Degumming Material Balance

Paper Number: MB04-304

Process Modeling Approach for Evaluating the Economic Feasibility of Biodiesel Production 1,2

Dhruv Tapasvi, 1Dennis Wiesenborn and 2Cole Gustafson 1

Department of Agricultural and Biosystems Engineering 2

Department of Agribusiness and Applied Economics North Dakota State University Fargo, ND 58105 Written for presentation at the 2004 North Central ASAE/CSAE Conference Sponsored by the Manitoba Section of CSAE Winnipeg, Manitoba, Canada September 24-25, 2004

Abstract. A biodiesel process model was developed by incorporating the process-engineering principles of mass and energy balances and with the help of spreadsheets and other software tools. The basis of the model is a production facility that uses a continuous transesterification process with two stirred tank reactors, methanol as alcohol and sodium methoxide as catalyst. The model quantifies all the inputs and outputs in the biodiesel production process. It can be used to compare various vegetable oils used for biodiesel production in terms of process inputs and outputs and it can be correlated to the economic cost data for performing various biodiesel economic feasibility studies. Keywords. Biodiesel, process model, spreadsheet, glycerol refining, methanol recovery, ester washing

The authors are solely responsible for the content of this technical presentation. The technical presentation does not necessarily reflect the official position of the American Society of Agricultural Engineers (ASAE), and its printing and distribution does not constitute an endorsement of views which may be expressed. Technical presentations are not subject to the formal peer review process by ASAE editorial committees; therefore, they are not to be presented as refereed publications. Citation of this work should state that it is from an ASAE meeting paper EXAMPLE: Author's Last Name, Initials. 2003. Title of Presentation. ASAE Paper No. 03xxxx. St. Joseph, Mich.: ASAE. For information about securing permission to reprint or reproduce a technical presentation, please contact ASAE at [email protected] or 269-429-0300 (2950 Niles Road, St. Joseph, MI 49085-9659 USA).

Introduction Numerous factors have pushed energy from biomass into the forefront of policy and industry discussions. Large harvests of traditional crops, low farm prices, dependence on foreign energy sources and environmental problems due to combustion of fossil fuels have increased interest in renewable energy sources such as biodiesel. Biodiesel is a diesel-fuel replacement produced from domestic renewable resources such as vegetable oils, animal fats and recyclable cooking oils. Chemically, biodiesel is defined as the mono-alkyl esters of long-chain fatty acids derived from lipid sources. The demand for biodiesel is constantly on the increase as more persons are becoming aware of its potential to address the above issues. In order to meet the anticipated demand, more biodiesel production plants have to be set-up in the near future. The foremost step for setting-up a biodiesel production plant in a particular region is the evaluation of economic feasibility of biodiesel production. The economic feasibility study must be based on authentic and reliable production data. Various process engineering principles, such as mass and energy balances, can be utilized to collect the production data by analyzing the biodiesel production process in terms of the various inputs, outputs and their compositions. Such information is, in turn, needed for quantifying utility requirements and sizing equipment. Thus, performing the mass and energy balance analysis of the full process is a key step towards obtaining economic cost data for performing the economic feasibility studies. Efforts have been reported to quantify the biodiesel production process (Sheehan et al., 1998; Kleber, M. 2004), but no clear and transparent approach has 1

been utilized to illustrate the detailed inputs and outputs involved in the process. Many economic feasibility study groups have applied the so-called “black box” approach by listing the data for various inputs and outputs of biodiesel production collected from private consultancy firms. Such approaches leave the readers with little understanding of the underlying calculations, and no basis for determining whether advances in process technology were taken into consideration. In order to resolve the above issues and in effort to provide a complete understanding of the biodiesel production, a detailed biodiesel process model is needed. Such a process-modeling approach has been utilized extensively by chemical and food process industries for detailed design and feasibility evaluation of a production plant. One excellent example of a process model was shown by Fryer et al., 1997 by modeling a cheese manufacturing plant using spreadsheets. The basic idea of development of a biodiesel process model on spreadsheets came heavily from the idea behind this cheese model. The objectives of this present study include: Objective 1: Selection of the biodiesel production process based on comprehensive literature survey and by visiting current biodiesel production plants. Objective 2: Developing a biodiesel process model based on the selected process by incorporating the process-engineering principles of mass and energy balances and with the help of computer spreadsheets and other software tools. This includes specifying user-inputs in the model, referenced assumptions made for mass and energy balance equations and the detailed calculations performed for quantifying various biodiesel process inputs and outputs.

