Modeling Considerations 5 for Control Purposes ‘The mathematical modeling of physical and chemical phenomena, encountered in processing systems of interest to chemical engineers, is a form of scientific art. Like every type of art it does not conform to, nor obey, clearly specified rules and recipes. It is an expression of all the educational background and previous practical experience of the modeler. Therefore, if previous sections have generated more questions to the reader than answers to specific modeling problems, he or she should not despair. Good and efficient modeling is acquired slowly with ever-increasing ability. In this chapter we attempt to focus the mathematical modeling on control purposes and needs. Thus we will examine the following issues: 1. Starting from the state-variables model, how one can develop an input-output model which is very convenient for control pur- poses? 2. Using the mathematical model of a process, how can one deter- mine the degrees of freedom inherent in the process, and conse- quently, identify the extent of the control problem to be solved? ‘We will close this chapter with some general guidelines that will help the control designer to formulate the scope of modeling for control purposes. chap. 8 Modeling Considerations for Contr! Purposes 81 5.1 The Input-Output Model Every chemical process and its associated variables can be described pictorially as shown in Figure 5.1. The main block represents the proc- ess, while the arrows indicate the inputs and outputs of the process. ‘A mathematical model that is convenient and useful to a control system designer should conform with the picture above, (i.e., be such that, given the values of the inputs, it provides directly the values of the outputs). In particular, the model should have the following general form for every output; output = flinput variables) Using Figure 5.1, the relationship above implies that, Y=flmy, My... smusdy, da. Such a model, describing directly the relationship between the input and output variables of a process, is called an input-output model. It is a very convenient form since it represents directly the cause-and-effect relationship in processing systems. For this reason it is also appealing to process engineers and control designers. ‘The mathematical models we learned to develop in Chapter 4 using state variables are not of the direct input-output type. Nevertheless, they constitute the basis for the development of an input-output model. This is particularly easy and straightforward when the state variables coincide completely with the output variables of a process. In such a case wwe can integrate the state model to produce the input-output model of the process. a) fori=1,2,....m Disturbances 4 Manipuated my 7 outputs Figure 5.1 A chemical process and its associated inputs and outputs. Example 5.1: Input-Output Model for the Stirred Tank Heater Consider again the stirred tank heater discussed in Example 4.4. Assume that F= F, which yields d¥/dr =0, leaving the total energy82 Modeling the Dynamic and Static Behawor of Chemical Processes Part il balance as the only equation of the state model, vile gr Ty+2 (4.50) a bey ‘The amount of heat Q supplied by the steam is given by Q=UA(Ts-T) where U is the overall heat transfer coefficient, 4, the arca of heat transfer, and 7, the temperature of the steam. In eg. (4.5b) replace by its equal and take V(r lr nn ry a Pg) ro opel na kre a mea where eelie LAR ath Sle cat Equation (5.1) is the mathematical model of the stirred tank heater with T the state variable, while T; and T,. are the input variables. Let us sce how we can develop the corresponding input-output model ‘At steady state, eq. (5.1) yields OF aT, a1 T+ Kas 52) where T;, Ti», and Ta, are the steady-state values of the corresponding. variables, Subtract (5.2) from (5.1) and take ar) +a(T-T,) (T.- Ty) + K(Ta = Ts) dt OP gata try + KTy (53) dt T where T= T~ T,, T)= T;~ Tiss and Ty = Tx Ty, indicate the devia- tions from the corresponding steady-state values. ‘The solution of (5.3) is en[t rie xr] (54) Tecetses f ‘Assuming that initially the heater is at steady state, [ie, at ¢= T’= 0), we easily find that c; = 0. Therefore, eq. (5.4) gives ‘chap 8 Moding Consideratons for Control Purposes 83 ~ 1 ro! Sriput tops i Input-outpt mel Figure 5.2 Input-output model of stirred tank heater. ~ «elas xr] re [lew [E rear] a (55) Equation (5.5) expresses the relationship between the inputs (T, T's) and the output (7°), and constitutes the input-output model for the tank heater. This relationship is also depicted pictorially in Figure 5.2 Example 5.2: Input-Output Model for a Mixing Process Consider again the process of mixing of two streams discussed in Example 4.11 (Figure 4.8). Assume that Fy + F)= Fy, which implies that AV /dt = 0 (ie., V = constant). The heats of solutions are independent of the concentration, which implies that (4,,- Af) =(47,-471,)-0 Then the state equations (eqs. (4.138) and (4.14a)] are reduced to the following: den, (Fi, Fs) a \yovs ey and as (FBV ey a \y v es A steady state. (5.6) and (56) sea Fy. Fs) o+(£14%) (30 and A o+ (Ee (B48) (3)‘Modeling the Dynamic and Static Behavior of Chemical Processes Parti Hey ' Figure $3 Input-outpu model ofa mixing process ‘Subtract (5.7a) from (5.6a) and (5.7b) from (5.6b) and take eas acy = Pen, + Fes, (5.8) a ay, aTy » , F " Fs aTy=— Ti+ Tie Ki (5.8b) 3 arya Fir PTs KO (58) where chy, Chy Chy Ti Tis Ti, and Q’ are deviation variables defined as follows: and Ye Ti- Ti The Ts-Try Tie T-Ty @=0-O Also, Aik 1 ape md Ka The solution of (5.8a) and (5.8b) yields " oe Fics 9 Ft ealt)= ce" (Ete Etey,|dt (5.98) ninco ser [ e*[ Bei, + ci] and Thy ~cre* ve [et [é rie Brie xe Ja (5.9) Outputs chap. 5 Modeing Consideratons for Contol Purposes 85 If the system is initially (Le., at = 0) at steady state, then Ca(t=0)=0 and TH(t=0) +0 and it results in c1 = c2 = 0. Equations (5.9a) and (5.9b) represent the input-output model for the ‘mixing process, and is shown schematically in Figure 5.3. Remarks 1, In Examples 5.1 and 5.2 the output variables coincide with the state variables of the two processes. Consequently, in order to develop the input-output model we need only solve the differen- tial equations of the mass and energy balances. This is not always ‘true, Take as an example the binary distillation column model (Example 4.13 and Figure 4.10). For this system we have: State variables: Liquid holdups: My, M2y..., Mj... Mv, Mao and Mp Liquid concentrations: x1, X2,..