A B C D of Process Dynamics-1
Uploaded by
Saurabh ChaurasiaA B C D of Process Dynamics-1
Uploaded by
Saurabh ChaurasiaABCD
OF
PROCESS DYNAMICS & CONTROLLER DESIGN
Subject: Process Control and Instrumentation
Subject Code: IC-503
Subject Coordinator
Dr. Om Prakash Verma
Assistant Professor
Department of Instrumentation and Control Engineering
Dr. B R Ambedkar National Institute of Technology Jalandhar
IC-503 Process Control and Instrumentation [3 0 0 3]
Review of Process Control Fundamentals: Process control principles, elements of process control system,
process characteristics, control system parameters, control system evaluation, process and instrumentation
symbols and diagrams
Transducers and Signal Conditioning: Definition and classification of sensors, working principles and salient
features of thermal sensor, optical sensors, displacement and location sensors, strain sensor, level sensor, motion
sensor, pressure sensor, flow sensor., principles of analog signal conditioning and design guidelines, passive
circuits, operational amplifier circuits in instrumentation,
Controller principles:Principles, applications and examples of discontinuous controller modes, continuous three
term controller, cascade control, over-ride control, split range control, feed forward control, ratio control, adaptive,
self-adaptive control, optimal control, supervisory and direct-digital control, final control operation: signal
conversion, actuators and final control element.
Programmable Logic Controllers: Principles, relative merits over hard-wired logic, relay and programming
languages, ladder diagrams.
Distributed Control Systems (DCS): Distributed process control, DCS-configurations, Control console
equipment, Video display, Overview display, detail & graphical displays. DCS-control unit, Controller file,
Communications between components, DCS-data highways, field buses, multiplexers and remote terminal units,
DCS-flow diagrams, generic nature of DCS, Supervisory Control techniques, introduction to SCADA.
Multivariable Control System: Interactions in multiples loops, RGA method for minimizing interactions.
Text Books
Stephanopoulos G, "Chemical process control: an introduction to theory and practice," Prentice Hall
1. Bartelt T, “Process control systems and instrumentation,” Cengage Learning
Reference Books
Seborg DE, Edgar TF and Mellichamp DA, "Process dynamics and control," Wiley
1. Smith CA and Corripio AB, "Principles and practice of automatic process control," Wiley
2. Johnson CD, "Process control instrumentation technology," Prentice Hall
3. Liptak GB, “Instrument Engineers' Handbook, vol.2: Process Control and Optimization,” CRC Press
ABOUT YOUR MENTOR:
DR. OM PRAKASH VERMA
Assistant Professor
Department of Instrumentation and Control Engineering,
Dr. B R Ambedkar National Institute of Technology Jalandhar, Punjab, India
Ph.D. (IIT Roorkee), M.Tech. (Dr B R Ambedkar NIT Jalandhar), B.E. (Dr B R Ambedkar University Agra)
Email: [email protected]; [email protected];
Google Scholar: https://scholar.google.co.in/citations?user=127RG7QAAAAJ&hl=en
YouTube Channel: http://www.youtube.com/c/DrOmPrakashVerma
+917579279839, +918279939052
Why Process Dynamics/Control ?
MATHEMATICAL MODELING
4
Why Process Dynamics ?
Fig: Control system for a stirred-tank heater
5
Why Process Dynamics ?
Fig: Block diagram of a simple control system.
6
Why Process Dynamics ?
Fig: Blending System Control Method 1. Fig: Blending System Control Method 2.
7
Why Process Dynamics ?
Fig: Blending System Control Method 1. Fig: Blending System Control Method 2.
8
Why Process Dynamics ?
Fig: Blending System Control Method 1. Fig: Blending System Control Method 2.
9
Time to Think
A Chemical Mixing Scenario
The process is running along at steady state. The concentration of A in stream
1 is 1 g/L and in stream 2 is 4 g/L. At 3:00 PM. the shift changes at the plant.
The new operator on our unit misreads the flowmeters for the process and
switches the flow rates of the two streams. Stream 1 is switched to 20 L/min,
and stream 2 is switched to 10 L/min. At 3:30 PM. the shift supervisor hurries to
the control room to determine the source of problem now being experiencing
with reactor.
Use your knowledge to determine what has happened to the exit
concentration from heating vessel over first half-hour of the shift. 10
Time to Answer
Let’s play with Modeling
• Model the mixing tee and blending tank using an unsteady-state mass
balance to predict the behavior of process since, shift change and the
unfortunate error by new operator.
• A balance on component A around mixing tee before and after change will
yield information on how the feed concentration to the heating vessel
changes. The component A balance around mixing tee is
Before the change: calculating at SS concentration into heating vessel
11
Time to Answer
Let’s play with Modeling
After the change, the new feed concentration to the heating vessel is
• So, the net result of the operator error is to decrease the feed
concentration to the heating vessel from 3 to 2 g/L.
• After analyzing process for a moment, it is apparent that the exit
concentration from the heating vessel will eventually fall from 3 to 2 g/L.
• if the process is left in its current configuration for a long enough time. So,
what is the situation?
12
Time to Answer
Let’s play with Modeling
13
Time to Answer
Let’s play with Modeling
Outlet concentration from the
heating vessel as a function of
time. 14
Think More !
Chemical mixing process flow
diagram showing initial
temperatures.
• Stream 1 (at 25°C) mixes with stream 2 (at 55°C), producing stream 3, the
feed to the heating vessel.
• The heater adds energy to vessel to bring the outlet stream to 80°C. Before
we look at the effect of the disturbance caused by the operator, it is
necessary to determine S.S process conditions prior to the upset.
• An energy balance around the mixing tee will enable us to calculate s.s
15
feed
temperature to the heating vessel T3 .
Think More !
Before the change
• So, the steady-state inlet temperature to the heating vessel is 45 °C.
• Determine SS heat input required from the heater by performing a SS
energy balance around heating vessel.
16
Think More !
After the change
17
Think More !
Outlet temperature
transient due to the
disturbance.
18
Concluding What we have studied !
Need to Study The process Dynamics and Control:
• Enhanced process safety
• Satisfying environmental constraints
• Meeting ever-stricter product quality specifications
• More efficient use of raw materials and energy
• Increased profitability
• Specified Production rate must be maintained
Need to develop the Process Dynamics Model:
• Improve understanding of the process
• Train Plant Operating Engineer
• Develop a control strategy for a new process
• Optimize process operating conditions
Types of System Identification/modelling
• Theoretical Modeling
• Empirical Modeling
• Semiempirical Modeling 19
Tools for modeling the process
20
Tools for modeling the process
• Theoretical models of chemical processes emanated from conservation laws:
Conservation of Mass
Conservation of Component ‘ith’
Conservation of Energy
21
Meaning of Modeling
George E. P. Box
“All Mathematical Models are wrong, but Some
are Useful”
22
TYPES OF MODELLING EQUATIONS
AND TRANSFER FUNCTION
• Algebraic equations (AEs)
• Ordinary differential equations (ODEs)
• Partial differential equations (PDEs)
• Linear Vs. NonLinear
• Deterministic vs. Probabilistic (Stochastic)
• Static Vs. Dynamic
• Lumped Parameters Vs. Distributed Parameters
• White box models: first principles
• Black box models: data driven (Empirical Modeling)
• Gray box models: with some unknown parameters
• Identification of model parameters – necessary step
– Assume known model structure
– Collect plant data: special experiment or normal operation
– Tweak model parameters to achieve a good fit
23
DEVELOPMENT OF MATHEMATICAL MODELS: EXAMPLE
Regulator problem: The goal of a
control system for this type of problem is
to enable system to compensate for
load changes and maintain controlled
variable at set point.
Servo problem: The goal of a control
system for this type of problem is to
force system to “track” requested set
point changes.
24
Meaning of Transfer Function
A dynamic system can be described by the following time-invariant differential equation:
Taking the Laplace transform and considering zero initial conditions we have:
25
Point to remember about: Transfer
Function
• The ratio of transform of output of component to transform of input.
• The component is assumed to be at rest prior to excitation, and all initial values are
assumed to be zero
• The transfer function is defined only for a linear time-invariant system. It is not
defined for nonlinear systems.
• The roots of N(s) are called poles of system and roots of M(s) are called zeros of
system.
• By setting denominator function to zero, is referred to as characteristic equation:
26
Basic Example of Transfer Function
27
Basic Example of Mechanical Systems
Spring: A spring applies a force against compression/expansion
Mass: A moving mass has inertia and resists changes in velocity
Viscous Damper: Resists motion (pure energy loss)
28
Meaning of Transfer Function
29
Meaning of Transfer Function
30
DEVELOPMENT OF MATHEMATICAL MODELS: EXAMPLE
• FIRST-ORDER SYSTEMS: Linear Model
Eq. 1
Eq. 2 Linear flow and head relationship
• Laminar range (Reynold Number: 1800 - 2100)
Eq. 3
• Turbulent flow (Reynold Number > 4000)
At steady state: Eq. 4
Subtracting Eq. 4 from Eq. 1 Eq. 5
Defining deviation variables
Eq. 6
Solving Eq. 5 and Eq. 6 Eq. 7
Standard Transfer Function: Eq. 8
Where is AR. 31
DEVELOPMENT OF MATHEMATICAL MODELS: EXAMPLE
• FIRST-ORDER SYSTEMS: NONLINEAR MODEL
If flow out of the tank follows a square root relationship: Eq. 1
Yields nonlinear differential equation: Eq. 2
LINEARIZATION: Taylor series expansion, the function qo(h) may be expanded
around the steady-state value hs
Eq. 3
Avoiding the higher derivatives:
Eq. 4
Evaluating q’0(hs)
Eq. 5
After solving
Eq. 6
Finally, at steady state flow entering tank equals flow leaving the tank: qs=qos
Hence, Where 32
LINEARIZATION OF NONLINEAR MODELS: GENEARLIZATION
Considering the general nonlinear model for input and output are:
33
LINEARIZATION OF NONLINEAR MODELS: GENEARLIZATION
34
EXAMPLE 2: MIXING PROCESS
At steady state
𝑥: (mass of salt/volume)
EXAMPLE 3: BLENDING PROCESS
An unsteady-state mass balance for the blending system:
The unsteady-state component balance
At steady state 35
EXAMPLE 2: BLENDING PROCESS
After canceling common terms and rearranging
36
EXAMPLE 3: HEATING PROCESS
At steady state
If we assume that T i is constant (and so T i = T is )
37
EXAMPLE 4: HEAT CONDUCTION IN A SOLID
Assumption and Notation: The conductivity k,
heat capacity C, and density r of the
conducting material are constant, independent
of temperature. Initially ( t < 0), the slab is at a
uniform steady-state temperature.
Therefore in deviation variables, which will be
used henceforth, the initial temperature is
zero. The cross-sectional area of the slab is A.
38
EXAMPLE 4: HEAT CONDUCTION IN A SOLID (Distributed-Parameter Model)
39
EXAMPLE 4: HEAT CONDUCTION IN A SOLID TRANSFER FUNCTION
Consider first the integral on the left side
Consider the integral on the right side
where T ( x, 0) is the initial temperature distribution in the solid
Initial Condition: T ( x, 0) = 0
40
State Space Techniques
Then the system is equivalently described by
the equations
41
State Space Techniques
And output to the state variables is
y = Cx 42
Model Development of Coupled Tank System
43
Model Development of some classical Example
Non interacting system
44
Model Development of some classical Example
45
Model Development of some classical Example
46
Model Development of some classical Example
Interacting system
47
Model Development of some classical Example
48
Model Development of some classical Example
• From this figure, it can be seen that interaction slows up the response.
This result can be understood on physical grounds in the following
way: If the same size step change is introduced into the two systems of
Fig., the flow from tank 1 ( q 1 ) for the noninteracting case will not be
reduced by the increase in level in tank 2.
• However, for the interacting case, the flow q 1 will be reduced by the
buildup of level in tank 2. At any time t 1 following the introduction of
the step input, q 1 for the interacting case will be less than for the
noninteracting case with the result that h 2 (or q 2 ) will increase at a
slower rate.
49
Time of Asses yourself?
50
Model Development of some classical Example
Liquid-Level Process with Constant-Flow Outlet
Taking the Laplace transform of each side and solving for H/Q give
Notice that the transfer function 1/ As in Eq. is equivalent to integration. Therefore, the
solution of Eq. is
Clearly, if we increase the inlet flow to the tank, the level will increase because the
outlet flow remains constant. The excess volumetric flow rate into the tank
accumulates, and the level rises.
if a step change Q ( t ) u ( t ) were applied to the system shown in Fig., the result would
be
51
Model Development of some classical Example
52
Model Development of some classical Example
MERCURY THERMOMETER.