2

Biodiesel production process description The continuous base-catalyzed process was selected for the process model development (Figure 1), based on a detailed literature review for biodiesel production (Freedman et al., 1984; Freedman et al., 1996; Noureddini et al., 1997; Darnoko et al., 2000; Van Gerpen et al., 2003; Sheehan et al., 1998) and by visiting current biodiesel production plants. This is the most widely used biodiesel process in the European Union and US. The biodiesel production process was divided into four main sections: (a) Crude oil degumming, refining and drying: The crude oil is subjected to acid degumming for removing hydratable and non-hydratable phosphatides followed by alkali refining for removing the free fatty acids. Crude oil (1) is heated to 70oC in the heater (A). Phosphoric acid (2) is added to the heated crude oil in the mixing tank (B) for converting the non-hydratable phosphatides to water-soluble phosphatidic acid. Soft water (4) is added to the mixing tank for the formation of gums from the hydratable phosphatides. The mixing tank outstream (5) is centrifuged (C) to separate the oil from the gums-water mix (6). Using a separator (D), the gums (7) are removed from water (8). The degummed oil (9) is sent to a refining tank (E) for alkali refining. Sodium hydroxide solution is added for converting the free fatty acids (FFA) present in the degummed oil to oil-insoluble soaps. Proper mixing is ensured for the reaction. This is followed by the addition of wash water for dissolving the soaps, which results in the formation of soapstock that is removed from the oil using a centrifuge (F). The resulting centrifuge outstream (14) goes to a vacuum oil dryer (G) to remove the

3

CRUDE OIL DEGUMMING

REFINING

10 3

11

15

4

Crude oil 1

A

2

B

5

C

G

E

9

12

14

F

16

H

17 Refined oil

6 13

D

7

8 28

TRANSESTERIFICATION 18 19

ESTER WASHING

23 24

34

M 29

17

I

20

22

K

25

N

27

J

31

L

21

26

33

O

30

P

35 Biodiesel

32

METHANOL RECOVERY 36 Q

37

T

S

R

40

42 Recycled Methanol

U

39

38

41 43

47

45

GLYCEROL REFINING V

44

W

46

48 X

Glycerol

Figure 1. Continuous base-catalyzed biodiesel production process A, heater; B, mixing tank; C, centrifuge; D, gums/water separator; E, refining tank; F, centrifuge; G, vacuum oil dryer; H, surge tank; I, continuous stirred tank reactor (CSTR) 1; J, decanter 1; K, CSTR 2; L, decanter 2; M, heater; N, wash columns; O, settler tank; P, vacuum ester dryer; Q, collecting tank; R, heater; S, glycerol-alcohol stripper; T, distillation column; U, condenser; V, glycerol hold tank; W, acidulation reactor; X, decanter. (Continued next page)