-5 Xp--+, XN. Xp and Xp Output variables: Distillate product flow rate and composition: Fo and xp Bottoms product flow rate and composition: Fy and Xp We notice that we have many more state variables than outputs. For such systems, the development of the input-output model is quite involved and difficult. Figure $.4 depicts pictorially the input-output model that we would like to develop for the binary ideal distillation column. 2. In subsequent chapters we will study the method of Laplace transforms, which allows a much simpler development of input- output models from the corresponding state models. Disturbancet | Foo a Masiaated Fr—e] tnpuroutt FMF ostpty varables Fed model Le ho -— x, Figure 4 Inputs and outputs of a binary distillation column,86 Modeling the Dynamic and Static Behavior of Chemical Processes Part i) 5.2 Degrees of Freedom The degrees of freedom of a processing system are the independent variables that must be specified in order to define the process com- pletely. Consequently, the desired control of a process will be achieved when and only when all the degrees of freedom have been specified. ‘A good understanding of how many degrees of freedom are inherent in a process, and which they are, is very crucial for the design of effective controllers. For a specified system, its mathematical model is the basis for finding the degrees of freedom under both dynamic and static conditions. Let us start with two characteristic examples. Example 5.3: Degrees of Freedom in a Stirred Tank Heater ‘The mathematical model of a stirred tank heater (Example 4.4) is siven by dh a (44a) Ah Le FT, ~1)+2 4.56) ai Pe ‘When eqs. (44a) and (4.5b) are solved simultaneously, we can find how h (liquid level) and T (liquid temperature) change with time when the inputs (7;, F;, Q) change. Let us ask though the following two questions: 1. Is solution ofthe equations possible? 2. If solution is possible, how many solutions exist? ‘To answer these questions, let us count equations and variables: number of equations =2; eqs. (4.4a) and (4.56) 3; AT, Fi, F, Tand Q ‘We have assumed that A, p, and c, are parameters with given constant values. We notice that number of variables number of variables > number of equations Consequently, the answer to the first question is: yes, there exists at least ‘one solution to the equations modeling the tank heater. With respect 10 the second question we easily conclude that there is an infinite number of solutions since we can specify arbitrarily the values of four variables (4=6~2) and solve eqs. (44a) and (4.5b) for the remaining two vari- ables. The arbitrarily specified variables are the degrees of freedom and their number is given by the following obvious relationship: ‘J'= (number of variables) ~ (number of equations) Suppose that we specify the values of the four variables Fi, Ti, F, and Q. Then we can integrate eqs. (4.4a) and (4.5b) and find how and T Cap. 5 Modeling Considerations for Control Purposes 87 change with time. If we give different values to F,, or T,, or F or Q, we find that and T change differently than before. Consequently, if we want h and T to change in a prescribed manner, we should not have any degrees of freedom (i.e., variables that can take arbitrary values). This leads us to the conclusion that in order 10 specify a process completely, the number of degrees of freedom should be zero. Example 5.4: Degrees of Freedom in an Ideal Binary Distillation Column Consider the model for an ideal binary distillation column developed in Example 4.13. We have: Number of | equations Origin N41 Equilibrium relationships (ea. (4.20)] Ny Hydraulic relationships [eq (4.21)] 2 Balances around the feed tray [ea (4.22a), (4.22b)] 2 Balances around the top tray feqs. (4.238), (4.238)} 2 Balances around the bottom tray [egs. (4.24), (4.24b)] 20 ~3) Balances around the ith ray; 7+ 1, N.f[eqs. (4.25a),(4.286)] 2 Balances around the reflux drum [eqs. (4.263), (4.26b)) 2 Balances around the column base leas. (4.27a),(4.276)) Totals ave Number of variables Type Ne2 0 x fav ND. B- liquid compositions Naty sfe--.N, B vapor compositions Ne. 0M Jone N, RD, B liquid holdups N Le bet Qeo fv N liquid flows 6 Fin Gn Eos Fs Fs Toial=4N+ 11 The number of degrees of freedom for the ideal binary distillation col- umn is S=(N +1) - GN +5)=6 ‘That is, we need to specify the values of six variables before we can solve the model of the binary distillation, The observations made and the conclusions drawn from the two examples above can now be generalized for any processing system described by a set of E independent equations (differential and/or algebraic) containing V independent variables. The number of degrees of freedom for such a system is given by SeV-E88 ‘Modal the Dynamic and Static Behavior of Chemvical Processes Parti According to the value of /, we can distinguish the following cases: Case I. If f = 0, we have a system of equations with equal number of variables. The solution of the E equations yields unique values for the V variables. In this case we say that the process is exactly specified. Case 2. If f > 0, we have more variables than equations. Multiple solutions result from the E equations since we can specify arbitrarily f of the variables. In this case we say that the process is underspecified by F equations (i.e., we need f additional equations to have a unique solution). Case 3. If f <0, we have more equations than variables and in general there is no solution to the £ equations. In this case we say that the system is overspecified by f equations (i.e., we need to remove f equa- tions to have a solution for the system). Remarks 1.It is clear from the analysis above that sloppy modeling of a process may lead to a model that does not include all the relevant equations and variables or includes redundant equations and variables. In either case we have an erroneous determination of the degrees of freedom, which may imply incorrectly that we have an infinite number of solutions or no solution at all. 2. The presence ofa control loop in a chemical process introduces an additional equation between the corresponding measured and manipulated