Processes • Process Model -- A mathematical abstraction of a real process. – The equation or set of equations that comprise the model are at best approximation to the true process. • Rationale for Mathematical Modeling – To improve understanding of the process – To train plant operating personnel – To design controller strategy for a new process – To select controller setting – To design the control law – To optimize process operating conditions • Classification of models - depending on how they are derived. 1. Theoretical Model (Transparent); developed using the principle of chemistry and physics.
2. Empirical Model (Black box Model); obtained from a
mathematical (statistical) analysis of process operating data. Process Identification!
3. Semi-empirical Model (Gray box Model); a compromise between
1 and 2 with one or more parameters to be evaluated from plant data. 2.1.1 Dynamic and Steady-State Models
• The stirred-tank heater
– the dynamic energy balance • dT Q VC = wC (Ti − T ) + Q (2.1) dt Equation (2.1) is a simple dynamic model for this process. At steady-state, dT dt can be set zero and the following equation is obtained.
0 = wC (Ti − T ) + Q (2.2)
Equation (2.2) represent the static model for this process.
2.1.2 General Modeling Principle • Fundamental law for Mass conservation. Rate of mass Rate of Rate of = − accumulation mass in mass out Rate of mass Rate of mass + − (2.3) generation disappearance • Fundamental law for Energy conservation. Rate of energy Rate of energyin = accumulation by flow or convection
Rate of energy out Net rate of heat additionto
− + (2.4) by flow or convection system from surrounings 2.1.3 Models of Several Processes • Liquid Storage Systems Assumption ; constant density – Mass Balance; [Accumulation]=[In]-[Out] d ( Ah) dh = qi − qo (2.5) A = (qi − qo ) (2.6) dt dt dh A = qi − qo = qi − CV h (2.7) dt
Figure 2.1. A liquid-level storage system.
•Development of Dynamic Models •Illustrative Example: A Blending Process
An unsteady-state mass balance for the blending system:
rate of accumulation rate of rate of
= − (2-1) of mass in the tank mass in mass out d (Vρ ) or = w1 + w2 − w (2-2) dt where w1, w2, and w are mass flow rates.
• The unsteady-state component balance is:
d (Vρx ) = w1x1 + w2 x2 − wx (2-3) dt
The corresponding steady-state model was derived in Ch. 1 (cf.
Eqs. 1-1 and 1-2). 0 = w1 + w2 − w (2-4) 0 = w1x1 + w2 x2 − wx (2-5) The Blending Process Revisited
For constant , Eqs. 2-2 and 2-3 become:
dV = w1 + w2 − w (2-12) dt d (Vx ) = w1x1 + w2 x2 − wx (2-13) dt Equation 2-13 can be simplified by expanding the accumulation term using the “chain rule” for differentiation of a product: d (Vx ) dx dV = V + x (2-14) dt dt dt Substitution of (2-14) into (2-13) gives: dx dV V + x = w1x1 + w2 x2 − wx (2-15) dt dt Substitution of the mass balance in (2-12) for dV/dt in (2-15) gives:
After canceling common terms and rearranging (2-12) and (2-16),
a more convenient model form is obtained: dV 1 = ( w1 + w2 − w ) (2-17) dt dx w1 w2 = ( x1 − x ) + ( x2 − x ) (2-18) dt V V Example Example. A stirred-tank blending process with constant liquid holdup of 2 m3 is used to blend two streams whose densities are both approximately 900 kg/m3. The density does not change during the mixing.
a) Assume that the process has been operating for a long period of time with flow rates of w1=500 kg/min and w2=200 kg/min, and the feed com position (mass fraction) of x1 = 0.4 and x2 = 0.75. What is the steady state value of x and w? • Constant liquid holdup means:
• To find the steady state outlet composition, x, set the time
derivative equals to zero:
Solve for x to have:
(b) Suppose w1 changes suddenly from 500 to 400 kg/min and remains at this new value, how is the response of the composition x(t), plot it ?