4

(Continued Figure 1). 1, crude oil; 2, heated crude oil; 3, phosphoric acid; 4, soft water; 5, mixing tank outstream; 6, gums-water mix; 7, gums; 8, water; 9, degummed oil; 10, NaOH solution; 11, wash water; 12, refining tank outstream; 13, soapstock; 14, refined oil; 15, water vapor; 16, hot oil; 17, dried refined oil; 18, sodium methoxide; 19, methanol; 20, CSTR 1 outstream; 21, glycerol phase; 22, ester phase; 23, sodium methoxide; 24, methanol; 25, CSTR 2 outstream; 26, glycerol phase; 27, ester phase; 28, wash water; 29, heated wash water; 30, waste stream; 31, washed esters; 32, aqueous phase; 33, esters; 34, water vapor; 35, biodiesel; 36, glycerol/aqueous phase; 37, heater outstream; 38, super heated steam; 39, saturated methanol vapors and saturated steam; 40, methanol vapor; 41, distillation column bottoms; 42, recycled methanol; 43, hot glycerol solution; 44, glycerine; 45, HCl solution; 46, acidulation reactor outstream; 47, waste; 48, product glycerol. remaining water in the oil. The dried degummed and refined oil is sent to a surge tank for cooling. (b) Transesterification reaction: The refined oil stream (17) enters the CSTR 1 (I) maintained at 65oC. Typically 100% excess methanol is added to the reactor along with the suitable amounts of the catalyst sodium methoxide. Transesterification reaction between triglycerides and methanol takes place in the presence of catalyst sodium methoxide to form methyl esters i.e. biodiesel and glycerol, a co-product. Also, the remaining trace amount of free fatty acids in the refined oil reacts with sodium methoxide to form soaps and methanol. The reaction products (20) are separated using decanter 1 (J) into a glycerol phase (glycerol, methanol, sodium methoxide, soaps) and ester phase (methyl esters, unreacted oil, methanol, soaps). Further, the ester phase (22) enters the CSTR 2 (K) and the glycerol phase goes to a collecting tank (Q). A similar process occurs in CSTR 2 (K) and decanter 2 (L), typically using 100% excess methanol, based upon the unreacted triglyceride present. The ester phase from decanter 2 (L) goes to the ester-washing section and the glycerol phase enters the collecting tank (Q).

5

(c) Ester washing: Impurities in the ester phase (27), such as methanol, soaps and free glycerol have to be separated from the methyl esters. This is done by washing the ester phase (27) by warm water (29) in wash columns. The resultant waste stream (30) is sent to the collecting tank (Q) and the washed ester stream (31) to a settler tank (O). The remaining aqueous phase is separated from the methyl esters in the settler tank (O). Finally, the ester stream (33) is vacuum dried to remove any trace amounts of moisture in the vacuum ester dryer (P). (d) Methanol recovery and glycerol refining: Streams 21, 26, 30 and 32 are combined in the collecting tank (Q). The resulting stream (36) is heated to the normal boiling point of methanol (65oC) in the heater (R). The methanol is stripped from the heated stream (37) using super heated steam (38) in the glycerol-alcohol stripper (S). The saturated methanol vapors and the steam are fed into a distillation column to recover pure methanol vapors as distillate (40). The methanol vapors (40) are condensed in a condenser (U) and are recycled back (42) to the CSTRs. Bottoms (41) of the distillation column contain the glycerol, condensed steam and other impurities. The hot glycerol solution (44) from the bottom of the stripper (S) is sent to a glycerol hold tank (V). The crude glycerol (44) from this hold tank is mixed with proper amounts of HCl solution in the acidulation reactor (W). Catalyst sodium methoxide in the stream (44) reacts with HCl to form methanol and NaCl and the soaps present in this stream react with HCl to form free fatty acids and NaCl in the acidulation reactor (W). Using a decanter (X), product glycerol (48) is separated from the free fatty acids and other impurities such as unreacted oil (47).

6

Biodiesel process model development User-specified inputs Following are the user-specified input parameters required by the spreadsheet model: (a) Desired transesterification reaction efficiency. Default value is 98%. (b) Amount of crude oil to be processed per day (kg/s). (c) Methanol/triglyceride mole ratio. Default value is 6 (100% excess methanol of that required by the reaction stoichiometry).

Mass balance assumptions (a) Degumming/refining section i. Mixing tank (B): Heated crude oil (2) is mixed with 0.1% percent of 0.85wt% phosphoric acid aqueous solution (3) (Hernandez et al., 1996) followed by the addition of soft wash water (4) equal to 75% of the phosphatide content in the crude oil (1) (Erickson., 1995). ii. Centrifuge (C): All phosphatide in the form of gums, all unreacted phosphoric acid and 99.5% of the stream 4 is recovered in stream 6 (Sheehan et al., 1998). iii. Refining tank (E): 9.5 wt% NaOH aqueous solution equal to 113% excess of that required for stoichiometric conversion of free fatty acids present in stream 9 is added. Wash water in the form of soft water equal to 15% of the mass flow rate of stream 9 is added (Sheehan et al., 1998). 99% of the free fatty acids are converted to soaps by reaction with NaOH. iv. Stream 13: contains soaps formed from the reaction of free fatty acids and NaOH in the form of soapstock, triglycerides equal to 2.5 times the amount of free fatty acid loss, unsaponifiable matter attached to triglycerides, in the same proportion