variables, thus reducing by one the initial number of degrees of freedom for the process. Example 5.5 The stirred tank heater is modeled by two equations containing six variables, thus yielding four degrees of freedom (Example 5.3). This is, true if the effluent flow rate F is determined by a pump, valve, and so on. Let us suppose that this is not the case and that the liquid flows out from the tank freely under the hydrostatic pressure of the liquid in the tank, In this case there is an additional equation relating F to h (eg., F = BVh), which reduces the number of degrees of freedom by one. Example 5.6 Consider again the stirred tank heater, but now under feedback con- trol (Figure 5.5). Control loop | maintains the liquid level at a desired value by measuring the level of the liquid and adjusting the value of the effluent flow rate. Therefore, control loop | introduces a relationship between F and h. Similarly, control loop 2 maintains the temperature of the liquid at the desired valuc by manipulating the flow of steam and thus the flow of heat @. Consequently, control loop 2 introduces a relationship hetween Oand yr Chap. Modeling Considerations for Control Purposes 89 Contoler Controle Figure 5.6 Feedback loops of a stirred tank heater It is clear from the analysis above that the two control loops introduce ‘two additional equations, thus reducing the degrees of freedom by two. 5.3. Degrees of Freedom and Process Controllers In general, a carefully modeled process will possess one or more degrees of freedom. Since for > 0 the process will have an infinite number of solutions, the following question arises: How do you reduce the number of degrees of freedom to zero so that you can have a completely specified system with unique behavior? It is clear that for an underspecified system with degrees of freedom, we need to introduce f additional equations to make the system com- pletely specified. There are two sources which provide the additional equations: (1) the external world and (2) the control system. Let us examine them closer using the stirred tank heater as our example. Recall from Example 5.3 that the stirred tank heater possesses four degrees of freedom. Therefore, we need four additional relationships, independent of the modeling equations (eas. (4.4a), (4.5b)]. These are provided from the following considerations: 1. The feed flow rate F, and feed temperature 7; are the main two disturbances for the stirred tank heater and they are both speci- fied by the external world (eg, the unit that precedes the tank heater), Although the equations that specify F; and T; may not be known to us, nevertheless they exist and remove two degrees of freedom. Thus we have 4 - 2 = 2 remaining degrees of freedom.90 ‘Modaling the Dynamic and Static Behavior of Chemical Processes Parti 2, Acceptable operation of the tank heater requires that the liquid evel and liquid temperature in the tank heater be maintained at desired values. These two control objectives can be achieved with the two control loops shown in Figure 5.5 and discussed in Exam- ple 5.6. But the introduction of the two control loops adds two equations (see Example 5.6), thus removing the remaining two degrees of freedom. ‘Summarizing the observations above we conclude the following: 1. The external world, by specifying the values of the disturbances, removes as many degrees of freedom as the number of distur- bances. 2. The control system required to achieve the control objectives removes as many degrees of freedom as the number of control objectives. During the reduction in the number of degrees of freedom for a chemical process, care must be exercised not to specify more control objectives than it is possible for the particular system. Thus we can have at most two control objectives for the stirred tank heater. When we attempt to have three control objectives, we are led to an overspecified system with f <0. Example 5.7: Reduce the Degrees of Freedom of an Ideal Binary Distillation Column Return to the ideal binary distillation column (Figure 4.10). The sys- tem possesses six degrees of freedom (see Example 5.4), which are speci- fied as follows: Specification of the Disturbances. Two are the main disturbances for the binary distillation column: the feed flow rate Fy and the feed composition ¢. Their values are specified by the external world (e.g., a reactor whose effluent stream is the feed to the distillation column). Although the equa- tions specifying F; and c; are not known to us, they do exist and remove two degrees of freedom, leaving four for additional specifications. Specification of the Control Objectives. We can have up to four control “objectives since there are four remaining degrees of freedom. Acceptable ‘operation of the binary column requires that the following variables be maintained at desired values: 1. Composition of the distillate stream, xp 2. Composition of the bottoms stream, x» 3, Liquid holdup in the reflux drum, Mro 4, Liquid holdup at the base of the column, My Specifications 1 and 2 characterize the two product streams. Specifica- tions 3 and 4 are required for operational feasibility (i.e., we do not want Cchap.§ Modeling Considerations for Control Purposes 91 = Eon wae Steg Ree Loop 3 Fos Ted Distiate Loop! |" 1 produce Tak Ise UL ~+fcontrater Figure $.6 Feedback loops of a binary distillation column to flood or dry up the reflux drum or the base of the column). Figure 5.6 shows the four control loops that satisfy the foregoing four objectives. ‘We must note that these four specifications of the control objectives may differ, according to the particular operating objectives. For example, ina different application we may impose the following control objectives: Keep at the desired values the distillate flow rate Fp, its composi mn Xp, and the two liquid holdups Map and My Keep at the desired values the bottoms flow rate Fs, its composi- tion xp, and the two liquid holdups, Myo, and My, Care must be exercised not to specify more control objectives than the available number of degrees of freedom. In such a case the system ‘becomes overspecified and it is impossible to design a control system that satisfies all the desired control objectives. Thus it is impossible to design @ control system for the ideal binary distillation column that can satisfy the