Cover Story Feature ReportEvaluating Capital Cost Estimation Programs Five capital-costestimation programs are compared using a set of case studies Ying Feng and Gade P. Rangaiah National University of Singapore apital cost assessment is an integral part of process design when building or expanding a chemical process plant. It is required to provide project analysis and evaluation, to select among alternative designs, to plan the appropriation of funds and to serve as a basis for project cost control. Capital cost estimation is also important for an engineerng student’s final design project. The accuracy of capital cost estimates depends on the available design details, the accuracy of the cost data, as well as the time available to prepare estimates. The generally accepted classification of capital cost estimates (Table 1) is published by the Association for the Advancement of Cost Engineering (AACE International). Currently, several methods and associated computer programs exist for estimating capital cost, and they provide mainly study or preliminary estimates (Table 1; Class 4). An evaluation and comparison of the computer programs available for capital cost estimation provides a better understanding of both their use and the methods behind them. This article arose out of a systematic effort by the authors to apply and analyze several programs available for capital cost estimation. The goals of this effort were to evaluate available cost estimation programs, apply the programs to case studies of plant design and to analyze 22 AACE International Recommended Practice No. 18R-97 cost estimate classification system Class / type Class 5: Order of magnitude estimates Class 4: Study or preliminary estimates Class 3: Definitive estimates Class 2: Detailed estimates Class 1: Check estimates Purpose Initial feasibility study or screening Concept study or feasibility Budget, authorization or control Control or bid/tender Check estimate or bid/tender Accuracy L:–20 to –50% H: 30 to 100% L: –15 to –30% H: 20 to 50% L: –10 to –20% H: 10 to 30% L: –5 to –15% H: 5 to 20% L: –3 to –10% H: 3 to 15% Level of project completion 0–2% 1–15% 10–40% 30–70% 50–100% TABLE 1. CLASSIFICATION OF CAPITAL COST ESTIMATES C TABLE 2. PURCHASE COST (CPCS) OF COMPRESSORS Compressor CapCost DFP CCEP Power, kW Material CPCS Δ% Δ% CS7-G-101 58 CS 217,350* 278%* –4%* CS4-C-101 183 CS 217,350* 226% –16% CS5-C-101 364 CS 217,350* 454% 78% CS6-C-101 3,100 CS 1,113,200 162% 128% *cost of minimum size, as size is less than the minimum size EconExpert Δ% –65%* –27% 92% 174% AspenPEA Δ% –27% –24% 33% –37% the cost estimates for different types of equipment. Five programs are discussed here — CapCost, EconExpert, AspenTech Process Economic Analyzer (AspenPEA), Detailed Factorial Method (DFP) and Capital Cost Estimation Program (CCEP). They were evaluated using seven case studies taken from textbooks. The selected case studies involve petroleum refining, petrochemical and biopharmaceutical processes. The cost estimates are compared and analyzed at both the equipment and plant levels to uncover the relative strengths of each of the programs used. Estimation methods Capital cost estimation methods in various process design reference books range from simple order-of-magnitude schemes to more detailed module costing and factorial methods. Common methods for preliminary capital cost estimation are based on the Lang factor method, the detailed factorial method and the module costing method. Most available cost estima- tion programs in the open literature use one of these methods. In the Lang factor method [1], plant capital cost is determined by multiplying the sum of the purchase costs of major equipment by a lumped factor called the Lang factor. Since the Lang factor is used across the whole plant, it applies the same installation factor to all equipment types in the plant, irrespective of material of construction and pressure. Sinnott and Towler [2] proposed a modified Lang factor method to estimate inside battery limit cost (CISBL) and fixed capital cost. This modified method accounts for varying installation factors in different processes, but not for different types of equipment. However, material of construction is considered through a material factor in calculating CISBL. Another cost estimation method is the module costing method originally introduced by Guthrie [3]. This method is well accepted for estimating the cost of a new chemical plant in the preliminary stage. It is detailed, and includes a breakdown of cost categories for de- CHEMICAL ENGINEERING WWW.CHE.COM AUGUST 2011 the purchase cost. Turton and colleagues proposed the following: o o CBM = CP × FBM = CP (B1 + B2FMFP) riving installed costs from purchase costs. The bare module cost (CBM) for a given piece of equipment i is thus defined as: CBM. Bare module cost (CBM) is defined as the sum of the direct and indirect expenses for purchasing and installing equipment.3 CS/CS 120.7 CS/CS 33. direct cost. CDE. when cost data are assembled from vendor quotes. CapCost is available with the process design book of Turton and coauthors [1]. Most programs use a para- (4) where S represents a parameter for the equipment size or capacity.140 5% –23% CS4-E-106 41 SS/SS 28.620 8% –16% CS4-E-101 405 CS/CS 87. PURCHASE COST (CPCS) OF HEAT EXCHANGERS DFP ESTIMATE FOR ALL SHELL-AND-TUBE HEAT EXCHANGERS IS FOR U-TUBE TYPE. they exhibit scatter due to different qualities of equipment design. For each piece of equipment. Detailed Factorial Program (DFP) is based on the detailed factorial estimates method described in Ref. CS=CARBON STEEL. These data are based on surveys of equipment vendors between May and September 2001.i. DFP. respectively. CM is the cost of field materials required for installation. written in Visual Basic.540 28% –52% CS1-E-101 20 CS/CS 26. 