CHEMICAL REACTION ENGINEERING HANDBOOK OF SOLVED PROBLEMS Stancey M. Wacas | Gordon and Breach Publishers eChemical Reaction Engineering Handbook of Solved Problems Stanley M. Walas University of Kansas, Lawrence Gordon and Breach Publishers ‘Australia + Austria * Belgium * China + France + Germany India + Japan * Luxembourg + Malaysia * ‘Netherlands * Russia * Singapore + Switzerland + Thailand * United Kingdom + United StatesCONTENTS Preface x List of Sources of Some Problems xi List of Symbols will CHAPTER 1. MATHEMATICAL AND NUMERICAL METHODS, seeienieennnennse L ‘THEORY 1 12 13 14 1s 16 7 18 Sonware.. sn Plotting and curvefitting.. Regression of tabular date Roots of numerical equations .. Integration and differentiation Ordinary differential equations... ‘Numerical solution of ODEs Partial differential equations Tables . PROBLEMS P'1.00.00 2 3 4 5. CHAPTER 2. REACTION RA’ Index of Reactions in Problem Section P2.02.XX .. Regression . Roots Integration and differentiation. Analytical solution of ordinary ditferential equations Numerical solution of ondinary differential equations. AND OPERATING MOD§S..... Index of Mechanisms in Problem Section P2.03.XX onan ee sone 33 THEORY 21 22 23 24 25 26 27 Rate of reaction... Concentration, mols, partial pressure, mol fraction Reaction time ou a Constants of the rate equation. Multipie reactions and stoichiometric balances Reactor operating modes Heterogeneous reactions, Rates of diffusion... Tebles ..vi CONTENTS, PROBLEMS 2.00.00 1. Stoichiometric balances 2. Integrated rate equations 3. Complex reaction mechanisms . 4. Numerical integrations . CHAPTER 3. TREATMENT OF EXPERIMENTAL DATA... ‘THEORY 3.1 Kinds of laboratory reactors . 3.2. Chemical composition... 3.3. Power law rate equations... ss 3.4 Rate equations for multiple mechanisms B.S Pressute $208 ve nsne 3.6 Other properties related to composition . 3.7. Temperature variation 3.8 Homogeneous catalysis. 3.9. CSTR data. Liquid or gas phase 3.10 Plug flow reactors. Laminar flow Figures PROBLEMS P3.00.00 1. Units g 5 2 Pressure changes Variables related to composition .. Half time and initial rate data .... ‘Temperature variation, Activation energy Homogeneous catalysis . Enzyme and solid catalysis Flow reactor data 10. CSTR data 11. Complex reactions .. Senawaen CHAPTER 4. IDEAL REACTORS ....... THEORY 4.1 Equilibrium constants .. 42. Temperature change and heat of reaction .. 4.3 Batch reactors ven 44 Continuous sitred tank reactors (CSTR) 4.5. Tubular and packed flow reactors 4.6 Recycle and separation modes. 4.7 Temperature changes 4.8 Laminar and power law flows ..... 4.9. Unsteady conditions with accumulation terms 4.10 Multiple steady states 4.11 Optima ... 4.12 Economic balance .. Figures.CONTENTS vii PROBLEMS P4.00.00 1, Equilibrium constants ..... 2. Temperature change and heat of reaction. 3. Batch reactors... 4. Continuous sitrred tank reactors (CSTR) 5. Tubular and packed flow reactors... 6 1, 8 98 283 Recycle and seperation modes... ‘Temperature changes sm Laminar and power law flows... Unsteady conditions with accumulation terms. 10, Multiple steady states 11. Optima 12. Economic balances... 370 383, 410 AIT 444 454 an CHAPTERS. REACTOR FFFICIENCY 1 489 THEORY $1 Introduction ... ve 1 89 5.2. Tracers, os nn 490) 53 Reactor efficiency 1 490 5.4 Tracer response vn 490 5.3. Tracer equations... 494 5.6 Characterization of curves 496 3.7 Chemical conversion... 498 5.8. The dispersion model nan nnn . 500 Figures and Tables... nnn a 503 PROBLEMS P5.00.00 L. Tracer response functions... 509 2) Correlations: Gamma, Gauss et. 331 3. Tracer response to combined elements ... 546 4, Conversion with known flow pattems. Laminar flow . 537 5. Segregated flow, part I. 568 6. Segregated flow, part II 584 7. Maximum mixed flow 602 8. Dispersion model 607 ‘CHAPTER 6. REACTIONS WITH SOLID CATALYSTS. fesse severe 3D THEORY 6.1 Catalytic processes. 6.2 Power law equations 63. Langmuir-Hinshelwood processes 64 Physical properties of granular catalysts. 65 Adsorption equations 66 Extensions of the Langmuir equation. os 6.7 Rate when adsorptive equilibrium is maintained .. 68 Rate when chemical equilibrium is maintained 69 Find the constants of a rate equation .... 6.10 Interpretation of data wnsnnnes 6.11 Severs! controlling steps... 6.12 With diffusional resistanceviii Figures and Tables «soso PROBLEMS P6.00.00 CHAPTER 7. REACTIONS WITH POROUS SOLID CATALYSTS... Physical properties of catalysts nr Adsorption «ers Reaction mechanisms... Finding rate equations Using rate equations, Diffusional resistance. THEORY 7A 72 13 14 15 16 17 18 PROBLEMS P7.00.00 Particles and pores ..... Diffusion and diffusivity Equations of diffusion and reaction ...00 Diffusion and reaction on nonpermeable catalysts. Diffusion and reaction in pores. Effectiveness... External and internal diffusion... . Variable temperature... Deactivation of catalysts . Particles and pores Diffusion Concentration profiles and effectiveness Conversion and reactor sizing . With external diffusion . Effects of temperature and deactivation CHAPTERS. MULTIPLE PHASE REACTIONS... 798 ‘THEORY 8.1 Axial and radial gradients in packed beds seven T9B 8.2 Gas-liquid reactions ....... 83 a4 as Liguid-liquid reactions... Gas-liquid-solid reactions... Biochemical reactors... Figures and Tables PROBLEMS P8.00.00 1 2. Gas-liquid reactions 3. Gascliquid-solid reactions... 4. Biochemical reactions ~ Index of Substances Index of Subjects secs Packed bedsPREFACE ‘This book isa collection of solved problems in elementary chemical reaction kinetics from an engineer 's point of view. Brief statements of definitions and theory begin each chapter, but detailed derivations of major formulas largely are left as problems to be solved. Some of the problems are original to the extent that “originality” is possible in an area that has been worked over for so many years. Those taken from the literature and textbooks often are modified and provided with solutions when they are not in the original, which is usually the case. The List of Sources of Some Problems may be consulted for additional problems, although solutions usually are not provided in these references. Standard textbooks also should be consulted for systematic parallel or fuller treatment of some of the theory covered by this collection, For the most part, the problems are quite short, ‘The object has been to present a large variety of problems and exercises, including some replication for purposes of drill. Accordingly, comprehensive design-type problems largely have been avoided—those requiring detailed knowledge and data of other areas of engineering such as mass transfer ‘or fluid dynamics or economics. The topics included are identified by chapter headings and subheadings. Briefly, they cover rate equations, analysis of rate data, sizes and performance of ideal reactors, residence time distributions and non-ideal models, solid catalyzed reactions, behavior of porous catalysts, and reactions involving mutliple phases. Solutions are presented in the form of equations, tables, and graphs—most often the last. Serious numerical results ‘generally have to be obtained with computersor powerful calculators. The introductory chapter describes the numerical procedures that are sequired. Inexpensive software bas been used here for integration, differentiation, nonlinear equations, simultaneous equations, systems of differential equations, data regression, curve fitting, and graphing. Students, other beginners in chemical engineering kinetics, and possibly experienced engineers who may wish to review their musty knowledge should be able to profit from the opportunites of “learning by doing” provided by this collection. Reactors are of course the basic equipment in any chemical plant, The large variety of substances that have been used in the research cited in the problems emphasize this point. Also cited are the many different kinds of equipment, analytical techniques, and methods of data analysis that have been used. The Indexes of Substances and Subjects are the keys to this information. Inspiration for the material in this book was provided by several generations of students; { bope that the current generation may find it of value. The word processor ChiWriter of Horstmann Software, Inc. was used. I appreciate the interest shown by Dr. Reza Shams and Professor Bala Subramaniam during the often tedious course of preparation of the manuscript. My parents Stanislaus and Apolonia and my wife Suzy Belle were with me in spiritSOURCES OF SOME PROBLEMS Aris, R, Introduction to the Analysis of Chemical Reactors, Prentice-Hall, Englewood Cliffs, NJ, 1965. Aris, R, Elementary Chemical Reactor Analysis, Prentice-Hall, Englewood Cliffs, NJ, 1969. Bares, J, Collection of Problems in Physical Chemisiry, Addison-Wesley, Reading, MA, 1962: Batalin, GT, Collection of Examples and Problems in Physical Chemistry, Kiev University Publishers, Russia, 1960. But, JB, Reaction Kinetics and Reactor Design, Prentice-Hall, Englewood Cliffs, NJ,1980. Capellos, C, BHI Bielski, Kinetic Systems: Mathematical Description of Chemical Kinetics in Solution, Wiley, New York, NY, 1972. Carberry, 15, Chemical and Catalytic Reaction Engineering, McGraw ill, New York, NY, 1976, Chen, NH, Process Reactor Design, Allyn & Bacon, 1983. Cooper, AR, GV Jeffieys, Chemical Kinetics and Reactor Design, Prentice-Hall, Englewood Cliffs, NJ, 1971 Espenson, JH, Chemical Kinetics and Reaction Mechanisms, McGraw-Hill, New York, NY, 1981. Fogler, HS, Elements of Chemical Reaction Engineering, Prentice-Hall, Englewood Cliffs, NJ, 1992. Froment, GF, KB