following six operational (control) objectives: Keep at the desired values the Fo, x0; Fas xe; Mro and Ma, Example 5.8: Degrees of Freedom of a Mixing Process Consider the nonisothermal mixing of two streams discussed inModeling the Dynamic and State Behawer of Chemical Processes Part! Example 4.11 (Figure 4.8). The mathematical model is given by the equa- tions Bae FFs (4.28) V Eo (eq, ~ en)Fit (em— Cu)Fs a 26 2 = ca Fils, - AAs) + em FilBils, - AAs) i (4.140) + pF ey Ti ~ Ts) + PFxep(T2 - TO number of variables = 16: Fy, Fa, Fy Cay Cay Cas Chap. 8 Modeling Considerations for Control Purposes 93 ‘Specification of the Disturbances. There are five main disturbances com- ing from the two feed streams: Feed stream: F,,T, Feed stream 2: Fs, Tzand ex; ‘The values of the disturbances are specified by the external worl. Specification of the Control Objectives. Considering five disturbances specified by the external world, we have only 8 ~ 5 = 3 degrees of freedom left. Therefore, we can specify up to three control objectives. These are: Keep the volume (V) of the mixture in the tank as well as the ‘temperature (7) and composition (c,,) of the effluent stream at desired values, Ty, Ta: To Q. Bey MAbs, SAlsy Mts number of state equations = 3 initial degrees of freedom = 16 — 3 = 13 Figure 5.7 shows three possible feedback control loops that satisfy the objectives above, Further limitations of the degrees of freedom are as follows: 5.4 Formulating the Scope of Modeling for Physical properties of the liquids are specified (ie., p and cy). Process Control The heats of solution AiTs;, AAs, and Ais, are functions of the corresponding concentrations, and the reference temperature To: It should be clear by now that efficient modeling of a chemical process is a nontrivial task but also very crucial for the design of a control system. Before closing the present chapter, let us emphasize some of the factors that will determine the scope of modeling for control purposes. Before attempting to model a process, we must pose the following questions and try to understand well their implications: AAs, =Suleay, Coy To) AAs, =fileay Coy To) Ally =filcay Cay TO ‘These three equations reduce the degrees of freedom by 3. Conse- quently, after the specifications above we have left, 13-5=8 degrees of freedom 1, What are the contro! objectives that we must satisfy? 2. What are the expected disturbances and their impact? 3. What are the dominant physical and chemical phenomena taking place in the process to be controlled? “The eight degrees of freedom are now specified as follows: stecam sucam 2 A clear understanding of these questions and their answers will help greatly to define and simplify: controller 1. The system that we will attempt to model tee 2. The mass, energy, and momentum balances that we should controller. develop 3. The additional equations that will be needed to complete the ‘mathematical model of the process (i.e., transport and kinetic rate expressions, reaction and phase equilibria relationships, etc.) They will also help to identify: Temperature controler 1. The state variables 2. The input variables (manipulated and disturbances) and 3. The output variables Figure 5.7 Feedback loops of a mixing process. ‘that the mathematical model should include.94 Modetng the Dynamic and Static Behavior of Chemical Processes Part Let us now examine each of the three questions posed above and see how they affect the modeling of a process for control purposes through a series of examples. Control objectives ‘As discussed in Chapter 2, the objectives that a control system is called on to satisfy may have to do with: Ensuring the stability in the operation of a process, or Suppressing the influence of external disturbances, or Optimizing the economic performance of a plant, or usually ‘A combination of the above All the objectives above are translated into quantitative expressions in terms of flow rates, temperatures, pressures, compositions, volumes, and so on, of the form variable x = desired value or variable x > or < bounding value where variable x = flow rate, temperature, pressure, volume, composi- tion, and so on. Itis clear, therefore, that if we have identified the variables x which define quantitatively our control objectives, the mathematical model that we will develop must describe how these variables change with time. ‘Also, it will help us determine what balances are needed for the devel- ‘opment of the mathematical model. Example 5.9 Consider the stirred tank heater discussed in Example 4.4. 1. Hour control objective is to keep the liquid level at a desired value, the only state variable of interest is the volume of the liquid in the tank (or equivalently the height of the liquid level) and conse- quently we need only consider the total mass balance. The disturb- lance of interest is the flow rate of the inlet stream, F;, while the ‘manipulated variables to be considered are the outlet flow rate F and the inlet flow rate Fi 2. If, on the other hand, our control objective is to keep the tempera ture of the outlet stream, T, at a desired value, we must consider both state variables: the temperature and the level of the liquid in the tank. This implies that we need write both total mass and energy balances. The disturbances of interest are the temperature and the flow rate of the inlet stream, while the available manipulated vari- How rate o' ene Cchap.5 Modeling Considerations for Control Purposes 95 3. Finally, if our control objectives are to keep the temperature of the effluent stream and the liquid level at desired values, we have a situation similar to case 2 above. Example 5.10 Consider the continuous mixing process discussed in Example 4.11 (Figure 4.8). We can distinguish the following control situations: 1. If our control objective is to keep the concentration of the effluent stream in A ata desired value, the state variables of interest are the volume of the mixture in the tank and its concentration in A. The relevant balances are those on total mass and on component A. The disturbances of interest are ca, (or Cx,), Fis C4; (0F Co,), and F>. ‘The available manipulated variables are Fy, F2, Fs, oF the ratio PulFi. 