56 and www.750* –3%* –28%* CS2-E-104 18. Computer programs The five progams used in this study are CapCost.990 34% –29% Heat exchanger – kettle reboiler CS2-E-102 85. CO is the cost of construction overhead and CE is the cost of contractor engineering expenses.945 0% –26% CS6-E-101 14.i is contingency fee and CFee.CHE.250 4% –45% CS6-E-103 1760 SS/SS 355. The Chemical Engineering Plant Cost Index (CEPCI. as well as correlations and plots for FBM. and can be used for estimating preliminary process cost.795 12% –17% CS5-E-106 76. fabrication. and therefore the estimating method based on these data can only be used for study or preliminary estimates. insulation and paint. controllers. equipmentspecific constants K1. concrete. of the major equipment items is esti23 CHEMICAL ENGINEERING WWW. Values for the constants B1 and B2. CP.650 19% –39% CS6-E-106 11.565 9% –19% CS7-E-102 100 CS/SS 34.500 14% –34% CS5-E-108 902 CS/CS 181. FP and CoP of different equipment can be found in the appendices in Ref.CapCost DFP CCEP ing-head type) Area. electrical hardware. FM.700 15% –38% CS5-E-104 2130 CS/CS 430.750 –58% –67% CS4-E-104 37. che. insurance and taxes.320 13% –16% CS5-E-109 680 CS/CS 138. It is based on the module costing method.2 CS/CS 26. For the whole plant. 2.i + CIDE. CL is the cost of labor to install equipment and materials. as well as the direct costs of field labor used for their installation.680 11% –18% 102/103/104 CS4-E-103 28.060 12% –15% CS6-E-102 61. o where CBM. i = CDE. market conditions.41 CS 4. which is useful in early conceptual estimates [5]. As discussed by Woods [6] and Walas [7].i = o [CP + CM + CL + CFIT + CO + CE]i (1) carbon steel material and atmospheric pressure).796 –35% –42% Heat exchanger – plate CS7-E-101 57 SS 95. K2 and K3. and the grassroots plant cost (CGR) is defined as the sum of the total module cost and the auxiliary facilities costs. CCEP.i and CP are.000 14% –34% CS1-E-102 740 CS/CS 149.290 8% –7% CS4-E-105 269 SS/SS 63.7 SS/SS 27. SS=STAINLESS STEEL Heat exchanger – shell and tube (Float. EconExpert and AspenPEA. indirect cost and purchase cost of equipment i in base conditions (that is. m2 Material CPCS Δ% Δ% CS4-E-102 4.com/pci for current value) value of 397 for this period can be used for escalating cost to a different time. vendor profit and other considerations.i for all equipment types by the factor 1. see p.350 31% –28% CS5-E-103 2010 SS/SS 405. Ulrich and Vasudevan [4] propose to find the total module cost (CTM) by multiplying CBM. as well as the indirect costs involved in installation.i. 1.450 –88% n/a+ *cost of minimum size as size is less than the minimum size + not available in this program and so taken from another program EconExpert Δ% –75%* –66% –59% –58% –43% –38% –37% 32% –39% –42% –43% –35% –42% –43% –73% –69% –52% –46% –38% –36% –36% –38% –40% –45% –41% –72% –64% –32% –16% –73% –78% 26% –80% AspenPEA Δ% –37% 50% –22% –24% –6% –3% –3% –3% –7% 11% –25% 0% -4% –2% –31% –29% –17% –13% –4% –1% 3% –5% –8% 12% –4% 49% –22% –15% –3% 67% –61% 21% –41% metric-cost model for the cost estimates. instruments. Hence. For o this program.6 SS/SS 27. such as insurance. CIDE. the accuracy of published equipment-cost data may be no better than ±25%.100 9% –40% CS5-E-105 2900 CS/CS 587.950 15% –27% Heat exchanger – shell and tube (Fixed-head type) CS2-E-101 14. bare module cost.025 12% –28% CS6-E-104 1090 SS/SS 224.i + CFee.TABLE 3. steel.940 9% –21% CS5-E-102 456 SS/SS 96.18 to account for contingency and contractor fee.2 CS/CS 26. construction overhead and contractor engineering expense.3 CS/CS 22. CFIT is the cost of freight.COM AUGUST 2011 . The cost equation constants used in these programs are obtained from vendor quotes or from past literature data.3 SS/SS 59.62 CS/CS 28.6 SS/SS 30.i (2) (3) o log CP = K1 + K2log(S) + K3[log(S)]2 where CCont.i) = 1.i is the contractor fee.4 CS/CS 26. the total module cost (CTM) is defined as the sum of the bare module cost. CTM = ∑ (CBM. To estimate the bare module cost and purchase cost of equipment.940 8% –15% CS1-E-103/104 150 CS/CS 44.820 11% –1% CS6-E-105 131 SS/SS 41.i + CCont.18 ∑ CBM.340 –36% –47% Heat exchanger – double pipe CS2-E-103 5. contingency and fee. Guthrie provided factors to estimate the direct costs of field materials.515 9% –4% CS6-E-107 192 SS/SS 51.680 10% –21% CS3-E-101/ 26.055 11% –26% CS5-E-101 541 CS/CS 111. These factors include material erection and equipment setting. such as piping.155 20% –36% CS7-E-103 240 CS/SS 48.120 10% –14% CS5-E-107 127 CS/CS 40. 375* –94% n/a+ CS7-M-104 2 SS 37. PURCHASE COST (CPCS) OF MIXERS Mixer Details CapCost DFP CCEP EconExpert Δ% –83%* –99% –81% –79% –74% –68% 56% AspenPEA Δ% –70%* –97% –67% –64% –57% –52% –13% Cover Story mated using the following: o CP = a + bSn (5) Cost constants a and b. kW Material CPCS Δ% Δ% CS3-M-101 0. Equipment examples Seven case studies based on process design data available in Refs. CHEMICAL ENGINEERING WWW. The cost data and correlations in EconExpert are for a CEPCI of 400 [4]. The ISBL cost and fixed capital cost are calculated according to the method in Ref. field-tested.. By analyzing the cost factors. other factors. the CEPCI seems likely to return to 575 in the near future. base condition purchase cost (CPCS) and total module cost (CTM) for equipment in the seven case studies are compared. if a piece of equipment has a size above the valid range.7 CS 37. whereas in EconExpert. only CISBL and CDirect. 