'-Classroom )

USING SPREADSHEETSAND VISUAL BASIC APPLICATIONS

As Teaching Aids for a Unit Operations Course

JUAN P. HINESTROZA*, KYRIAKOS P APADOPOULOS

Tulane University • New Orleans, LA 70118

The design of separation processes frequently uses in­tensive trial-and-error procedures as well as graphi­ca l method s suc h as McCabe-Thie le, Ponch on ­

Savarit, and Triangular diagrams. Using process simulationdesign packages such as ChemCAD, HYSYS, or Aspe nPLUSfacilitates the design of complex processes, but students of­ten treat simulators as black-boxes and tend to accept the re­sults they obtain without further analysis.[II Additionally simu­lators may not provide the user with know ledge of all calc u­lations that are performed or the respective algorit hms. Onthe other hand , manual step-by-step calculations and graphi­cal methods, while allow ing students to understand the fun­damentals of the design process, do not equip them with theability to adapt software tools to the solution of chemicalengineer ing prob lems or to critically use exis ting simulationpackages. Tools such as MS Exce l Macros and small VisualBasic Applications (VBA) brid ge the gap between the previ­ous altern ative s.

At Tulane University, spreadsheets have been intensivelyused as teachin g aid s in undergraduate courses. P'" Usin gspreadsheets and VBA to solve chemical engineering prob­lems requi res a deep understandi ng of the concepts behindthe calculations, while the extensive and time-consuming trial­and-error procedures are left to the computer. The interactivenature of the spreadshee ts and VBA programs allows "w hatif ' analyses in which the parameter values are changed andthe results are immediately displa yed.!" One adva ntage ofusing spreadsheets as teaching tools is that the instructor canspend significantly more time disc ussing the fundamentalsof mass transfer and the conceptual and quantitative descrip­tion of processes, as well as the engi neering insig ht that isneeded in designin g distillation, absorption, and other sepa­ratio ns, by spending less class time on the deta ils of solvingproblems graphically and by trial-and-error!"

• Currently an Assistallt Prof essor at North Caro lina State Universi ty.

3/6

During this course, we initially lectured on the fundamen­tals of the calc ulation methods and presented illustrative ex­amples. These first exa mples were solved using the trial-and­error and graphical proce dure s. Then we presented a solu­tion to the same prob lem using spreadsheets and VBA. Wedisc ussed the details on how to prog ram the spreadsheets andelaborate the macros. Once the students had learned how touse these tools, we asked them to develop their own ExcelMacros and VBA programs to solve problems for homework.We used a process simulator (ChemCA D) dur ing some ofthe lectures and compared results obta ined from both ap­proaches. As a final project, we asked the students to create amore complex algori thm for the design of an abso rber. Wewere very pleased to see that the students had developed verycreative, user-friendly computer programs. By the end of thecourse, 76% of the students used some kind of Exce l spread­sheets and Macros to solve their homework, compared to11.5% at the beginning of the semester.

Juan P.Hinestroza is Assistant Professor at NorthCarolina State University, Department of TextileEngineering, Chemistry, and Science. He receivedhis PhD from Tulane University in 2002. His re­search interests are in the development, testing,and modeling of novel protective clothing and bar­rier materials.

Ky rlakos D. Papadopoulos is Professor ofChemical Engineering at Tulane University,hav­ing joined its tecuny in 1981 and served as De­partment Chair from 1998 to 2001. He obtainedhis DEngSc from Columbia University in 1982.His research focuses on some of the phenom­ena that are important in the separa tion, trans­port, and reaction processes of particulate sys­tems, with emphasis on drug delivery, lubricant­technology, and environmental applications.

© Copyright ChE Division of A SEE 2003

Chemica l Engineering Edt/cation

It should be noted that although we have chose n Exce l,Version 2000, for all exa mples in this paper, other spread­sheet programs such as Quattro Pro and Lotus will performequally well.

