Scribd
Upload
102 views56 pages

Process Control Introduction To Process Dynamics: Niño B. Felices

Here are the key steps to solve this problem: 1. Write material balances for each tank. 2. Linearize the model around an operating point. 3. Express the linearized model in state space form. 4. Take the Laplace transform to obtain the transfer function. Let me know if you need help with any of these steps. Solving dynamic models is an important skill in process control.

Uploaded by

raven
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
102 views56 pages

Process Control Introduction To Process Dynamics: Niño B. Felices

Here are the key steps to solve this problem: 1. Write material balances for each tank. 2. Linearize the model around an operating point. 3. Express the linearized model in state space form. 4. Take the Laplace transform to obtain the transfer function. Let me know if you need help with any of these steps. Solving dynamic models is an important skill in process control.

Uploaded by

raven
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 56

PROCESS CONTROL

LECTURE 1:
INTRODUCTION TO
PROCESS DYNAMICS

Niño B. Felices
REVIEW
GENERAL BALANCE EQUATION

In – Out + Generation – Consumption = Accumulation

Applicable for mass and energy in accordance to the


conservation principles
REVIEW
STEADY STATE PROCESS – no build-up of conserved
quantities in the system
Accumulation = 0

UNSTEADY STATE OR TRANSIENT PROCESS


Accumulation ≠ 0
PROCESS DYNAMICS
Transient operations occur during:
Start-ups and shutdowns
Correction of process disturbances
Batch/semi-batch operations
Planned product grade transitions
Feedstock and/or product campaign changes

PROCESS DYNAMICS – study of the unsteady-state or


transient process behavior
Why study process dynamics?
Even at ‘steady state’, processes never truly run at steady
state design conditions.
• Feed and environmental disturbances, heat exchanger
fouling, catalytic degradation continuously upset the
conditions of a smooth running process
Why study process dynamics?
Processes must respond to both planned and unplanned
changes, leading to process control
Why study process dynamics?
Detailed understanding of dynamics is required for process
engineers running a plant.
• Previous operational experience is not enough to answer
questions
• Plant trial and error can be expensive and hazardous.
How to study process dynamics?
DYNAMIC MODELLING AND SIMULATION – predicts
how a process and its associated control system will
respond to various disturbances or inputs as a function of
time

General form of a DYNAMIC MODEL:

ẋ = f(t, x(t), y(t), u(t))


y = g(t, x(t), y(t), u(t))
DYNAMIC MODEL
ẋ = f(t, x(t), y(t), u(t))
y = g(t, x(t), y(t), u(t))

OUTPUT VARIABLES– variables of x, y


interest that are being measured or
being controlled
INPUT VARIABLES– variables that u
are specified or manipulated before a
problem can be solved or a process
can be operated
DYNAMIC MODEL
ẋ = f(t, x(t), y(t), u(t))
y = g(t, x(t), y(t), u(t))

STATE VARIABLES – output variables x


that arise naturally in the accumulation
term of a dynamic material or energy
balance
PARAMETERS – fixed physical or A, μ, ρ, k, ΔHr ,…
chemical properties that are known (e.g.
tank area, viscosity, density, reaction
rate constant, heat of reaction,
activation energy, etc.)
ẋ = f(t, x(t), y(t), u(t))
y = g(t, x(t), y(t), u(t))
EXPANDED FORM
ẋ = f(t, x(t), y(t), u(t))
y = g(t, x(t), y(t), u(t))

EXPANDED FORM
EXAMPLE: TWO TANKS IN SERIES
PROBLEM: How do changes in
F0 affect h1 and h2?

ASSUMPTIONS
1. Perfect mixing (constant ρ)
2. Bernoulli equation applies

DEFINE VARIABLES
Input variable: F0
Output (state) variables: h1, h2
EXAMPLE: TWO TANKS IN SERIES
•MATERIAL
  BALANCE
In tank 1

In tank 2

ρ
EXAMPLE: TWO TANKS IN SERIES
APPLYING
•   BERNOULLI’S
EQUATION

where β is a lumped parameter


containing
(1) Acceleration due to gravity
(2) Density
(3) Area of nozzle opening
EXAMPLE: TWO TANKS IN SERIES
•DYNAMIC
  MODEL

ẋ = f(t, x(t), y(t), u(t))


y = g(t, x(t), y(t), u(t))
  y=0
u = F0

Parameters: β1, β2, A1, A2


CLASSIFICATION OF DYNAMIC MODELS

1. Lumped Model vs. Distributed Model


Lumped: State variables are a function of time alone (ODEs)
Distributed: … a function of time and one or more spatial variables (PDEs)
2. Linear Model vs. Non-linear Model
Linear: Exhibit superposition; all variables and derivatives have exponent 1
Non-linear: If any one term is nonlinear, the model is nonlinear
3. First Principles Model vs. Empirical (Black-box) Model
First Principles: Based on physical and chemical laws
Empirical: Simpler models whose coefficients are fitted to the actual process
4. Continuous Variables vs. Discrete Variables (vs. Hybrid)
Continuous: Variables can assume any real value within an interval
Discrete: Variables can only take distinct values (integer, binary, etc.)
What type of model is this?
DYNAMIC
•   MODEL
What type of model is this?
DYNAMIC
•   MODEL

