Exercises 33

EXERCISES
2.1 A perfectly stirred, constant-volume tank has two input Notes:
streams, both consisting of the same liquid. The temperature
ρ and Cρ are constants.
and flow rate of each of the streams can vary with time.
U, the overall heat transfer coefficient, is constant.
As is the surface area for heat losses to ambient.
Ti > Ta (inlet temperature is higher than ambient tempera-
T1 T2
Stream 1 Stream 2 ture).
w1 w2
T3 2.3 Two tanks are connected together in the following
w3
Stream 3 unusual way in Fig. E2.3.

w1

h1 h2
w2
Figure E2.1 w3

Figure E2.3
(a) Derive a dynamic model that will describe transient oper-
ation. Make a degrees of freedom analysis assuming that both
Streams 1 and 2 come from upstream units (i.e., their flow rates (a) Develop a model for this system that can be used to find
and temperatures are known functions of time). h1 , h2 , w2 , and w3 as functions of time for any given variations
(b) Simplify your model, if possible, to one or more differ- in inputs.
ential equations by eliminating any algebraic equations. Also, (b) Perform a degrees of freedom analysis. Identify all input
simplify any derivatives of products of variables. and output variables.

Notes: Notes:

wi denotes mass flow rate for stream i. The density of the incoming liquid, ρ, is constant.
Liquid properties are constant (not functions of temperature). The cross-sectional areas of the two tanks are A1 and A2 .
w2 is positive for flow from Tank 1 to Tank 2.
2.2 A completely enclosed stirred-tank heating process is The two valves are linear with resistances R2 and R3 .
used to heat an incoming stream whose flow rate varies.
2.4 Consider a liquid flow system consisting of a sealed tank
with noncondensible gas above the liquid as shown in Fig. E2.4.
Derive an unsteady-state model relating the liquid level h to
the input flow rate qi . Is operation of this system independent
of the ambient pressure Pa ? What about for a system open to
Ti the atmosphere?
w You may make the following assumptions:
Q
Ta
(i) The gas obeys the ideal gas law. A constant amount of
mg /M moles of gas is present in the tank.
Heating
coil (ii) The operation is isothermal.
(iii) A square root relation holds for flow through the valve.
T

qi Pg Pa
Figure E2.2

The heating rate from this coil and the volume are both H
constant. h
(a) Develop a mathematical model (differential and algebraic
equations) that describes the exit temperature if heat losses to Cv
the ambient occur and if the ambient temperature (Ta ) and the q
incoming stream’s temperature (Ti ) both can vary. Cross-sectional
area = A
(b) Discuss qualitatively what you expect to happen as Ti and
w increase (or decrease). Justify by reference to your model. Figure E2.4
34 Chapter 2 Theoretical Models of Chemical Processes

Ra P1 Rb P2 Rc Process
Air Compressor furnace
supply Pd
V1 V2 Pf
wa wb wc Figure E2.5