Bischoff, Chemical Reactor Analysis and Design, Wiley, New York, NY, 1990. Fromherz, H, Physico Chemical Calculations in Science and Industry, Butterworths, Newton, MA, 1964 Brost, AA, KG Pearson, Kinetics and Mechanism, Wiley, New York, NY, 1961. Griffiths, PIF, JDR Thomas, Cafculations in Advanced Physical Chemistry, Edward Arnold, London, UK, 1983. Guggenheim, EA, JE Prue, Physicochemical Calculations, tnterscience, New York, NY, 1955. Hill, CG, An Introduction to Chemical Engineering Kinetics and Reactor Design, Wiley, New York, NY, 1977 Holland, CD, RG Anthony, Fundamentals of Chemical Reaction Engineering, Prentice-Hall, Englewood Cliffs, NJ, 1989, Levenspiel, 0, Chemical Reaction Engineering, Wiley, New York, NY, 1972. Pannetier, G, P Souchay, Chemical Kinetics, Elsevier, New York, NY, 1967, Pratt, GL, Gas Kinetics, Wiley, New York, NY, 1969, Sillen, LG, PW Lange, CO Gabrielson, Problems in Physical Chemistry, Prentice-Hall, Englewood Cliffs, NJ, 1952. Smith, IM, Chemical Engineering Kinetics, McGraw-Hill, New York, NY, 1981. Swinboume, ES, Analysis of Kinetic Data, Nelson, Chicago, TL, 1971. Walas, SM, Reaction Kinetics for Chemical Engineers, McGraw-Hill, New York, NY, 1959; Butterworths, Newton, MA, reprint 1989. Walas, SM, Modeling with Differential Equations in Chemical Engineering, Butterworth-Heinemann, Newton, MA, 1991, Westerterp, KR, WPM van Swaaij, AACM Beenackers, Chemical Reactor Design and Operation, Wiley, New York, NY, 1984. Wolfenden, JH, RE Richards, EE Richards, Numerical Problems in Advanced Physical Chemistry, Oxford, New York, NY, 1964.EO. E(t) FO. Kiko 2 names of substances, or their concentrations fee radical, as CH; concentration of substance A initial mean concentration in vessel, p. 506, 508 hheat capacity ‘Continuous Stirred Tank Reactor dispersion coefficient, p. $00, 617, 629 effective diffusivity, p, 721 Knudsen diffusivity, p. 721 residence time distribution, p. $07, 508 ‘normalized residence time distribution, p. 508 = CJC, of n/0,,, fraction of A remaining unconverted Age Function of tracer, p, 506, 508 Gibbs energy change, p. 249 Hatta number, p. 828, 829 heat of reaction chemical equilibrium constant, p. 248 specific rate of reaction, p. 35, 103 Jeagth of path in reactor parameter of Erlang or Gamma distribution, or number of stages in a CSTR battery, p. 497 number of mols of A present ‘umber of mols flowing per unit time, the prime (’) may be omitted when context is clear total number of molsxiv PB Pe PFR u(t) v ro 5H) subscript designating initial or inlet conditions, as in Cy, ayy V'gs --- partial pressure of substance A Peclet number for dispersion, p. 617 Plug Flow Reactor heat transfer rate radial position rate of reaction of A per unit volume, p. 34 radius of cylindrical vessel Reynolds number, p. 627, 717 Schmidt number, p. 627, 717 time mean residence time, p. 506, 508 = Ui, reduced time ‘Tubular Flow Reactor linear velocity unit step input, p. 492 volume of reactor contents volumetric flow rate volume of reactor axial position in a reactor =1- “CC 0F 1 ty ty, fraction of A converted = x/L, normalized axial position = VIR, normalized radial position, p. 410 skewness of distribution, p. $07, $08 unit impulse input, Dirac function, p. 492 fraction void space in a packed bed = vi, reduced time effectiveness of porous catalyst, p. 724 intensity function, p. 492, 508 viscosity SYMBOLSSYMBOLS ~ wp, kinematic viscosity total pressure density =1R, normalized radial position in a pore, p. 723 variance, p. 507, 508 normalized variance, p. 507, 508 = vi, reduced time tortuosity, p. 722 ‘Thiele modulus, p. 723 modified Thiele modulus, p. 731 vMN THEORY Software 1 Plotting and curvefitting 2 Regression of tabular data 2 Roots of nunerical equations 3 i. Single equations 2. Simultaneous Linear equations 3. Simultaneous nonlinear equations 1.5. Integration and differentiation 4 6. Ordinary differential equations § 2, Formulation Integration The second order OPE The Laplace transform method fumerical solution of ODEs 7 POLYMATH program CONSTANTINIDES program A. Second order equations with two-point boundary conditions 1.3. Partial differential equations, 9 fables 10 Boke s PROBLEMS Regression 15 Roots 18 Integration and differentiation 22 Analytical solution of ordinary differential eau Nuserical solution of ordinary differential equations tions The problens of this collection require numerical or graphical or sometimes analytical methods of solution. There is a large number of books and software on these topics. An outline with examples of these methods is the ail of this chapter. It is expected that the sludent wil! nave access to some equivalents of the software used here for the solved problems and listed subsequently. Some of the work can be done with a programmable calculator, but pot as easily as on a PC 4.1. SOFTHARE Of the following list, the first six items are the principal ones used here. All but TableCurve are currently availeble at reduced prices to students, but most of the required curvefitting also can be done with POLYMATH or AXUH. The last three items do analytical as well as numerical work 1. POLYMATH. AIChE Cache Corp, P Q Box 7939, Austin TX 78719-7939 Polynomial and cubie spline curvefitting, multiple linear regression, Simultaneous ODEs, simultaneous linear and nonlinear algebrate equations watrix qardpulations, integration and differentiation of tabular data by way of curve fit of the data. 2. CONSTANTINIDES, Applied Munerical Methods with Personal Computers, McCraw-Hill, 1987. Nonlinear regression, partial dirferential equations, matrix manipulations, and a mere Flexible program for similtaneous ODE: 3. CHAPRA & CANALE, Numerical Methods for Engineers, McGraw-Hill, 1988 Explains the methods but the diskette is limited. It does heve the trapezoidal 1rule for data in tabular or equation form. 4. SEQS simultaneous equation solver. CET, PO Box 2029, Norman OK 73070 Twenty simultaneous equations can be solved. The program seems to be superior to the one in POLYMATH S. AXUM. Trimétrix Inc, 444 NE Ravenna Blvd, Seattle WA 98115. High quality graphics and data analysis, including nonlinear regression. 6. TableCurve. Jandel Scientific, 65 Koch Road, Corte Madera CA 94925 Automatically fits 3000 different xy relations with statistical criteria 7. MathCad Numerical Supplement. MathSoft 201 Broadway, Cambridge MASS 02139 8. Maple V. Brooks/Cole Publishing Company, Pacific Grove CA 93950. 9. Mathematica. Wolfram Research, 100 Trade Center Drive, Champaign ILLINOIS 61820-9910. 1.2. PLOTTING AND CURVE FITTING Graphical representation and corresponding equations of tabular data are of value for interpolation, for revealing behavior patterns, and other purposes. The most complete software for this purpose is TableCurve. More limited but adequate for the present book are POLYMATH and MathCad. All spreadsheet software, for instance QuattroPro, can make plots of all kinds. In this book the most used software for making plots of tabular data and equations is AMUM which is also capable of fitting polynomials and noniinear equations to tabular data. All of these commercial packages have tutorials and user friendly instructions 1.3. REGRESSION OF TABULAR DATA Regression 1s a process of representing tabular data, say (C,t), in equation form. The purpose may be For interpolation of tnlermediate values of C or t To obtain an equation, C = f(t), that can be handled ‘readily in a computer progran ..To facilitate repeated differentiation or integration of the data The form of the equation is specified in advance, and the best correspnding values of any constants are found by least squares. The goodness of fit of several assumed equations are compared by statistical criteria such as the correlation coefficient or the F-test. When the number of sets of data equals the number of unknown constants in the equation, the constants are found by simultaneous solution. Otherwise a least squares regression is used. Linearity refers to the unknown constants. The following are linear in this sense, y = atbxtcx® and In y = atb In x + /x* but not b b) Ve za+ 2 = +P y= a+ 325 and y = exp(ar2) x although the last can be linearized as > Inysae2toks ny=a+2+c inx There are four main classes of data regression. ‘ 2 ; i 1. Polynomial, y = a + a,x + apx” + Linear regression is a special case. 2. Multilinear, y = ag + a,x + aX * ..., where %, xg, --. are independent variables or functions. 3. Nonlinear, say of two variables, x and y, 2Ye = £04: ag, aye Bare.) and the derivatives OE OE as ° Bag Jay ~ Ba, of the Least squares condition, Ely, — FOK389) a age... ))? > Minimum are not linear in the unknown constants a,. 4. Periodic data, using the Fast Fourier Transform. These topics are explained in specialist books and briefly by CHAPRA & CANALE. Polynomial and multilinear regression programs are in POLYMATH and AXUM, nonlinear in CONSTANTINIDES and AXUM. No periodic data are regressed in the present collection, Examples: Polynomlal regression is applied In problem P1.03.02. Several examples of POLYMATH multilinear regression are in sections P3.06, P3.08 and P3.10. A non-linear regression is worked out in P1.02.07 1.4. ROOTS OF NUMERICAL EQUATIONS Methods for finding roots of numerical nonlinear equations first involve making estimates and subsequently improving them by some systematic procedure ideal problems for implementation on computers. Many scientific calculators also have built-in capability for finding roots of single equations, for instance the HP-1SC or HP-32SII. 1.4.1 SINGLE EQUATIONS The Newton-Raphson method is based on the truncated Taylor expansion If xp is an estimate for f(x) = 0, an improved value 1s X = Xo ~ KF(%Q)/E? (xq) where f' (x) = df/dx and k is a positive fraction that is selected arbitrarily to speed convergence; k = 1 often is acceptable. When the derivative is awkward to find analytically, a finite difference equivalent can be used, for instance, ogy = £(1:0001K)-£ 6) £00 = “9 O00TK Then, 0. 