DESIGN OF ABSORPTION COLUMNSThe design of absorption using the McCabe-Thiele diagram

can be considere d as a graphical solution to a series of se­quential nonlinear equations.!" Spreadsheets have been usedin solving simultaneous nonl inear equations due to their in­corporation of a variety of mathemati ca l functio ns and theease of interactive programming, modificat ion, and rapidgraph generation.lSI

Example 10.3 from Geankopolis'f is used here to illustratethe use of spreadsheets in the design of absorp tion units. Theproblem requires removal of aceto ne from an acetone-air gasstream using water in a countercurrent stage tower. The pro­cess schematic and spreadsheet used for solving this prob-

lem are show n in Figure I. The initial data provided in theprob lem, such as the percentage of recov ery and the flowsand composi tion of the entering gas and liquid streams, areshow n in the upper portion of the spreadsheet under designparameters. Ass umptions include a constant molar overflowin the tower, negligible solubility of air in the water, and aphase equilibrium relationship that could be represented byHenry's Law. The compositions of ace tone in the liquid andvapor outlets, xN and Y N+l ' ca n be obtained from a ma ssbalance as shown in cells 0 8 and 0 9 of the spreadsheet;the eq uatio ns have been added to the respective co mmentbubb les on the gra ph.

Th e equilibrium and operating lines are plotted usingHenry 's Law and Equ ation 10.3-1 3 from Geankoplis, asshown in the 0 IS-E2S cell range of the spreadsheet and therespective comment bubb les. Using Excel' s "chart wiza rd,"an X-Y plot can be readily constructed showing the equilib­rium and operating lines.

NM

YN+l

LKH

0.0200

00050

of Stages13.61031

x

0.0100

0.0000 y--~-~~-~-~

0.0000 0.0020 0.0040 0.0060 0.0080 0.0100

> 0.0150

30

y Operating y Equilibrium0.00200.00600.01000.01400.01800.02200.02600.03000.03400.03800.0420

'--_ --=-:::::='-- ---:~0::::.0~1 0.02500.00200

COG

CALCULATING THE CONCENTRATION OF ACETONE PLATE BYPLATE

Concentration in Concentration in the Numbe rthe Uquid Phase Va por Phase of Stages

0.000000 0.00 0000

xO 0.000000 0.000000

0.000000 y1 0.002000

x1 0.00 0791 0.002000 =D32' (SDS6ISES6)+SES8

0.00 079 1 y2 0.004 372 2x2 0.001728 0.004372

0.00 1728 y3 0.007184 3x3 0.002839 0.007 184

0.002839 y4 0.0 10518 4x4 0 .004157 0.010518

0.004157 y5 0.014 472 5x5 0.005720 0.014472

0.005720 y6 0.019 161 60.007573 .019 16 1

0.000000.001330.002670.004000.005330.006670.008000.009330.010670.012000.01333

Absorption of Acetone in a Countercurrent Stage Tower

=C25'( SDS6/SES6)+SESB

CALCULAn ON OF THE OPERAn NG ANO EQUIU BRIUM U NE

B

456789

1011

12141516171819202122232425

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

Figure 1, Spreadsheet used for iJustrating absorption of acetone in a countercurrent stage tower.

Fall 2003 317

The next step involves the plate-by-plat e calculation of theconcentration of aceto ne in the liquid and vapor phases. Inhand calculations, this step is made graphically!" but in thiscase advantage is taken of the fact that both the operating andequilibrium lines are expressed as mathematical functions,so the concentration in the liquid and vapor phases can beeasily determined numerically, as shown in the D29-F42 cellrange of the spreadsheet. The (x,y) data series is added to theexisting grap h, thus comple ting a numerical McCabe-Thielediagram, also shown in Figure I.

Finally, the number of idea l stages required for the desiredpercentage of recovery is determined using the built-in"FORECAST' function, which operates as a linear interpola­tor. The series in the interpolatio n represe nts the number ofstages and the concentration of the vapor phase (columns start­ing at F29 and 029), while the value to be interpolated is thecalc ulated concentration at the exit of the tower (E7).