Component MB (eq. 2.18):
The component MB equation can be rearranged as:
(c) Repeat part (b) for the case where w2 (instead of w1) changes suddenly from 200 to 100 kg/min and remains at this new value, how is the response of the composition x(t), plot it? (d) Repeat part (c) for the case where x1 changes suddenly from 0.4 to 0.6 and remains at this new value, how is the response of the composition x(t), plot it ? Plot x versus t for different cases to see the dynamic response: (e) For part (b) through plot the normalized response xN(t) defined as: Plot xN(t) versus t for different cases to see the dynamic response: Classification and Definition
Processes and their dynamics can be classified in several ways:
1. Number of independent variables. a- Lumped if time is the only independent variable, described by ordinary differential equations. b- Distributed if time and spatial independent variables are required; described by partial differential equations. 2. Linearity a- Linear if all functions in the equations are linear functions. Basically a linear differential equation is one that contains variables only to the first power in any one term of equation. b- Nonlinear if not linear. If square roots, squares, exponentials, products of variables, etc., appear in the equation, it is nonlinear. dx Linear Example: 1 dt + a0 x = f (t ) a
Where a0 and a1 are constants or functions of time
only, not of dependent variables or their derivations. Nonlinear Examples: dx a1 + a0 x 0.5 = f (t ) dt dx a1 + a0 x 2 = f (t ) dt dx a1 + a0 e x = f (t ) dt dx a1 + a0 x1 (t ) x2 (t ) = f (t ) dt 4- Order. If a system is described by one ordinary differential eq uation with derivatives of order N, the system is calle d N-th order. dNx d N −1 x dx a N N + a N −1 N −1 + ... + a1 + a0 x = f (t ) dt dt dt Where ai are constants and f(t) is the forcing function or disturbance.
Two very important special cases are for N=1 and N=2 dx First-order: a1 + a0 x = f (t ) dt d 2x dx Second-order: a 2 2 + a1 + a0 x = f (t ) dt dt Linearization
• Linearize around some condition (usually steady state) and
then find the analytical solution.
• Linearization of non-linear term is performed using Taylor S
eries expansion around equilibrium (steady state):
• Where overbar denotes the steady-state
Table 2.1. A Systematic Approach for Developing Dynamic Models
1. State the modeling objectives and the end use of the model. They determine the required levels of model detail and model accuracy. 2. Draw a schematic diagram of the process and label all process variables. 3. List all of the assumptions that are involved in developing the model. Try for parsimony; the model should be no more complicated than necessary to meet the modeling objectives. 4. Determine whether spatial variations of process variables are important. If so, a partial differential equation model will be required. 5. Write appropriate conservation equations (mass, component, energy, and so forth). 6. Introduce equilibrium relations and other algebraic equations (from thermodynamics, transport phenomena, chemical kinetics, equipment geometry, etc.). Table 2.1. (continued)
6. Perform a degrees of freedom analysis (discussed later) to ensure that
the model equations can be solved. 7. Simplify the model. It is often possible to arrange the equations so that the dependent variables (outputs) appear on the left side and the independent variables (inputs) appear on the right side. This model form is convenient for computer simulation and subsequent analysis. 8. Classify inputs as disturbance variables or as manipulated variables. Table 2.2. Degrees of Freedom Analysis
1. List all quantities in the model that are known constants (or parameters that can be specified) on the basis of equipment dimensions, known physical properties, etc. 2. Determine the number of equations NE and the number of process variables, NV. Note that time t is not considered to be a process variable because it is neither a process input nor a process output. 3. Calculate the number of degrees of freedom, NF = NV - NE. 4. Identify the NE output variables that will be obtained by solving the process model. 5. Identify the NF input variables that must be specified as either disturbance variables or manipulated variables, in order to utilize the NF degrees of freedom. Example : Liquid storage tank