8. EconExpert is a Web-based interactive software for capital cost estimation [11]. are available from these programs. although any (6) Values of constants A0.3 and X = 0. For all programs. the equipment module costing method is used to calculate bare module cost and total module cost from the purchase cost of equipment.TABLE 4.2 SS 37. The purchase cost of the major equipment items is estimated using the following: 2 o CP = e{A0 + A1[ln(S) + A2[ln(S)] + . industry-standard cost modeling and scheduling methods [12]. CISBL and CDirect. Hence.. FM is the material factor and FP is the pressure factor. 4. CPCS in these programs is calculated by entering equipment details for the same condition as the one required using carbon steel as the material of construction and 1 barg as the operating pressure.375* –37% n/a+ CS7-M-103B 2.375* –37% n/a+ CS7-M-103A 4. Case 7 is a ß-galactosidase batch process via recombinant Escherichia coli [4]. AspenPEA claims to contain timeproven. Case 6 is an alternative synthesis of maleic anhydride via benzene in a shell-andtube reactor with catalyst [1]. 2 for different equipment items.CHE. 4. available in Ref. for flowrates less than the minimum amount required for costing. while the other two do not. based on available literature sources and vendor data. This causes some uncertainty in the predicted cost. as part of research projects supervised by the second author (these programs can be obtained from the authors). AspenPEA Version 7. AspenPEA is built on Aspen Icarus technology. Some process equipment may have sizes outside the valid range for the cost correlations or for the program. it is observed that CTM includes engineering design. In this textbook. Similar to CapCost. As the module costing technique is not used in DFP and AspenPEA.595 –36% n/a+ CS7-M-101 146 SS 103. The purchase cost data and bare module factors used can be found in Ref. A1 and A2 for various equipment items can be found in Ref. 2. the 24 Here.015 SS 37.} Power. are mainly for carbon-steel material. A CEPCI of 575 was used throughout to allow cost estimates by all programs to be compared meaningfully. and is designed to generate both conceptual and detailed estimates [12]. The CapCost program is used as the reference for comparison.375* –37% n/a+ CS7-M-102 7 SS 40. CCEP and AspenPEA. the minimum flowrate is used instead of the actual flowrate. where the CEPCI is 575). 1. Multiple regression is used to fit the data if the purchase cost is dependent on more than one variable. a petroleum refinery distillation column (the alkylate splitter) is employed to separate a mixture of C4 to C14 hydrocarbons into two streams.1 was used for this study (this version follows 2008 data. It takes a unique approach. 8. it is divided into multiple units of smaller size and the costing is done by summing up CTM of multiple smaller units. 8 and 13 were chosen for this study. the cost data are expressed in graphical form. The base condition (equipment made of carbon steel and operating near ambient pressures) purchase cost (CPCS) and total module cost (CTM) are compared and analyzed. In the following sections. calculated purchase cost is directly based on actual material of construction and pressure. Therefore.730 16% n/a+ *cost of minimum size as size is less than the minimum size + mixer cost not available for CCEP plots are represented as polynomial equations for calculation of the purchase cost. then its cost is estimated by taking the lower limit of the valid range. The CEPCI generally trends upward over time.5 SS 37. The material factor and Guthrie’s bare module factor are used thereafter to estimate the installed cost of that equipment. according to the factors stated in the detailed factorial method [2]. such as material of construction and operating pressure. The five computer programs were used to evaluate capital costs for each case. using CEPCI = 500. Hence. Similarly. A list of these cost correlations can be found in Ref.7). In DFP. In Case 1 from Ref. CISBL is the inside battery limit cost. A styrene production process in the same reference is Case 5. when the CEPCI dropped to 525 (from 575 in 2008) so. by Wong [9] and Huang [10]. are also taken into consideration in estimating the purchase cost. CTM = CISBL (1 + DE + X) CTM = CDirect (1 + DE + X) (8) (9) (7) where CP is the purchase cost of equipment. Case 2 is the monochlorobenzene (MCB) separation process in Ref.COM AUGUST 2011 . CCEP and DFP were developed in Microsoft Excel and Visual Basic environments. Case 4 is the formalin production process from methanol using silver catalyst [1]. the cost equations are often in the following form: o Cp = FM FP CP other program could have been used. Seider [8] developed the purchase cost correlations for common process equipment. 8. 13. CTM in DFP and AspenPEA is calculated by Equations (8) and (9). Case 3 is a crystallization process from the same source. including CTM. contract fees and contingency fees. representing equipment by comprehensive design-based installation models. respectively. and DE = 0.1.375* 38%* n/a+ CS7-M-105 0. If equipment has a specification below the valid range. In addition to size. respectively. respectively. Capital Cost Estimation Program (CCEP) uses cost correlations in Seider [8] for estimation of free-onboard purchase cost of equipment. An exception was 2009. and correspond to January 2007 (CEPCI of 509. 