2.5 Two surge tanks are used to dampen pressure fluctuations (c) In order to estimate the heat transfer coefficients, the reac-
caused by erratic operations of a large air compressor. (See tor will be tested with Ti much hotter than the exit tempera-
Fig. E2.5.) ture. Explain how your model would have to be modified to
(a) If the discharge pressure of the compressor is Pd (t) and the account for the presence of the exothermic reaction. (For pur-
operating pressure of the furnace is Pf (constant), develop a poses of this answer, assume the reaction is A → B and be as
dynamic model for the pressures in the two surge tanks as well specific as possible.)
as for the air mass flows at points a, b, and c. You may assume Notes:
that the valve resistances are constant, that the valve flow
characteristics are linear, for example, wb = (P1 − P2 )/Rb , that Ut , At : Overall heat transfer coefficient and surface
the surge processes operate isothermally, and that the ideal area between compartments.
gas law holds. Uc , Ac : Overall heat transfer coefficient and surface
(b) How would you modify your model if the surge tanks oper- area of cooling tube.
ated adiabatically? What if the ideal gas law were not a good V1 : Volume of Compartment 1.
approximation? V2 : Volume of Compartment 2.
2.6 A closed stirred-tank reactor with two compartments
is shown in Fig. E2.6. The basic idea is to feed the reactants 2.7 Recall the stirred-tank heating process with variable
continuously into the first compartment, where they will be holdup as described in Section 2.4.2. Recalculate the degrees
preheated by energy liberated in the exothermic reaction, of freedom for this example under the following separate
which is anticipated to occur primarily in the second com- circumstances (be sure to explain clearly your outputs,
partment. The wall separating the two compartments is manipulated inputs, and disturbance inputs):
quite thin, thus allowing heat transfer; the outside of the (a) The flow rate exiting the tank is through an orifice, and the
reactor is well insulated; and a cooling coil is built into overall rate is dictated by hydrostatic forces.
the second compartment to remove excess energy liberated in
(b) Two feedback controllers are added to the system, one
the reaction. Both tanks are completely full at all times.
adjusting the heating rate (Q) as a function of the temperature
Tests are to be conducted initially with a single-component
of the tank, and the other adjusting the exit flow rate (w) as a
feed (i.e., no reaction) to evaluate the reactor’s thermal
function of the volume of the tank. (Hint: the exact forms of
characteristics.
the controllers are not required to solve this part.)
(a) Develop a dynamic model for this process under the con-
ditions of no reaction. Assume that q0 , Ti , and Tc all may vary. 2.8 A jacketed vessel is used to cool a process stream as
(b) Make a degrees of freedom analysis for your model— shown in Fig. E2.8. The following information is
identifying all parameters, outputs, and inputs that must be available:
known functions of time in order to obtain a solution. (i) The volume of liquid in the tank V and the volume of
coolant in the jacket VJ remain constant. Volumetric flow rate
qF is constant, but qJ varies with time.
(ii) Heat losses from the jacketed vessel are negligible.
(iii) Both the tank contents and the jacket contents are well
mixed and have significant thermal capacitances.
(iv) The thermal capacitances of the tank wall and the jacket
wall are negligible.
Cooling (v) The overall heat transfer coefficient for transfer between
medium Tc the tank liquid and the coolant varies with coolant flow rate:

q0 q2 U = Kq0.8
J
Ti T2
V1 T1 V2 T2 where
q1
U [=] Btu/h ft2 ∘ F
qJ [=] ft3 /h
Figure E2.6 K = constant
 Exercises 35