0001x9f (x9) * = Xo ~ K FUT GOOI KS =F) The factor 1.0001 is arbitrary and can be changed to improve convergence or accuracy Wegstein method. This does not require a derivative and usually converges more rapidly than the Newton-Raphson. The equation is put in the form x = £(x) When xp is the estimated value of the root and x, is £(%), an Improved value Xof ()-£%q)® x of Ox? % © EGR )-aE OG) XytF OGIO, ‘The process is repeated with the improved estimate as a start Polynomials, a a,x azx* +a,x’ +... = 0. Procedures for finding all real and complex roots are on the diskettes of CONSTANTINIDES and Al-Khafaji & Tooley (Computerized Numerical Analysis, 1986). For instance the roots ofx® = 3x8 -ax4 ~9x + 10x? + 2x + 3 = 0 are x = -0.1438 4 0.1610j, -1.000 + 1.4145, 4.288, 1.000 1.4.2. SIMULTANEOUS LINEAR EQUATIONS These are solved by matrix manipulations. Programs are in POLYMATH, CONSTANTINIDES AND CHAPRA & CANALE. When the number of equations is not large, a manual procedure can be used to eliminate one variable at a time by reduction of the leading coefficients to unity and appropriate additions and subtractions of the equations 1.4.3, SIMULTANEOUS NONLINEAR EQUATIONS ‘The Newton-Raphson method for multiple equations also starts with truncated Taylor expansions. For the system f(x,y.2) = 0, g(x.y,z) = 0, blx,y,z) = 0 the corrections to initial estimates xg, yo, Zp are 2, b and ¢ found from the linear system fy + af, + bf, + xf, = 0 Bo + ag, + bey + cg, = 0 ho + ahy # bhy + ch, = 0 where the functions and all the derivatives, for instance f, = af/@x, are evaluated at (%9,¥o.%)). Programs are on the diskettes of POLYMATH and SEQS The speed of convergence and even the possibility of convergence depend on how close the estimates are to the correct values. It is of course possible that there may not be a solution in the range of Interest, or there may be several solutions INTEGRATION & DIFFERENTIATION Only a comparatively few forms of integrals are encountered in reaction kinetics, Some of the more complex ones are reducible with partial fractions. Table 1.1 of indefinite integrals and partial fractions supplies most needs When integrating between numerical limits it may be more convenient to proceed directly with a numerical integration rather than through an analytical integration, particularly when the coefficients and integration limits are long numbers, for example, the integral S88 —___*_____ ax = 1.2811, 50 trapezoidal steps 2.1-3. 78x+5. 86 1.2815, 200 trapezoidal steps by CHAPRA & CANALE which takes a very simple input. Such a result may well be more reliable than the laborious evaluation of the analytical integral by way of item 11 of Table 1.1 Softwares for numerical integration of equations include the calculator HP-32SI1, POLYMATH, CONSTANTINIDES AND CHAPRA & CANALE. The last of these also can handle tabular data with variable spacing. POLYMATH fits a polynomial to the tabular data and then integrates. A comparison is made in problem P1,03,.03 of the integration of an equation by the trapezoidal and Runge-Kutta rules. One hundred intervals with the trapezoidal rule takes little time and the result is usually accurate enough, so it 1s often convenient to standardize on this number For manual integration of tabular data, the trapezoidal or Simpson's rules usually are adequate. Between two points the trapezoidal rule is SI? yax = 0.50%5-%) (44a)and between n points spaced equally with Ax = x,"x%,; as interval, Yorn ax 5 + vityets-4¥nen) fie my vox Simpson's rule for three equally spaced points ts X17Xo iyet4y,t¥2) $2 ydx = * Differentiation. The main application of numerical differentiation in this book is to find a rate, dC/dt, from tabular data (C, t). In view of the availability of computer routines for fitting of polynomials and other equations to tabular data, this is the way to go, differentiating the curve fit equation analytically. POLYMATH can provide direct numerical values of derivatives and integrals from the curvefit result Numerical differentiation may be quite sensitive to the correlating equation. In problem P1,03.01, the results with four different curvefits do not agree well although the curvefits themselves are statistically satisfactory. In problem P1,0302, however, the agreement between the higher polynomial fits ls more nearly acceptable 1.6. ORDINARY DIFFERENTIAL EQUATIONS Differential equations arise in reaction kinetics through application of two main fundamental rate laws, which are 1. The rate of reaction 1s proportional to the concentrations of the participants. Law of mass action 2. The rate of transfer of mass ( or heat) 1s proportional to the concentration gradient. Fick’s law (or Newton-Fourier law). 1.6.1. FORMULATION ‘AS a basis, a differential element of volume, d¥, = dxdydz, or time, dt, is ldentified. In cylindrical geometry, the element of volume may be a thin disk, Adz, if changes occur only in an axial direction, or a ring 2urdrdz, if changes occur both radially and axially To this element of space the conservation law is applied in the form Inputs + Sources = Outputs + Sinks + Accumulations where cach of these terms may be a quantity or a rate. Inputs and Outputs are accomplished by crossing the boundary of the reference volume. In case of mass transfer this occurs by bulk flow and diffusion. Sources and Sinks are accretions and depletions of a species without crossing the boundaries. In a mass and energy balance, sinks are the rate of reaction, raV,, or a rate of enthalpy change, ~AH,pdC. Accumulation is the time derivative of the content of the species within the reference volume, for example, (8C/at)d¥, or C, (87/2 )pav, Conversion of the differential balance to a differential equation follows after division throughout by the differentials of space and tine. Boundary and initial conditions are specified to complete the physical description of the problem, With ordinary differential equations the number of such conditions is the same as the order of the equation. With partial differential equations, the safest rule to follow is that the problem be completely defined in a physical sense. For example, the condition of every entering and/or leaving stream must be defined, the initial state of the vessel must be known (or a final state if the initial is to be found), symmetry conditions (such as radial symmetry) must be recognized if they exist in the particular problem, and conditions at the boundary such as rate of heat transfer or rate of reaction or rate of diffusion must be formulatedmathematically. A detailed application of the conservation principle to the derivation of a second order equation is in problem P5.08.01. 1.6.2. INTEGRATION Differential equations of the first order arise with application of the law of mass action under either steady or unsteady conditions, and second order with Fick's or Newton-Fourier laws. A particular problen may be represented by one equation or severa) that must be solved simultaneously Solutions of the commonest first order equations are listed in Table 1.2. Variables separable and first order linear are most often encountered. Exercises dealing with first order equations are in problem P1.05.05 Sometimes an equation out of this classification can be altered to fit by change of variable. The equations with separable variables are solved with a table of integrals or by numerical means. Higher order linear equations with constant coefficients are solvable with the aid of Laplace Transforms. Some complex equations may be solvable by series expansions or in terms of higher functions, for Instance the Bessel equation encountered in problem P7.02.07, or the equations of problem P2.02.17. In most cases a numerical solution is possible. ‘Some problems with diffusion or dispersion give rise to the second order linear equation with constant coefficients, 2 SY s oky B+ uty = F(t) ate When £(t) = 0, the equation is called homogeneous. The general solution is the sum of the solution of the homogeneous equation and a complementary function which can be found by several means. The form of the homogeneous solution depends on the value of the discriminant, D = k*-1, as tabulated. minant Solution Positive ¥ = Cyexp(yt)+Czexp (At) Zero y = (Cy+Cat}exp(-kwt) Negative ye exp(-kut )[Cyeos (utd 1-1 *)sCosin(wtl 1 2 y = C exp(-kwt cos (wt! i-k® -¢) y = C exp(-kut )sin(wtl1-K® +9) where A,,Az = w(-kHB), ¢ = [c2c3, g = arctan(Cy/Cy) The complete solution of the non-homogeneous equation can be found directly in some cases by Laplace Transform, as the problems of section P1.04 show. Several examples of second order equations are in problem section PS.08 The Laplace Transform of a function f(t) is defined as Lif(t)) = f(s) = F = $8 £(texp(-st) at Several notations are in use, as indicated. Tables 1.3 and 1.4 summarize various operations and provide some pairs of functions and their transforms.Transformation replaces a differential equation with independent variable, t, to an algebraic equation in variable, s. The latter relation can be solved algebraically for the transform f(s). Then f(t) is found by inversion with Table 1.4. Problems P1.04.01 ff are examples. Cases of multiple reactions are treated this way in problems P2.02.08 and P2.02. 