• Consider the thermometer to be located in a flowing stream of fluid for which the
temperature x varies with time.
• Our problem is to calculate the response or the time variation of the thermometer
reading y for a particular change in x.
(The symbols x and y have been selected to represent surrounding temperature and thermometer
reading, respectively.)
Cross-sectional view of themometer
53
Temperature profiles in thermometer.
Model Development of some classical Example
MERCURY THERMOMETER.
• Assumptions
1. All the resistance to heat transfer resides in the film surrounding the bulb (i.e., the
resistance offered by the glass and mercury is neglected).
2. All the thermal capacity is in the mercury. Furthermore, at any instant the mercury
assumes a uniform temperature throughout.
(*Making these first two assumptions is often referred to as the lumping of parameters
because all the resistance is “lumped” into one location and all the capacitance into
another. These assumptions make it possible to represent the dynamics of the system
by an ODE. If such assumptions were not made, the analysis would lead to a PDE,
and the representation would be referred to as a distributed-parameter system.)
54
Model Development of some classical Example
MERCURY THERMOMETER.
• Assumptions
3. The glass wall containing mercury does not expand or contract during the
transient response.
(In an actual thermometer, the expansion of the wall has an additional effect on the
response of the thermometer reading. The glass initially expands and the cavity
containing the mercury grows, resulting in a mercury reading that initially falls. Once the
mercury warms and expands, the reading increases. This is an example of an inverse
response.)
55
Model Development of some classical Example
MERCURY THERMOMETER.
• Modeling
It is assumed that thermometer is initially at steady state.
This means that, before time 0, there is no change in temperature with time.
At time 0, the thermometer will be subjected to some change in surrounding
temperature x ( t ).
By applying the unsteady-state energy balance
56
Model Development of some classical Example
MERCURY THERMOMETER.
• Modeling
For the steady-state condition,
After Subtracting,
57
Model Development of some classical Example
SUMMARY
To summarize the procedure for determining the transfer function for a
process:
Step 1. Write the appropriate balance equations (usually mass or energy
balances for a chemical process).
Step 2. Linearize terms if necessary.
Step 3. Place balance equations in deviation variable form.
Step 4. Laplace-transform the linear balance equations.
Step 5. Solve the resulting transformed equations for the transfer function, the
output divided by the input.
58
BASICS OF EMPIRICAL MODELING
(INTRODUCTORY)
59
EMPIRICAL MODELING (INTRODUCTORY)
• Sometimes it is difficult or impossible to develop a mathematical model that
explains a situation.
• If data exists, we can often use these data as the sole basis for an empirical
model which consists of a function that fits the data.
60
EMPIRICAL MODELING (INTRODUCTORY)
61
EMPIRICAL MODELING (INTRODUCTORY)
62
EMPIRICAL MODELING (INTRODUCTORY)
63
EMPIRICAL MODELING (INTRODUCTORY)
64
EMPIRICAL MODELING (INTRODUCTORY)
65
EMPIRICAL MODELING (INTRODUCTORY)
66
EMPIRICAL MODELING (INTRODUCTORY)
67
EMPIRICAL MODELING (INTRODUCTORY)
68
EMPIRICAL MODELING (INTRODUCTORY)
69
EMPIRICAL MODELING (INTRODUCTORY)
70
Data
• Data to
• suggest model
• estimate parameters
• test model
• Crucial for an empirical model
• Empirical model
• based on data only
• to predict, not explain
96
Transforming the Data
Suppose y = c*ex, to find c, we can
(1) plot (ex1, y1), (ex2, y2), … (exn, yn), then find c ≈ slope of the line
that is the closest to the n points.
Or
(2) since ln(y) = ln(c) + x, we can also plot (x1, ln(y1)), (x2, ln(y2)),
…, (xn, ln(yn)), then find ln(c) ≈ y-intercept of the line that is the
closest to the n points.
97
Model Fitting Criteria
98
99
Example 1
• xLst = {0, 2, 4, 5, 6, 8}
• yLst = {5.3, 7.0, 9.4, 11.1, 12.3, 14.2}
• Form a list of ordered pairs.
• Plot points.
y
14
12
10
x
2 4 6 8
100
Example 1
• Fit function for least squares fit.
• Get y = 5.0449 + 1.16122 x
• Plot fitted function.
• Plot fitted function and data.
• Use the model to make predictions within the data range
(interpolations).
• Be careful not to use the model to make predictions that is outside
the data range (extrapolations).
y
14
12
10
x
2 4 6 8
101
Linear Regression
• Find a line y = ax + b that fits the data points the best using the
least sum of square of errors criterion.
102
Using elimination method to solve for a and b, we have
Most of computer software and calculators perform
these computations for us and find us the regression
line (or curves).
103
Linear Regression
To measure how close the regression line fits the data, we
use the following.
(1)Error sum of squares which reflects the variation about
The regression line
(2) Total corrected sum of squares
(3) Regression sum of squares SSR = SST - SSE
104
Linear Regression
(4) Coefficient of determination R2 = 1 – SSE/SST = SST/SSE
which is a measure of fit of the regression line.
(5) Another indicator is the residual plot where residual
ri = yi –(a*xi + b) vs the independent variable x are plotted.
1. residuals should be randomly distributed and contained in a
small band.
2. Large residual warrants further investigation.
3. A pattern indicates that model should be refined. The
pattern usually gives a clue.
105
Example
An experiment has been performed to determine the steady-state power delivered by a gas
turbine-driven generator as a function of fuel flow rate. The following normalized data were
obtained:
Develop its linear and quadratic model and analyze the result.
106
Example
Result:
107
DEVELOPING LINEAR MODELS FROM DATA
The deflection of a cantilever beam is the distance its end moves in response to a force
applied at the end. This distance is called the deflection and it is the output variable.
The applied force is the input. The following table gives the measured deflection x that
was produced in a particular beam by the given applied force f . Plot the data to see
whether a linear relation exists between f and x.
108
Example
The data lies close to a straight line, so we
can use the linear function x =a f to describe
the relation. The value of the constant a can
be determined from the slope of the line.
Choosing the origin and the last data point to
find the slope, we obtain
This relation is usually written as f = kx, where k is called the beam stiffness.
Thus, k = 1/a = 1025 lb/in.
109
FUNCTION IDENTIFICATION AND PARAMETER ESTIMATION
• Function identification, or function discovery, is the process of identifying or
discovering a function that can describe a particular set of data.
• The term curve fitting is also used to describe the process of finding a curve, and the
function generating the curve, to describe a given set of data.
• Parameter estimation is the process of obtaining values for the parameters, or
coefficients, in the function that describes the data.
The following three function types can often describe physical phenomena
• The linear function y(x) = mx + b. Note that y(0) = b.
• The power function 𝑌 𝑥 = 𝑏𝑋 𝑚 . Note that y(0) = 0 if m ≥ 0, and y(0) = ∞ if m < 0.
• The exponential function 𝑦 𝑥 = 𝑏(10)𝑚𝑥 or its equivalent form 𝑦 = 𝑒 𝑚𝑥 , where e is
the base of the natural logarithm (ln e = 1). Note that y(0) = b for both forms.
For example, the linear function describes the voltage-current relation for a resistor
(v = i R) and the velocity versus time relation for an object with constant acceleration
a (v = at +v0). The distance d traveled by a falling object versus time is described by
a power function (d = 0.5gt2). The temperature change T of a cooling object can be
described by an exponential function (∆T = ∆T0 e−ct ).
110
FUNCTION IDENTIFICATION AND PARAMETER ESTIMATION
Each function gives a straight line when plotted using a specific set of axes:
• The linear function y = mx + b gives a straight line when plotted on rectilinear axes. Its
slope is m and its y intercept is b.
• The power function y = bxm gives a straight line when plotted on log-log axes.
• The exponential function y = b(10)mx and its equivalent form, y = bemx, give a straight
line when plotted on semilog axes with a logarithmic y axis.
Using the following properties of base-ten logarithms, which are shared with natural
logarithms, we have:
log (ab) = log a + log b
log (am) = m log a
Take the logarithm of both sides of the power equation y = bxm to obtain
log y = log (bxm) = log b + m log x
This has the form Y = B +mX if we let Y = log y, X = log x, and B = log b. Thus if we plot
Y versus X on rectilinear scales, we will obtain a straight line whose slope is111m and
whose intercept is B.
STEPS FOR FUNCTION IDENTIFICATION
summary of the procedure to find a function that describes a given set of data.
We assume that the data can be described by one of the three function types given
above.
1. Examine the data near the origin. The exponential functions y = b(10)mx and y =
bemx can never pass through the origin (unless, of course b = 0, which is a trivial
case). The linear function y = mx + b can pass through the origin only if b = 0. The
power function y = bxm can pass through the origin but only if m > 0.
2. Plot the data using rectilinear scales. If it forms a straight line, then it can be
represented by the linear function, and you are finished. Otherwise, if you have data at
x = 0, then
a. If y(0) = 0, try the power function, or
b. If y(0) ≠ 0, try the exponential function.
If data is not given for x = 0, proceed to step 3.
3. If you suspect a power function, plot the data using log-log scales. Only a power
function will form a straight line. If you suspect an exponential function, plot it using
semilog scales. Only an exponential function will form a straight line. 112
OBTAINING THE COEFFICIENTS
113
OBTAINING THE COEFFICIENTS
114
OBTAINING THE COEFFICIENTS
If the data are scattered about a straight line to the extent that it is difficult to
identify a unique straight line that describes the data, we can use the least-
squares method to obtain the function. This method finds the coefficients of a
polynomial of specified degree n that best fits the data, in the so-called “least-
squares sense.”
The MATLAB implementation of this method uses the polyfit function,
115
PROBLEM
116
SOLUTION
Common sense tells us that the water temperature will eventually reach the air
temperature of 70◦. Thus we first subtract 70◦ from the temperature data T and seek to
obtain a functional description of the relative temperature, T = T − 70.
A plot of the relative temperature data is shown in Figure. We note that the plot has a
distinct curvature and that it does not pass through the origin. Thus we can rule out the
linear function and the power function as candidates.
Plot of relative
temperature versus time.
117
SOLUTION
To see if the data can be described by an exponential function, we plot the data on a
semilog plot, which is shown in Figure. The straight line shown can be drawn by aligning
a straightedge so that it passes near most of the data points (note that this line is
subjective; another person might draw a different line). The data lie close to a straight
line, so we can use the exponential function to describe the relative temperature.
Semilog plot of relative
temperature versus
time.
118
SOLUTION
119
PROBLEM
A hole 6 mm in diameter was made in a translucent milk container (Figure shown). A
series of marks 1 cm apart was made above the hole. While adjusting the tap flow to
keep the water height constant, the time for the outflow to fill a 250-ml cup was
measured (1 ml=10−6 m3). This was repeated for several heights. The data are given in
the following table.
Obtain a functional description of the volume outflow rate f as a function
of water height h above the hole.
120
SOLUTION
First obtain the flow rate data in ml/s by dividing the 250 ml volume by the time to fill:
250
𝑓=
𝑡
A plot of the resulting flow rate data is shown in above Figure. There is some curvature
in the plot, so we rule out the linear function. Common sense tells us that the outflow
rate will be zero when the height is zero, so we can rule out the exponential function
because it cannot pass through the origin.
The log-log plot shown in below Figure shows that the data lie close to a straight line, so
we can use the power function to describe the flow rate as a function of height. Thus we
can write 𝑓 = 𝑏ℎ𝑚
The straight line shown can be drawn by
aligning a straightedge so that it passes near
most of the data points (note that this line is
subjective; another person might draw a
different line). Next we select two points on the
straight line to find the values of b and m. The
two points indicated by an asterisk were
selected to minimize interpolation error because
they lie near grid lines. The accuracy of the
values read from the plot obviously depends on
the size of the plot. The values of the points as
121
read from the plot are (1, 9.4) and (8, 30).
SOLUTION
Thus the estimated function is f = 9.4h0.558, where f is the outflow rate in ml/s and the
water height h is in centimeters.
122
FITTING MODEL TO SCATTERED DATA
According to the least-squares criterion, the line that gives the best fit is the one that
minimizes J , the sum of the squares of the vertical differences between the line and the
data points.
