou , “Load forecasting”, Applied Mathematics for Restructured Electric Power Systems: Electric Power Systems: Optimization, Control, and Computational Intelligence (J. H. Chow, F.F. Wu, and J.J. Momoh, eds.), Spinger, USA, 2005, pp. 269-285 Figure 5 – Plot of Actual Vs Predicted values for the period of 01/01/2013 to 30/04/2013 7. Acknowledgement The authors would like to offer his appreciation and thanks to Eng. J. Nanthakumar, CEB System Control Centre for his assistance given during data collection, many helpful comments and suggestions. Also the guidance provided by Dr. Narendra de Silva of LECO, Dr. (Mrs) D.D.M. Ranasinghe and Dr. (Mrs) H.U.W. Rathnayake of OUSL is greatly appreciated. References 1. Spyros makridakis and Steven C. Wheelwright , Forecasting methods and Applications, Jhon Wiley & Sons,USA,1978. 2. Elaine Rich, Kevin Knight, Shivashankar B Nair, Artificial Intelligence , 3 rd ed , The McGraw- Hill Companies, New Delhi , India, 2009. 3. Kevin L. Priddy and Paul E. Keller, Artificial Neural Networks an Introduction, 1 st ed, Prentice- Hall of India Private Limited, New Delhi ,India, 2007. 5. Amral N., Özveren C. S., King D., “Short Term Load Forecasting using Multiple Linear Regression”, Proc of UPEC , 42 nd international, September, 2007, pp 1192- 1198. 6. Tilakasena D. K Bandula, “Country Presentation Sri Lanka”, Presentation of Technical Session-I: (USAID SARI/ Energy) Regional Clean Coal Partnership Programme, Kolkata, INDIA, September 16-19, 2008, p. 7. 7. Al-Shareef A. J., Mohamed E. A. and E.Al- Judaibi, “Next 24-Hours Load Forecasting using Artificial Neural Network (ANN) for the Western Area of Saudi Arabia”, J. Faculty of Eng.Sci, King Abdulaziz University (KAU) , Vol.19, No.2, 2008, pp. 25-40. 8. Adepoju G. A., Ogunjuyigbe S. O. A., and Alawode K. O., “Application of Neural Network to Load Forecasting in Nigerian Electrical Power System”, The Pacific journal of Science and Technology, Akamai University, Vol. 8, No 1, May, 2007. ENGINEER 57 ENGINEER 6 9 Alexandra KOTILLOVÁ, “Very Short- Term Load Forecasting using Exponential Smoothing and ARIMA models”, Journal of information, Control and Management Systems, Vol. 9, No. 2, 2011. 10. Howard Demuth, Mark Beale, “Neural Network Toolbox User’s Guide”, 6 th ed, Version 4 (Release 12), Copyright by The MathWorks, inc., September, 2000, chap. 2, 3, 5. 11. http://www.slideserve.com / Rita / regression - outliers, Visited, 13 th June 2011. 12. http://www.duke.edu / ~rnau / testing.htm Visited, 20 th June 2011. 13. http://www.wunderground.com, Visited and recorded the weather data of each day from 01/01/2008 to 30/04/2013. 14. http://www.ceb.lk , Visited and recorded the peak demand of each day from 01/01/2012 to 30/04/2013. ENGINEER 58 Appendix A1 Method of selecting best Multiple Regression Model Figure A1 – Algorithm for the development of best MRA model Figure A2 – Actual Vs Predicted values for year 2012 by MRA model Best Daily regression model Best Multiple Regression Model (with lowest M.A.P.E ) Daily Models Data divided as days {52 demands per year}. Seven models were treated giving one model per each day Data sets were taken as; 2008 to 2011, 2009 to 2011 & 2010 to 2011 Yearly models Data from year 2010 to 2011 were used. Since the war is ended in the middle of year 2009. Linear & Exponential Multiple Regression analysis Best yearly regression model Data Categorization Past electricity peak demands from 2008 to 2011 ENGINEER 59 ENGINEER 8 Appendix A2 Method of selecting best neural network structure Figure A1 - Algorithm for the development of best ANN Structure Figure A2 – Actual Vs Predicted values for year 2012 by ANN model Best 1 st ANN architecture Lowest M.A.P.E given architecture from 1 st ANN architectures Best Artificial Neural Network model Training & Testing data Training & Testing data 1 st ANN architecture Try different configurations, see what works Tan- Sigmoid (hidden layer) Pure linear (output layer) 2 nd ANN architecture Rule of thumb methods Tan- Sigmoid (hidden layer) Pure linear (output layer) Data pre-processing, Training and Testing of ANN architectures Best 2 nd ANN architecture Lowest M.A.P.E given architecture from 2 nd ANN architectures Past electricity peak demands from 2008 to 2011 ENGINEER 60 SECTION II 1 ENGINEER Floating Wetlands for Management of Algal Washout from Waste Stabilization Pond Effluent: Case Study at Hikkaduwa Waste Stabilization Ponds Sujatha Kalubowila, Mahesh Jayaweera, Chandrika M. Nanayakkara and Dhanesh N. De S. Gunatilleke Abstract: Waste stabilization ponds are advantageous wastewater treatment processes, especially for developing countries. Nevertheless, in spite of the well known advantages of the implementation of the stabilization pond system, the effluent of this system has a significant amount of algae and high nutrients. Disposing this effluent with high contents algae and nutrients to the receiving waters can hinder the water reuse for a wide range of different applications, it is essential to look for a post treatment method that can provide considerable removal of