Derive a dynamic model for this system. (State any additional 2.10 Example 2.1 plots responses for changes in input flows
assumptions that you make.) for the stirred tank blending system. Repeat part (b) and
plot it. Next, relax the assumption that V is constant, and
plot the response of x(t) and V(t) for the change in w1 for t = 0
to 15 minutes. Assume that w2 and w remain constant.
TF
qF TJ 2.11 A process tank has two input streams—Stream 1 at mass
flow rate w1 and Stream 2 at mass flow rate w2 . The tank’s
qJ effluent stream, at flow rate w, discharges through a fixed
valve to atmospheric pressure. Pressure drop across the valve
is proportional to the flow rate squared. The cross-sectional
Ti area of the tank, A, is 5 m2 , and the mass density of all streams
V
qJ is 940 kg/m3 .
(a) Draw a schematic diagram of the process and write an
T appropriate dynamic model for the tank level. What is the
q corresponding steady-state model?
(b) At initial steady-state conditions, with w1 = 2.0 kg/s and
Figure E2.8 w2 = 1.2 kg/s, the tank level is 2.25 m. What is the value of the
valve constant (give units)?
k1 k2 (c) A process control engineer decides to use a feed forward
2.9 Irreversible consecutive reactions A → B → C occur
controller to hold the level approximately constant at the
in a jacketed, stirred-tank reactor as shown in Fig. E2.9.
set-point value (hsp = 2.25 m) by measuring w1 and manipu-
Derive a dynamic model based on the following assumptions:
lating w2 . What is the mathematical relation that will be used
(i) The contents of the tank and cooling jacket are well in the controller? If the w1 measurement is not very accurate
mixed. The volumes of material in the jacket and in the tank and always supplies a value that is 1.1 times the actual flow
do not vary with time. rate, what can you conclude about the resulting level control?
(ii) The reaction rates are given by (Hint: Consider the process initially at the desired steady-state
level and with the feedforward controller turned on. Because
r1 = k1 e−E1 ∕RT cA [=] mol A∕h L the controller output is slightly in error, w2 ≠ 1.2, so the
r2 = k2 e−E2 ∕RT cB [=] mol B∕h L process will come to a new steady state. What is it?) What
conclusions can you draw concerning the need for accuracy in
(iii) The thermal capacitances of the tank contents and the
a steady-state model? For the accuracy of the measurement
jacket contents are significant relative to the thermal capaci-
device? For the accuracy of the control valve? Consider all of
tances of the jacket and tank walls, which can be neglected.
these with respect to their use in a feedforward control system.
(iv) Constant physical properties and heat transfer coeffi-
cients can be assumed. 2.12 The liquid storage tank shown in Fig. E2.12 has two inlet
streams with mass flow rates w1 and w2 and an exit stream with
Note: flow rate w3 . The cylindrical tank is 2.5 m tall and 2 m in diam-
All flow rates are volumetric flow rates in L/h. The concentra- eter. The liquid has a density of 800 kg/m3 . Normal operating
tions have units of mol/L. The heats of reaction are ΔH1 and procedure is to fill the tank until the liquid level reaches a nom-
ΔH2 . inal value of 1.75 m using constant flow rates: w1 = 120 kg/min,
w2 = 100 kg/min, and w3 = 200 kg/min. At that point, inlet