These differences are called the residuals.
123
FITTING MODEL TO SCATTERED DATA
124
GENERALIZED MODEL TO SCATTERED DATA
125
PROBLEM
126
PROBLEM
127
Modeling of Rigid-Body Mechanical Systems
128
TRANSLATIONAL MOTION
• A particle is a mass of negligible dimensions. We can consider a body to be a particle
if its dimensions are irrelevant for specifying its position and the forces acting on it.
For example, we normally need not know the size of an earth satellite to describe its
orbital path.
• Newton’s first law states that a particle originally at rest, or moving in a straight line
with a constant speed, will remain that way as long as it is not acted upon by an
unbalanced external force.
• Newton’s second law states that the acceleration of a mass particle is proportional to
the vector resultant force acting on it and is in the direction of this force.
• Newton’s third law states that the forces of action and reaction between interacting
bodies are equal in magnitude, opposite in direction, and collinear.
• The third law is summarized by the commonly used statement that every action is
opposed by an equal reaction.
129
TRANSLATIONAL MOTION
• For an object treated as a particle of mass m, the second law can be expressed as
𝑑𝑣
𝑚𝑎 = 𝑚 =𝑓
𝑑𝑡
where a and v are the acceleration and velocity vectors of the mass and f is the force
vector acting on the mass.
• Note that the acceleration vector and the force vector lie on the same line.
• If the mass is constrained to move in only on
direction, say along the direction of the coordinate x,
then the equation of motion is the scalar equation
𝑑𝑣
𝑚𝑎 = 𝑚 =𝑓
𝑑𝑡
𝑑𝑣 𝑓
𝑜𝑟, = =𝑎
𝑑𝑡 𝑚
𝑑𝑥
𝑥 𝑡 =
𝑑𝑡
𝑑2𝑥
𝑥 𝑡 = 2
𝑑𝑡 130
we can express the scalar form of Newton’s law as 𝑚𝑣 = 𝑓.
TRANSLATIONAL MOTION
• If we assume that the object is a rigid body and we neglect the force distribution
within the object, we can treat the object as if its mass were concentrated at its mass
center. This is the point mass assumption, which makes it easier to obtain the
translational equations of motion, because the object’s dimensions can be ignored
and all external forces can be treated as if they acted through the mass center.
MECHANICAL ENERGY
• Conservation of mechanical energy is a direct consequence of Newton’s second law.
Consider the scalar case, where the force f can be a function of displacement x.
𝑚𝑣 = 𝑓(𝑥)
Multiply both sides by v dt and use the fact that v = dx/dt.
𝑑𝑥
𝑣 𝑓 𝑥 𝑑𝑡 = 𝑓 𝑥 𝑑𝑡 = 𝑓 𝑥 𝑑𝑥
𝑑𝑡
131
TRANSLATIONAL MOTION
MECHANICAL ENERGY
This equation shows that V(x) has the same units as kinetic energy. V(x) is called the
potential energy (PE) function.
This Equation states that the sum of the kinetic and potential energies must be
constant, if no force other than the conservative force is applied.
132
TRANSLATIONAL MOTION
MECHANICAL ENERGY
133
MECHANICAL ENERGY
An object with a mass of m = 2 slugs drops from a height of 30 ft
above the ground (see Figure). Determine its speed after it drops
20 ft to a platform that is 10 ft above the ground.
134
TIME DOMAIN ANALYSIS
135
Introduction
• In time-domain analysis the response of a dynamic system to an input is
expressed as a function of time.
• It is possible to compute the time response of a system if the nature of input
and the mathematical model of the system are known.
• Usually, the input signals to control systems are not known fully ahead of time.
• It is therefore difficult to express the actual input signals mathematically by
simple equations.
136
Standard Test Signals
• The characteristics of actual input signals are a sudden shock, a
sudden change, a constant velocity, and constant acceleration.
• The dynamic behavior of a system is therefore judged and compared
under application of standard test signals – an impulse, a step, a
constant velocity, and constant acceleration.
• The other standard signal of great importance is a sinusoidal signal.
137
Standard Test Signals
• Impulse signal
• The impulse signal imitate the sudden shock
characteristic of actual input signal.
δ(t)
• If A=1, the impulse signal is called unit impulse
signal. A
A t0
(t )
0 t0
0 t
138
Standard Test Signals
• Step signal
• The step signal imitate the sudden
change characteristic of actual input
signal.
u(t)
• If A=1, the step signal is called unit step
signal
A t0 t
u(t ) 0
0 t0
139
Standard Test Signals
• Ramp signal r(t)
• The ramp signal imitate the constant
velocity characteristic of actual input
signal.
0 t
• If A=1, the ramp signal is called unit
ramp signal
At t0
r(t )
r(t)
0 t0 ramp signal with slope A
r(t)
unit ramp signal
140
Standard Test Signals
p(t)
• Parabolic signal
• The parabolic signal imitate the constant
acceleration characteristic of actual
input signal.
0 t
• If A=1, the parabolic signal is called unit
At 2
parabolic signal.
p(t)
t0
p(t ) 2
0 t0
parabolic signal with slope A
p(t)
Unit parabolic signal
141
Relation between standard Test Signals
A t0
•Impulse (t )
t0
d
0
dt
•Step A t0
u(t )
0 t0 d
dt
•Ramp At t0
r(t )
0 t0 d
•Parabolic At 2
dt
t0
p(t ) 2
0 t0
142
Laplace Transform of Test Signals
•Impulse
A t0
(t )
0 t0
L{ (t )} ( s ) A
•Step
A t0
u(t )
0 t0
A
L{u(t )} U ( s )
S
143
Laplace Transform of Test Signals
•Ramp At t0
r(t )
0 t0
A
L{ r(t )} R( s )
s2
•Parabolic
At 2
t0
p(t ) 2
0 t0
A
L{ p (t )} P( s ) 3
S 144
Time Response of Control Systems
• Time response of a dynamic system response to an input
expressed as a function of time.
System
• The time response of any system has two components
• Transient response
• Steady-state response.
145
Time Response of Control Systems
• When the response of the system is changed from equilibrium it
takes some time to settle down.
• This is called transient response.
-3
x 10 Step Response
6
Step Input
5
• The response of the
Steady State Response
system after the transient 4
Response
Amplitude
response is called steady 3
state response. 2
Transient Response
1
0
0 2 4 6 8 10 12 14 16 146 18 20
Time (sec)
Time Response of Control Systems
• Transient response depend upon the system poles only and not
on the type of input.
• It is therefore sufficient to analyze the transient response using a
step input.
• The steady-state response depends on system dynamics and the
input quantity.
• It is then examined using different test signals by final value
theorem.
147
Introduction
• The first order system has only one pole.
C( s ) K
R( s ) Ts 1
• Where K is the D.C gain and T is the time constant of the system.
• Time constant is a measure of how quickly a 1st order system responds to a unit
step input.
• D.C Gain of the system is ratio between the input signal and the steady state
value of output.
148
Introduction
• The first order system given below.
10
G( s )
3s 1
• D.C gain is 10 and time constant is 3 seconds.
• For the following system
3 3/ 5
G( s )
s 5 1/ 5s 1
• D.C Gain of the system is 3/5 and time constant is 1/5
seconds.
149
Impulse Response of 1st Order System
• Consider the following 1st order system
δ(t)
K
R(s ) C(s )
1
Ts 1
t
0
R( s ) ( s ) 1
K
C( s )
Ts 1
150
Impulse Response of 1st Order System
K
C( s )
Ts 1
• Re-arrange following equation as
K /T
C( s )
s 1/ T
• In order to compute the response of the system in time domain
we need to compute inverse Laplace transform of the above
equation.
K t / T
1
C at c( t ) e
L Ce T
sa 151
Impulse Response of 1st Order System
K t / T
• If K=3 and T=2s then c( t ) e
T
K/T*exp(-t/T)
1.5
1
c(t)
0.5
0
0 2 4 6 8 10
152
Time
Step Response of 1st Order System
• Consider the following 1st order system
K
R(s ) C(s )
Ts 1
1
R( s ) U ( s )
s
K
C( s )
s Ts 1
• In order to find out the inverse Laplace of the above equation, we
need to break it into partial fraction expansion (page 867 in the
Textbook) K KT
C( s )
Ts 1
153
s
Step Response of 1st Order System
1 T
C( s ) K
s Ts 1
• Taking Inverse Laplace of above equation
c(t ) K u(t ) e t / T
c(t ) K 1 e
• Where u(t)=1
t / T
• When t=T (time constant)
c(t ) K 1 e 1 0.632 K
154
Step Response of 1st Order System
• If K=10 and T=1.5s then
c(t ) K 1 e t / T
K*(1-exp(-t/T))
11
10
9 Step Response
8
steady state output 10
7 D.C Gain K
63 % Input 1
6
c(t)
2
Unit Step Input
1
0
0 1 2 3 4 5 6 7 8 9 10 155
Time
Step Response of 1st order System
• System takes five time constants to reach its final value.
156
Step Response of 1st Order System
• If K=10 and T=1, 3, 5, 7
c(t ) K 1 e t / T
K*(1-exp(-t/T))
11
10
T=1s
9
8 T=3s
7
T=5s
6
c(t)
5 T=7s
4
3
2
1
0
0 5 10 15
157
Time
Step Response of 1st Order System
• If K=1, 3, 5, 10 and T=1
c(t ) K 1 e t / T
K*(1-exp(-t/T))
11
10
K=10
9
8
7
6
K=5
c(t)
5
4
K=3
3
2
K=1
1
0
0 5 10 15 158
Time
Practical Determination of Transfer Function
of 1st Order Systems
• Often it is not possible or practical to obtain a system's
transfer function analytically.
• Perhaps the system is closed, and the component parts are
not easily identifiable.
• The system's step response can lead to a representation
even though the inner construction is not known.
• With a step input, we can measure the time constant and
the steady-state value, from which the transfer function can
be calculated.
159
Practical Determination of Transfer
Function of 1st Order Systems
• If we can identify T and K empirically we can obtain the
transfer function of the system.
C( s ) K
R( s ) Ts 1
160
Practical Determination of Transfer Function
of 1st Order Systems
• For example, assume the unit step
response given in figure.
K=0.72
• From the response, we can
measure the time constant, that
is, the time for the amplitude to
reach 63% of its final value.
• Since the final value is about T=0.13s
0.72 the time constant is
evaluated where the curve
reaches 0.63 x 0.72 = 0.45, or • Thus transfer function is
about 0.13 second. obtained as:
• K is simply steady state value. C( s ) 0.72 5.5
R( s ) 0.13s 1 s 7.7
161
Practical Determination of Transfer Function of 1st
Order Systems
Courtesy: Process Dynamics and Control, 4th Edition 162
Dale E. Seborg, Thomas F. Edgar, Duncan A. Mellichamp, Francis J. Doyle III
Practical Determination of Transfer Function of 1st
Order Systems
Courtesy: Process Dynamics and Control, 4th Edition 163
Dale E. Seborg, Thomas F. Edgar, Duncan A. Mellichamp, Francis J. Doyle III
First Order System with a Zero
C( s ) K (1 s )
R( s ) Ts 1
• Zero of the system lie at -1/α and pole at -1/T.
• Step response of the system would be:
K (1 s )
C( s )
sTs 1
K K ( T )
C( s )
s Ts 1
K
c(t ) K ( T )e t / T
T 164
First Order System With Delays
• Following transfer function is the generic representation of 1st order
system with time lag.
• Where td is the delay time.
C( s ) K
e std
R( s ) Ts 1
165
First Order System With Delays
C( s ) K std
e
R( s ) Ts 1
Unit Step
Step Response
t
td
166
TRANSPORTATION LAG
A phenomenon that is often present in flow systems is the transportation lag.
Synonyms
for this term are dead time and distance velocity lag.
167
TRANSPORTATION LAG
APPROXIMATION OF TRANSPORT LAG.
168
Second Order System
• We have already discussed the affect of location of poles and zeros
on the transient response of 1st order systems.
• Compared to the simplicity of a first-order system, a second-order
system exhibits a wide range of responses that must be analyzed
and described.
• Varying a first-order system's parameter (T, K) simply changes the
speed and offset of the response
• Whereas, changes in the parameters of a second-order system can
change the form of the response.