10. Step and impulse inputs. These discontinuous functions are used particularly in Chapter §. Their definitions and transforms are Step: H(t-a) = 0, tsa 1, tea Ltit(t-a)) = 2 exp(-as) Impulse: 8(t-a) =0, tea llesa L{3(t~a)} = expl-as) The inverse of a transform with exp(-as) as one of the factors is L{£(s)exp{-as)} = f(t-a) B(t-a) f(t-a), tea 0, tza Example. For the equation with a step input, SEs se = W(t-ad at f(s) = Sep s(s"4u*) and the solution is f(t) = SU - costw(t-a)) H(t~a) La = costuit-anl, tea a o, tsa With an impulse input, “oe 8(tra) at = exp (sas) f(s) = a situ s(t) = 2 sinfu(t-a)) m(t-a) 3 sintw(t-a)l, t= a 0. 0: For numerical processing, a first order ODE is arranged explicitly for the derivative,. aye y == fy) A procedure based on the truncated Taylor series is named after Fuler. With increment h = Ax, successive values of y are Yar = th £Oq,yy) A substantial ‘Improvement employs the following Lteration, eS yl o.sml eG yy) fC qs ¥r where the superscript denotes a trial number at a particular value of i. This is called the Modified Euler Method Both POLYMATH and CONSTANTINIDES use this method and also a fourth order Runge-Kutta method. Other methods are available in other software, but these two are adequate for the present book Higher order ODEs are reduced to a set of first order equations for solution by these softwares. Thus the third order equation, 3 2, ay = toay, &, SH dx dx’ becomes the set dy . dy" " se Sev, Bev, = fGayy ay") The independent variable is x and the dependent variables are y, y’, y" and y' Integration Is started with known values of the dependent variables at one value of the independent variable, except when the "shooting method" Is needed. Auxiliary algebraic equations can be entered to the program along with the differential equations and the boundary conditions 1.7.1. POLYMATH PROGRAM A variable name may have any combination of up to ten lower case letters and numbers. A typical equation entry is ax xy 0. 48x Ge Oraey 7 04% BOIL) = 0-0. Atte ~ xy 24 oo dly)/dtt) = -2% dz _ O1xy ous fe" ome d(z)/d(t) = 0.1% © = xty/(0.2+y) with intial values t = 0, %) = 0.1, Yo = 10, 2 Note the parentheses in the derivatives. There is no choice of the number of integration intervals; it is supposedly adjusted to obtain a good precision. The output can be a graph or a table with 20 divisions of the independent variable. The scale of the ordinate is selected automatically to fit the range of the abscissa. 1.7.2. CONSTANTINIDES PROGRAM The independent variable is always designated x and the dependent variables y(1), y(2), .... The derivative is designated dy() Sar = Fb yd, yl2)...., CO), C2),...) 6)where the Cli) are constants that are entered during the course of the solution and can be changed to obtain a range of solutions. In this nomenclature the equations of the POLYMATH example becone, xty(2) GO) = Biay ~ O-etyID) = © - O.atxty(a) =. 0.2442) 5g, ot) D2eyi2zy ~ 6 3) = OT KY(2) og gag GS) = “Gzeyrzy = OF oe = xty(2) O.ay(2i with initial values x = 0, y(1}g = 0.1, y(2)y = 10, y(3)p = 0 This program permits any number of intervals with a printout at each interval, and the ranges on the axes of the graph can be controlled, 1.7.3. SECOND ORDER EQUATIONS WITH THO-POINT BOUNDARY CONDITIONS Second order equations or a pair of first order equations require two conditions. When these are at the same value of x, say yo and (dy/dx)y at Xo, a numerical integration can be started. When one condition is yp at xo and the other is (dy/dx), at x, the shooting method can be applied. This consists of several steps 1, Assume a value of (dy/dx)g at xp 2. Do the integration from x) to * That result will provide both y, and (ay/ax),, 3. The correct integration will be attained when the specified and calculated values of (dy/d), agree. This procedure also 1s described in Section 5.8.2. Numerical examples are problems P1.05.02 and P5.08.23 , as well as P5.08.02, PS.08.03, PS.08.11 and PS. 08,12 1.8, PARTIAL DIFFERENTIAL EQUATIONS Common situations in reactors are when the conditions can vary with time and position as independent variables. Partial derivatives of dependent variables such as concentration and temperature with respect to each of the independent variables then are involved. Such an equation, for unsteady state dispersion in a cylindrical reactor, was derived in problem 5.08.01, namely we, #48 at © oy? ry or oy BE gem u SE ke! In this book PDEs appear primarily in Section 8.1 and problem section P8.01 Some simpler methods of solution are nentioned there: Separation of variables application of finite differences and method of lines. Analytical solutions can be made of some idealized cases, usually in terms of infinite series, but the main emphasis in this area is on numerical procedures. Beyond the brief statements in Chapter 8, this material is outside the range of this book Further examples are treated by WALAS (Modeling with Differential Equations in Chemica! Engineer ing, 1991)Table 1.1. Integrals and Partial Fractions 1. fier bres Say Ge te ax) 2. In (ex +b). 3, ~ itn (ax +). 4a aerey thine b), 5. 6. 5 dx * Swern eto 8. ¢__ ze ot Fb) Ox Fa * pal ax r 1 9. fe bt > 4 00 Suepueye je panby Vege OE jo. fii ~ 3. Siticye* rare 400 xdx Lime 3 Fore sg Or thet Z 1 A 8 c 2. pant hard "eee t eee tere here A= 1 B= 1 c= 1 whee A Gea" 0 ES OCT I 4 2 © > 3. wyaete xte ¥t6 uae’ ¥ed where d= “GE aay ae 34, t f CHD eee)” Jomo Va 1GTable 1.2. Solutions of Ordinary Differential Equations of the First Order 1. Variables separable, The differentia! equation hxlgstyi de + fabelgaty) dy = 0 hes the solution Aid aly) Led iy 4 f Se ay J Ao} os 2. Homogeneous equation, The equation dyldx = flyix} has the solution ov inx= f +6, wherev® 1” 3, The linear equation dyldx + filxly = galx} has the solution yeer|frinvreree], wneess [sie 4, Totel or exact differentia! equation. The equation f(x, y) dx + file, vl by = 0 Is exact when af, fay ~ af,/ax. The solution is Jove fla-2 Jaafar where y is constant when integrating with respect to x, and x is constant when integrating with respect to y. may be possible 10 find an integrating factor 4 that will make aluhHay = aluhMax. 5, Explicitly solvable for one of the variables'in Fx, y.p) = (2) Solvabie for p. Flu yale (y~ Mi yie~ fUX yi) 0, One incegral is abtained for each factor ayia ~ file.) (b) Solvable for y. Fle, y,p) = ly ~ flx. pilly ~ fle, pi] --- = 0. Etimin ste y from each factor by differentiation: dy ah aap pa Wh tee dx ax” op ax The solution will be Glx, p, C)= 0, It will be possible to eliminate the parameter p between the integral and the original differential equation, The final results will be 9, (x, y, C) = 0. Solvable for x. Fix, ¥.p) = Ix ~ fly, pillx = fale, pil + 0. Eliminate x from each factor by differentiation: tc 1M hee dy p sy apdy The solution will be Gly.p, Ch= 0. 1 will be possible to eliminate the paremeter p Between the integral and the original ditferential equation. The finad result will be glx, y, C) = 0. 6, Simuttaneous linear differential equations with constant coef: ficients. Take the pair Baar Baath) and Me asx + Byxs tb hlth a a Apply the Laplace transform to each with L{dx/at) = sf, ~ xj. Solve the resulting pair of linear algebraic equations for x = gy{s) and 2 = galsi; then invert to x, = y{0) and xs = hylel with the aid of a table of transforms. aLaplace Transform Operations Aa Fls}= Tish sa [lernoa 2 teri 0140 Sern 13 2 fhe PAs Nt amie sri) D 8m) 8. fit A= 0.060, ees Bena trey > [anes Sr # fnase- ne Fale testoen hore «ts srbasl = s 0, etl Fix, 8) = #0, 0) Prins! 2 efins)_ ao) ae rn) aay A ve nse tem ena ines 1S Sree oem nashTable 1.4, Laplace Transform Pairs Ae) a, ost Ble) nit ienpulee at Foedie) unit siep starting at =0 tH romp tonation ahr n= positive integer fait , 3 s. Hea) unitstep starting at? = @ 6. HO=HUt~ 2) rectangular pulse 2. e"" —exponentiat decay 1 = postive integer {stay p . 1 sista) yo, — ae ale # a a Ste Sts + alls + 6 2 — lew oits+o 3 —p (so ayer) te SE Bs ast Bi 1 e om & 8 i a a isralistole+) = Bm alle— ai oe 5) a 16 sta fa~ale , fan sie™ | te cle isFoltsroRs +a © alie~ ah" (e-dys Hl ta ello— v7 = sino 18. cost —s 48. sintot tg}, @ = arctan 20, sinhot + 8) 2 aie tot tg): pm arctan FE . 1 intut — a); db arctan® a . tera o%) 24 + aoe 1 Me Fy Beast wl as, S42 GrateTable 1.4 (continued) 26. 22. 2a, 28, 29, 30, 31 22, 3 34 38, 36. wrarre 1 e+ ero SIS + 2fens +o) SF a FO te Gralisra +57 od s(e+ elles a +B sist ois + a + BA Flere} 1 sis aye Bist sis + ais +b} sie + a) 64) sinlbe-s gy = arctan eM sin(or= gh é= arctan Fie sists Vim Fe ahs > arco y infbt+ oy = arctan " siniot— 6) Vicar +o? aE cle= a +O” Vea pVig wale 6 veesnveisun 2 $= arctan Hsinkbe + ah fe aretan 2 — arctan £~ arctan 2 Fleder Jeera Flo oo ato~ ate" 0, 294 ao peas BEF OO or 25° ela 8) Bia $84 Mat = BF aya + aa + b¥loy = 20)" 2 snot + 6 warctant; cheats ot 14PROBLEMS, CHAPTER 1 Vapor pressure data of toluene, in °C and Torr, are T -26.7 4.4 6.4 18.4 31.8 40.3 51.9 69.5 89.5 110.6 pot 5 10 20 40 60 100 200 400. 760 The Antoine equation is fitted to the data in the original and linearized forms, »’ T2732" The two TableCurve plots show that the constants are a = 18.62 and b = -4570, exactly the same in the two cases. The modified Antoine equation, however, B THD cannot be linearized. Although this form Is not in the TableCurve repertory, the following relation can be used, 2 atx, B = b ae p = expla + x7g7q-g) and In p= a inpoa- Trbx xD where A= c/b, D = ivb, B = D(A-a) 41 Ean 15, wesigalgsoomsewe form! sem 0 rot" 00) 0 ot 20 100 ° 3s 1000/{T#273.2) rey P1.01.02. POLYNOMIAL REGRESSION A quadratic equation will be fitted to N sets of data,(x,,y,), by least squares y = a9 + ax + agx® Ely, ~(agtayetagx?))