7

as present in the original oil (0.015:1), and 99.5% of the stream 11 (Sheehan et al., 1998). v. Vacuum oil dryer (G): 100% of the water removal from stream 14 is achieved. (b) Transesterification section i. CSTR 1 (I): sodium methoxide catalyst (18) equal to 1% of the stream 17 (Myron Danzer, personal communication. 2004.) is added in the form of a 10% solution in methanol. Methanol is added based upon the methanol: TG ratio being input. 85% transesterification reaction efficiency is assumed (Van Gerpen et al., 2003). Also, all the free fatty acids in the oil react with the catalyst to form soap and methanol. ii. Stream 21: contains 60% of the total methanol in stream 20, all glycerol and sodium methoxide in stream 20, and 10% of the total amount of soaps present in stream 20 formed from the reaction between free fatty acids and sodium methoxide in the CSTR 1 (I) (Van Gerpen et al., 2003). iii. CSTR 2 (K): sodium methoxide catalyst equal to 1% of the triglyceride left unreacted in the stream 22 is added in the form of a 10% solution in methanol (Myron Danzer. personal communication. 2004). Methanol is added based upon the methanol: TG ratio being input. Calculations are based upon the total transesterification reaction efficiency being input by the user. iv. Stream 26: contains 60% of the total methanol in stream 20, 10% of the total amount of soaps in stream 25, and all the glycerol and sodium methoxide in stream 25 (Van Gerpen et al., 2003).

8

(c) Ester washing section i. Stream 28: It is equal to the 20% of the methyl esters amount in stream 27 (Sheehan et al., 1998). ii. Stream 30: Contains 90% of the stream 28 and 100% of the methanol and soaps contained in stream 27. No ester is lost (Sheehan et al., 1998). iii. Stream 33: Only 0.5% of the stream 28 is remaining (Sheehan et al., 1998). iv. Vacuum ester dryer (P): 100% moisture removal. (d) Methanol recovery and glycerol refining section i. Glycerol-alcohol stripper (S): Super-heated steam is at 1 bar pressure and 180oC (Myron Danzer. personal communication. 2004). ii. Stream 39: 100% recovery of the saturated methanol vapors and saturated steam. iii. Methanol recovery distillation column (T): Stream 40 contains methanol vapors with 0.05% moisture level and stream 41 contains less than 0.5% methanol (Sheehan et al., 1998). iv. Stream 45: 10% aqueous HCl solution equal to 50% mass flow rate of the glycerine stream 44 (Sheehan et al., 1998). v. Acidulation reactor (X): Sodium methoxide reacts with HCl to form methanol and NaCl. Soaps react with HCl to form free fatty acids and NaCl. vi. Stream 48: Product glycerol solution is obtained from the acidulation decanter.

9

Sample mass balance calculations

Mass balance calculations involve the use of the basic conservation of mass equation, which states that the total mass of the components in a process has to be conserved. Mass balance equations were applied to each individual unit for identifying the components in each stream in the whole production process with the assumption that no net depletion or accumulation occurred and all mass was accounted for in Fig. 1 flows. Centrifuge: Refer to Unit C in Fig. 1. Stream 5 (mixing tank outstream) is separated into stream 6 (gums-water mix) and stream 9 (degummed oil). Overall mass balance equation is represented by M5 = M6 + M9. Component balances include (a) Water (w) balance: M5w = M6w + M9w, (b) Triglyceride (t) balance: M5t = M6t + M9t, and (c) Gums (gu) balance: M5gu = M6gu + M9gu. where Mi is the mass flow rate (Kg/s) of any stream i and Mij is the mass flow rate of any component j in stream i (Kg/s). Continuous stirred tank reactor (CSTR) 1: Refer to Unit I in Fig. 1. Stream 17 (refined oil) is mixed with stream 18 (catalyst sodium methoxide) and stream 19 (methanol). Following two reactions take place to yield the stream 20 (CSTR 1 outstream).