• A second-order system can display characteristics much like a first-
order system or, depending on component values, display damped
or pure oscillations for its transient response. 169
Introduction
• A general second-order system is characterized by the following
transfer function.
C( s ) n2
2
R( s ) s 2 n s n2
n un-damped natural frequency of the second order system,
which is the frequency of oscillation of the system without
damping.
damping ratio of the second order system, which is a measure
of the degree of resistance to change in the system output. 170
Example 2
• Determine the un-damped natural frequency and damping ratio
of the following second order system.
C( s ) 4
2
R( s ) s 2s 4
• Compare the numerator and denominator of the given transfer
function with the general 2nd order transfer function.
C( s ) n2
2
R( s ) s 2 n s n2
n2 4 n 2 2 n s 2s
n 1
s 2 2 n s n2 s 2 2s 4
0.5 171
Introduction
C( s ) n2
2
R( s ) s 2 n s n2
• Two poles of the system are
n n 2 1
n n 2 1
172
Introduction
n n 2 1
n n 2 1
• According the value of , a second-order system can be set into
one of the four categories (page 169 in the textbook):
1. Overdamped - when the system has two real distinct poles ( >1).
jω
δ
-c -b -a
173
Introduction
n n 2 1
n n 2 1
• According the value of , a second-order system can be set into
one of the four categories (page 169 in the textbook):
2. Underdamped - when the system has two complex conjugate poles (0 < <1)
jω
δ
-c -b -a
174
Introduction
n n 2 1
n n 2 1
• According the value of , a second-order system can be set into
one of the four categories (page 169 in the textbook):
3. Undamped - when the system has two imaginary poles ( = 0).
jω
δ
-c -b -a
175
Introduction
n n 2 1
n n 2 1
• According the value of , a second-order system can be set into
one of the four categories (page 169 in the textbook):
4. Critically damped - when the system has two real but equal poles ( = 1).
jω
δ
-c -b -a
176
Underdamped System
For 0< <1 and ωn > 0, the 2nd order system’s response due to a
unit step input is as follows.
Important timing characteristics: delay time, rise time, peak
time, maximum overshoot, and settling time.
177
Delay Time
• The delay (td) time is the time required for the response to
reach half the final value the very first time.
178
Rise Time
• The rise time is the time required for the response to rise from 10%
to 90%, 5% to 95%, or 0% to 100% of its final value.
• For underdamped second order systems, the 0% to 100% rise time is
normally used. For overdamped systems, the 10% to 90% rise time is
commonly used.
Peak Time
• The peak time is the time required for the response to reach
the first peak of the overshoot.
180
180
Maximum Overshoot
The maximum overshoot is the maximum peak value of the
response curve measured from unity. If the final steady-state
value of the response differs from unity, then it is common to
use the maximum percent overshoot. It is defined by
The amount of the maximum (percent) overshoot directly
indicates the relative stability of the system.
181
Settling Time
• The settling time is the time required for the response curve
to reach and stay within a range about the final value of size
specified by absolute percentage of the final value (usually 2%
or 5%).
182
Step Response of underdamped System
C( s ) n2 Step Response n2
2 C( s )
R( s ) s 2 n s n2
s s 2 2 n s n2
• The partial fraction expansion of above equation is given as
1 s 2 n
C( s ) 2
s s 2 n s n2
n2 1 2
1 s 2 n
C( s ) 2
s 2 n 2 s s 2 n s 2n2 n2 2n2
1 s 2 n
C( s )
s s n 2 n2 1 2 183
Step Response of underdamped System
1 s 2 n
C( s )
s s n 2 n2 1 2
• Above equation can be written as
1 s 2 n
C( s )
s s n 2 d2
• Where d n 1 2 , is the frequency of transient oscillations
and is called damped natural frequency.
• The inverse Laplace transform of above equation can be obtained
easily if C(s) is written in the following form:
1 s n n
C( s )
s s n d s n 2 d2
2 2
184
Step Response of underdamped System
1 s n n
C( s )
s s n d s n 2 d2
2 2
n 1 2
1 s n 1 2
C( s )
s s n d
2 2
s n 2 d2
1 s n d
C( s )
s s n d
2 2
1 2 s 2 2
n d
n t
c(t ) 1 e cos d t e nt sin d t
1 2
185
Step Response of underdamped System
n t
c(t ) 1 e cos d t e nt sin d t
1 2
c(t ) 1 e nt cos d t sin d t
1 2
• When 0
d n 1 2
n
c(t ) 1 cos n t
186
Step Response of underdamped System
c(t ) 1 e nt cos d t sin d t
1 2
if 0.1 and n 3
1.8
1.6
1.4
1.2
0.8
0.6
0.4
0.2
187
0
0 2 4 6 8 10
Step Response of underdamped System
c(t ) 1 e nt cos d t sin d t
1 2
if 0.5 and n 3
1.4
1.2
0.8
0.6
0.4
0.2
188
0
0 2 4 6 8 10
Step Response of underdamped System
c(t ) 1 e nt cos d t sin d t
1 2
if 0.9 and n 3
1.4
1.2
0.8
0.6
0.4
0.2
0 189
0 2 4 6 8 10
Introductory Concept of Stability
The stability of a cone.
Introductory Concept of Stability
When a process is "disturbed" from an initial steady state, say, by
implementing a change in the input forcing function, it will, in general, respond
in one of three ways:
• Response l: Proceed to a steady state and remain there, or
• Response 2: Fail to attain to steady-state conditions because its output
grows indefinitely, or
• Response 3: Fail to attain steady-state conditions because the process
oscillates indefinitely with a constant amplitude.
191
Introductory Concept of Stability
output grows indefinitely, or
output grows indefinitely, or
oscillates indefinitely with a constant
amplitude
192
Introductory Concept of Stability
Stability in the s-plane.
193
Introductory Concept of Stability
It now seems perfectly logical to consider the process that ultimately settles
down after having been disturbed as being “Stable" and the one that fails to
settle down as being "unstable."
If in response to a bounded input, the dynamic trajectory of a system
remains bounded as t→ ∞, then the system is said to be stable; otherwise,
it is said to be unstable
*For the study of linear systems in process control, it is normally the concept of BIBO
stability that we shall use.
194
Practical Issues in Process Dynamics and Control
There are several practical questions concerning system stability that arise in
process dynamics and control. Some of the most important are:
• When is a linear system stable?
• When is a nonlinear system stable?
• Can a system that is unstable by itself (i.e., a "naturally unstable"
system) be made stable by addition of a control system?
• Can a "naturally stable" system be made unstable by the addition of a
control system?
195
Concept of Stability
A linear system is stable if and only if
all its poles have negative real parts
so that they lie in the left half of the
complex plane; otherwise it is
unstable.
196
Concept of Stability
• If all the poles of the system lie in left half plane the
system is said to be Stable.
• If any of the poles lie in right half plane the system is
said to be unstable.
• If pole(s) lie on imaginary axis the system is said to be
marginally stable. j
LHP RHP
s-plane
197
Concept of Stability
• The system is said to be stable if for any bounded
input the output of the system is also bounded
(BIBO).
• Thus for any bounded input the output either
remain constant or decrease with time.
u(t) overshoot
y(t)
1
Plant 1
t
t
Unit Step Input
Output
198
Concept of Stability
• If for any bounded input the output is not bounded the system is said to
be unstable.
u(t)
y(t)
1
e at
Plant
t
t
Unit Step Input
Output
199
BIBO vs Transfer Function
• For example
Y (s) 1 Y ( s) 1
G1 ( s) G2 ( s)
U (s) s 3 U ( s) s 3
Pole-Zero Map Pole-Zero Map
4 4
unstable
3 stable 3
2 2
1 1
Imaginary Axis
Imaginary Axis
0 0
-1 -1
-2 -2
-3 -3
-4 -4
-4 -2 0 2 4 -4 -2 0 2 4
200
Real Axis Real Axis
BIBO vs Transfer Function
• For example
Y (s) 1 Y ( s) 1
G1 ( s) G2 ( s)
U (s) s 3 U ( s) s 3
1 Y (s)
1 1 1 Y (s) 1
G1 ( s ) 1
G2 ( s ) 1
1
U (s) s3 U (s) s 3
y (t ) e 3t u (t ) y (t ) e3t u (t )
201
BIBO vs Transfer Function
• For example
3t
y(t ) e u (t ) y(t ) e3t u (t )
12
exp(-3t)*u(t) x 10 exp(3t)*u(t)
1 12
10
0.8
8
0.6
6
0.4
4
0.2
2
0 0
0 1 2 3 4 0 2 4 6 8 10
202
203
BIBO vs Transfer Function
• Whenever one or more than one poles are in RHP the solution of dynamic
equations contains increasing exponential terms.
• That makes the response of the system unbounded and hence the overall
response of the system is unstable.
204
205
Concept of Stability
A linear system in the form of state space form is stable if and only if all of its
eigenvalues, λi have negative real parts so that they lie in the left half of the
complex plane.
206
207
Conclusion:
• Stability analysis for a closed-loop system boils down to investigating the
roots of the characteristic equation.
• This characteristic equation is a polynomial in s whose order is, of course,
determined by the order of the contributing transfer functions.
• It should not be too difficult to see from the above example that for low-order
polynomial characteristic equations, the roots are easy to find, and stability
analysis is therefore easy to carry out.
• For higher order polynomials, however, finding these roots explicitly can be
quite tedious. Thus less tedious methods of stability analysis, which cleverly
sidetrack the issue of actually calculating these roots, are to be preferred.
208
INTRODUCTION TO INDUSTRIAL
CONTROLLERS
209
Introduction to Industrial Controllers
210
Introduction to Industrial Controllers
211
Industrial Controllers
• Proportional Controller
• Integral Controller
• Derivative Controller
• Proportional Plus Integral Controller
• Proportional Plus Derivative Controller
• Proportional Plus Integral Plus Derivative Controller
212
Proportional Controller
213
Proportional Controller
• For digital implementation, p and e often expressed as numbers between 0 and
100%.
• This representation is usually used for compute
control software.
• Some controllers have a proportional band setting
instead of controller gain.
• The proportional band PB (in %) is defined as
• Note that a small (narrow) proportional band corresponds to a large controller gain,
whereas a large (wide) PB value implies a small value of Kc
214
Proportional Controller
• A more realistic representation is
• The controller saturates when its output reaches
a physical limit, either Pmax or Pmin
• Transfer function for P-Control:
• Disadvantage of Proportional Control:
Steady-state error (or offset) occurs after a set-point change or a sustained disturbance.
215
Integral Controller
• The controller output depends on the integral of the error signal over time,
• Integral controller eliminates the Offset
216
Integral Controller
217
Proportional + Integral Controller
218
219
Proportional + Integral Controller
220
Proportional + Integral Controller
221
Proportional + Integral Controller
222
Derivative Controller
223
Derivative Controller
D-Control action is never used alone; it is always used in
conjunction with proportional or PI control.
224
Proportional + Derivative Controller
• An ideal PD controller has the transfer function
• Derivative control action also tends to improve the
dynamic response of the controlled variable by
decreasing the process settling time.
• If the process measurement is noisy, that is, if it contains
high frequency, random fluctuations, then the derivative
of the measured variable will change widely and
derivative action will amplify the noise unless the
measurement if filtered.
225
Proportional + Derivative Controller
• Derivative action is seldom used for flow control because
flow control loops respond quickly and flow
measurement tend to be noisy.
• Unfortunately, the ideal PD control algorithm is physically
unrealizable because it cannot be implemented exactly
using either analog or digital components.
• For an analog controllers, the TF can be approximated by
• Where α typically has a value between 0.05 and 0.2, with
0.1 being a common choice. 226
Proportional + Derivative Controller
• This equation is also known as Derivative
mode filter and it reduces the sensitivity of
the control calculations to high frequency
noise in the measurement.
227
Proportional + Integral + Derivative Controller
228
Proportional + Integral + Derivative Controller
229
Derivative kick
• One disadvantage of the PID controllers is that a sudden change in
set point (and hence the error, e) will cause the derivative terms
momentarily to become very large and thus provide a derivative kick
to FCE.
• This sudden change is undesirable and can be avoided by basing the
derivative action on the measurement, ym rather than on the error
signal, e.
• Eliminating the derivative kick by considering the parallel form PID
control by replacing de/dt by –dym/dt gives
230
Derivative kick
• A more flexible PID control algorithm can be obtained by weighting the set point in the
proportional term, as well as in the derivative term. This modification eliminates the
Proportional kick. That also occurs after a step change in set point.