* » Minimum On setting 8 | Of aE Ga, Bay Gay the resulting “normal equations” are zy = EX; 2 agN + ayEx, + ag&xp = 2 3 agEXy + ayExy + agExy 15gE, ajExp ta,Bx} = Edy, The various sums are tabulated below so the normal equations become llag + 6.0la, + 4.6545a, = 5.905 6.Olay + 4.6545a, + 4.1150, = 2.1899 4.6945ag + 4.1150a, + 3.9161a, = 1.3357 and the correlation becones y = 0,998 - 1.018x +0.225x" x [0.05 O11 O15 031 046 052 0.70 0.74 082 098 117 ¥, | 0.956 0.890 0.832 0.717 0.571 0.539 0.378 0.370 0.306 0.242 0.104 Dx = 6.01 N= Dx = 46545 LY = 5.905 Dd = 4150 Dxy = 2.1839 Y xf = 3.9161 Day = 13357 P1.01.03. MULTILINEAR REGRESSION The data of the first three columns of the table are to be fitted to the equation ¥ = ay + ayy + agxy The least squares criterion is Ely, - Cagtayxytayxg)]? 2 Minimum Applying 8 oz Jag Ba; makes the normal equations aN + xy + ax, = BY, AoE + EEL + aE ve = Bay On substituting the various summations from the table, the equations are a9 + 16.5a, + Mag = 54 16.5a9 + 76.25a, + 48a, = 243.5 lay + 48a, + Sda, = 100 The solution is y = 5 + Ox, -3x.. ae ee 500 0 0 0 0 0 6 3 4 8 5 3 ew ee ee a a gS ok 8 BF 3 8 % How 8 Bose 16s #7825 5k B25 100 1.01.04. NONLINEAR REGRESSION The normal equations will be written for the least squares regression of the nonlinear equation for N sets of data, by.y,), 16b eat2 Yr 88 sre The least squares criterion is by )2 Ely, “Cat g)1? > Minimum &_ 8 _ 8g 3a ~ Bb > Bc ~ Be _ -a-? = & = atzty,-a-2r1 = 0 an + pe t= 5, a mre 7 er wa Pye 1 - 5 7 UEC -agead > gagh) = 0 1 1 ye EA ab ake + PEG ey)” = Ege (2) =e Fe = 2B 1 yz i Ms atl, )° + bE(—_— —s 3) ee re Ge i Equations 1, 2, and 3 are nonlinear in a, b and c. In principle they could be solved simultaneously, but an improved technique is explained in problem P1.01.05. Take the case of a function of one variable and two constants, yy = fOys a, b) Estimates of the values of the constants are made. For the j-th iteration they Will be called ay and b, and corrected values will be aya = ay # Aa, by = dy + Ab a Represent the function by the truncated Taylor series, af (x4) af (x,) Fay = FOq), + AG Ds dat (GE, ob =f + fyna + f,Ab (2 The least squares criterion is Ely, “FOG yaql? = Ely — (fda tyOb)]? » Minimum The function and the derivatives are evaluated at the trial values a, and by of the constants. On placing o_o da ab the normal equations become ~f ,ha-f,Ab) (-f,)1 -£,ha-f,Ab) (-f,)] 0 or babte + AbUff, = Lyf) fy (3) qTbakf,f, + AbEf, = Ely~f)f, @ where 8 by) af (x) oy fe = as (s) Solve for da and Ab from (3) and (4), substitute into (1) and repeat the procedure. Continue until values of 4a and Ab are near enough to zero £2 £64)), fa Data of the first two columns of the table are to be fitted to the equation . b y = £(x) = expla) Trial values of the constants are a, = 18.6, by = -4.57 The derivatives are af f, = Fy = explatb/x,) = Of L fy = 3p = ) exp(a+ b/x;) Corresponding values of the function and the derivatives are tabulated. The various summations are Efe = 0.61(10°), Beg = 4.26(10°), Eff, = 1.61(10%), E(y-f)f, = $720, E(y-f)f, = 15624 Eqs (3) and (4) of problem P1.01.05 become 0.614 + 1,614b = $720(10°°) (3) 1.61da + 4.26db = 15624(10°°) @ Accordingly fa = -0.1084, Ab = 0.04498 and corrected trial values of the constants are a = 18,6-0.1084 = 18.492, b = ~4.57+0.045 = ~4.525 ROOTS OF A SINGLE EQUATION The equation is y = -1 + 0.7% expl0.5(1-x)*] + 1.2(1-x) exp(.5x*) = 0 or 2 ) 1.2(1-x) exp: 0.7 exp{.S(1-x)*} The derivative is x =x +1)40.7 exp(0.5(1-x)?] - 1.2 exp(0.5x7)} 18The plot shows that the roots are approximately -0.47 and +0.65. Hi ai 1 oO 1 2 3 4 5 6 1.02.02 Programs are written in QUICKBASIC for the Newton-Raphson method with analytical or numerical derivatives and the Wegstein method. The particular equation jowever, ssumed -1 and +1. rial ~0.7687857 0.76534 ~0.4971118]9. 66673 ~0.4721097|0. 65458 ~0.4706308]0.65442 -0, 4706259|0, 65442 -0. 4706259 P1.02.01 EN-R SOLUTION OF Y=F(X)=0 WITH ANALYTICAL DERIVATIVE, SHORT X INPUT X PRINT X 19 GosUB 30 HEYYL XoX-H PRINT X BE ABS(H/X)<~.0001 THEN 29 Goro Iv END SR FOR Y AND YL Yon ToXAEXP(Se{ =X) 241.24 1 -X)MEXP (502) YI (K2AKE Io HEXPC SH{I-XP2)~ 1.2 EXP(50X2)) RETURN 20 30 NCR SOLUTION OF YFOQ=0 WITH NUMERICAL DERIVATIVE SHORT X INPUT X PRINT X 10 Gosup 30° yey Xr 1.00014% Gosun 30 ~0, 495743] 0.95931 -0.479811/0.73556 -0, 470626 |0. 66884 0.65492, 0.65442 is entered following line 30. 20 30 10 20 to test the Newton-Raphson and Wegstein methods, starting values were The successive evaluations are tabulated. Wegstein is a little faster for the negative root but a little slower for the positive one. PROGRAMS FOR ROOTS OF SINGLE EQUATIONS The one used here is from problem yey, 001+ X¥1/(¥2-¥1) X=X/1.0001-# PRINT X_ IF ABS(H/X)<=.0001 THEN 20 Goro 10 END USRFOR YF) YD) LAs XSEXP {| WEGSTEIN SOL'N OF X=F00 SHORT X PUT xt PRINT XL XeX! x2ex ‘GosuB30 xvi Gosus30 Xe (RIeVINX 2YKIFYI-20) IF ABS(X-YI/(X+Y1))S=.0001 THEN 20 PRINT X Xinx Goro 10 END ESR FOR Yi=X=F00) Yim(1=1.24(1- KEXP S0X2Y,T/EXPL Se (=x)°2) RETURN P1.02.03. MULTIPLE LINEAR EQUATIONS ‘A set of linear equations can be solved by a variety of procedures. In principle the method of determinants is applicable to any number of equations but for large systems other methods require much less numerical effort. The method of Gauss illustrated here eliminates one variable at a time, ends up with a single variable and finds all the roots by a reverse procedure. In the example, each of the four equations are divided by the leading coefficient, 2.0, of the first equation. By appropriate additions and 19subtractions the variable x, is eliminated from the second, third and fourth equations. Similarly x, 1s eliminated from the third and fourth equations, and so on. The second group of equations is the result. By a reverse procedure, all variables are evaluated by the third group of equations. 2.0%, 10x, 0.12, +1.0:,=2.7, 0.48, 4 05e,-$4.05,—85x, 221.9, a 03e-— 1.02, 4 10%, -+5.22,——39, Ly, $0.2¢)42.52,—1.02, 09.9. 2, 4$0.8x,-0.05:, +055, 35, 1 Fe#1S.AOe— 29.002, 71.20, (2) 1.12296, = 4.72286, 1.11998x, = 1.11588. =~ 1.00000, y= 4.72298 — 1.72298 =3.00000, @) 2,=71.20—13.40-3-4 29,0 =2.00000, #1 1.35-- 05-24 0.05-5440.5m1.00000, P1.02.04. SOLUTION AFTER LINEARIZATION For the rate equation r= xctcB data are Ca 0.7 0.9 1.6 Linearize the equation and substitute data ln 0.4567 = Ink + @ In 0.7 + 6 In 0.2 in 1.0707 = Ink + @ 1n 0.9 +8 In 0.3 ln 2.5917 = Ink +@ In 1.6 + 8 In 0.2 The solution by determinants is r= 3.scz cp Data of the first three columns of the table are to be used to find the constants of the rate equation r= kchcd In the linearized form, Inr=Ink+p Inc, +q Inc, or in the simplified notation, you+ep + Bq POLYMATH multilinear regression could be used to find the constants w, p and q but the problen will be solved directly. The least squares condition is lw + ap + Byq - y)* » Minimum Apply a= _ ah _ ah _ oy” Op © Bq to obtain the normal equations Nw + pEay + azB, = Ey, mo 20WEa, + pay + q¥o,By = Lay, (2) WEB, + pEayBy + EB; = EBs (3) ‘The various summations are evaluated from the data of the table and substituted into (1), (2) and (3). The following groups of equations are for successive eliminations of the variables w, p and q. The coefficients are listed. From the last line, q = -1.2160/0.8625 = -1.4098, Then p = 1.0815+0.7388(- 1.4098) = 0.03994. Finally, w= Ink = 3,492, and k = 32.85 so the rate equation is r= 92.8809 c,h? Values from this equation are listed in the last column of the table. The agreement is only fair. The pair of equations sinx+y=2 and expx+y?=3 are plotted to estimate the solution. SEQS is used to obtain accurate values of the roots. x 7 Converged Values 3 2 WwW Does not converge -4 1 3.42267, 1.72261 45 1 ~ 3.42267, 1.72261 2 “5 1 ~9.69606, 1.73203 y -5 W7 6.01118, 1.73134 “6 1 ~ 6.01119, 1.73134 + —10 i ~8.69606, 1.73203 Which roots are found apparently denends en the prior esti: 0 mated values. =10 -5 * 0 21 woP1.02.07. THO NONLINEAR EQUATIONS In Newton’s method for a set of nonlinear equations, each equation is expanded in a truncated Taylor series. The result is a set of linear equations in corrections to previous estimates. Repetition of the process ultimately may converge to correct roots provided initial estimates are sufficiently close Suppose the equations are F(x,y) = 0, Glx,y) = 0 mm and estimates of the roots are x, and y,. The successive approximations then are computed by the following formulas. tegen [Fen HF yl am) eat TE WG (sn 4056 94) [~The | og ay EG Flee wale, ap Ines YF Fd] Clery U2) G0, gn) 2* Te a" where apn] Ete In) Fy len Ye gon a|Fe lene He) Fem In) "1G bus Yn) Gyn ta” PVG vs) Glen 90) and the Jacobian Fee, y) Pyle 4) Gite, uy) Gylx, 9) Example 1. Find the real roots of the system F (x, y= 28 —y'—1 0, Ge, yxy —y—4=0. Solution, Find graphically the approximate. values x,=1.2 and Je, y= leo Computing the Jacobian at the point (1.2; 1.7), we have From formulas (2) we get Lact [One 3.40 *—e7310| 0.1956 9.40 8.64 —0.434 = 17 —era75/ 4/91 0.1986 |~ 1-70-9980 = Continuing this process with the obtained values x, and y, we get &_ = 1.2343, 6615 and so forth. Joey) J(L% 17, x 1,2+0.0349 = 1.2349, 6610. P1.03.01. DI NTIATION OF FITTED EQUATIONS Four different equations are fitted by TableCurve to the data of the table. They all have correlation coefficients > 0.995. The derivatives are evaluated at selected points, and mostly do not agree closely. Presumably tne derivatives from the equation with four constants are the best. Vig = 2(atbxtex™ eax?) (b+2ex+3dx”) ¥3, = béZex Yar = explatbxtex?) (b+2cx) 22Yao > ee WONmasan~ % eceeeekece vente a-cencoe ge 8 e2ke Catalytic oxidation of carbon monoxide is studied in problem P3.07.06. The change in pressure is given as a function of the time in the first two columns of the table. Polynomials of several degrees are fitted to the data and the derivatives, -dP/dt, are found at selected points by POLYMATH. Values derived fron degrees 3, 4 and 5 agree fairly well. aP/at, 0.7555 |0.7789|0.7722|0.7710 0.4539 |0.3864]0. 