10

1. Reaction a: Transesterification reaction between triglyceride and methanol to produce methyl esters known as biodiesel and glycerol in the presence of catalyst sodium methoxide. Reaction efficiency is user-defined. CH2COOR1

CH2OH

CHCOOR2

+

3 CH3OH

3 RCOOCH3

+

CH2COOR3 (Triglyceride)

CHOH CH2OH

(Methanol)

(Methyl esters)

(Glycerol)

2. Reaction b: Reaction between free fatty acids (FFA) in the refined oil with sodium methoxide to produce soaps and methanol. Reaction efficiency is assumed to be 100% (transesterification section assumption i). RCOOH + NaOCH3

RCOONa + CH30H

(FFA)

(Soaps)

(Sodium methoxide)

(Methanol)

Overall mass balance equation: M17 + M18 + M19 = M20

(1)

Components balances: Triglyceride (t): M20t = M17t - [(Reaction a efficiency/100)*(M17t)]

(2)

Methyl esters (me): M20me = {[(M.Wme*3)/(M.Wt)]*[(Reaction a efficiency/100)*(M17t)]}

(3)

Glycerol (g): M20g = {[(M.Wg)/(M.Wt)]*[(Reaction a efficiency/100)*(M17t)]}

(4)

Free fatty acids (f): M20f = M17f - [(Reaction b efficiency/100)*(M17f)]

(5)

11

Methanol (m): M20m = [M18m + M19m] - [(M.Wm*3)/(M.Wt)]*[(Reaction a efficiency/100)*(M17t)] - {[(M.Wm)/(M.Wf)]*[(Reaction b efficiency/100)*(M17f)]} (6) Sodium methoxide (sm): M20sm = M18sm - {[(M.Wsm)/(M.Wf)]*[(Reaction b efficiency/100)*(M17f)]} (7) Soaps (s): M20s = {[(M.Ws)/(M.Wf)]*[(Reaction b efficiency/100)*(M17f)]}

(7)

where

Mi is the mass flow rate (Kg/s) of any stream i, Mij is the mass flow rate of any component j in stream I (Kg/s) and M.Wj be the molecular weight of component i. Mass flow rates of all other components remain the same.

Super-heated steam (su) calculations in glycerol-alcohol stripper: Refer to Unit S in Fig. 1. Stream 38 (Super-heated steam) is required to vaporize all the water and methanol from glycerol present in stream 37. Stream 39 contains the saturated methanol vapors, water vapor and saturated steam vapors. Stream 44 is crude glycerol solution with some impurities (sodium methoxide and soaps). M38su = {[(M37w)*(Cpw)*(

tw)] + [(M37w)*(Lw)] +[(M37m)*(Cpm)*(

[(M37m)*( Lw)]}/(Hsu - Hsa)

tm)] +

(8)

where Cpw, Cpm are the specific heats of water and methanol respectively,

12

tw,

tm are the temperature differences required for water and methanol to reach

their respective normal boiling points, Lw, Lw are the latent heats of vaporization of water and methanol respectively and, Hsu and Hsa are the enthalpies of super-heated steam and saturated steam respectively.

Discussion Figure 2 shows the basic structure of the model as seen from a portion of the spreadsheet screen. The specified user inputs of the desired transesterification efficiency, crude oil to be processed and methanol: triglyceride ratio is given at the top left hand side of the screen. This portion of the spreadsheet lists only 10 from a total of 48 streams identified in the actual model whereas all the components in the various streams have been shown. An example of the various process inputs and outputs identified by this model have been illustrated in Table 1 by choosing an arbitrary basis of 100 kg/hr crude oil entering the production plant. All the calculations change according to the specified inputs. It provides the reader with a clear understanding of how the amounts of various process inputs and outputs are inter-related. As the model has been based upon the compositional data of vegetable oils, it can be used for comparisons between different vegetable oils used for biodiesel production in terms of various process inputs and outputs. Table 1 also shows an example of this type of comparison between soybean oil and canola oil.