231
Derivative kick
• Where β and ϒ are non negative constants. This control algorithm is known
as the parallel PID controller with proportional and derivative mode
weighting, or as the beta-gamma controller.
232
PID controllers
233
PID controllers Response
234
PID controllers Response
235
PID controllers realization using OP-AMP
236
PID controllers realization using OP-AMP
237
PID controllers realization using OP-AMP
238
The verb tuning in radio contexts means
adjusting the radio receiver to receive the
desired radio signal carrier frequency that a
particular radio station uses. 239
WHAT MOTIVATES CONTROL ENGINEERS TO
TUNE INDUSTRIAL CONTROLLERS
240
WHAT MOTIVATES CONTROL ENGINEERS TO
TUNE INDUSTRIAL CONTROLLERS
241
WHAT MOTIVATES CONTROL ENGINEERS TO
TUNE INDUSTRIAL CONTROLLERS
242
WHAT MOTIVATES CONTROL ENGINEERS TO
TUNE INDUSTRIAL CONTROLLERS
243
WHAT MOTIVATES CONTROL ENGINEERS TO
TUNE INDUSTRIAL CONTROLLERS
Choice of value of the P, I and D parameters is
very much process dependent.
244
WHAT MOTIVATES CONTROL ENGINEERS TO
TUNE INDUSTRIAL CONTROLLERS
To choose/compute optimal
controller settings (Kp, Ki and Kd)
for satisfactory response
245
WHAT MOTIVATES CONTROL ENGINEERS TO
TUNE INDUSTRIAL CONTROLLERS
Satisfactory Response
Single Performance criteria
• Minimum Overshoot
• Settling Time as small as possible
• Rise Time as small as possible
• Minimum Steady State Error
(ess or e(∞)= 0; ideally)
• ¼ (QAD) decay ratio
• Satisfactory Setpoint Tracking with
high disturbance rejection
246
WHAT MOTIVATES CONTROL ENGINEERS TO
TUNE INDUSTRIAL CONTROLLERS
QAD decay ratio
• it’s a poor choice for process
stability.
• The idea behind QAD is to
eliminate any error between
s.p and p.v very fast.
247
WHAT MOTIVATES CONTROL ENGINEERS TO
TUNE INDUSTRIAL CONTROLLERS
QAD decay ratio
Although the QAD tuning objective provides very fast rejection of disturbances,
it creates three problems:
1.It makes loop very oscillatory, often causing interactions with similarly tuned
loops. If control loops in a highly interactive process, such as a paper machine,
power plant boiler, or hydrodealkylation process, are tuned for QAD,
oscillations affecting the entire process often occur.
2.It causes a loop to overshoot its setpoint when recovering from a process
disturbance and after a setpoint change. Many processes cannot tolerate
overshoot.
3.QAD-tuned loops are not very stable and have low robustness. They can
very easily become completely unstable if process characteristics change. For
example, such a loop will become unstable if its process gain doubles,248which
can happen very easily in industrial processes.
WHAT MOTIVATES CONTROL ENGINEERS TO
TUNE INDUSTRIAL CONTROLLERS
Satisfactory Response
Performance Criteria
Time Integral
249
WHAT MOTIVATES CONTROL ENGINEERS TO
TUNE INDUSTRIAL CONTROLLERS
250
WHAT MOTIVATES CONTROL ENGINEERS TO
TUNE INDUSTRIAL CONTROLLERS
251
Summary
• The adjustment of control parameters to achieve satisfactory control is called
tuning.
• The selection of controller parameters is essentially an optimization problem
in which designer of control system attempts to satisfy some criterion of
optimality, result of which is often referred to as “good control”.
• The process of tuning can vary from a trial and error attempts to find suitable
control parameters for “good” control to an elaborate optimization calculation
based on model of the process and a specific criterion of the optimal control.
• In many applications, there is no control model of the process and the
criterion for good control is only vaguely defined.
• A typical criterion for good control is that the response of the system to a step
change in set point or load have overshoot and one quarter decay ratio.
• Min. rise time, settling time, overshoot, decay time etc.
252
Tuning Method:
• Trial and Error Method
• Process reaction curve technique
• Zeigler-Nichols method:
• Cohen-Coon Methods
253
254
• Trial and error is a fundamental method of
problem Solving.
• It is characterized by repeated, varied
attempts which are continued until
success, or until the practitioner stops trying.
• Trial and error is also a heuristic method of problem
solving, repair, tuning, or obtaining knowledge.
• In the field of CSE, the method is called generate and test
(Brute force).
• In elementary algebra, when solving equations, it is guess
and check. 255
• Consider a third-order plant model given by
G(s) = 1/(s + 1)3.
If a proportional control strategy is selected, i.e., Ti → ∞ and Td → 0 in
the PID control strategy, for different values of Kp, the closed-loop
responses of the system can be obtained using the following MATLAB
G=tf(1,[1,3,3,1]);
for Kp=[0.1:0.1:1]
G_C=feedback(Kp*G,1);
step(G_C), hold on;
end
figure; 256
257
258
G=tf(1,[1,3,3,1]);
Kp=1;
s=tf('s');
for Ti=[0.7:0.1:1.5]
Gc=Kp*(1+1/Ti/s);
G_C=feedback(G*Gc,1);
step(G_C), hold on;
end
figure;
259
260
G=tf(1,[1,3,3,1]);
Kp=1; Ti=1; s=tf('s');
for Td=[0.1:0.2:2]
Gc=Kp*(1+1/Ti/s+Td*s);
G_c=feedback(G*Gc,1); step(G_c),
hold on
end
figure;
261
• In practical applications, the pure derivative action is never
used, due to the “derivative kick” produced in the control
signal for a step input, and to the undesirable noise
amplification.
• It is usually replaced by a first-order low pass filter.
G=tf(1,[1,3,3,1]);
Kp=1; Ti=1; s=tf('s');
Td=1;
Gc=Kp*(1+1/Ti/s+Td*s);
step(feedback(G*Gc,1)), hold on
for N=[100,1000,10000,1:10]
Gc=Kp*(1+1/Ti/s+Td*s/(1+Td*s/N))
; G_c=feedback(G*Gc,1);
step(G_c)
end
figure; [y,t]=step(G_c); err=1-
y; plot(t,err) 262
It can be seen that with N = 10, the
approximation is fairly satisfactory
263
Error Signal, The error signal e(t) when N = 10
It can be seen that with N = 10, the
approximation is fairly satisfactory
264
265
THE ZIEGLER-NICHOLS METHODS
• In the late 1930s, Nathaniel Nichols’s worked at Taylor
Instruments Company in Rochester as a process control
engineer.
• There he met chemical engineer John Ziegler, and together
they developed and published simple techniques for adjusting
automatic controllers for process control systems, which
became known as the Ziegler-Nichols tuning rules.
http://web.mit.edu/klund/www/papers/CSM25v26n6.pdf
• Ziegler and Nichols published in 1942 a paper
J. G. Ziegler and N. B. Nichols: Optimum Settings for
Automatic Controllers, Trans. ASME, Vol. 64, 1942, s. 759-768
• They described two methods for tuning parameters of P-, PI-
and PID controllers:
1. Ziegler-Nichols’ open loop method
266
2. Ziegler-Nichols’ closed loop method
THE ZIEGLER-NICHOLS METHODS
In Summary, the ideal response curve should be
267
THE ZIEGLER-NICHOLS METHODS
• In order to account for higher-order dynamics that are
neglected in a first order model, a time delay term can
be included.
• This modification can improve the agreement between
model and experimental responses.
• The fitting of a first-order plus time-delay model
(FOPTD)
to a unit step input.
268
Process Reaction Curve Method: Z-N Open
loop Method
• In this method, we obtain experimentally the open loop
response of the FOPDT to a unit step input.
• This method only applied if the response to a step input exhibits
an s-shaped curve as shown in figure.
• This means that if the plant involves integrators (like 2nd order
prototypes system) or complex-conjugate poles (general 2nd
order system), then this method can’t be applied since s-
shaped will not be obtained.
• The Ziegler-Nichols open-loop method is also referred to as a
process reaction method, because it tests the open-loop
reaction of the process to a change in the control variable
output.
269
Process Reaction Curve Method: Z-N Open
loop Method
270
Process Reaction Curve Method: Z-N Open
loop Method
To use Ziegler-Nichols open-loop tuning method, perform the
following steps:
• The process gain K is found by calculating ratio of steady-
state change in y to the size of input step change M.
• A tangent is drawn at the point of inflection of the step
response; the intersection of the tangent line and the time
axis (where y = 0) is the time delay.
• If the tangent is extended to intersect the steady state
response line (Where y = KM) , the point of intersection
corresponds to time t = θ + τ therefore, τ can be found by
subtracting θ from the point of intersection.
271
Process Reaction Curve Method: Z-N Open
loop Method
Open-loop Calculation of (𝐾𝑝, 𝑇𝑖 ,𝑇𝑑)
Approximate M constrained integral gain Optimization, AMIGO 272
Process Reaction Curve Method: Z-N Open
loop Method
• The major disadvantage of the time delay estimation
method is to compute the point of inflection
• The method of Sundaresan and Krishnaswamy (1978) avoids use of the
point of inflection construction entirely to estimate the time delay.
• They proposed that two times t1 and t2 be estimated from a step
response curve.
• These times correspond to the 35.3 and 85.3% response times,
respectively
• The time delay and time constant are then calculated from the following
equations:
𝜃 = 1.3𝑡1 − 0.29𝑡2 and 𝜏 = 0.67(𝑡2 − 𝑡1 ) 273
Process Reaction Curve Method: Z-N Open
loop Method
• The method of Sundaresan and Krishnaswamy (1978) avoids use of the
point of inflection construction entirely to estimate the time delay.
• They proposed that two times t1 and t2 be estimated from a step
response curve.
• These times correspond to the 35.3 and 85.3% response times,
respectively
• The time delay and time constant are then calculated from the following
equations:
𝜃 = 1.3𝑡1 − 0.29𝑡2 and 𝜏 = 0.67(𝑡2 − 𝑡1 )
274
Limitation of Open Loop Tuning
The main disadvantage of the open loop tuning
methods is that it is performed with the controller
switched to manual, i.e. leaving the output
uncontrolled in open loop.
The Ziegler-Nichols continuous cycling method
described next is a well-known closed loop tuning
strategy used to address this problem,
275
ZIEGLER-NICHOLS CONTINUOUS CYCLING METHOD
To tune PID controller using closed-loop Z-N method, do the
following:
• After the process reaches steady state at the normal level of operation,
remove the integral and derivative modes of the controller, leaving only
proportional control. On some PID controllers, this requires that the integral
time 𝜏𝑖 be set to its maximum value and the derivative time 𝜏𝐷 to its minimum
value. On computer-based controllers, the integral and derivative modes can
be removed completely from the controller.
i
• Select a value of proportional gain Kc, disturb the system, and observe the
transient response. If the response decays, select a higher value of Kc and
again observe the response of the system. Continue increasing the gain in
small steps until the response first exhibits a sustained oscillation. The value
of gain and the period of oscillation that correspond to the sustained
oscillation are the ultimate gain Kcu and the ultimate period Pu.
• From the values of Kcu and Pu found in step 2, use the Z-N rules given in
Table to determine controller settings (Kc, tI, tD). 276
ZIEGLER-NICHOLS CONTINUOUS CYCLING METHOD
277
ZIEGLER-NICHOLS CONTINUOUS CYCLING METHOD
Defining the ultimate gain Ku
as:
Where, A is the overall
amplitude ratio determined at
the crossover frequency, ωco
The ultimate period Pu is defined
as the period of the sustained
cycling that would occur if a
proportional controller with gain
Ku were used.
Open-loop Bode diagram for
a typical control system. 278
ZIEGLER-NICHOLS CONTINUOUS CYCLING METHOD
279
ZIEGLER-NICHOLS CONTINUOUS CYCLING METHOD
Using the Ziegler-Nichols rules, determine Kc and tI for the
control system shown in Fig.
280
ZIEGLER-NICHOLS CONTINUOUS CYCLING METHOD
281
Cohen and Coon (C-C) Rules
• Cohen Coon open-loop method: in this tuning, the control action is removed from the
controller by placing it in manual mode and an open-loop transient is induced by a
step change in the signal to the valve.
• This method was proposed by Cohen and Coon (1953) and is often used as an
alternative to the Z-N method.