3854) 0.3865 -0,1494]0.0889]0.0511]0.0570 23sam (uble Satine P1,03.02, TRAPEZOIDAL AND RUNGE-KUTTA INTEGRATION The trapezoidal program of CMAPRA & CANALE and the Runge-Kutta routine of CONTANTINIDES are compared for the integral ys Sy aes x J e-22erx020 with several intervals as tabulated. The Runge-Kutta is superior Intervals RK 2 1.6488 10 1.6498 50 1.6498 100 1000 1.6498 1.6498 P1.03.04. ANALYTICAL AND NUMERICAL INTEGRATION The integral to be evaluated is p22? 1 Yo SCs Gales) bas. o1ao.sy The integrand is transformed into partial fractions by item 13 of Table 1.1. 20.2 ( _ 0.2892 , 0.1538 _ 0.02339 | 0.1587 Jus (" san2 * ees "55.0 * Saas)? & = 0.009326, by analytic evaluation 0.009377, by trapezoidal rule with 100 intervals With the aid of the computer, the numerical evaluation is much less laborious. y 24P0491. A ANALYTICAL SOLUTION OF FIRST ORDER EQUATIONS Separable Variables (ol x?y' cosy + yon xaex (b) 2y' scos2yet, yo Pa, xe (1 Oy =sinyst, po bn xe (d) (18 xy’ feos? 2ye0, yd, xo-= le} e” = ey’ +1, y being bounded when x» += (0) be + tly’ =y— Wy bsing bounded when x (al y= 2sln# yl. y being bounded when x» Oo) ty a sinayet, yo la, xa te Answers to ya aren) 56 t yoauma(eh) on V3! (©) y= 2arctan(s = + arctan x) 4 22 (a) y=} arctan| an 2 (el yeo yet (el y= =m (h) y= arceot La Sp ae B. Homogeneous (a) 2xy dx + b+ yh dy = 0 to) (+ VP = xyl dy —y de (cl t+ yd Yaymo (2) xy = y= xsin\yix) = 0 (eh (2073 Ph des by? = 216 (8) Px + WV =a ~ xy dy 0 (9) Leos hax — (EsinY + cost) dy «0 2 My OR Answers (a) 3x77 4 Pac, Ol ysce"V" y>ox y. (e) Acetaniyii) ~§ fools? + y" (9) y= 2rArctan ox el xi + log xy 0, x #0, #0. (0 0x2 VAR, oF equivalently, ty + Vy aay 7 >a {g) ysiniyixl =e. 25, (3) ¥ cosx-y sin x™ 2% Yo (6) xy = 2y =2%c0s x My ~ytanx= costs, een (0) ying y= 3° Int {h) @e= Aly © 2y Ai) y+ y008%= €08x; Yleeo 1 Answers (a) pax a2, Wb) ye (ore, (0) yaiC+xie™, (a) p= x*Hleos x) (eh ys OF + x4sin x. (1 y= (sin xiticos*x} (a) y= (C +20) Inx (n) xem cy = v2, a Simultaneous wo toh & a tel ta 2 fo) & 3 ws Answers (a) Ts tte (0) fo) @ He Seahy ee ae pe ee taette zm CSE H (ey FD (oy eal + 05 le) xm etbey varie te, beat” (9 xm ae! Slsineteosts yet 1 tee"!P1.04.02. LT SOLUTION WITH INITIAL VALUES UNKNOWN ‘Transforms of the first two derivatives are it) nsr-m, (EB ete-a- ‘The initial values, yo and ys, can be made part of the integration constants The equation is y* + 3y' + 2y = dt, and its transform is 29~sy0~ 5+ 9107~ vol 2V= 4 Isolating the tanstarm, pe [4 egten tea ys Pr emaiera (tO +8) lavert with the data of Table 1.4 Yabo 0% 4 2-34 (207 em Myo +(e"! ois, (4+ Ayo + ysle™" = 1 yo vile" + 20-3 — Cie" + Coe + 20-3 This solution alsa is found readily by the method of undeter- mined coefficients ‘When the conditions are y{1) = 1 and y{2) = 2, the integra. tion constants are found from = 0.86700, + 0.13530; + 2-3 2 0.1989C, + 0.01830, + 4-3 Cy 4 yo + Y= 8.5287 Ce= HU) yy + y6) = 8.4071 ‘The initist conditions are found from these equations as Yo= 2879, y= 10.2061 P.04.03, LT SOLUTION WITH CONDITIONS AT TWO POINTS The equation is aya? +0.25y=0, with conditions 10) = 0, yln) = 0.5, The transform is S9~ ayo yo + 0287= p--tiy, fo with yo = 0 F028" * Investing, y= 2ybsin 0.8 Since Yin} = 0.5 = 26 sin(x/2) = 2¥6, y= 0.26, and the solution y= 0Ssinast 26P].04.04. LT SOLUTION WITH SQUARE PULSE INPUT (cos 1)=(1=c05 1ult-3) Ng The transform ie SY sI07 yb WY With the initial conditions yo = ye 0, pe -- 1 TOF er wa) Lo-cor nutes) / yf The second term corresponds to @ delayed response. The in- eA verse ie ° 3 v0 15 1 ya dott cose - G1 cos ule 3) Te two contibutions of the response and the net response are piloted for w= 1 and 9~5. Pi.04.05. IMPULSE AND STEP INPUTS The solutions of second order equations with impulse or step inputs are to be found y" +y = 6(t-a), with y'() =0 am 2" +z = H(t-a), with 2(0) = z'(0) = 0 (2) The transforms are ye exp(-as) set z a> exp(-as) s(s*41) The inverses are found from Table 1.4, y(t) = sin(t-a) u(t-a) sin(t-a), tea Oo, tsa z(t) = 1 - cos(t-a) ult-a) 1 - cos(t~a), tea Oo, tsa The plots with the two inputs are shown for a = 1. Fonction 1.04.06. USE OF THE CONVOLUTION THEOREM The convolution theorem, item 8 of Table 1.3, is of value when the transform is a product of two factors whose inverses are known individually. Take the case of the equation, y" + By’ + 2y = £(t) = exp(-3t), with y(o) = y'(0) = 0 The transform is (s%* 3s + 2) ¥ = F(s) 333 or yls) = £(s) ED £(s)g(s) where 271 a) © ERT g(t) = exp(-t) - exp(-2t) Also g(t-8) = exp[-(t-9)] ~ exp[-2(t-9)} Since f(8)} = exp(-38), application of the convolution theorem gives y(t) = Sif (o)g(t-o)ds = SSexp(-30) {exp{- (1-0) J-expl-2(t-2) 14a = 0.5exp(-t) -exp(-2t) + 0.Sexp(-3t) This can be checked by finding the inverse of 1 s+1) (s+2) (s+3) with item 29 of Table 1.4. y(s) = P1.05.01. POLYMATH AND CONSTANTINIDES Three simultaneous ODEs are to be solved. 8 2 ay, 8% = styzy, MH = aete, at tity, ae S(y-z), at Btt+z with to = 0, ty = 1, Yo= 1, 29 = Wo = The POLYMATH input is ay) a(z) a = 5*(y-z), 2 acey = t™2+y. Sey = S*Grz) " d(t) Note the parentheses around the variables in the derivative. The CONSTANTINIDES input is G(1) = x@y(1), G(2) = S*(y(1)-y(2)), G(3) = S4xty(2) Note that the independent variable is always designated x in this program. The two graphs agree. = Sttez 5-887 poLynarll Kev: sat rey ge 4 CONSTANTINIOES Wa 3.58! 2.88 1 5.89, 408 828 6M 868 BE t 281.05.02. SHOOTING METHOD FOR EQUATIONS WITH ‘The higher oller ODEs are reduced to systems of first-order equations and solved by the Runge-Kutta method. The missing condition at the intial point is estimated vatil the condition at the other and is satisfied. After twa trials, linear interpolation is applied; after three o: more, Lagrange interpolation is applied. {a) Fluid enters an immersed cooling coil 10m long at 200° and is required to leave at 40". The cooling medium is at 20% The heat balance is nos a00, wn no wich Is equi lent to the pair of equations & 1 = o.0n7 ~ 20) ‘The table end the graph show that the exk point condition is satisfied when 70) « ~46.515. {b) The equation yaxys (2 02Kyy" with yt} = 2; 3} =0 Is equivalent to the system a Was Bavy st —o2xye gree i ‘The value y'{0) = ~1.493 makes the solution satisfy the con: ition {3 = 0, The xy plat for that condition ig the midaile one. {c) The third-order equation yom yh= ae xty, with WO) =0; yl) = 15 yi)» =10 js solved as the system The curves and the tabulation show that y(Z] = ~10 is satisfied with the assumption ¥"(0) = ~3923. {d) The equation of part (b} is solved for the terminal con. dion, yi3) + 2/13) = 0. The results of several trials are vn yea yi) + 2918) Graphs of ylx] and y'{x] are shown for the coreect value, y'(1) = 1.8708. 29 TWO-POINT BOUNDARY CONDITIONS O39) 209 mS pes GijeePL.0S.03. DIRECT AND LAPLACE TRANSFORM SOLUTIONS (a) A pait of consecutive ceactions, Reactions: a8, Rate equations: Pe — tang at 08 6 kyo, ~ kane at ‘Material balance: Pao t Poet Aco My +8 + Me The solution of the first rate equation is direct 14 ge Substitution into the other rate equation results in @lineac differ. tential equation of the first order: 108 6 bang = karage Tht Hate kat of which the solution is ate, Ae = age + Ieo ks letters Also, Ina 90 Meg ~ age = age ateeggnbi hy Ay =i ) (0) Solstion of part (2) by Leplace transformation toy fate ts” tt silky 8} The inverse transforms are found in Table 1.4. and give the same results as part (a). 30P1.05,04. LAPLACE TRANSFORM AND NUMERICAL SOLUTIONS A set of reversible-consecutive reactions, A $! B 3 = “kA + kge k,A ~Ckgtky)B & a ~ Laplace transformations are made and rearranged to (stk, JA = KgB = Ap wh +(s#kgtg)B = By ~KgBtst= Co linear equations D= 8 + (kytk tky)s + kykz 5 = fo8t kathy) AoreaBo These are solved for the transforms as = _ Bostky (Ap*Bp) Be = _ keBeCy cH 3 Formulas 24a and 27a of Table 1.4 can be used for ‘transforms to obtain A, B and C as functions of t. A numerical solution is found, alternately, by 42 c, has the rate a) (2) (3) (a) () (6) mM inversion of these application of the CONSTANTINIDES program for these values of the constants: k, = 1, k, = 9.5, ky = 0,3, Ay = 1, By = Cy = 0. 1 ” 2 2 c 16 5 4 13 8 2 A a “gaa s 31CHAPTER 2 REACTION RATES AND OPERATING MODES THEORY A. Rate of reaction 34 2. Concentration, mols, partial pressure, mol fraction 35 3. Reaction time 36 4. Constants of the rate equation 36 1. Linear regression of the rate data 2. From the integrated equations 3. Plots of the data 5. Multiple reactions and stoichiometric balances 37 1, Multiple reactions 2. Stoichiometric balances 6. Reactor operating modes 39 1. The stirred vessel material balance 1. A filling operation 2. An emptying operation 3. A batch operation 4, Steady state operation (CSTR) 2. The stirred vessel heat balance 40 3. Tubular flow reactor 41 4. Packed bed catalytic reactor 41 7. Heterogeneous reactions. Rates of diffusion 41 Tables 43 PROBLEMS 1, Stoichionetric balances 50 2. Integrated rate equations 54 3. Complex reaction mechanisms 68 4 Numerical integrations. 