577 3. The wealth of data results in relatively more-accurate cost estimation. While other programs calculate piping cost as a small fraction of purchase cost.807 4.12 CS1-P-102 1.730 n/a 45%* CS7-P-103 1.3 Material CS CS CS CS CS CS CS CS CS CS CS CS CS CS SS SS SS CapCost CPCS 3. DFP gives a very high cost compared to all other programs.257 3.7 CS5-P-102/103 2 CS2-P-101 2. CCEP and EconExpert predict lower costs than the other three programs for all sizes of fixed-head heat exchanger.4 CS 47. Since floating-head heat exchangers are generally costlier than the U-tube type. all five programs predict total module cost for floating-head exchangers with less than 50% deviation using the CapCost value as a reference. the deviations become more significant. In these and subsequent tables.934 8. air-cooled fin fan. For double-pipe heat exchangers. the floating-head heat exchanger cost by DFP is expected to be slightly higher than that in Figure 1. 2. CapCost predicts 50 to 80% higher purchase and total module costs than the other four pro- grams.75 CS2-P-102 1. Since only the U-tube shell-and-tube heat exchanger cost equation is given in Ref.24 CS 7. In the case studies.46 CS5-P-105 0.7 SS 17. including plate-and-frame and spiral plate types. which is mainly due to the cost equation given in Ref. and compact heat exchangers.025 4. (up to 300 m2) the percent deviation is relatively small. diaphragm compressors.7 CS4-P-102 1. such that Δ = 100 × (cost using other method – cost using CapCost) / cost using CapCost).4 SS 5.24 CS5-P-106 2.554* 3.726 3. Hence. Because heat exchangers are common. kW CS4-P-101 0.69 SS 4.176 n/a+ –29%* Diaphragm pumps + CS7-P-105 1 SS 10. However. most of the fixedhead heat exchangers have areas below 300 m2.38 CS6-P-101 0.050.CHE. AspenPEA model calculates piping cost based on respective equipment. centrifugal compressors and axial compressors.258 n/a+ –55% CS3-P-101 22. the DFP estimate in Figure 1 is for this type.807 3. Hence.726 3. Figure 2 shows the variation in the total module cost of fixed-head heat exchangers calculated by all five programs. a positive (or negative) Δ value means the cost estimate by that method is more (or less) than that determined by CapCost.613 n/a+ –24%* CS3-P-103 0. the data available from vendors and other sources are voluminous.16 SS 7. For double-pipe heat exchangers. In the seven case studies. Results in Table 2 indicate that compressor purchase cost estimates by CapCost. for smaller areas. all programs predict comparable purchase cost. and purchase costs of these compressors at base conditions are presented in Table 2. all five programs predicted similar costs. From this equation. Case studies 4–7 involve compressors. while the purchase costs given by EconExpert increase very fast with increasing size. For smaller heat transfer areas (up to 400 m2).659 n/a+ –25%* CS3-P-105 0. all five programs predicted a similar cost. double-pipe. when the heat transfer area is large.500 n/a+ –35% CS6-P-103 0.554* 3. which is due to differences in installation and other factors used for CTM. and the cost estimate by other methods is expressed as the percent difference from that by CapCost. many of the floating-head heat exchangers have areas below 400 m2.4 CS1-P-103 24 CS4-P-103 0.876 3. As the area increases. However.8 CS6-P-102 5.554* 3. with somewhat lower cost by CCEP and EconExpert (Table 3).2 CS 7.554* 3. In fact.84 CS5-P-101 6.455 n/a+ 182%* CS7-P-102 3.380 n/a+ –21% CS3-P-102 29. the DFP estimate in Figure 1 is for this type.COM AUGUST 2011 . 2. As the area 25 CHEMICAL ENGINEERING WWW.08 CS7-P-101 2. Since only the U-tube shell-and-tube heat exchanger cost equation is given in Ref.554 3. even the lower limit (power of 75 kW) gives a cost as high as $714.565 3. For kettle reboilers.65 CS1-P-101 2.887 DFP Δ% 152% 156%* 152%* 165% 164% 165% 110% 166% 176% 209% 251% 297% 289% 204% 146% 143% 163% CCEP Δ% 50%* 49%* 62%* 46% 48%* 27%* 46%* 54%* 28%* 53% 26% 120% 75% 24% 59%* 223% 51%* EconExpert Δ% 7% 14%* 21% 39% 57% 73% 73% 79% 83% 89% 91% 110% 112% 117% 24% 56% 84% 99% 118% 11% 11% 22% 22% 22% 4% 26% 27% 27% 25% AspenPEA Δ% 55% 57%* 57% 60% 35% 53% 53% 50% 37% 76% 89% 123% 107% 10% 57% 57% 50% 132% 58% –9%* –8%* –24%* –38%* –49%* –3%* –50% –50% –51% –87% CS6-P-106 3. Heat exchangers Common heat-exchanger types are shell-and-tube. Such a high cost may be due to the data source for calculation of empirical constants.105 n/a+ 17% *cost of minimum size as size is less than the minimum size + not available in this program and so taken from another program Compressors Major compressor types are trunkpiston and crosshead reciprocating compressors.9 SS 13. which contributes 10×CPCS to CTM. [2]. The most significant factor contributing to the high total module cost in AspenPEA is the piping cost. the cost estimate by CapCost is given.979 4.808 438% 140% Reciprocating pumps CS3-P-104 0. Hence. Figure 1 shows the total module cost of floating-head heat exchangers calculated in all five programs as a function of heat transfer area. Total module cost is shown in online table I. However. the piping cost is around three times the purchase cost. with heat transfer area. PURCHASE COST (CPCS) OF PUMPS Centrifugal pumps Power.6 CS 9. Differences in total module cost of compressors by different programs are larger than those in purchase costs (Table online).267 145% 294% CS6-P-104 10. For example. CapCost and DFP predict a relatively higher cost than the other programs.6 CS 57. the purchase cost and total module costs of floatinghead and fixed-head heat exchangers predicted by all programs are comparable.3 CS5-P-104 0. CCEP and AspenPEA are comparable. AspenPEA predicted a total module cost three times that by CapCost (online table II).710 n/a+ –36% CS7-P-104 60 SS 95.5 CS6-P-105 1.TABLE 5. For low-power pumps. One reason for this difference may be due to using different input data.000 800. the