Block diagram of a control loop for measurement of a process
reaction curve.
282
Cohen and Coon (C-C) Rules
• This Fig. shows a typical control
loop in which the control action is
removed and the loop opened for
the purpose of introducing a step
change (M/s) to the valve.
• The step response is recorded at
the output of the measuring
element. Block diagram of a control loop for
measurement of a process reaction curve.
• The step change to the valve is conveniently provided by the output from the
controller, which is in manual mode.
• The response of the system (including the valve, process, and measuring
element) is called the process reaction curve; a typical process reaction curve
exhibits an S shape.
283
Cohen and Coon (C-C) Rules
• The response of the system (including the valve, process, and measuring
element) is called the process reaction curve; a typical process reaction curve
exhibits an S shape.
• C-C method is summarized in the
following steps:
After the process reaches steady
state at the normal level of operation,
switch the controller to manual. In a
modern controller, the controller
output will remain at the same value
after switching as it had before
switching. (This is called “bumpless”
transfer.)
With the controller in manual,
introduce a small step change in the
Typical process reaction curve showing
controller output that goes to the
graphical construction to determine first-
valve and record the transient, which order with transport lag model.
is the process reaction curve (Fig.).
284
Cohen and Coon (C-C) Rules
Draw a straight line tangent to the
curve at the point of inflection, as
shown in Fig.. The intersection of the
tangent line with the time axis is the
apparent transport lag Td; the
apparent first-order time constant T
is obtained from
where Bu is the ultimate value of B at large t
and S is the slope of the tangent line. The
steady-state gain that relates B to M in Fig. Typical process reaction curve showing
is given by graphical construction to determine first-
order with transport lag model.
Using the values of Kp, T, and Td from step 3,
the controller settings are found from the
relations given in Table
285
Cohen and Coon (C-C) Rules
286
FINAL CONTROL ELEMENT
(Control Valve)
287
• The control system shown in Fig. is called a closed-loop system or a feedback
system.
• Because the measured value of the controlled variable is returned or “fedback” to
a device called the comparator.
• In the comparator, the controlled variable is compared with the desired value or set
point.
• If there is any difference between the measured variable and the set point, an
error is generated. This error enters a controller, which in turn adjusts the final
control element to return the controlled variable to the set point. 288
Valves positioner
Valves body
flange
• In the above figure, both are the same control valves but looking differently in
shapes and size.
• The type we choose the valve depend on the process variables, i.e., what we
have to control.
• The design of valve depend on the control variables ex: Either we have to control
289
flow, temperature of steam or level of liquid.
flange
Brain of the valve
• A flange is a method of connecting pipes, valves, pumps and other equipment to form a
piping system. It also provides easy access for cleaning, inspection or
modification. Flanges are usually welded or screwed. Flanged joints are made by bolting
together two flanges with a gasket between them to provide a seal.
290
291
292
293
294
295
296
297
Control Valve Mechanism
Pneumatic control valves. (a) Air to close; (b) air to open.
298
Control Valve Mechanism
• The control valve shown in Fig. contains a pneumatic device (valve motor) that moves
the valve stem as the pressure on a spring-loaded diaphragm changes. The stem
positions a plug in the orifice of the valve body.
• In the air-to-close valve, as the air pressure increases, the plug moves downward and
restricts the flow of fluid through the valve.
• In the air-to-open valve, the valve opens and allows greater flow as the valve-top air
pressure increases.
• The choice between air-to-open and air-to-close is usually made based on safety
considerations.
• If the instrument air pressure fails, we would like the valve to fail in a safe position for
the process.
• For example, if the control valve were on the cooling water inlet to a cooling jacket for
an exothermic chemical reactor, we would want the valve to fail open so that we do
not lose cooling water flow to the reactor. In such a situation, we would choose an air-
to-close valve.
299
Control Valve Mechanism
• Valve motors are often constructed so that the valve stem position is proportional to
the valve-top pressure.
• Most commercial valves move from fully open to fully closed as the valve-top
pressure changes from 3 to 15 psig.
• In general, the flow rate of fluid through the valve depends upon the upstream and
downstream fluid pressures and the size of the opening through the valve. The plug
and seat (or orifice) can be shaped so that various relationships between stem
position and size of opening (hence, flow rate) are obtained.
• In our example, we assume for simplicity that at steady state the flow (for fixed
upstream and downstream fluid pressures) is proportional to the valve-top pneumatic
pressure.
300
Control Valve Mechanism
• Valve motors are often constructed so that the valve stem position is proportional to
the valve-top pressure.
• Most commercial valves move from fully open to fully closed as the valve-top
pressure changes from 3 to 15 psig.
• In general, the flow rate of fluid through the valve depends upon the upstream and
downstream fluid pressures and the size of the opening through the valve. The plug
and seat (or orifice) can be shaped so that various relationships between stem
position and size of opening (hence, flow rate) are obtained.
• In our example, we assume for simplicity that at steady state the flow (for fixed
upstream and downstream fluid pressures) is proportional to the valve-top pneumatic
pressure.
• A valve having this relation is called a linear valve.
301
Control Valve Numerical Example
Example
302
Pneumatic Control Valve
303
Pneumatic Control Valve
304
305
306
307
308
309
310
311
312
313
314
315
In hydraulics the pressure port is designated P and the return port R or T (for tank). The
two other ports are designated A and B.
In hydraulics the pressure port is designated P and the return port R or T (for tank). The
two other ports are designated A and B.
In hydraulics the pressure port is designated P
and the return port R or T (for tank). The two
other ports are designated A and B.
In hydraulics the pressure port is designated P
and the return port R or T (for tank). The two
other ports are designated A and B.
WHAT IS A SPOOL VALVE?
• A Spool valve can be found in almost every industry you can imagine performing a
wide range of tasks, but what is a spool valve and how does it work?
• Spool valves can be used in both hydraulics (where the oil is the energy source) or
Pneumatics (where the air is the energy source) and their job is to control the flow
direction of the energy source by combining or switching the paths through which the
oil or air can travel.
In hydraulics the pressure port
is designated P and the return
port R or T (for tank). The two
other ports are designated A
and B.
What is a Spool Valve Exactly?
• Essentially a spool valve is a cylinder inside a sealed outer case.
• There are several chambers drilled through the case from one side to the other,
these are commonly known as “ports”.
The function of the spool is to move within the sealed case and provide the function
of either blocking or opening these ports depending on the position of the spool.
What are the Spool Valve Actuators?
The spool can be moved in many different ways;
– Manually using a button or a lever
– As part of a larger control system using a solenoid actuator
• Whatever the method is chosen to control the valve, all that we are doing is
simply pushing the spool to move within the housing, and by doing this we are
allowing or blocking the path between the ports.
• In some cases, there may be a solenoid at each end of the valve, whereas in others
there is a spring return so when the solenoid is no longer pushing the actuator, the
spring will return the spool to its “normal” or rest position.
How does a Spool Valve Work?
Imagine the example of extending a pneumatic cylinder using a spool valve.
• The Pneumatic air supply is
connected to the inlet port
1.
• Port 2 is connected to
pneumatic cylinder extend
connection.
• Port 3 is unconnected as it
is the exhaust port.
• In the normal state without the actuator being operated nothing will happen as the
spool is physically blocking the airflow at port 1 and the cylinder is retraced because
ports 2 and 3 are connected due to the spool position.
• So any air already within the system would escape to the atmosphere through the
exhaust port. Let’s consider this the “normal” or “rest position”.
Now imagine we operate
the actuator.
• We would see the spool position changes and the cylinder extends.
• This is because the spool has moved position unblocking port 1 and simultaneously
blocking port 3 resulting in allowing air flow from port 1 to port 2 while blocking the
exhaust.
• This is known as our “working position”.
• The cylinder will remain extended until the spool is moved back to the normal
position, either by way of another actuator at the opposite end of the valve or by an
internal spring return.
In our example, this type of valve would be known as a 3/2 (three by two)
valve because we have 3 ports and 2 spool positions.
What are the different Types of Spool Valves?
• Spool valves come in a wide variety of different types and configurations, some
having more ports and able to control multiple items of equipment at once.
• Imagine how a 4/2 (four by two) or even a 5/3 (five by three) spool valve might be
used.
• It is also entirely possible to have a standard outer casing with the ports in the same
configuration and fit a spool inside of this which have different profiles.
The blocking parts of the spool are known as “lands”, which are raised above the spool
core; if these are made different in configurations they can change the way the port
pathways interconnect.
STABILTY ANALYSIS
335
CONCEPTS OF STABILITY: A linear time-invariant system is said to be stable, if it
satisfy the following two criteria
i) When the system is excited by a bounded input, the output is bounded (BIBO)
ii) In the absence of the input, the output tends towards zero irrespective of the initial
conditions. This stability is called asymptotic stability
Bounded-Input Bounded-Output (BIBO) Stability
Let r(t), c(t) and g(t) be the input, output and impulse response of a linear time-
invariant system. Therefore C(s) = R(s) ∙ G(s)
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
(a) Using τ1= 1, τ2= ½ and τ3= 1/3,
determine the values of Kc for which the
control system in Fig.
(b) For the value of Kc for which the system
is on the threshold of instability,
determine the roots of the characteristic
equation.
372
373
Root Locus
Contents
Introduction of Root Locus
Properties of Root Locus
Root Locus Sketching Rules
Root Locus Based System Analysis and
Design
375
Introduction of Root Locus
What we have known?
• We have had a quick look at modeling and analysis for
specifying system performance.
(Time domain method for first and second order systems)
• We will now look at a graphical approach, known
as the root locus method, for analyzing and designing
control systems
• As we will see, the root locations are important in
determining the nature of the system response
376
Proportional controller
R (s ) C (s )
K G (s )
-
H (s )
We have seen how this form of feedback is able to minimize the
effect of disturbances.
For investigating the performance of a system, we have to solve the
output response of the system.
Limitations:
(1) It is difficult to solve the output response of a system –
Root Locus
especially for a higher order system.
(2) We can’t easily investigate the changes of a system’s
performance from the time-domain
--especially when parameters of the system vary in a given range,
or, when devices are added to the system.
377
History:
Root locus method was conceived by Evans in 1948.
Root locus method and Frequency Response method,
which was conceived by Nyquist in 1938 and Bode in 1945,
make up of the cores of the classical control theory for
designing and analyzing control systems.
By using root locus method，we can：
analysis the performance of the system
determine the structure and parameter of the system
design the compensator for control system
378
[Definition]：The path traced by the roots of the
characteristic equation of the closed-loop system as
the gain K is varied from 0 to +∞ is called root locus.
Example：The second order systemR (s ) K C (s )
with the open-loop transfer - s(0.5s 1)
function
K
Gk ( s)
s(0.5s 1)
2K
Closed-loop Transfer function： (s) 2
s 2s 2 K
Characteristic equation: s 2 2s 2 K 0
Roots of characteristic equation： s1, 2 1 1 2K
379
Roots： s1, 2 1 1 2K
[Discussion]：
K 5
① For K=0，s1=0 and s2=-2, K 1
K 0
j1
are the poles of closed-loop system. 1 K 0
② For K=0.32，s1=-0.4,s2=-1.6
2 0
j1
③ For K=0.5，s1=-1,s2=-1
④ For K=1，s1=-1+j,s2=-1-j
⑤ For K=5，s1=-1+3j,s2=-1-3j
⑥ For K=∞，s1=-1+∞j,s2=-1-∞j
[Terminology]：In the root locus， “ ”denotes the poles of the
open-loop transfer function， “ ” zeros of the open-loop
transfer function. The bode line represents the root locus，arrow
shows the root locus direction along some parameters.
380
Properties of Root Locus
Control System： Closed-loop transfer ( s ) G( s)
function: 1 G( s) H ( s)
R (s ) C (s )
G (s )
- Open-loop transfer
Gk ( s) G( s) H ( s)
function:
H (s )
m
(s z )
i
Canonical form of Gk (s) Gk ( s) K g i 1
n
(s p )
j 1
j
K g root locus gain
zi and p j are zeros and poles of open - loop transfer function
381
5.2.1 Two canonical form of open-loop
transfer function
m
K ( s 1)
Gk ( s ) G ( s ) H ( s )
nm （4-1）
i 1
i
(T s 1)
j 1
j
m Time constant
K g ( s zi ) m canonical form: K is
n
i 1 z i open-loop gain
K Kg i 1
(s p ) n
（4-2）
j
j 1 pj
j 1
Zero-pole canonical form: Kg is
root locus gain
In root locus method, we use (4-2）.