86 1. qA 9 Products at constant V or constant P 2 AeB 3 A+BSC+D 4. A+ Be C+D, in terms of equilibrium conversion 5. aA + bBo cC + dD, at constant P 6. A+ B+ Co Products 7. A933, Bac 8. ASBSCSD 9. Hydrogenation of cottonseed oi} 10. A@BeC 11. Ae Bec 12. 2Ae BHC 13. A 9B, A+B9C 14.4 +B9C,B+C 4D 1S. A+BeCH+E, A+CHD+E 16. A+B 2C, A+C > D, Numerical 17. AB, 2B eC; 2A= Bs C; 2A9 2B, 2B C. Analytical INDEX OF MECHANISMS IN SECTION 3 1. Diffusion and n-th order reaction 2. Enzyme kinetics, Michaelis-Menten equation 333. Enzyme kinetics with reversible decomposition 4, Cell growth rate. Monod equation 5. Solid catalyzed reaction. Langmuir-Hinshelwood mechanism 6. Chain polymerization 7. Rice-Herzfeld mechanisms of thermal decompositions 8. A+3B > C+D by a three step process 9. ABCD from the components by a three step process 10. A+2B » C+2D by a two step sequence 11. Acetaldehyde thermal decomposition 12. Pyrolysis of ethyl nitrate. Data from a CSTR 13. Decomposition of ethane 14. Nitrogen peroxide decomposition 15. Hydrogen and bromine reaction 16. Phosgene from CO and Cl, 17. Photochemical oxidation of phosgene 18. Decomposition of ozone catalyzed by chlorine 19. Ozone decomposition 20. Decomposition of diethyl ether 21. Chlorination of n-heptane 22. Decomposition of nitrous oxide 23. Phosphorous acid 24, Acetone decomposition 25, Thermal cracking of propane 26. A kinetic reforming model This chapter deals with rate equations of chemical reactions carried out under batch or flow conditions. The applications are to identifying mechanisms of reactions, to sizing equipment, and to evaluating its performance. Tables 2.1 to 2.6 are an overview of the substance of this chapter. The participant A is identified by subscript a. Thus the concentration Js Cy, the number of mols is n,, the fractional conversion is x,, the partial pressure is p,, the rate of decomposition is r,. The capital letter A also is used on occasion instead of C,. The flow rate in terms of mols is ny but the prime (’) is left off when the meaning is clear, The volumetric flow rate 1s V', reactor volume is ¥, or simply V of batch reactors, the total pressure is om 2.1. RATE OF REACTION The term “rate of reaction" means the rate of decomposition per unit volume, , mols/(unit time)(unit volume) (2.1) > = Pao Xa _ 2 FP Ge! Ma = aol I-x,) (2.2) where x, is the fractional conversion of substance A. A rate of formation will have the opposite sign. The negative sign is required for the rate of decomposition to be a positive number. When the volume is constant, ac, - we only at constant volume (2.3) Law of mass action. The effect of concentrations on the rate is isolated as 34ry = KE(C, Ge.) (2.4) where the specific rate, k, 1s independent of concentrations but does depend on temperature, catalysts and other factors. The law of mass action states that the rate is proportional to the concentrations. For the reaction aA+ B+ cC +... 9 IL + m+... (2.5) the rate equation is 14M, Payp(PeyaPeyr r Vae RGGI. (2.6) = k.CRCICE..., at constant volume (2.7) The exponents p, q, T, ... are empirical, but they are identical with the stoichiometric coefficients a, b, c, ... when the stoichiometric equation truly represents the mechanism of reaction. The first group of exponents identify the order of the reaction, the second group the molecuiarity. Effect of temperature. The Arrhenius equation relates the specific rate to the absolute temperature, k = kgexp(-E/RT) (2.8) = exp(A-B/T) (2.9) ink =a-% (2.10) E is called the energy of activation and kg the preexponential factor. From experimental data, the constants are found with the linearized plot, Eq 2.10. 2.2. CONCENTRATION, MOLS, PARTIAL PRESSURE, MOL FRACTION Any property of a reacting system that changes regularly as the reaction proceeds can be formulated as a rate equation which should be convertible to the fundamental form, Eq 2.6. Examples are rate of change of electrical conductivity or of pH, or of optical rotation. The commonest other variables are partia) pressure p, and mol fraction N,. The relations between these units are, n Pa (2.11) ny = VC, = aN = % where subscript t denotes total and m the total pressure. For ideal gases, La vy ry, ny = TEC, = oy = ae = FM (2.12) Other volune explicit equations of state are sometimes necessary, such as the compressibility factor equation, V = 2RT/P, or the truncated virial equation, v = Btisp). the quantities z and B are not constants so some kind of averaging may be required. More accurate equations of state are even more difficult to use but are not often justified for kinetic work. Designate 8, as the increase {n the number of mols per mol decrease of substance A according to stoichiometric Eq 2.5, by ateror? (2.43) The total number of mols present is De = Myo*SA(NGO-Ma) = Nyt Xe = NegtSpXy T° (2.14) It then follows that -o Se _ 5 Ste (2.15) at vat ~ Peat . 35At constant volume, any ac, oy apy aN, any a Yacma "ha * Mae (2.18) At constant pressure” dn, dc, dn, ny (dC, /dt) 1. AT), ar 1), BufaCy at 7 rae “aE = qaRTE Cc, (2.17) dn. n, (dp,/at) ze Sp (2.18) any ny (aN, /at) ae Ta (2.19) An equation like Eq 2.7 can be written in terms of partial pressures and mol fraction as Fy = KeC'= kPa = Kye (2.20) The relation between the specific rates is k= (Rte, = Ey %y (2.21) Typical units of a first order reaction in terms of different variables are: Variable Rate equation ‘Typical units of k Concentration r= KC sect Partial pressure r = kp nols/(Liter) (sec) (atm) Mol fraction r= kN mols/ (liter) (sec) 2.3. REACTION TIME Flow reactors usually operate more nearly at constant pressure, thus at variable volume with gases. An apparent residence time is defined as the ratio of the reactor volume to the inlet volumetric rate, Papp = VeZVo (2.22) The true residence time is found by integration of the rate equation, t =f dv,/v' =f dn/rv’ (2.23) Both the reaction rate, r, and the volumetric flow rate, V’, must be known in terms of the number of mols, n, of the reference component. The apparent residence time is popularly used to indicate the loading of a reactor. A related concept is that of space velocity which 1s the ratio of a flow rate at STP (usually 60 F, 1 atm) to the size of the reactor. The most common versions in typical units are, GHSV (gas hourly space velocity) = (volumes of feed as gas at STP/hr}/(volume of the reactor or its content of catalyst) = SCFH gas feed/cuft LHSV (liquid hourly space velocity) = (Volume of liquid feed at 60 F/hr)/cuf t= SCFH liquid feed/cuft. WHSV (weight hourly space velocity) -= (1b of feed/hr)/(1b of catalyst) 2.4 CONSTANTS OF THE RATE EQUATION The problem is to find the constants of rate equations such as aC on =- Hake (2.24) 36or pe — Ek exp(a-p/t) cm (2.28) or ra = kCRCR. (2.26) Experimental data that are most easily obtained are of (C, t), (p.t), (r,t) or (C,T,t). Values of the rate are obtainable directly from measurements on a continuous stirred tank reactor (CSTR), or they may be obtained from (C,t) data by curve fitting and then differentiating. Several methods are conmonly used to find the constants. (4) Linear regression of the rate data. Take logarithns of Eq 2.25 Inr = ink +*A~-B/T +n In€ (2.27) The linear variables are In r, 1/T and In C. The term (In k + A) Is a single constant. A multilinear regression program can be used to find the constants, or only three sets of data can be used and the constants found by simultaneous solution of three linear equations. For a linearized Eq 2.26 the linear variables are the logarithms of r, C,, Cp,....The logarithmic form of Eq 2.24 has only two constants so the data can be plotted and the constants read off from the slope and intercept of the best straight line. ii). From the integrated equations. The integral of Eq 2.24 is - c Le (2.28) cot ot Tee wed yo)" 1), when n #1 (2.29) A value of n is assumed and values of k are calculated from the data. The correct value of n has been chosen when the values of k are nearly constant Integrals of some common rate equations are summarized in Table 2.2, Others are developed in problem Section 2 of this chapter. An(C/Cy) = k(t=to), when n = 1 (2.30) azey* = (1c) + ktn-1)(t-tg), when n #1 (2.31) A value of n is assumed and values of the left hand side are plotted against (t-to). The correct value of n has been found when the data are collinear. The correct value of k is found from the slope of the best straight line Before an expression such as Eq 2.26 can be integrated, the RHS must be expressed in terms of the single variable C,. From the stoichiometric Eq 2.5, Bao-Ma _ Peo~Py _ MoT | Bao"Pe a > ny _ Ayonx Co Ry - (2.32) Rygubx/a ~v Rggrex/a (2.33) . Vv Also, 37Cro = BC yo-Cy) Ce = Cop ~ ElCyo-Cy) (2.34) Thus the rate equation becomes ree bok 2 1 Pe Vat Vodt Nyo~* p Myo-b¥/A, 4 Neg-Cx/a. = kOY ae eae (2.35) In terms of concentrations, Ta =~ ae = kCRCECE. = KOZ (Cyo~B(CyonCy) 39 [Ceo ElCaomCa) > +: (2.36) 2.5.1, Multiple reactions. When a substance participates in several reactions at the same time, its net rate of decomposition is the algebraic sum of its rates in the individual reactions. The rates of the individual steps are identified with subscripts: (dC/dt),, (dC/dt)2,.... Take this group of reactions: c 1 A+B> 2 A+CSD+E 3 3 where itis not necessary true that 2 Yy 4. Ateonstant volume = 4) a 1 nf CHtse= DaPesNGre= Ge ° vere 0 LRT balan a CCompated integrals for some values of and Bargin Tobie 2-2 5, Idea! gases at censtont pressure aM net ing rk 6, Temparature ofect onthe specie eo oxplE/AT = xa 6/7) igebraie sur of the take the theee Fa tae Fant as = Cla” Oly + Cae Clo hCala Fe EyCaa + Ea FEC Fema ‘The number of independent rato equations is the sama a the umber of independant stoichiomettie relations. In the present ‘example, reactions 1 and 2 ore a cevarsible action and are not indapendent. Accordingly, C and Cy, for example, can be eliminated from the equations for r, and f, which then bacome an Integrable. system, 43TABLE 2.1 Ccontinued? 