cost CHEMICAL ENGINEERING WWW. and so the former two predict significantly higher cost than CapCost for large mixers. the total module cost of CS7-M-101 by DFP and EconExpert are. kneaders and mullers. they have rather different results. fixed-head heat exchangers are slightly less expensive than U-tube heat exchangers [8] (Table 3. For diaphragm pumps with small power. m2 600 700 800 DFP CapCost AspenPEA CCEP EconExpert '' FIGURE 1.000 800. TABLE 6. respectively.Total module cost. Hence. with DFP predicting a significantly higher cost than all the others.0 SS 565.105 –9% 249%* n/a+ 0% CS2-T-101 0. respectively. which requires its costs to be estimated as that of two smaller mixers. but different costs in the other three programs.CHE.0 SS 78. with DFP giving costs up to three times of those predicted by CapCost (Table 5). will have the same cost in CapCost and EconExpert. This difference is due to the variation in cost data used in these two programs to calculate purchase cost.S. when the power is large.9 10. The deviation in total module cost (online table III) is similar to that for the purchase cost for EconExpert and AspenPEA. Pumps Mixers Of the huge variety in mixer designs available.000 400.800 –5% –57% n/a+ –51% Packed Tower CS4-T-101 0. Reciprocating pumps are not available in DFP. PURCHASE COST (CPCS) OF TOWERS Tray Towers CapCost DFP CCEP EconExpert AspenPEA Dia. CCEP predicts 50% lower total module cost than CapCost (online table). For example. and hence not compared terials of construction is not available in CapCost. m2 800 1.650 6% 16% n/a+ –18% CS6-T-102 1.000 500.000 0 0 200 400 600 Area. However. selection of ma26 Commonly used pumps in chemical processing plants are radial-centrifugal. which results in similar costs for mixers of different materials in that program. dollars Total module cost.000 1. but deviations in estimates still increases with area increases. the total module costs of centrifugal pumps are plotted against their power. DFP and AspenPEA predict similar costs. Also. common types are ribbon and tumbler mixers.0 CS 702. the four other programs predict higher purchase and total module costs than CapCost. mainly due to the higher cost of piping predicted). U. but different flowrates due to design pressures.135 –19% 121% n/a+ –4% CS1-T-101 3. The cost estimate by DFP and EconExpert increases faster than that by CapCost. Mixer costing is not included in the CCEP program. In Figure 4. However.200 DFP CapCost AspenPEA CCEP EconExpert 900. However. Matem m rial CPCS Δ% Δ% Δ% Δ% CS2-T-102 0.000 100. 193% and 160% of that by CapCost. CapCost gives the lowest. from those of CapCost (Table 5). For double-pipe exchangers.9 21. purchase and total module costs by AspenPEA deviate by 50–100% and 100–400%. in addition to pump power.9 CS 63. the differences become more significant. Such a mapping increases the cost significantly. Mixer purchase costs differ by 50–100% among different programs (Table 4).1 18. and CCEP and AspenPEA require flowrate and pump head.000 400.000 26% 29% n/a+ 76% CS5-T-101 3.S.000 24% 48% n/a+ 21% CS6-T-101 4.000 200. even though the purchase costs are comparable. the total module cost given by DFP is higher due to a large installation factor for mixers in this method.2 10. This is due to the limitation of mixer size in DFP.0 30. and EconExpert predicts a cost in the middle range. Compared to other programs..775 25% 89% n/a+ 87% CS4-T-102 2. diaphragm and external-rotator-gear pumps. the deviation between different programs increases further (online table IV). plunger-reciprocating. For centrifugal pumps.8 CS 26. DFP and AspenPEA predict higher cost than the others. AspenPEA predicts a significantly higher cost than all other programs (again.5 19. DFP requires flowrate and liquid density. The other four programs give similar purchase costs (Table 5 and online table).0 SS 230.000.COM AUGUST 2011 . Fixed-head heat exchangers are generally less expensive. This is due to different module factors used in the two programs. This may be partially due to the mapping of fixed-head heat exchanger to U-tube heat exchanger in DFP. In general. Length. even though both CapCost and EconExpert use pump power to predict the cost. while DFP and CCEP predicted slightly lower cost (Figure 3).200.000 200. Figures 1–2 and online table II).000 700.045 63% 284%* n/a+ 135% *cost of minimum size as size is less than the minimum size + cost trays and packing are not included in EconExpert.000 1. two pumps with the same power. U. This observation coincides with the results obtained in case studies above. dollars 1.000 300.0 CS 21.000 600. The high total module cost is mainly due to the higher piping cost given by AspenPEA.5 12. In general.000 600. However.000 0 0 100 200 300 400 500 Area. DFP predicts a higher cost than AspenPEA. as the power increases. the deviation in total module cost between different programs is within 70%.6 28.0 CS 483. EconExpert and CapCost predicted similar costs. As the area of floating-head heat exchangers increases the deviation in total module cost estimates grows FIGURE 2. However. The total module cost of double-pipe heat exchangers is plotted against heat transfer area in Figure 3. For centrifugal pumps. While CapCost and EconExpert programs require only pump power for cost prediction. 