algae, nutrients and organic matter from the effluent and at the same time, assure that the treatment system as a whole will maintain the advantages of the pond treatment processes. In this context, this research study was planned and intended to introduce a floating treatment wetland in which water hyacinth plants (Eichhornia crassipes) were used as macrophyte or vegetation in the part of the maturation pond area to control algae and nutrients in the effluent. With the application of the floating wetland the removal efficiencies were found to have increased in the maturation pond in terms of BOD and COD from 13.3% to 62.9% and 13.6% to 57.5%, respectively. In the case of TP and TN there were no significant reductions achieved prior to the establishment of the wetland but, reductions of 74.8% for TP and 55.8% for TN were achieved since the establishment of floating wetland. It was also possible to achieve a reduction of algal cell densities of 900 units/ml to zero unit/ml for the algal species of Spirulina and for Oscillatoria, the reduction was from 290 units/ml to 0 units/ml. In case of Chlorella and Pandorina, density reductions were 830,000 units/ml to 68,000 units/ml and 4300 units/ml to 280 units/ml respectively. Accordingly, the reduction efficiencies for Spirulina, Oscilltoria, Chlorella and Pandorina were reported to be improved from 31.8% to 100% and 4.5% to 100%, 34.2% to 91.8% and 42.2% to 93.5%, respectively. Application of this research can therefore be possible to polish waste stabilization pond effluent economically in order to re-use for various beneficial uses except potable use. Key words: Algae; Nutrients; Macrophyte; Wetland 1. Introduction Waste Stabilization Ponds (WSP) are large, shallow basins in which raw sewage is treated entirely by natural processes involving mainly both algae and bacteria. They are used for sewage treatment in tropical climates, and represent one of the most cost-effective, reliable and easily-operated methods for treating domestic wastewater. WSPs are very effective in the removal of faecal coliform bacteria. Sunlight energy is the only requirement for its operation. Further, it requires minimum supervision for daily operation, in terms of simple cleaning of the outlets and inlet works. The temperature and duration of sunlight in tropical countries offer an excellent opportunity for high efficiency and satisfactory performance for this type of water-cleaning system. They are well-suited for low-income tropical countries where conventional wastewater treatment cannot be achieved due to the lack of a reliable energy source. Further, the advantage of these systems, in terms of removal of pathogens, is one of the most important reasons for its use. WSP systems comprise a single string of anaerobic, facultative and maturation ponds in series, or several such series in parallel. Eng. (Ms) Sujatha Kalubowila, , C. Eng., MIE(Sri Lanka), MSc Env Eng & Mgt (Moratuwa). Civil Engineer, Water Supply & Drainage Board. Eng.(Dr.) Mahesh Jayaweera, B.Sc. Eng.(Hons) (Moratuwa), PhD. (Japan), CEng., MIE(Sri Lanka), Senior Lecture, Department of Civil Engineering , University of Moratuwa, Sri Lanka. Dr (Mrs.) Chandrika M. Nanayakkara, B.Sc. (Hons)(Colombo), M.Sc. (Kelaniya), PhD (Aberdeen), Senior Lecturer, Department of Plant Sciences, University of Colombo, Sri Lanka. Eng. Dhanesh N. De S. Gunatilleke B.Sc. Eng. (Hons) (Moratuwa), P.G. Dip (Moratuwa), MSc Sanitary Engineering (UNESCO-IHE, Delft), Diploma (Lund University), FIE(Sri Lanka), MCIWEM (UK), Int PE, CEng, Specialist (Sewerage Designs), National Water Supply &Drainage Board, Sri Lanka. ENGINEER - Vol. XXXXVI, No. 04, pp. [63-74], 2013 ©The Institution of Engineers, Sri Lanka ENGINEER 63 ENGINEER 2 In essence, anaerobic and facultative ponds are designed for removal of Biochemical Oxygen Demand (BOD), and maturation ponds for pathogen removal, although a considerable BOD removal also occurs in maturation ponds and some pathogen removal in anaerobic and facultative ponds. In most cases, only anaerobic and facultative ponds will be needed for BOD removal when the effluent is to be used for restricted crop irrigation and fish pond fertilization, as well as when weak sewage is to be treated prior to its discharge to surface waters. Maturation ponds are only required when the effluent is to be used for unrestricted irrigation, thereby having to comply with the WHO guideline of 1000 faecal coliform bacteria /100 ml. The WSP does not require mechanical mixing, needing only sunlight to supply most of its oxygenation. Its performance may be measured in terms of its removal of BOD and faecal coliform bacteria. 