Once the sprea dshee t is built, several "what-if' scenarios

ca n be analyzed. For instance, in this example an increase inthe recovery requirement as well as a moderate increase inthe gas flow will readily show that the number of stages re­quired will increase significantly. Also, a large decrease inthe liquid flow rate or an increase in the gas flow rate willdemonstrate that the separation is impossible to achie ve asthe operating line and the equilibrium line cross each other.

INTERFACIAL COMPOSITIONS IN MASSTRANSPORT BETWEEN TWO PHASES

Tr ial-and-error iterati ve procedures for determining thecomposition of the interface between immiscible phases arefrequently required in mass-transfer-based separa tion pro­cesses. To demonstrate the versati lity of spreadsheets in ac­co mp lis hing th is task , we use Examp le 10.4-1 fromGeankoplis. t" as shown in Figure 2. The objective of this prob­lem is to determine the interfacia l concentrations of the va­por and liquid phases yAi and X

Ai' respectively, in a wetted-

COG K

Equili brium Datalea ya

0.000 0.0000.050 0.022

0.100 0.052

0.150 0.087

0.200 0.1310.250 0 .1870.300 0.2650.350 0.385

,----

ICLICK HERE TO

SOLVE PERFORM THEITERATION

CALCULATION

Error between the va lue of yfrom Equation 1Q-4-8 and thevalue predicted by theequ ilibrium correlation .This is the target cell for theSOLVER rout ine and theobject ive is to min imize it.

0 .19776 70.000000

0.1978050.000000

0.197928

-1.1584330.25717 1

ration Third Iteratio n Fourth Iteration0.302346 0258376 0.2571710.275350 0.200427 0.1979280.794537 0.818259 0.8189020670966 0.705984 0.7071 34

Value obtained fromsolving nume rica lly theEquilibrium equat ion andEquaUon 10-4-8

SUb INTERPHASE_COMPOSITION ()

'IN TERPHASE COMPOSITION Macro, Macro recorded 9/22J2002 by JPH

Range("G2T1SeleetSoIverOk SetCell:="$ESl 9". MaxMmVal= 3. ValueOf:=-o". ByChang e :="SESl 6"SoIverSotveSo/verOk SetC e ll:="$FSl9" , MaxMinVa l:=3, ValueOf:=V , ByChange:="SFS16"SoIverSo lveSoIverOk SetCell :="$GS19". MaxMinVal:=3. ValueOf:-V, ByChange:="SGSl 6"SoIverSotveSoiverOk SetCel l= "SHSl9" , MaxMmVal:=3. Va lueOf:=V , ByChang e :="$HS16"SolverSolve

End Sub

ky=kx=

0.340

Firs t Iteration Seccn0.40.9

0.73989 10.285002

0.290

ITERATIVE CALCULATION

0.240

x

y =8.1414x 3 .1.4766)(2+ 0.6184x · 0.00 22

R' = 0.9994

0.190

Form ulaInitial Guess

Init ial Gue ss=« 1-S0S 6Hl -El l ))/LN« 1-S0 S6)/(1-El l ))=«I-E12H l -S0S7 »)/LN«1-E12)/(1-S0S7 »

=-{SFS5IE13 )/(SFS4/E14)

=+E15"(EI6-S0S6)+SOS7=8 .14 14"E16"3-1.4766"E 16"2+O.6184"E16­0.0022=+ E17-E18 "2

Iterative calculation of interface composit ions in interphase mass transfer

Gas phase film mass t rans fer coeffi ci entLI uld phase mass transfer coefffcient

••,.11 xi12 yi

'3 (l-xai),. {t -ya t }Calcu lated Slope from

rs Equat ion 10-4-816 X equilib ium

y equi (From slope17 calculations )

y equ i (from Equilibrium

" correl ation)19 Error2.21

22

23

0.400

0.350

0.300

0.250

>- 0 .200

0.150

0.100

0.050

0.000

0.090 0.140

Figure 2. Spreadsheet and Macro used for interative calculation of interface compositions in int erph ase mass transfer.