382
Magnitude and Phase Equations
R(s) + E(S) C(s)
G(s)
-
H(s)
Characteristic equation of closed-loop transfer function
1 G(s) H s 0 G ( s ) H s 1 （4-3）
The objective is to determine values of s that
satisfy the equation. Such s are the on the root
locus.
383
Definition
m
(s z )i
Gk ( s) 1 or k g i 1
n
1 is the root-locus equation.
(s p )
j 1
j
Remembering that s may be complex, it root-locus
equation can be equivalently written into
m
| (s z ) |
i
Magnitude equation
Kg i 1
n
1
| (s p ) |
j 1
j Phase equation
Argument equation
m n
( s zi ) ( s p j ) (2k 1) , k 0,1,2...
i 1 j 1
384
Root locus method is performed in two stages:
1.Finding all values of s satisfying the argument equation
2.Finding particular values of s that satisfy the magnitude
equation
Phase equation is the sufficient and necessary condition
for root locus.
—— All the values of s satisfying the argument equation
constitute the root locus.
——Magnitude equation is used to determine the gain Kg
for a particular root.
——Argument equation is independent on Kg.
385
Trial guess method：
Example 5.1 k
Gk ( s)
s(2s 1)
Solution：
1）Find zeros and poles of open-loop transfer function on
s plane 0.5k K
Gk (s) K g 0.5k
g
,
s(s 0.5) s(s 0.5)
jω
Poles:
-p2 -p1 p1 0, p2 0.5
×
-0.5
×
0 σ
There are no zeros.
386
2）Inspecting the real axis：
s1>0 (s1 p1 ) (s1 p2 ) 0 Argument equation
is not satisfied.
-< s1 <-p2 (s p ) (s p ) 180
1 1 1 2 jω
-p2 S1 -p1
×
-0.5
×
0 σ
-p2< s1 <-p1 (s1 p1 ) (s1 p2 ) 180 0 180
The all values between p1 ~ p2 of the real axis are
the points of root locus
For s1=-0.1 K g s1 p1 s1 p2 0.1 0.4 0.04
387
3）Examine locations off the real axis: arbitrarily
choose an s1 S1
• jω
In order to satisfy
(s1 p1 ) (s1 p2 ) 180
-p2 -p1
s1 should be on the ×
-0.5 -0.25
×
0 σ
perpendicular bisector of the
line between the two poles.
Let s1=-0.25 + j0.25
(s1 p1 ) (s1 p2 ) 135 0 45 0 180 0
2
K g s1 p1 s1 p2 0.25 2 0.125
388
Example: For the second order system，draw the root
locus with respect to Kg from 0→∞
kg
Solution: ( s) 2
s s k g R (s ) kg C (s )
Characteristic equation of closed-loop transfer - s ( s 1)
function： 1 1
s 2 s k g 0，s1, 2 1 4k g
2 2
[Discussion]：
① For k g 0，we have s1, 2 0 and -1.
k g ，s1 goes to the leftalong the negative realaxis，
②
and s2 goes to the right along negative real axis from 1.
389
R (s ) kg C (s )
- s ( s 1)
1 1
③ For k g ，s1, 2 .
4 1 2
For any 0 k g ，s1, 2 are on the negative real axis.
4
1
④ For k g ，s1, 2 are complex numbers .
4
As Kg increases, location of roots branch out along the
vertical line.
⑤ k g 时，s1, 2 1 j
2
390
j1
kg
s 1 s
[Summary]
A A'
When Kg changes from zero to infinity,
there are two segments starting from the A1
kg 0
open-loop poles to infinity. A2
For higher order system, it is very
B 1 0.5 0 k g 0
difficult to sketch all the root loci. But
graphical method shows its advantage. kg
Are Points A and A’ on the root locus？(Trial guess j1
method)
kg
s (s 1) OA BA A1 A2 (2k 1)
s(s 1)
Obviously，A is on the root locus, but A’ is not on the root locus.
391
Example 5-2: The open-loop transfer function is given by
k ( s 4)
G( s) H ( s)
s ( s 2)( s 6.6)
Check if s1= -1.5+j2.5 is on the root locus. If yes, please
determine the parameter gain k
1)Via argument equation
( s1 z1 ) ( s1 p1 )
( s1 p2 ) ( s1 p3 )
S1 45 - 120 - 79 - 26 180
k=12.15 2)Via magnitude equation
s1 p1 s1 p2 s1 p3
k s1 z1
26O 45O 79O 120O 2.9 2.6 5.8
-P3 -Z1 -P2 -P1 12 .15
3.6
392
We can rapidly sketch the root-loci of a control system.
Sketch the root-loci for the following open-loop transfer functions:
K ( s 2)( s 6) K ( s 2)( s 4) K ( s 3)
(1) G ( s) (2) G ( s) (3) G ( s)
s( s 4) s( s 6) s( s 2)
K (s 2)(s 3) K ( s 2 6s 13) K ( s 1)( s 2)
(4) G(s) (5) G( s) (6) G ( s)
s(s 1) s( s 2) ( s 2 6s 13)
Im Im Im
Re Re Re
Im Im Im
Re Re Re
393
Summary
• The preceding angle and magnitude criteria can be used to
verify which points in the s-plane form part of the root locus
• It is not practical to evaluate all points in the s-plane to find
the root locus
• We can formulate a number of rules that allow us to sketch
the root locus
394
Model development using Data set
Linear Regression Method
Bungee-jumping
where v = downward vertical velocity (m/s), t = time (s), g = the
acceleration due to gravity (= 9.81 m/s2), cd = a lumped drag coefficient
(kg/m), and m = the jumper’s mass (kg).
In this table, force is the dependent variable (y)
and velocity is the independent variable (x).
395
Linear Regression Method
The mathematical expression for the straight line is
Criteria for a “Best” Fit
where n = total number of points. 396
Linear Regression Method
• However, this is an inadequate criterion, as illustrated
by Fig., which depicts the fit of a st. line to two points.
• Obviously, the best fit is line connecting points.
• However, any straight line passing through midpoint of
connecting line (except a perfectly vertical line) results
in a minimum value of eqn. equal to zero because
positive and negative errors cancel.
• One way to remove effect of signs might be to
minimize sum of absolute values of discrepancies, as
in
397
Linear Regression Method
• Fig. b demonstrates why this criterion is also
inadequate.
• For four points shown, any straight line falling within
dashed lines will minimize sum of absolute values of
residuals.
• Thus, this criterion also does not yield a unique best
fit.
• A third strategy for fitting a best line is minimax criterion.
In this technique, line is chosen that minimizes the
maximum distance that an individual point falls from
line.
• From Fig. c, this strategy is ill-suited for regression
because it gives undue influence to an outlier, i.e., a
single point with a large error.
• It should be noted that minimax principle is sometimes
well-suited for fitting a simple function to a complicated
function
398
Linear Regression Method
This criterion, is called least squares, has a number of advantages, including that it
yields a unique line for a given set of data.
Least-Squares Fit of a Straight Line
399
Linear Regression Method
Least-Squares Fit of a Straight Line
This result can then be used in conjunction with Eq. to solve for
400
Linear Regression Method
The slope and the intercept can then be calculated
401
Practice Problem:
402
ANALYSIS AND DESIGN OF
ADVANCED CONTROL SYSTEMS
CASCADE CONTROL
INTRODUCTION: ANALYSIS AND DESIGN OF ADVANCED CONTROL
SYSTEMS
• Feedback control is most used control systems in Chemical Process
but, it is not only one.
• A disadvantage of conventional feedback control is that corrective
action for disturbances does not begin until after the controlled
variable deviates from the set point.
• There exists situations where feedback control action is insufficient to
produce the desired response of a given process.
• In such cases, other control configurations may be used including
Feedforward
Cascade
Ratio
Override
Split Range
Multi-variable
Adaptive and Inferential Control
CONTROL SYSTEMS WITH MULTIPLE LOOPS
• Feedback control configurations involves one measurement (O/P) and
one manipulated variable in a single loop.
• However, other control configurations may be:
More than one measurement and one manipulation,
One measurement and more than one manipulated variables
• In such cases control systems with multiple loops may arise.
• Typical examples of such configurations:
Cascade Control
Various types of Selective Control
Split-range control
CASCADE CONTROL SYSTEM
CASCADE CONTROL SYSTEM
ANALYSIS OF CASCADE CONTROL
• Cascade control is an alternative approach that can significantly improve the
dynamic response to disturbances by employing a secondary measurement and a
secondary feedback controller.
• The secondary measurement point is located so that it recognizes the upset
condition sooner than the controlled variable, but the disturbance is not necessarily
measured.
• The outer loop and its controller are called master loop and master controller
whereas the inner loop and its controller are called slave loop and slave controller
respectively.
ANALYSIS OF CASCADE CONTROL
ANALYSIS OF CASCADE CONTROL
Now see what happens if the secondary loop is absent. In that case:
ANALYSIS OF CASCADE CONTROL
ANALYSIS OF CASCADE CONTROL
Response time of a cascade Controller
First analyze the secondary loop:
In absence of the secondary loop and the slave controller
In presence of the secondary loop and the slave controller (say a pure gain controller)
Clearly there is a decrease in time constant in presence of slave loop. This will guarantee
that the time constant of the overall process decreases when a secondary loop is
introduced to the system. Eventually that decreases the overall response time.
ANALYSIS OF CASCADE CONTROL
Stability of a Cascade Controller:
let us take a simple example:
In absence of the secondary loop the characteristic equation of the overall process is:
In presence of the secondary loop the characteristic equation of the overall process is:
Now apply Routh-Hurwitz criteria to find the limits of both these cases:
First case: (In absence of secondary loop)
ANALYSIS OF CASCADE CONTROL
For a stable closed loop process,
Second case: (In presence of secondary loop)
ANALYSIS OF CASCADE CONTROL
ANALYSIS OF CASCADE CONTROL
Few points to remember on Cascade controller:
• The slave loop should be tuned before the master loop. After the slave loop is tuned
and closed, the master loop should be designed based on the dynamics of inner loop.
• There is little or no advantage to using cascade control if the secondary process is
not significantly faster than the primary process dynamics. In particular, if there is long
dead time in the secondary process, it is unlikely that the cascade controller will be
better than the standard feedback control.
• The most common cascade control loop involves flow controller (eg. TC/FC example
in distillation column) as the inner loop. This loop easily rejects the disturbances in
fluid steam pressure, either upstream or downstream of the valve.
FEEDFORWRD CONTROL
FEEDFORWRD CONTROL
Advantages of Feedback control
1. Corrective action occurs as soon as the controlled variable deviates from the set point,
regardless of the source and type of disturbance.
2. Feedback control requires minimal knowledge about the process to be controlled; in particular,
a mathematical model of the process is not required, although it can be very useful for control
system design.
3. The ubiquitous PID controller is both versatile and robust. If process conditions change, re-
tuning the controller usually produces satisfactory control.
Certain Inherent Disadvantages of Feedback Control:
1. No corrective action is taken until after a deviation in the controlled variable occurs. Thus,
perfect control, where the controlled variable does not deviate from the set point during
disturbance or set-point changes, is theoretically impossible.
2. It does not provide predictive control action to compensate for the effects of known or
measurable disturbances.
3. It may not be satisfactory for processes with large time constants and/or long time delays. If
large and frequent disturbances occur, the process may operate continuously in a transient
state and never attain the desired steady state.
4. In some situations, the controlled variable cannot be measured on-line, so feedback control is
not feasible.
FEEDFORWRD CONTROL
For situations in which feedback control by itself is not satisfactory,
significant improvement can be achieved by adding Feed-Forward
Control.
For situations in which feedback control by itself is not satisfactory, significant
improvement can be achieved by adding feed-forward control. But feed-forward control
requires that the disturbances be measured (or estimated) on-line.
The basic concept of feed-forward control is to measure important disturbance variables
and take corrective action before they upset the process.
FEEDFORWRD CONTROL
• The basic concept of feed-forward control is to measure important disturbance
variables and take corrective action before they upset the process.
• In contrast, a feedback controller does not take corrective action until after the
disturbance has upset the process and generated a nonzero error signal.
Simplified block diagrams for feed-forward
and feedback control.
FEEDFORWRD CONTROL
Feed-forward control has several disadvantages:
1. The disturbance variables must be measured online. In many applications, this is not
feasible.