1.1 Processes in Stabilization ponds Anaerobic ponds are commonly 2-5 m deep and receive wastewater with high organic loads (i.e., usually greater than 100 g BOD/m3.day equivalent to more than 3000 kg/ha.day for a depth of 3 m). They normally do not contain dissolved oxygen or algae. In anaerobic ponds, BOD removal is achieved by sedimentation of solids and subsequent anaerobic digestion in the resulting sludge. The process of anaerobic digestion is more intense at temperatures above 150C. The anaerobic bacteria are usually sensitive to pH less than 6.2. Thus, Acidic wastewater must be neutralized prior to its treatment in anaerobic ponds. A properly- designed anaerobic pond will achieve about a 40% removal of BOD at 10 0 C, and more than 60% at 200C. A shorter retention time of 1.0-1.5 days is commonly used [5]. Facultative ponds (1-2 m deep) are of two types: primary facultative ponds that receive raw wastewater, and secondary facultative ponds that receive particle-free wastewater (usually from anaerobic ponds, septic tanks, primary facultative ponds, and shallow sewerage systems). The process of oxidation of organic matter by aerobic bacteria is usually dominant in primary facultative ponds or secondary facultative ponds. The processes in anaerobic and secondary facultative ponds occur simultaneously in primary facultative ponds. It is estimated that about 30% of the influent BOD leaves the primary facultative pond in the form of methane [9]. A high proportion of the BOD that does not leave the pond as methane ends up in algae. This process requires more time, more land area, and possibly 2-3 weeks of water retention time, rather than 2-3 days in the anaerobic pond. In the secondary facultative pond and the upper layers of primary facultative ponds, sewage BOD is converted into “Algal BOD,” and has implications for effluent quality requirements. About 70-90% of the BOD of the final effluent from a series of well-designed WSPs are related to the algae they contain. In secondary facultative ponds that receive particle-free sewage (anaerobic effluent), the remaining non-settlable BOD is oxidized by heterotrophic bacteria such as (Pseudomonas, Flavobacterium, Archromobacter and Alcaligenes spp.). The oxygen required for oxidation of BOD is obtained from the photosynthetic activity of the micro-algae that grow naturally and profusely in facultative ponds. Facultative ponds are designed for BOD removal on the basis of a relatively low surface loading (100-400 kg BOD/ha.day), in order to allow for the development of a healthy algal population, since the oxygen for BOD removal by the pond bacteria is generated primarily via algal photosynthesis. The facultative pond relies on naturally-growing algae. The facultative ponds are usually dark-green in colour because of the algae they contain. The algal concentration in the pond depends on nutrient loading, temperature and sunlight, but is usually in the range of 500-2000 μg chlorophyll-a/liter [5]. The maturation ponds, usually 1-1.5 m deep, receive the effluent from the facultative ponds. Their primary function is to remove excreted pathogens. Although maturation ponds achieve only a small degree of BOD removal, their contribution to nutrient removal can also be significant. Maturation ponds usually show less vertical biological and physicochemical stratification, and are well-oxygenated throughout the day. The algal population in maturation ponds is much more diverse than that of the facultative ponds, with non-motile genera predominantly be more common. The algal diversity generally increases from pond to pond along the series [5]. ENGINEER 64 3 ENGINEER Time and temperature are the two principal parameters used in designing maturation ponds. Faecal bacteria die-off in ponds increases with both time and temperature. High pH values (above 9) occur in ponds, due to rapid photosynthesis by pond algae, which consumes CO2 faster than that can be replaced by bacterial respiration. The resulting CO2 is fixed by the algae, and the hydroxyl ions accumulate, often raising the pH to values even above 10. Faecal bacteria (with the notable exception of Vibro cholera) die very quickly at pH values higher than 9 [13]. The role of high light intensity and high dissolved oxygen concentration has recently been elucidated in the literature. Light wavelengths between 425-700 nm can damage faecal bacteria from being absorbed by the humic substances ubiquitous in wastewater. They remain in an excited state sufficiently long to damage the cell. Light-mediated die-off is completely dependent on the presence of oxygen, as well as being enhanced at high pH values. Thus, the sun plays a threefold role in directly promoting the faecal bacterial removal in WSP by increasing the pond temperature, and more indirectly by providing the energy for rapid algal photosynthesis. This not only raises the pond pH value above 9, but also results in high dissolved oxygen concentrations, which are necessary for its third role; namely, promoting photo-oxidative damage. WSPs are widely used as natural treatment systems because of their low cost and simplicity of construction, operation, and maintenance. However, the major operational problem encountered in WSPs is the excessive discharge of particles in the effluent caused by algal activity. The algal cells are produced extensively in facultative ponds and flow progressively to the maturation ponds and are ultimately discharged into inland surface water bodies such