Feed w1 w2
cAi, cBi
qi, Ti

Coolant out
V
qc, Tc

Coolant in q4
qci, Tci
Jacketed 1m
reactor h
Product
cA, cB, cC, T, q w3

Figure E2.9 Figure E2.12

flow rate w1 is adjusted so that the level remains constant. washout and the value that provides the maximum produc-
However, on this particular day, corrosion of the tank has tion rate. The parameter values are: μm = 0.20 h−1 ; KS = 1.0 g/l,
opened up a hole in the wall at a height of 1 m, producing and YX/S = 0.5 g/g. The steady-state condition is D = 0.1 h−1 ,
a leak whose volumetric flow rate q4 (m3 /min) can be X = 2.25 g∕L, S = 1.0 g∕L, and Sf = 10 g∕L.
approximated by
√ 2.15 In medical applications the chief objectives for drug
q4 = 0.025 h − 1 delivery are (i) to deliver the drug to the correct location in
the patient’s body, and (ii) to obtain a specified drug con-
where h is height in meters.
centration profile in the body through a controlled release of
(a) If the tank was initially empty, how long did it take for the the drug over time. Drugs are often administered as pills. In
liquid level to reach the corrosion point? order to derive a simple dynamic model of pill dissolution,
(b) If mass flow rates w1 , w2 , and w3 are kept constant indefi- assume that the rate of dissolution rd of the pill in a patient
nitely, will the tank eventually overflow? Justify your answer. is proportional to the product of the pill surface area and the
concentration driving force:
2.13 Consider a blending tank that has the same dimensions
and nominal flow rates as the storage tank in Exercise rd = kA(cs − caq )
2.13 but that incorporates a valve on the outflow line that
is used to establish flow rate w3 . (For this exercise, there is where caq is the concentration of the dissolved drug in the
no leak in the tank as in Exercise 2.13.) In addition, the aqueous medium, cs is the saturation value, A is the surface
nominal inlet stream mass fractions of component A are area of the pill, and k is the mass transfer coefficient. Because
x1 = x2 = 0.5. cs ≫ caq , even if the pill dissolves completely, the rate of
The process has been operating for a long time with constant dissolution reduces to rd = kAcs .
flow rates and inlet concentrations. Under these conditions, (a) Derive a dynamic model that can be used to calculate pill
it has come to steady state with exit mass fraction x = 0.5 and mass M as a function of time. You can make the following sim-
level h = 1.75 m. Using the information below, answer the fol- plifying assumptions:
lowing questions: (i) The rate of dissolution of the pill is given by rd = kAcs .
(a) What is the value of w3 ? The constant, Cv ? (ii) The pill can be approximated as a cylinder with radius
(b) If x1 is suddenly changed from 0.5 to 0.6 without changing r and height h. It can be assumed that h/r ≫ 1. Thus the
the inlet flow rates (of course, x2 must change as well), what is pill surface area can be approximated as A = 2πrh.
the final value of x3 ? How long does it take to come within 1% (b) For the conditions given below, how much time is required
of this final value? for the pill radius r to be reduced by 90% from its initial value
(c) If w1 is changed from 120 kg/min to 100 kg/min without of r0 ?
changing the inlet concentrations, what will be the final value ρ = 1.2 g∕ml r0 = 0.4 cm h = 1.8 cm
of the tank level? How long will it take to come within 1% of cs = 500 g∕L k = 0.016 cm∕min
this final value?
(d) Would it have made any difference in part (c) if the con- 2.16 Bioreactions are often carried out in batch reactors. The
centrations had changed at the same time the flow rate was fed-batch bioreactor model in Section 2.4.9 is also appli-
changed? cable to batch reactors if the feed flow rate F is set equal
Useful information: The tank is perfectly stirred. to zero. Using the available information shown below,
determine how much time is required to achieve a 90% conver-
√
w3 = Cv h sion of the substrate. Assume that the volume V of the reactor
contents is constant.
2.14 Suppose that the fed-batch bioreactor in Fig. 2.11 is con- Available information:
verted to a continuous, stirred-tank bioreactor (also called a (i) Initial conditions:
chemostat) by adding an exit stream. Assume that the inlet and
exit streams have the same mass flow rate F and thus the vol- X(0) = 0.05 g∕L, S(0) = 10 g∕L, P(0) = 0 g∕L.
ume of liquid V in the chemostat is constant.
(ii) Parameter values:
(a) Derive a dynamic model for this chemostat by modifying
the fed-batch reactor model in Section 2.4.9. V = 1 L, μm = 0.20 h−1 , KS = 1.0 g∕L,
(b) Derive the steady-state relationship between growth rate YX∕S = 0.5 g∕g, YP∕X = 0.2 g∕g.
μ in Eq. 2-80 and dilution rate D where by definition, D = F/V.
Suggest a simple control strategy for controlling the growth 2.17 Sketch the level response for a bathtub with cross-
rate based on this result. sectional area of 8 ft2 as a function of time for the following
(c) An undesirable situation called washout occurs when all sequence of events; assume an initial level of 0.5 ft with the
of the cells are washed out of the bioreactor and thus cell drain open. The inflow and outflow are initially equal to
mass X becomes zero. Determine the values of D that result 2 ft3 /min.
in washout. (Hint: Washout occurs if dX/dt is negative for an (a) The drain is suddenly closed, and the inflow remains con-
extended period of time, until X = 0.) stant for 3 min (0 ≤ t ≤ 3).
(d) For the numerical values given below, plot the steady-state (b) The drain is opened for 15 min; assume a time constant in a
cell production rate DX as a function of dilution rate D. Dis- linear transfer function of 3 min, so a steady state is essentially
cuss the relationship between the values of D that result in reached (3 ≤ t ≤ 18).
 Exercises 37