8 SS 124. which is three times CPCS.992 629% 441%* CS3-V-101 1.9 1.8 SS 6.2 3.2 SS 15. settlers. For tray towers. chemical reactors. although the purchase cost from CapCost is comparable or even lower than in other programs. For small-dia. whereas CCEP is higher when the diameter is small (or lower when large). the pump is operating in one stage and vertical split case. When CTM for reciprocating pumps is plotted against shaft power. and the cost is higher. while AspenPEA predicts the lowest (reciprocating pumps are not available in DFP). the cost given 27 of a centrifugal pump in CCEP depends on an additional factor: the stages and split-case orientation. They contain trays or packing. for horizontal vessels with low design pressure and small diameter. Hence. C2 and C3 are constants that can be found in Ref.2P) + CA] (10) Here.135 –14% 130% CS5-V-104 4. towers.5 2. and their costs follow these trends. the designed pressure is provided in the equipment data. a jump is observed for the CCEP cost in Figure 4 due to the change of stages.5 SS 75.5 7. When volume is large.841 430%* 1301%* CS7-V-101 0.All vessel pressures less than 10 barg except CS1-V-101 with pressure = 60 barg TABLE 7.0 4. In general. EconExpert gives the highest cost of the four programs. the purchase cost given in EconExpert does not include trays and packing. Since in most cases. for towers of large diameter.5 CS 9. CapCost and EconExpert predict similar purchase and total module costs while the others give much higher costs (Table 7). Hence. shell thickness is required for cost estimation in DFP. CapCost predicts a much lower cost than the others. However. When the total module cost of the tower is compared. Matem m rial CPCS Δ% Δ% CS2-V-101 0.3 4.3 3.787 552%* 180%* CS4-V-101 1.8 3. when the design pressure of the vessel is very high (CS1-V-101). For vertical vessels. PURCHASE COST (CPCS) OF VESSELS Vessel – horizontal CapCost DFP CCEP Dia. while those are included in the total module cost. The total module cost in AspenPEA is exceptionally high due to the high instrumentation cost. the total module cost in CapCost is much higher than that by the other four programs.400 –56% –13% CS7-V-102 0.8 SS 59.5 6.025 291% 923%* CS7-VS-102 0. This result is similar to what is observed in case studies. plus manholes and nozzles. CapCost predicts a very high pressure factor according to: log10 FP = C1 + C2 log10P + C3(log10P)2 (11) where C1.5 1.5 SS 97.4 6.9 CS 10.014 155%* 372%* CS6-V-101 1.. following equation from Ref.405 –44% 85% Vessel – jacketed CS7-VS-110 2.850 260% –14% Vessel – vertical CS2-V-102 0.0 SS 8.COM AUGUST 2011 .5 5. E is the weld efficiency and CA is the corrosion allowance (0.4 CS 40.9 3. CapCost predicts lower purchase and total module costs than the others (Figure 6.2 1.0 SS 4.300+ n/a+ n/a+ CS7-VS-109 2.670 19% 135% CS6-V-102 4.761 238% 765%* CS7-VS-104 1.7 8. knock-out drums. Hence.2 SS 136.7 CS 14. and the cost is lower. followed by CapCost and CCEP. DFP predicts a slightly higher cost. DFP gives a purchase cost similar to CapCost.0 SS 4.7 6. Table 7 and online table VI). CapCost. while CCEP predicts a higher cost.985 12% 191% CS5-V-101 2. CCEP’s cost of horizontal vessels increases faster with increasing volume than that by AspenPEA. A packed tower is also evaluated (Table 6). CCEP shows a relatively higher total module cost (+59%). EconExpert’s purchase cost is not compared with CapCost (Table 6). and the deviation becomes more significant as the volume and size increase.295 44% 189% CS7-VS-106 2. the CHEMICAL ENGINEERING WWW.5 CS 386.500+ n/a+ n/a+ CS7-VS-107 3.CHE. t = [PD ÷ (2Smax E – 1. such as absorption.7 CS 15. whereas AspenPEA and CCEP predict the highest costs. flash drums.100+ n/a+ n/a+ *cost of minimum size as size is less than the minimum size + not available in this program and so taken from another program EconEx. CapCost gives significantly lower costs than the other four programs. P is the design pressure in bars. for large-diameter vertical vessels (CS7-VS-101). This contributes to CTM as ten times CPCS (online table VI).6 4.591 93% 414% CS7-VS-101 3.490 38% 87% CS1-V-101 2.0 CS 38.0 34.9 SS 9.603 52% 183% CS1-V-102 1.4 CS 5.348 141% 527% CS7-VS-103 1.0 SS 2. the pump operates in two or more stages and horizontal split case. t is the shell thickness in meters. Length.5 SS 199.0035 m). CapCost predicts a much lower cost than the rest. Also.2 8. For higher power and flowrates. 1.640 –10% 93% CS6-V-103 1.4 13.9 SS 10. For both towers and vessels. The CCEP program warns the user if the pump is not operating in the appropriate stage. distillation and stripping.3 3. For vertical vessels of small diameter. The horizontal vessels evaluated in the case studies are all below 100 m3 (Table 7). all programs give similar results (online table V).212 308%* 188%* CS5-V-102/103 1.600+ n/a+ n/a+ CS7-VS-108 3. Towers Towers are vertical-pressure vessels for separation operations. CCEP predicts a much higher cost than AspenPEA. 1 is used to estimate the shell thickness.Aspenpert PEA Δ% 1% 1% 2% 4% –4% 2% 5% 27% 30% –45% 102% 58% 35% –31% 276% 71%* 194% 11% 17% –15% n/a+ n/a+ n/a+ n/a+ Δ% 21%* 89% 105% 78% 14% 104% 111% 80% 43% –28% 40%* 72% 17% –27% 151% 127% 144% 56% 26% –54% n/a+ n/a+ n/a+ n/a+ Vessels Vessels are used in chemical processing plants as reflux drums. D is the diameter of the vessel (meters).7 CS 6. For horizontal vessels of large diameter.0 3. DFP and EconExpert give similar results. For lower-power pumps (up to 55 kW and flowrates up to 57 L/s).0 3. mixing vessels and storage drums.1 SS 29. Smax is the maximum allowable working pressure of the material (bars). However. The plot of total module cost of horizontal vessels against volume (Figure 5) shows that CapCost and EconExpert predict similar total module costs for horizontal vessels.718 173% 345% CS7-VS-105 2.5 10. Both CCEP and AspenPEA predict the cost of this packed tower to be significantly higher than that by CapCost. Solano Beach. P. the costs predicted deviate from one another. In case study seven. New Jersey.” 