8, Mass tronserroslstance: Cus= interacial concentration of rexctant A Sense, coeter=i(e,-Z) om , =n “The relation between 1, and C must be establishes (numerically it ‘eed be) trom the second tine Belore the integration con be competed solid catalyzed rections. These wvotve simultaneous diffusion, adsorption and surface reaction. See Chapter o. 10. A continuous sted tank reactor batery (CSTR Matecalbatances: Mom nb ra Mayer yr forthe jth stage For 8 fist order tesetion, with 4 = KC, Su, CTF RUIN FEED FED 1 TRY" {or tants in series with he same temperatures ond residence times y= Ml Vi, where Vs the volumetric How rata 11, Plug tow reactor (PFA ng Sb sex Sha sczet Hoy VIMO v oy, 12, Motarie! snd eneray balances for batch, CSTR and PFR are in Tables 2.3, 2.4 and 2.5 13, Notation A.BR,S ace panticipants in the reaction; the leitersslso are used to represent concenvations G.= Al, 0F nf/V, concenteation = mols of componeat jin the reactor 1} = molal flow rate of component | % = volume of reactor Vie volumetric low rate y= swoichiometric coefficient Gm rate of reaction of substance # {molffonittimeliunit volums}] if m eenpirial exponents in # rate equation 44TABLE 2.2. Some Isothermal Rate Equations and their Integrals 1 A+ products: Ae ae A Mbo epl-Ktty, Get 1 ere. aed 5 THCq-AD ASTER tos 2. A¥B- producte: FAB=HIA+ By A) AdA+ BoA) Be Het) 3. Reversible reaction A by A~ fg Ba ~ AD + DA ily Be) byAg~ be Tht RIAq bat Bah (eee +s 4, Reversible reaction, second order, A+ 8 & AB ~ KAS = KALA + By — Ah = kale + Ry — Allg # So Al aA + BA y oak hy Br bylBy~ Aa) + kel2Ay # Ro + Se) UA Ralls + 5) oir Zara + B aca p ere Ly [(2adokBoa\(204e 090) 2 ell Neearaza)) 9° Kt okt Bt q)\2oAw pf 5, The reaction v/A+ 8-2 v/R t vl* w Qt ° © Solve £9 (6}10 ind T= f1C,}; combine Eqs. (1) and (2) and integrate os @ 4g CP ersar semen yature and time a8 a function of composition are ehawn fortwo values of UA/Y, or a paniculat ease represented by at walsoo = Ty Fa [sooo srs AEE r= exp(te-s000/7) T3860 Cont on : i i os 46TABLE 2.4. Material and Energy Balances of a CSIR ‘The sketch identifies the nomenclature Mian residence tine few eo ‘Temperature dependence kseaple'— 270 a Rate equation pea kOT = det = 15 x= (Coe CNG © Matai batece: 6 Enthalpy balance Deity E notes O— Atma ni) © (Ler n eh = thant | AG er © For the rection #4 +B (A +55, 8G, = Gye 5Cpu — Cpu Ga oy \When the heat eapaciies are equal and constant, the heaCbalance is GoW (T= Te OAH ICo~ 6) 100) qs (2) and (5) combine to 500 Fes and Eq. (10) becomes, "= 2{0.9|7—350)—25x1,, Meal/nr “The tomperature ane the rate of heat input @" are tabulated as functions ‘ofthe residence time for conversions of 30 aad 85%| TABLE 2.5. Material and Energy Balances of a Plug Flow Reactor CPFRD The balances are made over 8 different Fate equation: volume d¥, of the reactor Dianeter, 0 ayant o = a enestea' vin o Enthspy balan: ath thet [86.67 @ co=uin nas, Merny sui =n, = hay o 60+ AH, dni Sm dh =D ny oT @ BE AH AUT = TD a7, Fd o z a Ta [At constant Ty, Ea, (7} may be integrated numerically to yield the emperatere a¢ 2 function of the number of mols T= oto @ ‘Then the resctor volume is found by integration ° Adiabate process: ao=0 to The balance around ane end of he reaetor is Trobe Ehaloia=oid= Lo = Soa [Ger on in tren tempaature aT entapis Ha 0 ora Sa f Guat wa, Aciabaeproznts with AC, =0 and with constant heat capaci Mea ray btlaltio= aw . Tiny “Ti exprsson i ubeiute instead of Ea (8) 10 fing he volume wth at 4BTSBLE 2.6. Material ana Energy Balances of facked Bed Reactors Difusivity and thermal conductivity are taken appreciable only in the redial direction ‘Mataral balance equation: Be) dene 0 Enecgy balance equation: aT_k (27,197) she, Be ae (Sate ac, "9 a ithe net —_—__ x10, 7) = x0 * 3) 10, = Te rR ‘a z za ithe center: § 5 « roa, fo 6 0 > Alte walt % 2» & @ aru Zelir-n ” When the temperature T° of the host 1ransfer medium is not constant, snother enthalpy balance must be formelated to relate T” with the process temperature. ‘A.vumerical solution uf these equations msy bs obtained in terms of Fite diferonce equivalents, taking m radial increments and ‘ones. With the following equivalents for the derivatives, the solution ray be caried out by direct aeration: ®) a 0) a Expressions for the x-dervatives are of the same form: ‘res eate of cbaction, a function of sand T G= mass flow rate, mazs/timel(superficial cross section} um linear velocity itsivity ice thormal conductivity 49PROBLEMS, CHAPTER 2 P2.01.01. STOICHIOMETRIC BALANCES The reactions are 3A # 2B and 8 + 2C 4D, two reactions involving fou components. Express C and D in terms of A and B and formulate a pair of solvable ODEs. Consider the reactions to proceed consecutively, and identify intermediate concentrations of those substances that participate in more thay one reaction, in this case substance B. Eliminate the intermediate concentration B, from these two equations 2 By = BytS(Ag-A) CorC = Be -S = B+D-Do C= Cor $lAgrA)-2(By-B) a 2 D = Dg +5 (Ag~A}+ (Bo-B) (2) The differential equations are | dA 3. 2 ae 7 KA +KB (3) aB 2. 2. Se = KA gB?-KkBC™ = kyAP kg BK BECg-G(AgA)~2 (By-B)]? w A and B are found as functions of t by integration. Then C and D are found algebra. “The reactions are AaB Awe B+2C3D They are assumed to proceed consecutively. Accordingly the relations betwee the conversions are omy = B,-By a Ay-Ay = Cy-Co (2) B,-B = 0.5(C,-C) = D-Dy (3) Only D needs to be found algebraically. Eliminate A, making AorA = By-By + Cy-Co @ From (3), B, = B+ (D-) cy = C+ (2(D°Dg) Substitute into (4) and solve. so(Ag-A) + (By-B) + (Cy-C) p=n+ereree (5) The differential equations are GA ~OytkglA g = kyA-k BC" SE = ak spc After integration for A, B and C, D = f(t) is found with (5). P2.01.03. PAIR OF SECOND ORDER EQUATIONS The reactions are AtB = CHE Atl + DHE The relations between the changes in concentration are wm (2) D-Dy Combine these relations so as to eliminate the concentrations with subscript 1 and to find C, D and E in terms of A and B. = CotBy-B (AgrA) -2(By-B) (3) =(By-B)-(Dy-D) ~ (By-B) = (Ag~A}#2 (By-B) = = (A-A)+ (BoB) (4) =(AgrA) (5) Eys (3), (4) and (5) are the desired stoichiometric relations. The rate equations for A and B are, aA at ae at = -K,AB - kAC = -A(k,Btk, [Co~ (Ag“A) #2 (Bo-B) ]} (6) K,AB 7 P2,01.04. THE REACTIONS 2A » B, 2B + € The relations between the changes in concentration are = By-By @ = Cy @ Combining, A B, = Bot < B + 2(C-Co) "A. Bo-B Cy = * +> a) The differential equations are SL= k,A-k, After integration, C is found from (3). THREE REACTIONS, FIVE COMPONENTS The reactions are 2A > BIC, AXC > D, AWD FE D and E are to be found in terms of A, B and C. The stoichiometric balances are = C,Co mM D,-Dy (2) AjrA = D,-D = E-Eg (3) Eliminating A, and Aj, Ap-A = 2(By-B) - (Dg-D) - 2(Eo-E) (4) Also, AgrA = 2(Cy-Co) +Cy~C+E-Ey C#(Ag-A)#(Eq-E) #2Cy ge ee = Co“ (BoB) Rearranging, Co-C = (Ag-A)43(By~B)+(Eg-E) (5) and EgrE = ~(Ag~A)-3(Bg-B) + (CC) (6) Then, Dg-D = ~(Ag~A)~2( (By-B)-2 ( (EE) «) = (Ag"A) #4 (BoB) -2(Cy-C) mM The differential equations are dA 2, HSE = iAP agAC HAD = ALKA #kaCtkg[ (Ag~A) +4 (By-B)-2(Co-C) +Do] } dB_ Kia ~ae a ac Ly Kt Ge 7 Alp A + RC) After these three equations have been integrated, D and E are found from Eqs (6) and (7). 2.01.06. THE SHIFT REACTION Carbon monoxide is shifted to the dioxide for ease of removal by the reactions CO+3H, @ CHy+H{O, A+9B 4 C+D COFH;O # Hy+COp, AMD § BYE Components C, D and E are to be expressed in terms of A and B. The stoichiometric relations are 52AoAy = CCy = Di-Dy a@ A,-A = D,-D = B-B, = E-Ey (2) Eliminating Ay, (Ag-A) = =(Cy-C) - (EgE) (3) Solving for By, B, = B+3(Cp-C) = B4(E-E) which becomes, (By-B) = ~3(Cg-C) + (EgrE) @ Combining (3) and (4), (CoC) = tpn) = 408,-8) (s) Substituting into (3), (Eq-E) = = (Ap-A)~(Co-C)4 3 1 = -FlAg-A)+G(Bo-B) (6) Solving for Dy, Dy = DotC-Cy D+E-Ey from which Do-D = (Cy-C)=(Eg-E} = FA ADF (8-8) 7) Items (5), (6) and (7) are the required stoichiometric relations. The differential equations of the two reversible reactions are GA AB? ,CD-KgAD+K BE TK AB“+K,CD+K3AD-K,BE Substitute for C, D and E before integrating. REACTIONS. For the reactions, A+2B ® C and A+2C » D+2E, only A and B are present initially. Find the minimum number of differential equations that will give the composition of the mixture as a function of time, Let x = amount of A converted by the first reaction y = amount of A converted by the seccnd reaction A + 2B oe C Agr By-2x x A +20 9 D+ 2E Agx-y x-2y yay The final values are A= boox-y B = Bo-2x C= x-2y Dey E = 2y Elimination of x and y results in 933(Bp-B)-4(Ag-A) ss te qa 2fAg~A)~(By-B) = z (2) E= 2D (3) The differential equations are . dA _ 2 ~98 = kyaB™ + aac’ 2, Ke 2, = ACK:B'+ Z13(ByB)-4(Ay~A) 173 aB “in = 2ky. AB? After these have been integrated, the other compositions are found from (1) (2) and (3). 2.02.01. REACTION, gA » Products, AT CONSTANT ¥ OR CONSTANT P ‘At constant volume. the rate equation 1s sc ae and the integral is = kct . 1 egia k(toty) = giyt€ c Ca¥ (q-t)kCH? (t= tg)? exp[-k(t-tg)] q 1 At constant pressure qa = vB + vc + (vytygt.. I-a 3, = Wetter a q The total mols and the ideal gas volume are Dy = Neg + Sy (qo) v= BE and the rate equation, [ngot8g(ngg-n) The variables are separable, ot pao [Megt8qlMag-,)] K (toto) (He? = se? 0 < dn, ‘a ng When q = 2 for instance, sa