318 Chemical Engineering Education

wall tower. Exper imen tal equil ibr ium data are provided aswell as the gas and liquid phase film mass-transfer coeffi­cien ts. In this pro blem, the solute A diffuses throu gh stag­nant B in the gas phase and then through a liquid film.

The first step in solving this problem invo lves initial guessesfor X

Aiand yAi ' In solving the probl em by hand , these guesses

are cruc ial to the rapid con vergence of the iterative process.Spreadsheet s are less sensitive to the initial guesses as a largenumber of iterati ons can be processed and visualize d in frac­tions of seconds.

In Figure 2, cells D 13-D 19 display the equations and col­umn s E - H displ ay results from four iterations. Once theiniti al guesses are selected (ce lls E l l and EI 2), the slope forthe line connec ting the bulk conce ntration and the ass umedinterfacial concentra tions is ca lculated, as show n in ce ll E 15.With the slope from EI5 and point P (the bulk concentrationin cells D6 and D7) on the x-y plot in the lower-right come rof Figure 2, an equation for a straight line is dedu ced as show nin cell D 17. A third-order polynomial was used to fit the equi­librium data (cell D 18).

Th e Excel fun ction "SOLVER' is used to solve simulta-

neou sly the equations in ce lls 0 17 and 01 8 by mini mizingthe error between the values of ce lls E I6 and EI7. SOLVERcan use a Newton or a conjugate num erical procedure to findthe answer; the defaul t Newton procedure was chosen forthis example. When co mpariso n betwee n the values for X

A1

and yAi fro m this procedure (cells E 15 and E 16) and thein iti al guesses (cells E 12 and E 13) shows a di screpanc y,an additio na l iteration is required. The latest calculatedva lues fo r xAi and yAi (cell s E 15 and E 16) are used as th enew init ia l guesses .

Due to the ease of modificat ion of spreadshee ts, the ce llscontaining the equations can be copied and pas ted into thenext co lumns as many times as necessary. In th is example,four iteration s provide a rel iable answer (less than 0.1% be­tween the latest and pen ultima te ca lculated values) .

Wh at-if ' sce narios in this example include how an increasein the liqu id-film mass-transfer coefficient will readily showthat the value for the interfacia l conce ntrations X

Aiand yAi

increase and how a large decrease in the bulk co ncentrationwill produce a significan t decrease in X

Aiand yAi' In order to

automate the iterat ion process, a MA CRO was cre ated using

c 0 E

1

1.00.8

q nne

0.60.40.2

H

1 1.0I

I0.9

0.8

0.7

0.6

I 0.5

I0.4

I0.3

0.2

0.1

SOLVING THE MASS AND CO MPONENT BALANCES TO OBA TIN D AND WSoIli erOk SetCel l:=MSF$lBM, Max MlnVaI:: 2, ValueOf ,="O", ByChange:: M$C$16$C$17"SoIver$ollieSOLVI NG THE INTERCEPT 8 ETVv'EEN THE q UN E AND THE ENR ICHING UNESoIli srOk SetCell ""-SF$27" , MaxMinVal :: 2, ValueOf.:'"O", ByChange::"$F$24"SoIli erSollie

End Sub

Stripp in Une=SCS32"B31+SCS33

-0.0330404330 .01339 19130.29 115606 5

0 81220 56611,0119189881291363898

" FOR ECAST(D7.M35. M. 37.373 3-45252 ,135:143)