(c) The inflow rate is doubled for 6 min (18 ≤ t ≤ 24). of 1.0 ft with the drain open, and that level and outflow rate
(d) The inflow rate is returned to its original value for 16 min are linearly related. The steady-state inflow and outflow are
(24 ≤ t ≤ 40). initially equal to 2 ft3 /min. The graph should show numerical
values of level vs. time.
2.18 Perform a degrees of freedom analysis for the model in (a) The drain is suddenly closed, and the inflow remains con-
Eqs. 2-64 through 2-68. Identify parameters, output variables, stant for 3 min (0 ≤ t ≤ 3).
and inputs (manipulated and disturbance variables).
(b) The drain is opened for 15 min, keeping the inflow
2.19 Consider the PCM distillation column module of at 2 ft3 /min, where a steady state is essentially reached
Appendix E, in which a 50%–50% mixture of methanol (3 ≤ t ≤ 18).
PCM
(MeOH) and ethanol is separated. (c) The inflow rate is doubled to 4 ft3 /min for 15 min
Do the following sequence of simulations: (18 ≤ t ≤ 33).
(a) Change the Vapor Flow Rate from the initialized value to (d) The inflow rate is returned to its original value of 2 ft3 /min
a new value of 0.045 m3 /s, and start the simulation. Generate for 17 min (33 ≤ t ≤ 50).
plots for the Overhead MeOH Composition and the Bottom 2.22 Consider the unusual piping diagram for the four tanks
MeOH Composition. Record the starting values and final in Fig. E2.22 in which both the flow rates F1 and F2 are split
(steady-state) values for these variables. between two streams entering the upper and lower tanks
(b) Once the system has reached a new steady state (and you (denoted by the fractions in the diagram). For the exit lines
have collected the information requested in part (a)), increase leaving the bottom of each of the four tanks through an
the Feed MeOH Composition to 0.55. Generate plots for the orifice, you may assume that flow through that orifice obeys
Overhead MeOH Composition and the Bottom MeOH Com- the square root dependence on the height of liquid in the
position. Record the starting values and final (steady-state) val- tanks, as described in Section 2.4.5.
ues for these variables. (a) Derive the mass balances for each of the four tanks, and
(c) Comment on the relative differences between these two express them as simple equations (four total) with one deriva-
input changes (strength of effect, timing of response). tive term on the left hand side.
2.20 Consider the PCM Furnace module of Appendix E, (b) For the case of γ1 = 0.5 and γ2 = 0.5 (i.e., equal splitting of
which is used to preheat a high-molecular-weight hydro- each stream), what is the resulting form of the equations? Can
PCM
carbon feed (C16 –C26 ) to a cracking unit at a petroleum the levels be solved independently? Can the flow rates be used
refinery. independently to adjust the heights in the tanks?
(c) For the extreme case of γ1 = 0 and γ2 = 0, what is the result-
Do the following sequence of simulations: ing form of the equations? Does this make sense in terms of the
(a) Change the Fuel Gas Purity from the initialized value of process schematic?
1.0 to a new value of 0.95, and start the simulation. Generate
plots for the Oxygen Exit Concentration and the Hydrocar-
bon Outlet Temperature. Record the starting values and final Fraction 1 − γ2
(steady-state) values for these variables.
Fraction 1 − γ1
(b) Once the system has reached a new steady state (and you
have collected the information requested in part (a)), increase
the Hydrocarbon Flow Rate by 10% over the current value.
Generate plots for the Oxygen Exit Concentration and the Tank 3 Tank 4
Flow rate F1 Flow rate F1
Hydrocarbon Outlet Temperature. Record the starting values
and final (steady-state) values for these variables.
Fraction γ1 Fraction γ2
(c) Comment on the relative differences between these two
input changes (strength of effect, timing of response). Tank 1 Tank 2
2.21 Plot the level response for a tank with constant
cross-sectional area of 4 ft2 as a function of time for
the following sequence of events; assume an initial level Figure E2.22