2. To make effective use of feed-forward control, at least an approximate process model
should be available. In particular, we need to know how the controlled variable
responds to changes in both the disturbance variable and the manipulated variable.
The quality of feed-forward control depends on the accuracy of the process model.
3. Ideal feed-forward controllers that are theoretically capable of achieving perfect
control may not be physically realizable. Fortunately, practical approximations of
these ideal controllers can provide very effective control.
FEEDFORWRD CONTROL
Fig1: Feedback control of the liquid level in a boiler drum.
• A boiler drum with a conventional feedback control system is shown in Fig. 1. The level of the
boiling liquid is measured and used to adjust the feed water flow rate. This control system tends
to be quite sensitive to rapid changes in the disturbance variable, steam flow rate, as a result of
the small liquid capacity of the boiler drum. Rapid disturbance changes are produced by steam
demands made by downstream processing units.
• Another difficulty is that large controller gains cannot be used because level measurements
exhibit rapid fluctuations for boiling liquids. Thus a high controller gain would tend to amplify the
measurement noise and produce unacceptable variations in the feed water flow rate.
FEEDFORWRD CONTROL
Fig 2: Feedforward control of the liquid level in a boiler drum.
• The feedforward control scheme in Fig. 2 can provide better control of the liquid level. The
steam flow rate is measured, and the feedforward controller adjusts the feedwater flow rate so
as to balance the steam demand. Note that the controlled variable, liquid level, is not measured.
As an alternative, steam pressure could be measured instead of steam flow rate.
FEEDFORWRD CONTROL
• Feedforward control can also be used advantageously for level control problems
where the objective is surge control (or averaging control), rather than tight level
control.
• For example, the input streams to a surge tank will be intermittent if they are effluent
streams from batch operations, but the tank exit stream can be continuous.
Feedforward-feedback control of the boiler
drum level.
• In practical applications, feedforward control is normally used in combination with
feedback control. Feedforward control is used to reduce the effects of measurable
disturbances, while feedback trim compensates for inaccuracies in the process
model, measurement errors, and unmeasured disturbances.
FEEDFORWRD CONTROL
The closed-loop transfer function
Ideally, we would like the control system to produce perfect control, where the
controlled variable remains exactly at the set point despite arbitrary changes in the
disturbance variable, D.
FEEDFORWRD CONTROL
Thus, if the set point is constant (Ysp(s) = 0), we want Y(s) = 0, even though D(s) ≠ 0.
This condition can be satisfied by setting the numerator equal to zero and solving for Gf
RATIO CONTROL
RATIO CONTROL
• Ratio control is a special type of feedforward control where two disturbances (Loads)
are measured and held in a constant ratio to each other.
• Hence, its objective is to maintain the ratio of two process variables at a specified
value.
• Its mostly used to control the ratio of flow rates of two streams. Both flow rates are
measured but only one can be controlled.
• The two variables are usually flow rates, a manipulated variable u and a disturbance
variable d.
• Thus, the ratio is controlled rather than the individual variables.
• Where, u and d are physical variables, not deviation variables.
• The stream whose flow rates is not under control is usually referred to as WILD
STREAM.
• https://www.youtube.com/watch?v=MCf83Ipd-us
RATIO CONTROL
• Typical applications of ratio control include
specifying the relative amounts of components in blending operations,
maintaining a stoichiometric ratio of reactants to a reactor,
keeping a specified reflux ratio for a distillation column, and
holding the fuel-air ratio to a furnace at the optimum value.
RATIO CONTROL
Ratio control, Method I
Source: Process Dynamics and Control, 3rd Edition Dale E. Seborg, University of California, Santa Barbara; Thomas F. Edgar,
University of Texas at Austin; Duncan A. Mellichamp, University of California, Santa Barbara; Francis J. Doyle III, University of
California, Santa Barbara, Wiley Publication
RATIO CONTROL
Ratio control, Method I
• The flow rates for both the disturbance stream
and the manipulated stream are measured, and
the measured ratio, is calculated.
• The output of the divider element is sent to a
Ratio Controller (RC) that compares the
calculated ratio Rm to the desired ratio Rd and
adjusts the manipulated flow rate u
accordingly.
• The ratio controller is typically a PI controller
with the desired ratio as its set point.
Source: Process Dynamics and Control, 3rd Edition Dale E. Seborg, University of California, Santa Barbara; Thomas F. Edgar,
University of Texas at Austin; Duncan A. Mellichamp, University of California, Santa Barbara; Francis J. Doyle III, University of
California, Santa Barbara, Wiley Publication
RATIO CONTROL
Ratio control, Method I
• In this configuration, both
the flow rates are measured
and ratio has been taken.
• This ratio is compared to the
desired ratio (Set Point) and
the deviation (Error)
between the measured and
desired ratios constitutes the
actuating signals for the
ratio controller.
Source: Chemical Process Control An Introduction to Theory and Practice by George Stephanopoulos, P T R Prentice Hall,
Englewood Cliffs, New Jersey.
RATIO CONTROL
Ratio control, Method I
• The main advantage of Method I is that the measured ratio Rm is calculated.
• A key disadvantage is that a divider element must be included in the loop, and this
element makes the process gain vary in a nonlinear fashion.
• As we know that the ratio
• Hence, the process gain
Inversely related to the disturbance flow rate d.
Because of this significant disadvantage, the preferred scheme for implementing ratio control is
Method II,
RATIO CONTROL
Ratio control, Method II
Source: Process Dynamics and Control, 3rd Edition
Dale E. Seborg, University of California, Santa Source: Chemical Process Control An Introduction
Barbara; Thomas F. Edgar, University of Texas at to Theory and Practice by George Stephanopoulos,
Austin; Duncan A. Mellichamp, University of P T R Prentice Hall, Englewood Cliffs, New Jersey.
California, Santa Barbara; Francis J. Doyle III,
University of California, Santa Barbara, Wiley
Publication
RATIO CONTROL
Ratio control, Method II
• In this Method II, the flow rate of the
disturbance/wild stream A is measured and
multiply it by the desired ratio.
• The result is the flow rate that the stream B
should have and constitutes the set point value
which is compared to the measured flow rate of
stream B.
• The deviation constitutes the actuating signal for
the controller, which adjusts appropriately the
flow of B.
RATIO CONTROL
Ratio control, Method II
• In Method II the flow rate of the disturbance
stream is measured and transmitted to the ratio
station (RS), which multiplies this signal by an
adjustable gain, KR, whose value is the desired
ratio.
• The output signal from the ratio station is then
used as the set point Usp for the flow controller,
which adjusts the flow rate of the manipulated
stream, u.
• Note that disturbance variable d is measured
in both Methods I and II. Thus, ratio control
is, in essence, a simple type of feedforward
control.
RATIO CONTROL
Ratio control, Method II
• Ratio Control is used extensively in chemical process with the following most commonly
encountered examples:
Keep a constant ratio between the feed flow rate and the stream in the reboiler of a
distillation column.
Hold constant the reflux ratio in a distillation column.
Control the ratio o two streams entering a reactor at a desired value.
Hold the ratio of two blended streams constant, in order to maintain the composition
of the blend at the desired value.
Hold the ratio of a purge stream to the recycle streams constant.
Keep the ratio of fuel/air in a burner at its optimum value (most efficient
combustion)
RATIO CONTROL
Ratio control, Method II
• Regardless of how ratio control is implemented, the process variables must be scaled
appropriately.
• For example, in Method II the gain setting for the ratio station Ka must take into account the
spans of the two flow transmitters. Thus, the correct gain for the ratio station is
• where Ra is the desired ratio, and Su and Sa are the spans of the flow transmitters for the
manipulated and disturbance streams, respectively.
RATIO CONTROL
RATIO CONTROL
RATIO CONTROL
449
450
Thank
You!
Mail at;
[email protected]
YouTube channel:
https://www.youtube.co
m/feed/my_videos?disa
ble_polymer=true%27
You might also like
- Advanced Temperature Measurement and Control, Second EditionFrom EverandAdvanced Temperature Measurement and Control, Second EditionNo ratings yet
- Tce5002 Introduction To PDC, Modelling Simulation & ControlNo ratings yetTce5002 Introduction To PDC, Modelling Simulation & Control15 pages
- Process Dynamics & Control: Muhammad Rashed JavedNo ratings yetProcess Dynamics & Control: Muhammad Rashed Javed33 pages
- Introduction To Process Dynamics and ControlNo ratings yetIntroduction To Process Dynamics and Control45 pages
- GATE Chemical Instrumentation and Process Control-IPCNo ratings yetGATE Chemical Instrumentation and Process Control-IPC29 pages
- Process Dynamics & Control: Muhammad Rashed JavedNo ratings yetProcess Dynamics & Control: Muhammad Rashed Javed25 pages
- Instrumentation and Process Control: Dr. Iftikhar A. SalarzaiNo ratings yetInstrumentation and Process Control: Dr. Iftikhar A. Salarzai81 pages
- Process Dynamics & Control: BSC, Chemical Engineering 8 SemesterNo ratings yetProcess Dynamics & Control: BSC, Chemical Engineering 8 Semester72 pages
- Process Dynamics & Control: Muhammad Rashed JavedNo ratings yetProcess Dynamics & Control: Muhammad Rashed Javed32 pages
- Advanced Process Control: Instructor: Jayant M Modak Course: CH 246No ratings yetAdvanced Process Control: Instructor: Jayant M Modak Course: CH 24624 pages
- Process Dynamics and Control: Fourth YearNo ratings yetProcess Dynamics and Control: Fourth Year14 pages
- Chapter 1-Revision of Feedback Control System-Part 1 PDFNo ratings yetChapter 1-Revision of Feedback Control System-Part 1 PDF141 pages
- Modeling The Dynamic and Static Behavior of Chemical ProcessesNo ratings yetModeling The Dynamic and Static Behavior of Chemical Processes4 pages
- Block Diagram of A Control System: 11 August 2021 12:48No ratings yetBlock Diagram of A Control System: 11 August 2021 12:4868 pages
- Fundamental of Process Control - Chapter 3 - OriginalNo ratings yetFundamental of Process Control - Chapter 3 - Original88 pages
- ENCH 442: Chemical Process Control and SafetyNo ratings yetENCH 442: Chemical Process Control and Safety16 pages
- 3014all - 7th Semester - Mineral EngineeringNo ratings yet3014all - 7th Semester - Mineral Engineering62 pages
- Introduction To Process Instrumentation & ControlNo ratings yetIntroduction To Process Instrumentation & Control44 pages
- Introduction To Process Dynamics and ControlNo ratings yetIntroduction To Process Dynamics and Control44 pages
- Class Rules: CENG 4120: Process Dynamics and ControlNo ratings yetClass Rules: CENG 4120: Process Dynamics and Control14 pages
- CHE412 Process Dynamics Control BSC Engg 7th SemesterNo ratings yetCHE412 Process Dynamics Control BSC Engg 7th Semester17 pages
- MD - Tanweer Khan Vector Solved Problem 9830359247: Excellence Is Created HereNo ratings yetMD - Tanweer Khan Vector Solved Problem 9830359247: Excellence Is Created Here18 pages
- (FREE PDF Sample) Structural Analysis Third Edition Coates EbooksNo ratings yet(FREE PDF Sample) Structural Analysis Third Edition Coates Ebooks52 pages
- Design of Small Bulb Turbines With Unequal Specific Work Distribution of The Runner'S Elementary StagesNo ratings yetDesign of Small Bulb Turbines With Unequal Specific Work Distribution of The Runner'S Elementary Stages13 pages
- Supply Network Coordination 03 Nim 01192162No ratings yetSupply Network Coordination 03 Nim 0119216212 pages
- NEET 2 (09 June 2024) (EH) Paper _croppedNo ratings yetNEET 2 (09 June 2024) (EH) Paper _cropped46 pages
- Machines and Mechanisms Applied Kinematic Analysis 4th Edition by David Myszka 0132157802 9780132157803 - Get the ebook instantly with just one clickNo ratings yetMachines and Mechanisms Applied Kinematic Analysis 4th Edition by David Myszka 0132157802 9780132157803 - Get the ebook instantly with just one click49 pages
- Preboard - Class 12, Maths Indu 22-23, ANo ratings yetPreboard - Class 12, Maths Indu 22-23, A6 pages
- MODULE 1_Introduction to Coordinate SystemsNo ratings yetMODULE 1_Introduction to Coordinate Systems77 pages