as lakes and rivers. These effluents contribute to eutrophication and eventually leading to loss of water resources. Therefore, it is essential to polish the effluent from the WSPs by removing over-discharged suspended solids (SS), BOD, and nutrients. An effective method to separate algae and other particles from the effluent of WSPs is the use of floating wetland with water hyacinth plants (Eichhornia crassipes). Water hyacinths can remove particles through sedimentation and filtration due to their dense root system. The leaves and stems also help to control algal growth by limiting the sunlight from reaching the water surface. Beyond their ability to remove suspended matter from the wastewater, several researchers in their studies, have recognized the water hyacinths’ role as an additional treatment step to reducing organic matter and nutrients from an effluent stream. This study examines the potential of developing and applying a novel “floating wetland” concept for the provision of enhanced effluent polishing particularly with regards to prevention of escaping algal cells from WSP effluent. This concept was therefore tested at the existing sewage treatment plant (STP) in Hikkaduwa of Southern Province of Sri Lanka. It is about 100 km away from Colombo towards the south of Sri Lanka. The sewage treatment method adopted in this STP has been the WSP system. 2. Literature Review The term “waste stabilization pond” in its simplest form is applied to a body of water, artificially or naturally employed with the intention of retaining sewage or organic wastewater until they are rendered pollution free and inoffensive status for discharge into receiving waters or on land, mainly relying on physical, chemical and biological processes commonly referred to as “self purification” involving the actions of algae and bacteria under the influence of sunlight (photosynthesis) and air. Organic matter contained in the wastewater is stabilized and converted into more stable matter by algal cells simply providing oxygen required for mineralization which thrive in large numbers and find their way with the effluent and hence the term “stabilization pond” [5]. [8] defines WSP as shallow basins into which wastewater continuously flows and from which treated effluent is discharged. [3] explains that the degree of treatment is a function of the number of ponds in series and the retention time of the wastewater in each pond. Although the number of ponds and retention time have a major effect on the quality of the effluent, it is possible to manipulate each individual pond to achieve a desired function [4]. A wide range of pond types therefore exists, allowing flexibility in different configurations to suit different conditions and discharge standards [13]. ENGINEER 65 ENGINEER 4 According to [13] correctly designed ponds can match the effluent quality that could otherwise be achieved by other conventional wastewater treatment technologies. 2.1 Specialized Pond Types While the above-mentioned three pond types are the most common WSPs in use, there are, however, specialized ponds that are sometimes used for wastewater treatment. These include high rate algal ponds (HRAP) and macrophyte ponds. HRAP is a shallow, paddle wheel mixed pond, which designed to enhance exposure of the algae to sunlight and avoid thermal stratification, thereby maximizing growth, photosynthesis and productivity [11], [12], [3]. This results in surplus dissolved oxygen, high pH as well as a high rate of carbon assimilation and nutrient uptake [11]. All of these contribute to the efficacy of HRAP as a combined secondary and tertiary treatment operation. The function and performance of HRAP will be examined in more detail throughout this study. Macrophyte ponds remove suspended algae and thus Macrophyte ponds are ponds where aquatic plants are grown, either on the pond surface (free floating macrophyte) such as water hyacinth, submerged such as hydrilla (Hydrilla spp.) or attached to the bottom (emergent) such as common reed (Phragmites australis) to further reduce BOD and in WSP effluent. Faecal coliform removal is however, negligible [3]. 2.2 Roles of Water hyacinth and their roots in effluent polishing Performance of WSPs depends on the effective use of bacteria for degradation of organic matters, efficient use of algae for maintaining an adequate level of oxygen in the system and especially, separation of algal biomass from the effluent. Excessive loss of algae from the ponds deteriorates the effluent quality. When proper hydraulic residence time is not provided for the WSPs, the content of organic matters in the effluent can even be higher than that of the influent. This has been recognized as one of the most troublesome operational problems [14]. Thus, separation of the algae is essential to produce lower concentrations of SS, and nutrients as well. For removing algal particles from pond effluent, various methods have been proposed in the literatures such as maturation or polishing ponds, [12] fishing ponds, land or wetland treatment with microstraining [2]. From a pilot plant study, [6] examined the individual effects of the water hyacinth leaves, stems and root mat on the algal concentration. The results showed that filtration and settling almost equally contributed to the separation of algal particles. The canopy effect of the leaves and stems, which suppresses algal growth, is equivalent to a considerable amount of algae removed by gravity settling. A portion of the SS in the influent sewage is also removed by settling in the root zone, as it flows through the water hyacinth channel. Another portion of the SS that will not settle by gravity is screened by filtration as wastewater flows through the root mat of the water hyacinth. Because filtration is of great importance in removal mechanism, transport of the wastewater to the root zone is a critical designing consideration in water hyacinth treatment systems. [4] reported that the key separation phenomenon of algae particles by root surface is similar to adsorption processes. In other words, there is a maximum capacity in a given weight of roots, but effluent algal concentration does not increase at saturation due to the sloughing off of attached particles as a clump from the roots and the continuous reproduction of new attachment sites caused by the growth of roots. In many cases, water hyacinth ponds remain unsaturated, providing thus an efficient method for particle removal [7]. How algal and other suspended particles are retained on the surface of roots is not clearly known yet, the attachment mechanism may include electrostatic interactions and chemical bridging, or specific adsorption, all of which would be affected by chemical characteristics. Microscopic examination of the roots revealed that a gelatinous matter, which covered the root surface, surrounds algal particles. These materials are only assumed to provide an attachment force between roots and particles [6]. 3. Study Area This concept was therefore tested at the existing sewage treatment plant (STP) in Hikkaduwa of Southern Province of Sri Lanka. It is about 100 km away from Colombo towards the south of Sri Lanka. The sewage treatment method adopted in this STP has been the WSP system. It consists of inlet structure, three treatment lagoons (two facultative ponds and one ENGINEER 66 5 ENGINEER maturation pond) and a decanting structure at the outfall for discharge of treated effluent. STP at Hikkaduwa treats the sewage that is collected by the sewerage system as well as by the septage from tankers. It consists of a number of treatment processes. Inlet works remove grit and gross nonbiodegradable materials such as plastic. In facultative ponds, settlable solids are removed. Polishing and disinfection occurs in maturation ponds. Treated effluent from STP at Hikkaduwa is discharged during the receding tide every day by manual decanting from the maturation lagoon to the Hikkaduwa river. SPT at Hikkaduwa is equipped with two facultative lagoons with a capacity of 6500 m 3 and 6200 m 3 and one maturation pond with the capacity of 7000 m 3 . The normal process train is to operate the three lagoons in series so as to pass sewage after going through the inlet works enters lagoon 1 and is discharged eventually from lagoon 3. Finally all sewage leaves the plant via the manually operated decant structure. All lagoons are equipped with baffles to support inflow distribution and to avoid short circuiting. The flow from the inlet works (flow splitter) to the facultative lagoons is bifurcated to support inflow distribution in facultative ponds. Floating booms are attached to the top of the baffles so that any floating scum, grease or debris from the sewage feed is trapped and contained for ease of daily removal by STP operators. 4. Research Methodology This study basically consisted of two components such as establishment of floating wetland and the maintenance of it respectively and the water quality measurements obtained before and after establishment of wetland were then analyzed by means of statistical method to generalize the findings for population. 4.1 Establishment of floating wetland A wetland was created with wetland plant covering the maturation pond from the inlet up to the first baffle wall. It is approximately 1355 m 2 in extent and selected area is shown in figure 4.1. The macrophyte used for this wetland was water hyacinth (Eichhornia crassipes) which is fast-growing plant in fresh waters. They were collected from Bolgoda lake nearby University of Moratuwa premises and were selected as they hold up high nutrient levels, grow faster, contain deeply grown root systems. Three truck loads were transported to the Hikkaduwa SPT from predefined areas of Bolgoda Lake taking into account the precautions not to dump them elsewhere. Once established, it took almost two month period to get acclimatized to the new environment without any multiplication of plants. 