2.If,on the other hand, our control objectives are to keep the compo- sition and the temperature of effluent stream at desired values we need consider all three state variables (cas, V, Ts) and formulate all three balances (total mass, component A, total energy). In this case the important disturbances are cy, Fi, Ti, Cay Fx, and Ta. ‘The available manipulated variables are Fi, Fs, Fs, the ratio Fi/F;, and Q. Expected disturbances and their impact, The external disturbances which are expected to appear and affect the operation of a process will influence the mathematical model that we need to develop. Furthermore, disturbances with a very small impact on the operation of the process can be neglected, whereas distur- bances with significant impact on the process must be included in the model. This will determine the complexity of the model needed: that is, what balances and what state variables should be included in the model. State variables that are affected very little by the expected distur- bances can be eliminated from the model and, together with them, the corresponding balances. Example 5.11 Let us return to the stirred tank heater (Example 4.4). Ifthe feed flow rate (disturbance) is not expected to vary significantly, the volume of the liquid in the tank will remain almost constant. In this case aV/dt = A dh/dt ~ 0 and we can neglect the total mass balance and the associated state variable /. The mathematical model of interest for von- trol purposes is given by the total energy balance alone [eq. (4.5b)}, with temperature the only state variable. Remarks. Note that if the feed temperature 7; is not expected to vary significantly, but the feed flow rate F; is expected to change substantially, the mathematical model cannot be simplified as above, but it will be given by both eqs. (4.4a) and (4.5b).96 Modeling the Dynamic and Static Behavior of Chemical Processes Partil Similar results to those above can be deduced for the CSTR system (Example 4.10). If the feed flow rate is not expected to change signifi- cantly, the model can be simplified (i.e., d¥/di = 0) and it is given by ‘eqs. (4.9a) and (4.10). Physical-chemical phenomena in a process A good understanding of the physical-chemical phenomena taking place in a process can lead to significant model simplifications for control purposes. Such simplifications can be done by excluding from the balances (model) those terms that have small contributions. Example 5.12 Let us return to the continuous mixing process discussed in Exam- ple 4.11. Assume that for the particular components A and B of the mix- ture, the heat of solution does not depend significantly on the composi- tion of A and B. In this case (see Example 4.11) (Ais, ~ AAs) ~ 10s, Affe] ~ 0 and the total energy balance [eq. (4.4a)] can be simplified tothe fllow- ing 2ep¥ Ee pref, = 15) + pFagT:~ Ts) + In other words, we see that the nature of the mixing phenomenon leads to a simplification of the model. Furthermore, assuming that from all possible disturbances only the feed compositions c4, and cq, are expected to change significantly whereas the feed flow rates F, and F; and feed temperatures T, and T are expected to remain almost the same, we can omit from the mathemat- ical model the total mass and energy balances and from the set of state variables volume V and temperature Ts. Thus the simplified model is siven only by the balance on componeat A [eq. (4.13a)]. Example 5.13 Consider again the CSTR discussed in Example 4.10. If the heat of reaction for the particular reaction A — Bis very small and the temper- ature of the feed stream is not expected to change significantly, the tem- perature of the reacting mixture will not change appreciably. In this case the reactor can be assumed isothermal. We can exclude the total energy balance from the mathematical-model and the temperature from the set of state variables. Examples 5.9 through 5.13 demonstrate very simply but vividly how the mathematical model of a process can be simplified when we take Cchap.8 Modeling Considerations for Control Purposes 97 into account various considerations related to the nature of the process and the characteristics of the control problems. The control designer always looks out for such simplifications. THINGS TO THINK ABOUT 1, What is an input-output model, and how can you develop it from a state ‘model? When is this possible? 2. Describe a procedure that would allow you to develop the input-output ‘model for an ideal, binary distillation column. 3. Define the concept of degrees of freedom and relate it to the solution of E equations with V variables. 4. How many degrees of freedom do you have in a system composed of P phases with C components? (Recall Gibbs’ rule.) 5. How many degrees of freedom do you have in a system composed of P phases with C components if the mass of each phase is given (ic, Mi, Ms, -, Mp)? (Recall Duhem’s rule.) 6. How does the number of degrees of freedom affect the number and the selection of the control objectives in a chemical process? 7. Why do we claim that d disturbances reduce the number of degrees of freedom by a? 8. Why can a control system not be designed for an overspecified process? 9. Can you have the desired operation for an underspecified process? If yes, explain why. If no, explain how can you lift the underspecification, 10, Consider a system modeled by the following set of state equations: ax, a f(s X2 My, M2, Ms, dy, da) Bo fy 25 my de) Aaa, X35 Xs, ms, ms, dr, do, ds) where x,,x3, and x3 are the state variables, my, ms, and ms are the manipulated variables, and dj, da, and d; are the external disturbances. (a) How many degrees of freedom does the system possess? (b) How many control objectives can you specify at most? (©) Consider this system at steady state. How many degrees of freedom does it possess?98 Modeling the Dynamic and Static Behavior of Chemical Processes Part ll 11, A system is described by the following set of state equations: xy dt Find the degrees of freedom for the system at its dynamic state and steady state, Are they equal? If not, why? What are the implications on control in this case? dey jm, ma, di, a) Ea flim, ma ds fl didi) and 2 = films, ms ds) 12, What are the principal control considerations that affect the scope of math- ‘ematical modeling of a chemical process? 13, In what sense do the control considerations affect the mathematical model- ing of a chemical process? 