Lumped model
Non-linear model
First-principles model
Continuous model
EXAMPLE: TWO TANKS IN SERIES
What do we need to solve the
•DYNAMIC
  MODEL model?
1. Value for all parameters
A1 = ? A2 = ? β1 = ? β2 = ?
2. Initial steady state conditions
At t = 0, h1 = ? h2 = ? F0 = ?
3. Solver
Numerical methods?
Available software?
4. A change in input variables
A sudden increase in F0
A small disturbance in F0
At what time, t?
DYNAMIC SIMULATION
Simple models can be solved by
1. Euler method in MS Excel
2. Maplesoft Maple
3. Wolfram Research Mathematica

More difficult models can be solved by


programming languages
4. Implement numerical methods in any
language: Java, C++, FORTRAN,
MATLAB.
5. Use built-in libraries from languages:
MATLAB, POLYMATH, ACSL, IMSL,
MathCad
6. Use special packages: DASSL, LSODAR,
DOPR15, RADAU IIA
DYNAMIC SIMULATION
•DYNAMIC
  MODEL Initial conditions CHANGE IN F0
h1(0) = h2(0) = 1.5 m
F0 = 7.3 m3/s
Parameters
A1 = A 2 = 8 m 2
β1 = β2 = 6 m2.5 s-1
Inputs in dynamic models
Inputs can be any or a combination of the following
EXAMPLE: TWO CONSTANT
HOLD-UP TANKS IN SERIES
PROBLEM:
How do changes in Co affect C1
and C2?

State variables: C1, C2


Input variable: C0
Invariant: F1, F2, F3, h1 and h2
ASSUMPTIONS:
1. Perfect mixing (constant ρ)
2. Constant hold-up
(constant h1 and h2; F0 = F1 = F2 )
F = F0 = F1 = F2
EXAMPLE: TWO CONSTANT
HOLD-UP TANKS IN SERIES
•Component
  mass balance:
EXAMPLE: TWO CONSTANT
HOLD-UP TANKS IN SERIES
•Dynamic
  model:
DYNAMIC SIMULATION
What equations can we write
for dynamic models?
• CONSERVATION LAWS
What equations can we write
for dynamic models?
• CONSTITUTIVE EQUATIONS
Blending tank with variable
hold-up
Blending tank with variable
hold-up
Heated mixing tank
Heated mixing tank
Isothermal CSTR
Isothermal CSTR
Exercise 1

• A perfectly-stirred variable-
volume tank has two input
streams both consisting of the
same liquid. The temperature and
flow rate of each the streams can
vary with time. Derive a dynamic
model that shows the effect of T1
and T2 to T3. Write the state
variables, input variables and
invariants.
Exercise 2
• A  first order reaction of species A takes place in a heated CSTR with
constant hold-up. The rate of consumption of A per unit volume is
given by
where k = k0e-E/RT

The heat of reaction is denoted as ΔH (energy per unit mole A


reacted). Assume that the volumetric flow rate F, heat capacity Cp and
density ρ are constant across time.

Control objective: It is desired to control the species A concentration


by manipulating heater input Q, as well as monitor tank T.
a) Write down the input variables, state variables and invariants.
b) Derive the dynamic model of the system
Exercise 3
• A completely enclosed stirred-tank
heating process is used to heat an
incoming stream whose flow rate varies.
The heating rate from the coil and the
volume of fluid inside the tank is
constant. Contact area of tank with
surrounding is As. Develop a dynamic
model that describes the exit
temperature if heat losses to ambient
temperature (Ta) and the incoming
stream’s temperature can both vary (Ti).
Identify state variables, input variables,
and invariants.
Linearization of models
Linearization of models
Linearization of models
Linearization of models
State-space models
State-space models
State-space models
Transfer functions
Transfer functions
Transfer functions
Transfer functions
State space to Transfer
function transformation
Dynamic model: two tanks in
series
Dynamic model simulation in
MATLAB
Class Exercise
Laboratory Exercise 1
• A
  CSTR with a single first-order chemical reaction is cooled by
a coolant at Tc. The dynamic model of the system is given by
the following equations.

where k is defined by the Arrhenius expression below

a) Write the dynamic models in deviation form.


b) Express the state-space model for the given system.
c) Derive the following transfer functions:
Laboratory Exercise 2
Laboratory Exercise 3
• An isothermal irreversible reaction AB takes place in the
liquid phase in a constant volume reactor. Observation
patterns indicates that a two-tank mixing system with back
mixing should approximate imperfect mixing. Each of the two
tanks are perfectly mixed.
V1 = V2 = 50 gal
F = 20 gal/min
Fr = 10 gal/min
k = 15 min-1
a) Derive the dynamic models describing the system.

b) Derive the transfer function relating C’A2 to C’A0.

You might also like