4.2 Sampling and testing water quality before and after establishment of wetland To study the variation of influent and effluent water quality parameters and algal diversity and density, six sampling points were selected as shown in figure 4.2. Samples were collected and analyzed every two weeks for a period of approximately six months in order to study the temporal variation of different parameters. For the elucidation of spatial variation, samples from all six locations in the ponds were analyzed at the same time as well. Sample collection and analysis were performed before and after the creation of wetland. Figure 4.1 -Wetland area Figure 4.2 -Location map of sampling points ENGINEER 67 ENGINEER 6 The physiochemical parameters and biological measured were as follows; i. BOD ii. COD iii. Total Dissolved Solids iv. Total Phosphorus v. Dissolved Oxygen vi. Conductivity vii. pH viii. Temperature ix. Salinity x. Algal diversity and xi. Algal density. Testing for TP, TN, BOD and COD were carried out ex situ in the Environmental Engineering Laboratory, Department of Civil Engineering at the University of Moratuwa and DO, turbidity conductivity, temperature, pH, TDS and salinity were measured in situ by using portable Water Quality Meter (Sensor module – WM5 – 24 – 1 - 01) respectively. Algae enumeration was carried out in the Plant Science Laboratory, Department of Botany at University of Colombo. A riding boat was used to collect samples from sampling points L-3 and L-5 which were located in the middle of the ponds. Depth samples were not taken mainly due to the fact that all ponds were too shallow (1.2-1.5 m) so as not to cause any depth variation predominately because of wind mixing. Thus, one sample from each sampling point was collected from the top of the water column. Identification of algae was accomplished with the aid of a microscope, a counting chamber called “Sedgewick rafter” and a tally counter. A sample of 1 ml was placed in the Sedgewick Rafter for counting cells and the enumeration of the organisms was made with the aid of a compound microscope. The magnification used was 100X obtained by means a 10X ocular and 10X objective magnifications. The plankton organisms appearing in 10 fields were counted and from their total, the number of organisms per milliliter of the water sample was calculated. Quantitative records for each genus (isolated cells plus colonies) were reported separately. 4.3 Statistical methods adopted In order to analyze the rate of reduction of the water quality parameters and algal density due to the introduction of floating wetland, a non- parametric statistical analysis method was used. Since the values of water quality parameters and algal diversity and density measured were not the same as they change due to prevailing climatic conditions and influent quality such as organic loading intensity, pH etc. two sampling points namely, L-4 (just before the entrance to the wetland) and L-6 (final exit point after the wetland) were selected for the statistical analysis. Hypothesis tests were performed for all parameters collected from these two points using the Mann Whitney test. 5. Results and Discussion The test results of the water quality parameters and algal diversity and density on the wastewater in the Hikkaduwa SPT before and after establishing the wetland are shown in Table 5.1 and Table 5.2 respectively. Temporal variations according to the sampling dates and spatial variations along the sampling points are represented from Figure 5-1 to Figure 5-15 respectively. Even though pH, DO, temperature, TDS and turbidity were measured, they showed no signs of considerable changes statistically for conditions prevailing before and after the establishment of wetland. However, salinity was monitored throughout the study period, since an increase in the level of salinity above 2 ppt would affect the existence of water hyacinth plants due to toxicity effects. It was possible to achieve a reduction of algal cell densities of 900 units/ml to zero unit/ml for the algal species of Spirulina and for Oscillatoria, the reduction was from 290 units/ml to zero units/ml. In case of Chlorella and Pandorina, density reductions were 830,000 units/ml to 68,000 units/ml and 4300 units/ml to 280 units/ml respectively. Accordingly, the reduction efficiencies for Spirulina, Oscilltoria, Chlorella and Pandorina after the addition of wetland were reported to be improved from 31.8% to 100% and 4.5% to 100%, 34.2% to 91.8% and 42.2% to 93.5%, respectively. The removal efficiencies of organic matter were found to have increased in the maturation pond in terms of BOD and COD from 13.3% to 62.9% and 13.6% to 57.5%, respectively. In the case of nutrients, the reduction percentage of 74.8% for TP and 55.8% for TN were achieved since the establishment of floating wetland. ENGINEER 68 7 ENGINEER In general, the growth rate of algae depends on temperature, photosynthetically active light intensity and limiting nutrient levels respectively. The introduction of floating wetland made the light intensity available for algal growth minimal as shading effect was pronounced due to wetland plants. Further the limiting nutrients such as phosphorous were also utilized by wetland plants making them further