Assumptions : Adiabatic and isothermal proces Incompressible (fluid Liquid) Friction in the outlet pipe segments is negligible (short segment Steady-state Mechanical Energy Balance (MEB) is applicable for this unsteady flow . The valve loss coefficient , Cv, does not vary with flow rate . Example : CSTR Energy Balance CSTR
d (VCP (T − TR )} = qCP (Ti − TR ) − qCP (T − TR ) dt Accumulation In − Out + ( − H )VkC A + UAC (TC − T ) (2.9) Net rate of heat addition Responses of simple linear systems (Solution of different equation) • First Order Linear ODE dC A + F + k C = F C A A0 Consider the general first order ODE dt V V dx + p(t ) x = Q(t ) Eq. 1 dt
Multiply both sides of the above equation by the integrating factor,
the general solution is pdt x=e − pdt Q(t ) e dt + c1 Eq. 2
Where C1 is a constant of integration and can be evaluated by using
the boundary or initial conditions. Example 1
• Consider an isothermal constant holdup, constant -
throughput CSTR with a first order irreversible reaction, which described by a component continuity equation that is a first-order linear ODE: dC A F F + + k C A = C A0 Eq. 3 dt V V
Find the time dependent response of concentration
CA to the step disturbance in feed concentration. Solution
Initial condition C A0 (0) = 0
At time equal to zero a step change in feed concentration is made from
zero to a constant value C A0 Forcing function C A0 (t ) = C A0 Comparing equations 1 and 3 F F x = CA; P= + k; Q = C A0 V V pdt x=e − pdt The solution is Q (t ) e dt + c1 F F Therefore: pdt V + k dt + k t e =e =e V Solution
F F C A0 1 V + k t F F + k t Q(t ) e dt = C A0 e e pdt V dt = + c1 V V F + k V The solution to eq. 3 is according to eq.2 F F − + k t F C A0 1 V + k t F C A0 F − + k t C A (t ) = e V e + c1 = + c1e V V F + k F + kV V V If = residence time of the vessel F 1 C A0 − + k t C A (t ) = + c1e 1 + k Solution The initial condition is now used to find the value of C1 C A0 C A (0) = 0 = + c1 1 1 + k − C A0 c1 = 1 + k Therefore the time-dependent response to the step disturbance in feed concentration is:
C A0 C A0 1 − + k t C A0 − + k t 1 C A (t ) = − e = 1 − e 1 + k 1 + k 1 + k Eq. 4 Solution The response is sketched in the following figure and is the classical first order exponential rise to the new steady-state.
Figure: Step response of a first-order system
C A0 − + k t 1 C A (t ) = 1 − e 1 + k Does the solution consistent with the initial conditions? At t=0, equation 4 becomes:
C A (t = 0) = C A0 1 − 1 = 0 1 + k So the initial condition is satisfied. Does the solution make sense from a steady-state point of view? The new steady-state value of CA can be found from the solution (eq. 4), letting time (t) go to infinity, or from the original ODE (eq. 1), setting the time derivative equal to zero. C A0 dC A C A (t → ) = C A = or = 0 in the original ODE 1 + k dt C A0 − + k t 1 C A (t ) = 1 − e 1 + k The ratio of the change in the steady-state value of the output divided by the magnitude of the step change made in the input is called the steady- state gain of the process Kp. CA 1 Kp = = C A0 1 + k These steady-state gains will be extremely important in our dynamic studies and in controller design. If k is zero (no reaction) the final steady-state value of CA will be e qual to the feed concentration CA0 as it should be. Good Chemical Engineering Sense C A0 − + k t 1 C A (t ) = 1 − e 1 + k Does the solution make sense dynamically? The rate of rise will be determined by the magnitude of the 1 +k term The bigger this term, the faster the exponential term will decay to zero as time increases. The smaller this term, the slower the decay will be. Therefore, The dynamics are set by 1 +k 1 The reciprocal of this term is called the process time constant 1 +k 1 p = The Time Constant 1 k+ The bigger the time constant, the slower the dynamic response will be. Small k or big τ will give a large process time constant If there is no reaction, the time constant is just equal to the residence time τ