14, What are the usual, general quantitative representations of the control objectives? In terms of what variables are they expressed? 15, How can the expected impact of the disturbances simplify the model of a process? Give an example other than that discussed in the text. 16. Give examples to demonstrate how you can simplify the model of a process by disregarding physical and chemical phenomena with a limited impact ‘on the behavior of the process. 17. Outline the steps that you should take during the development of a mathe- ‘matical model for chemical process control purposes. REFERENCES FOR PART I Chapter 4, Three exceptional references with a large number of process modeling examples are: 1. Process Dynamics and Control, Nol. 1, by J. M. Douglas, Prentice-Hall, Inc., Englewood Cliffs, N.J. (1972). 2. Process Modeling, Simulation, and Control for Chemical Engineers, by W. L. Luyben, McGraw-Hill Book Company, New York (1973). 3. Dynamic Behavior of Processes, by J. C. Friedly, Prentice-Hall, Inc., Engle- ‘wood Cliffs, N.J. (1972). For the development of the dynamic material and energy balances, the reader ‘could also consult the following book, from which Examples 4.10 and 4.11 are adapted: 4. Introduction to Chemical Engineering Analysis, by T. W. F Russell and M. M. Denn, John Wiley & Sons, Inc., New York, (1972). Par ll_ Modeling the Dynamic and Static Behavior of Cherncal Processes 99 ‘Additional references for material and energy balances at 5. Basic Principles and Calculations in Chemical Engineering, 3rd ed., by D. M, Himmelblau, Prentice-Hall, Inc., Englewood Cliffs, N.J. (1974) 6. Elementary Principles of Chemical Processes, by R. M. Felder and R. W. Rousseau, John Wiley & Sons, Inc., New York (1978). For the modeling of specific unit operations and reactors, there exist a large ‘number of textbooks that the reader could use. Not all models included in these books are convenient for process control purposes, but they could help you to develop simplified and useful models. Among all the available references the following constitute a partial list: For the modeling of chemical reactors: 7. Chemical Reaction Engineering, by O. Levenspiel, John Wiley & Sons, Inc., New York, (1962), 8. An Introduction to Chemical Engineering Kinetics and Reactor Design, by C.G. Hill, Jr, John Wiley & Sons, Inc., New York (1977) 9. Elementary Chemical Reactor Analysis, by R. Aris, Prentice-Hall, Inc., Englewood Cliffs, N.J. (1969). 10. Chemical and Catalytic Reaction Engineering, by J. J. Carberry, MeGraw- Hill Book Company, New York (1976). For the modeling of transport processes: 11, Transport Phenomena, by R. B. Bird, W. E. Stewart, and E. N. Lightfoot, John Wiley & Sons, Inc., New York (1960). 12. Mass-Transfer Operations, 2nd ed., by R. E. Treybal, McGraw-Hill Book ‘Company, New York (1968). 13. Heat and Mass Transfer, 2nd ed., by E. R. G. Eckert and R. M, Drake, Jr, McGraw-Hill Book Company, New York (1959). For more on reaction equilibria and phase equilibria the reader may consult the following books: 14, Introduction to Chemical Engineering Thermodynamics, 3rd ed., by J. M. ‘Smith and H. C. Van Ness, McGraw-Hill Book Company, New York (1975). 15. Chemical Engineering Kinetics, 2nd ed., by J. M. Smith, McGraw-Hill Book Company, New York, (1976). For an extensive discussion of the mathematical modeling of an ideal, binary distillation column and of a nonideal multicomponent column, the reader can consult the books by Douglas (Ref. 1], Luyben [Ref. 2], and Friedly [Ref. 3]. An100 ‘Modeling the Dynamic and Static Behavior of Chemical Processes Part interesting discussion of the difficulties encountered during the modeling of ‘chemical processes can be found in: 16. “Critique of Chemical Process Control Theory,” by A. S. Foss, AICHE J., 19, 209 (1973). 17, “Advanced Control Practice in the Chemical Process Industry: A View from Industry,” by W. Lee and V. W. Weekman, Jr., AIChE J., 22, 27 (1976). Chapter S. For additional study on the degrees of freedom and their impact con the design of process control systems, the reader should consult the follow- ing references: 18, Automatic Control of Processes, by P. W. Murrill, Intext Educational Pub- lishers, Scranton, Pa. (1967) 19. The Chemical Engineers Handbook, Sth ed., J. H. Perry (ed.), McGraw- Hill Book Company, New York (1974). PROBLEMS FOR PART I Chapter 4 ILA Consider the two systems shown in Figure PIl.1, System 1 differs from system 2 by the fact that the level of liquid in tank 2 does not affect the effluent flow rate from tank 1, which is the case for system 2. (a) Develop the mathematical model for each of the two systems. (b) What are the state variables for each system, and what type of balance equations have you used? (© Which mathematical model is easier to solve, that for system 1 or that for system 2? Why? Assume that the flow rate of an effluent stream from a tank is propor- tional to the hydrostatic liquid pressure that causes the flow of liquid. The cross-sectional area of tank 1 is 4, (f?) and of tank 2 is 42 (ft?Mfor both systems), The flow rates Fy, F, and Fy are in f/min. ine system | System 2 Figure PIL Peat m2 Modeling the Dynamic and Static Behavior of Chemical Processes 101 Develop the mathematical model for the system shown in Figure PIL.2. What are the state variables for this system and what type of balance equations have you used? All the flow rates are volumetric, and the cross-sectional areas of the three tanks are Ai, 43, and A3 (f", respec- tively. The flow rate Fis constant and does not depend on fs, while all other effluent flow rates are proportional to the corresponding hydro- static liquid pressures that cause the flow. Figure PIL2 Consider the two stirred tank heaters shown in Figure PIL.3. (a) Identify the state variables of the system. (b) Determine what balances you should perform. (©) Develop the state model that describes the dynamic behavior of the system. (@) How would you express the heats given by the two steam flows in terms of other variables? The flow rates of the effluent streams are assumed to be proportional to the liquid static pressure that causes the flow of the liquid. The cross- sectional areas of the two tanks are 4, and 2 (ft?) and the flow rates are volumetric. No vapor is produced either in the first or the second tank. Aa and A. are the heat exchange areas for the two