restricted for algal growth. Also noted was the reduction of carbonaceous compounds due to microbial activity taking place in the root zone vegetation. These phenomena therefore helped to reduce algal densities drastically in the maturation pond. Table 5.1 Reduction percentages from inlet to outlet in WSP on 24/4/2012 (before establishing the wetland) L-4 L-6 Reduction % BOD 90 78 13.3 COD 220 190 13.6 TP 4.29 4.91 -1.7 TN 10.29 12.83 -24.7 Chlorella 698,000 459,000 34.2 Pandorina 9000 5200 42.2 Spirulina 2200 1500 31.8 Oscillatoria 1100 1050 4.5 Table 5.2 Reduction percentages from inlet to outlet in WSP on 28/9/2012 (after establishing the wetland) L-4 L-6 Reduction % BOD 70 26 62.9 COD 160 68 57.5 TP 1.55 0.39 74.8 TN 16.35 7.23 55.8 Chlorella 830,000 68,000 91.8 Pandorina 4300 280 93.5 Spirulina 930 0 100 Oscillatoria 290 0 100 Table 5.3: Results obtained by statistical analysis Median Zw P value Before After BOD 14.56 60.56 3.0 0.0407 COD 17.54 55.56 3.0 0.0407 TP -25.61 74.84 3.0 0.0407 TN -12.20 51.43 3.0 0.0407 Chlorella density 13.85 78.34 3.0 0.0407 Pandorina density 44.95 73.79 3.0 0.0407 Note: P value – Probability; Zw – Normal approximation for test statistics 90 95 80 64 64 71 70 220 280 210 230.4 140.8 158.4 160 0 50 100 150 200 250 300 2 4 / 4 / 2 0 1 2 1 5 / 5 / 2 0 1 2 3 / 8 / 2 0 1 2 1 7 / 8 / 2 0 1 2 3 0 / 8 / 2 0 1 2 1 4 / 9 / 2 0 1 2 2 8 / 9 / 2 0 1 2 BOD (mg/l) COD (mg/l) Before establishing the wetland Figure 5.1- Temporal variation of BOD and COD at Location L-4 After establishing the wetland ENGINEER 69 ENGINEER 8 78 80 32 24 34 28 26 190 220 110 76.8 110 70.4 68 0 50 100 150 200 250 2 4 / 4 / 2 0 1 2 1 5 / 5 / 2 0 1 2 3 / 8 / 2 0 1 2 1 7 / 8 / 2 0 1 2 3 0 / 8 / 2 0 1 2 1 4 / 9 / 2 0 1 2 2 8 / 9 / 2 0 1 2 BOD (mg/l) COD (mg/l) 110 95 93 90 80 78 280 260 210 220 220 190 0 50 100 150 200 250 300 L o c a t i o n 1 L o c a t i o n 2 L o c a t i o n 3 L o c a t i o n 4 L o c a t i o n 5 L o c a t i o n 6 BOD (mg/l) COD (mg/l) 90 80 75 70 32 26 190 186 160 160 80 68 0 20 40 60 80 100 120 140 160 180 200 L o c a t i o n 1 L o c a t i o n 2 L o c a t i o n 3 L o c a t i o n 4 L o c a t i o n 5 L o c a t i o n 6 BOD (mg/l) COD (mg/l) 4.828 4.806 6.717 1.65 1.52 1.63 1.55 18.89 20.67 10.384 17.24 18.86 17.5 16.35 0 5 10 15 20 25 2 4 / 4 / 2 0 1 2 1 5 / 5 / 2 0 1 2 3 / 8 / 2 0 1 2 1 7 / 8 / 2 0 1 2 3 0 / 8 / 2 0 1 2 1 4 / 9 / 2 0 1 2 2 8 / 9 / 2 0 1 2 TP (mg/l) TN (mg/l) Before establishing the wetland After establishing the wetland Figure 5.2 - Temporal variation of BOD and COD at Location L-6 Figure 5.3 - Spatial variation of BOD and COD on 24/4/2012 Figure 5.4 - Spatial variation of BOD and COD on 28/9/2012 Figure 5.5 - Temporal variation of TP and TN at Location L-4 Before establishing the wetland After establishing the wetland ENGINEER 70 9 ENGINEER 5.757 6.343 1.753 0.369 0.389 0.41 0.39 13.38 12.9 11.3 9.6 10.6 8.5 7.23 0 2 4 6 8 10 12 14 16 2 4 / 4 / 2 0 1 2 1 5 / 5 / 2 0 1 2 3 / 8 / 2 0 1 2 1 7 / 8 / 2 0 1 2 3 0 / 8 / 2 0 1 2 1 4 / 9 / 2 0 1 2 2 8 / 9 / 2 0 1 2 TP (mg/l) TN (mg/l) 5.62 6.981 5.72 4.828 4.908 5.757 11.88 9.65 9.38 10.29 12.83 13.38 0 2 4 6 8 10 12 14 16 L o c a t i o n 1 L o c a t i o n 2 L o c a t i o n 3 L o c a t i o n 4 L o c a t i o n 5 L o c a t i o n 6 TP (mg/l) TN (mg/l) 1.42 1.81 1.75 1.55 0.35 0.39 15.4 16.2 17.9 16.35 8.2 7.23 0 2 4 6 8 10 12 14 16 18 20 L o c a t i o n 1 L o c a t i o n 2 L o c a t i o n 3 L o c a t i o n 4 L o c a t i o n 5 L o c a t i o n 6 TP (mg/l) TN (mg/l) 553,000 420,250 650,000 1,025,000 530,000 760,000 830,000 0 0 0 94,000 32,080 51,00048,000 0 200,000 400,000 600,000 800,000 1,000,000 1,200,000 2 4 / 4 / 2 0 1 2 1 5 / 5 / 2 0 1 2 3 / 8 / 2 0 1 2 1 7 / 8 / 2 0 1 2 3 0 / 8 / 2 0 1 2 1 4 / 9 / 2 0 1 2 2 8 / 9 / 2 0 1 2 Chlorella (units/ml) Ankistrod esmus (units/ml) Figure 5.6 - Temporal variation of TP and TN at Location L-6 After establishing the wetland Before establishing the wetland Figure 5.7 - Spatial variation of TP and TN on 24/4/2012 Figure 5.8 - Spatial variation of TP and TN on 28/9/2012 Figure 5.9 - Temporal variation of Chlorella and Ankistrodesmus at Location L-4 Before establishing the wetland After establishing the wetland ENGINEER 71 ENGINEER 10 459,400 675,000 200,350 222,000 218,000 80,000 68,000 0 0 0 5,100 2,000 2,500 1,700 0 100,000 200,000 300,000 400,000 500,000 600,000 700,000 800,000 2 4 / 4 / 2 0 1 2 1 5 / 5 / 2 0 1 2 3 / 8 / 2 0 1 2 1 7 / 8 / 2 0 1 2 3 0 / 8 / 2 0 1 2 1 4 / 9 / 2 0 1 2 2 8 / 9 / 2 0 1 2 Chlorella (units/ml) Ankistrodesmu s (units/ml) 547,000 336,000 267,400 153,000 810,200 759400 0 100,000 200,000 300,000 400,000 500,000 600,000 700,000 800,000 900,000 L o c a t i o n 1 L o c a t i o n 2 L o c a t i o n 3 L o c a t i o n 4 L o c a t i o n 5 L o c a t i o n 6 Chlorella (units/ml) 720,000 890,000 810,000 830,000 79,000 68,000 65,000 210,000 100,000 48,000 14,300 1,700 0 100,000 200,000 300,000 400,000 500,000 600,000 700,000 800,000 900,000 1,000,000 L o c a t i o n 1 L o c a t i o n 2 L o c a t i o n 3 L o c a t i o n 4 L o c a t i o n 5 L o c a t i o n 6 Chlorella (units/ml ) 14,800 