13

Figure 2: Screen view of the first 10 screens in the biodiesel process model spreadsheet

14

Table 1: Biodiesel process inputs and outputs provided by the model with basis of 100 Kg/hr crude oil processed Soybean oil (Kg/hr)1

Canola oil (Kg/hr)2

Process Inputs Crude oil (1)

100.00

100.00

Methanol (19 + 24)

13.84

14.31

Sodium methoxide (18 + 23)

10.84

10.98

NaOH (9.5 wt% aqueous

14.93

14.86

HCl (10% aqueous soln.) (45)

6.24

6.12

Process water (4 + 11 + 28)

34.69

34.49

Biodiesel (35)

92.81

94.04

Methanol recycled (42)

13.36

13.70

Glycerol (48)

10.28

10.53

Waste (6 + 13 + 41 + 47)

55.57

54.05

soln.)(10)

Process Outputs

1

Composition: 96.0% triglycerides, 0.5% free fatty acids, 2.0% phosphatides, and 1.5% others (unsaponifiable matter) (Erickson.1995). 2 Composition: 97.25% triglycerides, 0.5% free fatty acids, 1.25% phosphatides, and 1.0% others (unsaponifiable matter) (Gunstone.2002).

Conclusion The present biodiesel production model was developed for analyzing the biodiesel production and ultimately designing its production unit. The model can be utilized in performing economic feasibility studies of biodiesel production in different regions simply by linking it to the economic cost data. This paper deals mainly with the mass balances calculations including the super-heated steam calculations in the 15

glycerol-alcohol stripper (S) that were necessary for satisfying the conservation of mass principle. A separate spreadsheet has been included in the model that deals with the energy balance calculations based on the conservation of energy principle (energy is conserved here similar to mass conservation principle) for calculating the process steam requirements of the biodiesel production process but is out of the scope for the present paper. Once a particular capacity of the biodiesel production unit has been selected (say 15 million gallons per year), the model can also be utilized in the process equipment design of the various equipments involved in the production process based on the various stream mass flow rates and the desired process conditions, ultimately leading to the detailed biodiesel process plant design.

References 1. Danzer, M. 2004. Personal Communication. Ralston, Iowa.: West Central Cooperative. 2. Darnoko, D., and M. Cheryan. 2000. Kinetics of Palm Oil Transesterification in a Batch Reactor. Journal of American Oil Chemists Society. 77(12): 1263-1267. 3. Erickson, D.R. 1995. Practical Handbook of soybean processing and utilization. AOCS press. 4. Freedman, B., E.H. Pryde, and T.L. Mounts. 1984. Variables affecting the yields of fatty esters from transesterified vegetable oils. Journal of American Oil Chemists Society. 61(10): 1638-1643. 5. Freedman, B., R.O. Butterfield, and E.H. Pryde. 1996. Transesterification Kinetics of Soybean Oil. Journal of American Oil Chemists Society, Vol. 63(10): 1375-1380. 6. Fryer, P.J., D.L. Pyle, and C.D. Rielly. [1997]. Chemical engineering for the food industry. Blackie Academic & Professional. New York.

16

7. Gunstone, F.D. 2002. Vegetable oils in food technology: Compostion, properties and uses. Blackwell Publishing, CRC Press, UK. 8. Hernandez, E., and E.W. Lusas. 1996. Practical short course in Processing of Vegetable Oils. College station, Texas.: The Texas A&M University System. 9. Kleber, M. 2004. Economics of a biodiesel plant. Presentation, American oil chemists’ society annual meeting 2004. Available at: http://www.lurgipsi.com/files/Lurgi%20AOCS%202004%20Presentation.PDF 10. Noureddini, H., and D. Zhu. 1997. Kinetics of Transesterification of Soybean Oil. Journal of American Oil Chemists Society. 74(11): 1457-1463. 11. Sheehan, J., V. Camobreco, J. Duffield, M. Graboski, and H. Shapouri. 1998. Life cycle inventory of biodiesel and petroleum diesel for use in an urban bus. NREL/SR-580-24089. Golden, Colorado.: National Renewable Laboratory. 12. Van Gerpen, J., D. Clemens, G. Knothe, B. Shanks, and R. Pruszko. Biodiesel Production Technology. 2003. Biodiesel Workshop. Ames, Iowa.: Iowa State University.

17