Enrich ing Woe"'SCS29"B31+SCS30

0,190.2~

0,31105180-8220119

~1C;Ul.ltI09 the Inte rce pt be'tWelln tM Enr ic hi~ Indtn.qUne

.- -~

C~~~~C~~~~SS~;~::~ERAT:~ ITHE INTERCEPT BElWEEN THE a

AND ENRICHING LINE I

'--- - ~I

C.. lcu~l lng the Oi5.1Il111l<lindBon oms Flowr;lt ll

Overa. Mas.s Balance S600g8E-O~I:"OS-C'5-C16c omponent Mass Balance 2.3696E-<l5 =OS"06 ·C15"01 -e16-oaError HI9a59E-09 aE 15A 2·E16A2

xlnl en:e pty from q lineV from me enrich ing ~nll

E~,

Number of PI..,es

U de grell i line- B38

o0.080 25

0 ..850.19,

o023

051"0 ,73

""1

70. 681 210401129.41173

1.3304 04331 " (08 -E24 )1(08 ·E23 )-0 ,03 304 04 33 " E24-e31 "E23

DESIGN PARAMETE RS

Rectifi cat ion of a Benzene-Toluene Mixtu re

,", 0,wRUl lin! 1'1111 1

C~

T,T,

o00802 5

0 4850.79

1

"" 0" w17

11 qUne

~ q~ SIOpll qlin ll

21 Intercept

----,,--f,----fr----f,-

26 Enriching Line

x Y Equ ilibr ium X Stripping X Enric hing SII es"'IF{K35>l35.K35,l35) .... 3392 "130"5 · 12.539 ' 130A

" -(J3Q. SCS33 )1SC$32 =(C3Q.SJ$.30YSCS29+ 1"228"130"3 -

8."9 62 '130"2 t 3," 699 "130

--.!.!.....- stopeInlercep l

21 Str ippi ng Line

W~30 Intercept

"

~0 .1 02150 ..3-492 0.23 15716"9 0.106 30" 365 1--#-

--#- 0.23 157164; 0 " 91"09159 0.394203612 0 316 162199 2

~0.39" 2036 12 0,65753506 9 0.5HI01t 899 0,58.... , 883 7 3

+ 0 58«18637 0.79813279<t 0.62520 3331 0 ,1609 159 93 ·0,7609 15993 0.8919145" 3 0.6952" 3509 08713;3178 5-f,- 0,8773;3178 0.93 799581 0,7298805" ; 093" 99" 763 e--ij-- 09J.<t99<t763 0.96 268"833 0.143"36082 O,965856Q.41 7---f,- 0.96 58 5604 1 0.9185 9585 5 0.160 391621 0.9851 " .819 •-f,- 0.9857 ....81 9 0 99043 9152 0 ,7693001 1,000 5" 969 ,~L...!L

Figure 3. Spreadsheet and Macro used for dis tillation of a benzene-toluene mixture.

Fal/2003 319

VBA ; its text is also show n in Figure 2.

To create a user-friendly interface, a button is inserted intothe spreadsheet using the ' FORM TOOLBAR' menu fromExcel and assigning the Macro to it. The button allows theuser to run several "what-if ' scenarios by changing the de­sign param eters.

DISTILLATION

Whil e interfacial composi tion calculations used a VBAprogram and the absorption exa mple was based on cell andformula manipulation of the spreadsheet, in this exampl e acombination of both approaches is used for the design of adistillation unit. Such design is made using the McCabe-Thielediagram with special con siderations for the location of thefeed and the types of condenser and reboil er.!"

Exampl e 11.4-2 from Geank opli s' book is chosen to illus­trate use of spreadsheets in the design of distillation towers.The problem requ ires the rectification of a benzene-toluenemixture. Initial data of the problem include the flow and con­dition of the feed strea m as well as its composition. The re­flux ratio and the compositions of the distillate and bottomsare also specified. These design parameters are located in theupper portion of Figure 3 under design param eters. It is as­sumed that a constant molar overflow is present in the tower.

Solving the overall mass balance (cell F15) and a benzenemass balance (cell F16) simultaneously with SOLVER pro­vides the values for the distillate and bottom- stream flow­rates (cells CI5 and C16). In this exa mple, we take adva n­tage of the capabilities of SOLVER for mult ivariable calc u­lation s. The error cell (cell F 17) is set as the tarset cell ande ,

the SOLVER should change the values of cells C 15 and C 16until the value ofF16 becom es negligible. The multi variableoptimization capabilities of SOLVER are implicit , which isvery useful since no additional programming is required.