10,400 7,600 9,000 5,900 5200 0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 L o c a t i o n 1 L o c a t i o n 2 L o c a t i o n 3 L o c a t i o n 4 L o c a t i o n 5 L o c a t i o n 6 Pandorina (units/ml) Figure 5.10 - Temporal variation of Chlorella and Ankistrodesmus at Location L-6 After establishing the wetland Before establishing the wetland Figure 5.11 - Spatial variation of Chlorella on 24/4/2012 Figure 5.12 - Spatial variation of Chlorella on 28/9/2012 Figure 5.13 - Spatial variation of Pandorina on 24/4/2012 ENGINEER 72 11 ENGINEER 11,500 5,200 4,050 4,300 310 280 0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 L o c a t i o n 1 L o c a t i o n 2 L o c a t i o n 3 L o c a t i o n 4 L o c a t i o n 5 L o c a t i o n 6 Pandorina (units/ml) 1,500 900 0 0 0 0 0 1050 870 0 0 0 0 0 0 200 400 600 800 1,000 1,200 1,400 1,600 2 4 / 4 / 2 0 1 2 1 5 / 5 / 2 0 1 2 3 / 8 / 2 0 1 2 1 7 / 8 / 2 0 1 2 3 0 / 8 / 2 0 1 2 1 4 / 9 / 2 0 1 2 2 8 / 9 / 2 0 1 2 Spirulina (units/ml) Oscillatoria (units/ml) 5. Conclusions and Recommendations The following conclusions were inferred from the study. 1. It was clear that all types of algae got reduced remarkably with the introduction of the floating wetland. The introduction of wetland causes less solar intensity to be available for algal growth hence reduction in algal densities. Therefore, it can be concluded that a floating wetland consisting of water hyacinth plants is ideal to control washout of algae from waste stabilization ponds. 2. Further, root zone of vegetation cover activated microbial assimilation of carbonaceous compounds thereby reduction of BOD and COD. Hence the introduction of the wetland has proved to be a good technique even to lessen the organic loading flowing into the receiving water body or in other words wetland seems to be good option for polishing effluent quality prior to discharge into water bodies. 3. It has also proved that this is an ideal technique to cut off excessive nutrients from escaping to receiving water bodies which would otherwise contribute to eutrophication incidences in a great deal. The introduction of carefully designed and maintained floating wetlands inhabiting water hyacinth plants would be recommended for other WSPs of Sri Lanka in order to reduce the algal washouts, to polish or refine the effluent water quality and to reduce the escape of nutrient such as N and P to nearby water bodies with high percentage of success. Acknowledgement Authors wish to acknowledge the National Water Supply & Drainage Board for granting necessary financial support for making the study possible and both University of Moratuwa and University of Colombo for providing laboratory facilities. References 1. Arceivala S. J., 1981, Wastewater Treatment and Disposal, Engineering and Ecology in Pollution Control, Marcel Dekker, Inc, Newyork,. pp 787-859. 2. AWWA (1991) Introduction to Water treatment, Vol. 2, Denver, CO. 3. Horan, N. J. (1996). Biological Wastewater Treatment Systems : Theory and Operation. John Wiley and Sons. 4. Hosetti, B. and Frost, S. (1998). A Review of the Control of the Biological Waste Treatment. 5. Kayombo.S, Mbwette.T. S. A, Katima. J. H. Y, Ladegaard N, Jorgensen S. E., “Waste Stabilization Ponds and Construction Wetlands Design Manual”, UNEP-IETC, Danida. 6. Kim, Y. and Kim, W. J., “Roles of Water Hyacinths and their Roots for Reducing Algal Concentrations in the Effluent from Figure 5.14 - Spatial variation of Pandorina on 28/9/2012 Figure 5.15 - Temporal variation of Spirulina and Oscillatoria at Location L-6 Before establishing the wetland After establishing the wetland ENGINEER 73 ENGINEER 12 Waste Stabilization Ponds,” Water Res., 34(1), 3285-3294 (2000). 7. Kumar, P. and Garde, R. J., “Stratification in Laboratory Simulation of Water Hyacinth Ponds,” J. Envion. Eng.,125(4), 382-384 (1999). 8. Mara, D. D. and Pearson, H. (1986). Artificial Freshwater Environment: Waste Stabilisation Ponds in Biotechnology – A Comprehensive Treatise 8. Rehm H. J. and Red G. (Eds). Weinheim and Verlagsgesellschaft. 9. Marais, G. V. R. (1974). “Faecal Bacterial Kinetics in Waste Stabilization Ponds.” Journal of the Environmental Engineering Division, American Society of Civil Engineers 100 (EEI): 119-139. 10. Oswald W. J. (1988a). Micro-Algae and Wastewater Treatment. In: Mico-Algae Biotechnology. M. A. Borowitzka and L. J. Borowitska (eds). Cambridge University Press pp 357-394. 11. Oswald W. J. (1988b). Large-Scale Algal Culture Systems (Engineering Aspects). In: Micro-Algae Biotechnology. M. A. Borowitzka and L. J. Borowitska (eds). Cambridge University Press pp 357-394. 12. Oswald W. J. (1991). Introduction of Advanced Integrated Wastewater Ponding System. Water Sci. Tech. 24(5), 1-7. 13. Pearson, H. W. (1996). Expanding the Horizons of Pond Technology and Application in an Environmentally Conscious World. Wat. Sci. Tech. 33.7 pp 1- 9. 14. WEF and ASCE (1992) Designing of Municipal Wastewater Treatment Plants, Vol. 2\quad\quad. Chapter 13, P. 834. Alexandria, VA. ENGINEER 74 Printed by Karunaratne & Sons (Pvt) Ltd.