3rd ed. Sharpen Your Capital-Cost-Estimation Skills. DFP predicts a higher total module cost for double-pipe heat exchangers Total module cost. Due to limited data.800.M.. AspenTech. For Further Reading: Brennan. Walas. Aspen Technology Inc. Aspen Process Evaluator User’s Guide. Analysis of Capital Cost Estimation Methods and Computer Programs. However.C.. L.” 2nd ed. 9. Prentice Hall (2009). Bailie. Simplified Approach to Preliminary Cost Estimates.. EconExpert and AspenPEA.000 90. Chem. (1975). The de- viation is mainly due to differences in the CPCS predicted based on different correlations in the programs. S. “Process Plant Estimating. pp.. Turton. 254256 (1999). and Agrawal.” Trans.. W.” 3rd ed.000 500. Inc.Y. Eng. New York (1984). 11. dollars 2. Filters and filtering centrifuges can only be found explicitly in CapCost. R. and Design of Chemical Processes.R. and Shaeiwitz. This is mainly caused by the different approach in calculating the purchase cost of the mill in different programs. (2009). 114-142 (1969).000 1. Comparing the results in Table 8 (and online table VII). m2 600 700 800 FIGURE 4. Dysert.” 5th ed. Seader..S. “Analysis.000. the cost of CS6-H-101 shows significant deviation for different programs. the trend of fired heater cost is not clear. Chemical Engineering Education.COM AUGUST 2011 .000 160.000 0 0 100 200 400 500 300 Volume. London (1988). C. and Golonka. 2010. B.400.000. While CapCost’s cost is low for small volumes.S. New Hampshire (2004). U.T. dollars 100.1.000 1. 80A (6) pp. 3. The rate of cost increase by CapCost is faster than that by other programs.000 1. 13. V7. AspenPEA gives high purchase and total module costs compared to CapCost and CCEP for both crystallizers and centrifuges. compared to fewer choices for CapCost and EconExpert. Guthrie K.. Huang.J. filters and jacketed and agitated reactors are discussed (Table 8). Seider. fired heaters.CHE.J.Total module cost. Dissertation. Evaluation. B. “A Guide to Capital Cost Estimation.D.000 30. Englewood Cliffs.” Butterworth.000 0 0 1 2 3 4 5 6 Area. 70-81 (2001)...000 DFP DFP CapCost EconExpert AspenPEA CCEP 2. Due to the mapping to different type of filters.000 0 0 5 CapCost EconExpert AspenPEA CCEP CCEP vertical split case CCEP horizontal split case 10 20 25 15 Power.A. Sinnott R. U. The first is a mill. New Factors for Capital Cost Estimation in Evolving Process Designs.000 1. A.000 40. K.T.. pp. Crystallizers and centrifuges are not listed in DFP and EconExpert.” 4th ed.000 1. J. Development and Application of a Capital Cost Estimation Program. G. other pieces of equipment. and Towler. dollars 80. IChemE.000.J. (2009). “Chemical Engineering Process Design and Economics – A Practical Guide. 2010. Eng. G.Eng. N. of Chemical & Biomolecular Engineering. Other equipment In this section.M. In case studies five and six. K. Chem. However. mills. Furthermore. “Product and Process Design Principles. G.600. “Financial Decision Making in the Process Industry. its predicted cost is the highest for volumes more than 200 m3.000 1. jacketed References 1. both purchase and total module costs predicted by EconExpert are half that predicted by AspenPEA. Burlington.000 1. 5.000 140.S.000 100.000 10. Y. Chem.M.000 80. whose costing is available only in EconExpert and AspenPEA.200.A. the deviation is within the acceptable range for those which are within the size range. Butterworth and Heinemann. New York (2010). Eng.. 112113 (1948). D. 55(6).000 600.500.000 70.. 28 CHEMICAL ENGINEERING WWW. W.000 20.000 20. Process Publishing.000 60.D. crystallizers.000 400. kW 30 35 40 FIGURE 3..000 0 CapCost EconExpert AspenPEA CCEP 0 100 200 400 300 Volume..B. “Chemical Engineering Design. D. National University of Singapore. Capital Cost Estimating. there are equipment types that are not present in earlier case studies. and Lewin. Wong. 10.. Ulrich.000 120. California (1974). Lang. A Software Package for Capital Cost Estimation.” John Wiley & Sons. P. National University of Singapore. 6. 8.” Craftsman. Mass.000 800. “A Guide to Chemical Engineering Process Design and Economics.500.S. D.000 60. H.000 180. all programs give similar purchase and total module costs with DFP giving a relatively higher cost than the rest. 76(6). J. DFP and AspenPEA predict higher total module costs of centrifugal pumps than other programs Total module cost. Total module cost predictions for horizontal vessels are lowest for CapCost FIGURE 6. Ulrich. Synthesis. D. m2 7 8 9 10 DFP CapCost AspenPEA CCEP EconExpert 200. Gerrard. and Vasudevan. Whiting. For CS5-H101. 33(3). Institution of Chemical Engineers (IChemE) (2000).000 40. Dissertation. Woods. fired heater cost is evaluated. Guthrie.Eng.. Dept. 7. R.000 50.000 200. 579-586 (2002). pp. D. 2. Due to its treatment of piping. Vasudevan. K.” Prentice Hall.R.D.. of Chemical & Biomolecular Engineering.L.000 DFP Cost. “Chemical Process Equipment. John Wiley & Sons Inc. m2 500 600 FIGURE 5. there are more types of filters and centrifuges to which to map the process equipment. U. Dept. dollars 2. 108 (11).M..000. such as centrifuges.. pp. In AspenPEA. The rate of increase in predicted total module cost for vertical vessels is greatest with CapCost by CapCost is lower than CCEP only. and Control. 4. U.. 12. 