steam coils, E Steam 100 pt Gremio) Steam 300 Fe (5 din Figure P13102 Modeling the Dynamic and Static Behavior of Chemical Processes Part ILA. Do the same work as in Problem I1.3 for the stirred tank heaters’ system shown in Figure PIL.4, For tank 1, the steam is injected directly in the liquid water. Water vapor is produced in the second tank. A, and As are the cross-sectional areas of the two tanks, Assume that the effluent flow rates are proportional to the liquid static pressure that causes their flow. ‘Avis the heat transfer area for the steam coil Steam, 40 ps ee by Water Ar ‘Steam, 200 psi ain) Figure PIA 5 Consider the mixing process taking place in a two-tank system (Figure PUL5). (a) Identify the state variables of the system. (b) Determine what balances you should perform. (c) Develop the state model that describes the dynamic behavior of the process, assuming that the heats of solution are strong functions of the composition. (é) How is the mode! simplified if the heats of solution are very weak functions of the composition? ‘Assume that the flow rates are volumetric and the compositions are in ‘moles per volume. The effluent flow rates are proportional to the liquid static pressure that causes their flow. 4, and 42 are the cross-sectional areas of the two tanks and A, is the heat transfer area for the steam coil Fa Ta tay 70 FT cay 0 | 4 Fu Tyas Steam, 50s er honey a Figure PILS IL6 Develop the state model for the batch mixing of two solutions (Figure PIL.6). Initially, the tank is empty. The volume of the tank is ¥ Ponti 17 ‘Modeling the Dynamic and Static Behavior of Chemical Processes 103 we, The flow rates are volumetric andthe concentrations are in moles (a) How long does it take to fill up the tank? (b) Show how would you find the composition and temperature of the mixture in the tank during the time that the tank is being filled up. ‘Assume that the flow rates are volumetric, the compositions in moles per volume, and that the heat of solution depends on the composition. Solution 2 Figure PIL6 Develop the state model for a batch reactor where the following reac- tions take place: atop tic atnp {All reactions are endothermic and have first-order kinetics. The react- ing mixture is heated by steam of 150 psig, which flows through a jacket around the reactor with a rate of Q (\b/min). me Consider the continuous stirred tank react reactor system shown in Figure PIL7. Stream 1 is a mixture of A and B with composition ca, and ¢», (moles/volume) and has a volumetric flow rate Fy and a tempera- ture T,, Stream 2 is pure R. The reactions taking place are: Reaction I: A+R + Py Reaction 2: B+2R +2. p; Fa. Ty pute R fr 2 E,Tocuseu me Oy Py FuTa tay em samt ton 4 Figure PULT Both reactions are endothermic and have second-order (reaction 1) and third-order (reaction 2) kinetics. Heat is supplied to the reaction mix- ture by steam which flows through a coil, immersed in the reactor’s content, with a heat transfer area A, (@) What are the state variables describing the natural state of the system?104 ‘Modeling the Dynamic and Static Behavior of Chemical Processes Part! Hose my Cee a te mde in atrttve ‘igure PII.8. A To ery ee aoe fa SS ai a Fg "constant Tae Fea ey Figure PILS 10 Assuming plug flow conditions for a jacketed tubular reactor (Figure PILS), develop its state model. simple exothermic reaction AB with first-order kinetics takes place. Due to the very large heat of reaction, molten salt, which flows countercurrently tothe reaction min- ture around the tube of the reactor, is used as coolant to keep th reaction temperature at acceptable levels. Assume constant temperature for the molten salt along the length ofthe reactor. The reaction takes place in the gaseous phase, The flow rates are volumetric and compositions are in moles per volume. The internal diameter of the ‘tube is d (in.) and its length /. . (@) Is the system a lumped parameter (described by ordinary differen tial equations) or a distributed parameter system? Develop the del. (by Does the reactor possess deal time between inputs and outputs? Explain. Te Fan éay °0y Ty Fis €ay "= ten salt Figure PLS Parl Modeling the Dynamic and Static Behavior of Chemical Processes 105 1 Figure PII.10 shows a simplified representation of a drum boiler. Feed water enters the boiler with a flow rate F (mass/hr) and a temperature T and itis heated by an amount of heat Q (Btu/hr) which is supplied by burned fuel. The generated steam flows out from the top of the boiler, with a flow rate F: (mass/hr) and a pressure p (psig). A simple feedback control system has been installed to keep the level of the water in the drum boiler constant by manipulating the flow rate of the feedwa- ter stream, (a) What are the state variables describing the system? (b) What balances are appropriate for the drum boiler system? (©) Develop the state model of the system. For the feedback control system, use a relationship of the form F\ = f(h — haeies) Produced steam Fup | ZB Fu eed water Combistion oF tel Figure PI.10 IL12 Consider a pipe of length Z (in.) with an internal diameter D (in.). 143 Water flows through the pipe with a volumetric flow rate F. Let p, be the pressure at the entrance of the pipe and p, the pressure at the exit. (a) Identify the appropriate state variables to describe the system. (b) What are the relevant balances for the system? (©) Develop the state model for this laminar flow system. Consider turbulence in the flow system of the Problem 11.12. At time 1 = 0a stream containing pure component A is mixed with the entering, water, causing @ concentration c, (moles/volume) as the water enters the pipe. Assuming that A does not diffuse along the tength of the pipe: (@) Develop a state model that describes how the concentration of A changes with time and along the length of the pipe. (b) Show that the system possesses dead time between input and output, ‘and compute the value of the dead time. 11.14 A liquid stream is a mixture of two components A and B and has a volumetric (volume/hr) flow rate F, temperature 7), and pressure py. Let cx and cp be the mole fractions of A and B in the liquid stream. Itis ‘assumed that the pressure p, is larger than the bubble point pressure of the mixture A and B, so that there is no vapor present. The liquid stream passes through an isenthalpic expansion