After calculating all flowrat es, the next step is to build theequilibrium and operating lines. The equilibrium line is con­structed using experimental equilibrium data and fitted to afifth-d egree polynomi al using the TRENDLINE option ofExcel. The "q line" is calcul ated by using a boil ing point dia­gram and the physical properties of the feed strea m. CellsB 18 to D21 show the calculations performed to obtain thevalue of q and hence the slope and intercept of the "q line."The enri ching line is con structed usin g Eq. 11.4-8 fromGeankoplis,!" as shown in cell range B26 to D27. Once theslope and intercept of the q and enriching lines are deter­mined, a numerical method is used to calculate the interceptbetwe en these two lines. SOLVER is agai n used as show n incell range E21 to 025 . Since this problem requires the use ofSOLVER twice, a VBA program is buil t and assigned to abutton so these calculations are automated with a sinale clicke

by the user. The stripping line is construc ted using the init ialconditions of the probl em and the intercept between the qand the enriching lines as shown in cell range B28 to D 30.

320

The table containing the data as well as the formu las usedto determine the equilibrium, enriching, stripping, and q linesis shown on cell range B32 to 0 38. To calculate the numb erof plates required for the recti fication, the followin g proce­dure is followed, as shown in cell range B40 to F50. Theinitial point (cell B42) corresponds to the bottoms concentra­tion, xw' yEquis calcul ated using the equilibrium equation (cellC42 ), and the equ ations for the enriching and stripping linesare used for cells D42 to E50. For every iteration an IF state­ment is used to select the larger value for x. Th is IF statementinitially selects the stripping line as the opera ting line, butonce the "q line" is reached, the enriching line becomes theoperating line. The number of plates is calculated using theFORECAST function as shown in cells E29 to 0 29. Basedon the spreadsheet, "what-if ' scenarios can be considered andthe student is able to visualize the effect of chan ges in thedesign param eters such as concentrations, flowrates, reflu xratios, etc., on the numb er of plates required for a desiredseparation. Concept s such as the pinch point and the mini ­mum reflux ratio can also be analyzed.

CONCLUSIONS

MS Excel Macros and Visual Basic for Appli cations en­hanced the educational experience of students in a j unior­level separation processes course, teaching them to developsimple software and providing them with an intermediate stepbetwe en doing hand calculations and using commerciallyavailabl e packages. Distill ation, absorption, and interfacialmass-transfer problems were solved using spreadsheets andwere incorporat ed into a web-based learnin g platform .

In addition to analyzing several "w hat-if' scenarios, theseteachin g tools can also be slightly modifie d to solve the in­verse problems. For example, in the absorption case, the num­ber of stages as well as the inlet flowrates and concentrationscan be given as design param eters, and then the students canbe asked to determine the concentration of the outlet streams.Also, in the distillation case, the numb er of plate s in the en­richin g and stripping sections can be fixed, and the studentscan be asked to determ ine the appropriate reflu x ratio andinlet flowrates to achieve a certai n degree of purity in the topor bottom streams.

REFERENCESI. Wankat, P.• "Teaching Separa tions: Why, What, When, and How?,

Chem. Eng. Ed., 35,1 68 (200 1)

2. Rives, c..and D. Lacks, "Teaching Process Control with a Numerica lApproach Based on Spread sheets, Chem. Eng. Ed.. 36, 242 (2002)

3. Mitchell. B.S., "Use of Spreadsheets in Introductory Statistics andProbability: ' Chem. Eng. Ed.. 31. 194 (1997)

4. Burns, M.. and J. Sung. "Design of Separation Units Using Spread­sheets," Chem. Eng. Ed.. 29. (1995)

5. Mackenzie, J., and M. Allen, "Mathematical Power Tools, Chem. Eng.Ed.. 32. (1998)

6. Geankoplis, c.. Transport Processes and Unit Operations, 3rd ed.,Prentice Hall PTR ( 1993)

Chemical Engineering Educat ion