500 67.P.0 94% 14% 2% –11% study (Table 9 and online 7 9. Authors Ying Feng is currently a technologist at Shell Chemicals Seraya Pte. equipment capital costs for different programs may not be comparable. There is generally good agreement for the purchase costs of floating-head heat exchangers.000 170. and also received a Baden-Württemberg Scholarship during her student exchange program at Karlsruhe Institute of Technology. and is currently professor and deputy head for student and academic affairs in the Dept.9 Fired heater CS6-H-101 26. Since material.00035 Filtering centrifuges CS7-S-103 0. it is not evaluated in case four. capital cost estimates may differ.in some cases.3 27% –23% –28% –36% whole plant in each case 6 17. Singapore 119077. In case studies four and six. If a piece of in DFP are mainly due to cost differequipment had a specification below ences in vessels and packing towers. the valid range. In case study two. This is mainly due to the high costs Total purchase and module cost of towers and vessels in AspenPEA. Rangaiah has received several teaching awards. then its cost was estimated by taking the lower limit of the Conclusions valid range.600* 18. if a piece of equipment had a size the high costs of crystallizers and vesabove the valid range. IIT Kanpur and Monash University. For Total modEconAspensimilar reasons. He received bachelor’s.00002 Filters CS7-S-104 0.4 CS3-R-102 28.5 CS7-S-102 55 CS7-S-105 80 CS7-S-106 90 CS7-S-107 90 Reactor-jacketed agitated CS3-R-101 28.900+ 18.1 79% 19% –29% 28% total module costs for the 5 33. in Germany. if the flowrate Although based on different methods was less than the minimum required and developed in different platforms. for costing. and a B. Since layers of silver and are useful tools for estimating wire gauze inside the reactor cannot be the capital cost of chemical process TABLE 9.com). (economics) from the National University of Singapore in 2010. comparable (Table 8).5 6% –11% –14% 81% comparable purchase and 4 4.200 184.S. G. PURCHASE COST (CPCS) OF OTHER TYPES OF EQUIPMENT Centrifuges Dia. modeling and optimization of chemical.chase and total module costs due to grams. the into multiple units of smaller size and high deviations (more than 80%) in the costing was done by summing CTM both purchase and total module cost for multiple smaller units. the minimum flowrate was all five programs are user friendly used (not actual). For all pro. 2 1.600 303. DFP is the most limited in that regard. master’s and doctoral degrees in chemical engineering. pressure and installation factors vary in different methods in calculating total module cost. which shows higher purby the different programs. with first-class honors in both degrees. both able in CapCost and EconExpert. Δ% Δ% Δ% Δ% Case study $million excluded in the total fixed 1 4. Rangaiah worked at Engineers India Ltd.405. Ltd.700 6. His research interests are in control.sg) since 1982. there is greater deviation in both purchase and total module costs among different programs studied. Based on our analysis.CHE. TOTAL MODULE COST (CTM) OF EQUIPMENT plants.COM AUGUST 2011 29 .E. the overall plant cost does not deviate much among the different programs studied. Email: chegpr@ nus.078. for two years before his doctoral study. Hence.000 109. Edited by Scott Jenkins Editor’s note: Additional tables for total module cost of equipment are included in the online version of this article (www. He has edited two books. Also.000 Mills CS7-C-101 0. Phone: +65 6516 2187. it was divided sels. The the purchase and total module costs results obtained from these two are obtained by ApenPEA are more than 60% higher than those by CapCost. respectively. CHEMICAL ENGINEERING WWW.800 CS5-H-101 360..135 DFP Δ% n/a+ n/a+ n/a+ 63% –23% n/a+ n/a+ n/a+ n/a+ n/a+ n/a+ n/a+ n/a+ n/a+ n/a+ CCEP Δ% n/a+ n/a+ –76% 112% –15% n/a+ n/a+ n/a+ n/a+ n/a+ n/a+ n/a+ n/a+ n/a+ n/a+ EconExpert Δ% n/a+ n/a+ n/a+ –46% –13% –90% 181% –99%* –89% n/a# n/a# n/a# n/a# –70% –70% AspenPEA Δ% 49% 49% 210% 59% –35% n/a+ n/a+ –98% –47% –37% 31% 33% 33% n/a+ n/a+ *cost of minimum size as size is less than the minimum size + not available in this program and so taken from another program # purchase cost not available in EconExpert mapped to any equipment in the programs. from Andhra University. Feng’s research interests are capital cost estimation and optimization of petrochemical processes.4 Material CS CS CS CS CS SS SS SS SS SS SS SS SS CS CS CapCost CPCS 303. Rangaiah has supervised ten research fellows and 35 graduate theses. published 120 papers in international journals and presented 100 papers in conferences. Similarly.4 34% 7% 8% 14% table VIII). However.che.TABLE 8.600 139. m CS3-Ct-101 2 CS3-Ct-102 2 Crystallizers CS3-Cr-101 21. She received a scholarship for her performance at the university. We also compared total purchase and In case three.S. there is good agreement total module costs for all pieces of among different programs except for equipment in each case study given AspenPEA.600* 131.edu. petrochemical and related processes. there is significant deviation for most vessels and towers in total module cost. it is important to use only one program for cost evaluation of process design options to maintain consistent results..000 184.150 1.0075 CS7-S-101 1. with DFP giving a slightly higher cost and agitated reactors are only avail.135 109.100+ 53. including the Annual Teaching Excellence Awards from the National University of Singapore for four consecutive years.3 –6% –4% –36% –3% capital cost in case five. She received a B.0 16% 71% –5% 68% The five programs give 3 2. For fixed-head heat exchangers and pumps. the cost ule cost CapCost DFP CCEP Expert PEA of catalyst pellets is also CTM.Ch. AspenPEA has the most equipment types available out of the five programs. Rangaiah has been with the National University of Singapore (21 Lower Kent Ridge Rd. of Chemical & Biomolecular Engineering. while evaluating plant design alternatives.