valve106 Modeling the Dynamic and Static Behavior of Chemical Processes Part and is “flashed” into a flash drum (Figure 4.6). The pressure p in the drum is assumed to be lower than the bubble point pressure of the liquid mixture at 7j, As a result, two phases at equilibrium with each other appear in the flash drum: a vapor phase with a composition ya land y» (molar fractions) which is drawn with a volumetric flow rate V, and a liquid phase with 2 composition x, and xe (molar fractions) ‘drawn with a volumetric flow rate L. Let 7 be the temperature of the two phases at equilibrium in the flash drum. (a) What are the fundamental dependent quantities whose values describe the natural state of the flash drum? (b) What are the boundaries of the system(s) around which you will perform the various balances? (c) What are the relevant balances? (@) Besides the balance equations, what additional relationships do you need to complete the state model for the flash drum? (@) Identify the state variables and the input variables (manipulations, disturbances) of the system. (®) Develop the complete state model of the system. Chapter 5 ILAS Starting with the state models for the two systems of Problem 1.1 (Figure PIL.1), develop the corresponding input-output models. Also: (@) Find the degrees of freedom for each system. (b) Specify how many control objectives you can have for each system. 11.16 For each of the following systems, find: (a) The number of degrees of freedom. (b) The number of independent control objectives you can specify. (©) The input-output model. ‘System A: The three-tank system of Figure PI.2 (Problem II.2). ‘System B: The two stirred tank heaters system of Figure PIL.4 (Problem 11.4), System C: The two-tank mixing process of Figure PILLS (Prob- Jem I.5). IL.17 For the CSTR system of Figure PII.7 (Problem 11.8), find: (a) The number of degrees of freedom. (b) The number of independent control objectives that you can specify. (©) For the number of control abjectives you have specified, do you have an equal number of manipulated variables so that you can achieve your control objectives? (@) Identify what you would use for control objectives. (©) Show how you would form the corresponding input-output model, ‘but do not develop its analytical form. Partll_ Modeling the Dynamic and Static Behawor of Chemical Processes 107 TLAB Do the same work a8 in Problem TL.17 forthe two CSTR 5 Figure PII.8 (Problem I1.9). S system of 11.19 For the flash drum system of Problem I14 (see also Fi mI. igure 4.6), find: (2) The sumber of depres of reedom. e number of independent control objectives you can specify. (©) Identify a set of control objectives that are meaningful irom a practical operation point of view. (@) Show how you would develop an input-output model for the flash drum, but do not derive its complete analytical form, 11.20 A simple chemical reaction, A — B, with first-order kinetics, takes place in a CSTR. The effluent of the reactor enters an ideal binary distillation column where the unreacted A is taken as the overhead product with a composition y4 (molar fraction) and is recycled back to the reactor after it has been mixed with fresh feed (Figure PUI.1!). ‘Assume that the mixer and the CSTR are both isothermal. (a) Identify the fundamental dependent variables that describe the nat- ural state of the plant. (b) What are the relevant balances, and what are the boundaries of the systems around which you will perform the balances? (c) Formulate all the relevant balance equations. (4) Identify the state variables of the plant. ia Fd) wiser Fresh eed «ure AY mo le i Distiation Figure PILI 11.21 Consider the small plant described in Problem 11.20 (Figure PII!) (2) Determine the number of degrees of freedom for the plant. (b) If (aumber of degrees of freedom) > 0, how would you specify the additional equations needed to render an exactly specified system (ie., number of degrees of freedom = 0)? (©) How many disturbance specifications do you have, and how many control objectives can you identify?108 “Modeling the Dynamic and Static Behavior of Chemical Processes Part (@) Specify the control objectives that have a practical meaning for the plant. (©) Develop a simple input-output model for the plant without per- forming the required analytic computations. 11.22 Develop the state model for an ideal binary batch distillation column ith N ideal plates (Figure PII.12). At ¢ =0, the composition of the initial mixture is cx and cy (molar fractions), and its total mass is M (moles). (a) List all assumptions that you make for modeling the system. (b) Identify the relevant balance equations. (©) In addition to the balance equations, what other relationships do you need to complete the state model? (@ Show how you can find the time when the composition of the ‘overhead product has reached a composition y (molar fraction). (e) How many degrees of freedom does the system possess? (0) How many control objectives can you specify? Additional questions: (g) Does the vapor flow rate remain constant if the heat input Q remains constant with time? (h) What about Fx; does it remain constant? Under what conditions ‘would you change i Overbead product t— |i Fy : Initiay Steam = ais = Seam no Ld ompostion cg, (molt ration) Figure PILI? 11.23 Develop the state model for a gas absorption column (Figure PII.13) with V ideal transfer units. The flow rate (mol/hr) of the entering gas stream is F, with a composition (molar fraction) ya, while the composi tion of the effluent gas stream should be ys (molar fraction). The flow rate of the liquid absorbing stream is F, (mol/hr) and at its entrance it is free of component A. Partll_ Modeling the Dynamic and Static Behavior of Chemical Processes 109 guid steam Liquid + ‘Gaseous stream Figure PILI3 11.24 Develop the state model for a multicomponent (C components) nonideal distillation column with N trays. Use the general nomencla- ture developed in Example 4.13 for the ideal binary distillation. (@) State your assumptions. (b) Include the dynamics of the condenser and reboiler considering them as a perfectly stirred tank cooler and heater, respectively. (©) What relationships do you need in addition to the balance equa- tions? (@) How many degrees of freedom does the system possess? (e) How many control objectives can you specify?