Process Instrumentation 3A (PINA301)

TOPIC 2 NOTES - CONTROL THEORY

N.B Prescribed/Recommended Books
Instrument Engineers Handbook (Vol. 1): Process Measurement & Analysis
Instrument Engineers Handbook (Vol. 2): Process Control and Optimization
Process Industrial Control Handbook
PINA NOTES
PAST EXAM PAPERS

Mthabiseni Mbokazi
 CONTROL THEORY
1.1 CONTROL BASICS (Section 2.1 Vol.2 Page 167)

Final Control Element - Valve

Condensate Water

▪ Cool water come in the bottom of the water heater tank.

▪ The aim of this is to heat up the water.
▪ The operator is controlling the steam valve.
▪ The steam enters the water heater, the water is not in contact with the steam.
▪ Once the steam enters the heat exchanger or a water heater it will condense, turn into normal water out of the
system.
▪ The operator looks at the temperature indicator.
▪ If the cool water is too cold, the temperature indicator will show too low temperature.
▪ The operator manually opens the valve.
▪ If the temperature exceeds the set point or the desired temperature, the operator will manually close the valve.
➢ The valve is manual closed/opened to control the temperature of the water.
➢ Controlled Variable (C.V) – Temperature of the water
➢ Manipulated Variable (M.V) – Flow rate of the steam.
➢ Final Control Element (F.C.E) is the steam valve.

Basic Feedback Control

u – Controlled
d Variable
𝑒 =𝑟−𝑢
r - Error
 ➢ Controller resides in a microcontroller, or a microprocessor.
➢ It looks at all the inputs coming in.
➢ Input comes from a sensor.
➢ The plant is lumped transfer function.
▪ The Set Point, r, comes in, the error is e.
▪ If there is no error, it means the process output is running in accordance with the set point.
▪ The controller will keep the output close to the Set point.
▪ If the error increases the controller will increase the value of u, by opening the valve to change the system
dynamics.
▪ If it is a negative error (output much higher than the set point) the controller will cut back on the value of u.
▪ The signal u will go to the final control element.
▪ The output will be the process output, which is the controlled value c like temperature.
▪ Then it is fed back, negative feedback because an error is generated from it.
▪ This is reactive system; it is reacting to process changes.
Feedback page 12

Simple Tank Level Control

S.P - 80% Liquid Level

▪ P & ID - loop diagram of the plant or process Piping & Instrument Diagram.
▪ Level Transmitter gives feedback to the controller, LC - Level Controller.
▪ PID - Proportional Integral Derivative it is a type of control algorithm which runs inside the process control system
and adjust the output.
▪ The controller reads a set point from the output which is 80%.
▪ The controller sends a signal to the valve Final Control Element- Valve A
▪ If the liquid is above 80%, the valve will cut back or close.
▪ If the level is too low below the set point Valve A will open fully to maintain level.
1.2 PROCESS DYNAMICS
Resistance In Pipelines
▪ Any pipe will have resistance.
▪ The transfer function for particular pipeline will be given by
Input(m), Resistance R, and Output (c).
▪ The output is given as c = Rm, c is the output which is the head-
Head is given in terms of pressure.
▪ A certain pressure or flow of liquid is put in a pipeline, there will
be a pressure drop because of the internal resistances of the
pipe.
▪ The new head will be different.
▪ If the flow rate is increased, the output variable c will increase
to certain point depending on what R & m is.

➢ Flow is proportional to the pressure drop.

➢ Described by the steady-state Gain where Resistance (R)
➢ System input (m = flow) changes from zero to m
➢ Output (c = head) will go into an instantaneous state from zero to c
Capacitance In a Tank
▪ The amount of flow rate or flow of the liquid coming in is linked to the level (h).
▪ This will be affected by flow rate, diameter of the pipe, and the diameter of the tank and also the resistance of the
output.
▪ If the flow rate of the liquid coming in is greater than the flow rate of the liquid going out there will be an overflow.
System Gain and Time Constant

▪ The liquid is sent out through a series of pipes with resistance.

▪ If there is no liquid on the tank and the flow rate is increased, opening the valve fully from 0% to 100%.
▪ The liquid will flow into the tank, the head will be increased.
▪ Combines capacitance type process element (tank) with a resistance type process component (valve)
▪ If the tank was initially empty, then the inflow was going to start at constant rate of m, the level on the tank would
rise and would eventually rise to the steady state height of [ c = Rm] in the tank.

Reaction Curve – System gain and the time constant.

▪ Indicate the level in the tank, if the level increases the pressure increases.
▪ It is shown by the solid line, then tap out because it will reach equilibrium.
▪ Because of the resistances of the of pipe and internal capacitances of the physical tank the liquid will rise at a
certain rate this is given in terms of Time Constant.
▪ Time Constant is derived from the first point that the first time constant in terms of time will be one minute.
▪ The Time Constant of any process can be calculated by working out 0.632 of the steady state value or 63.2% of
the final steady state value.
Typical 1st Order Response

Transportation Lag (Dead Time)

▪ Steam will come in; TC will regulate the steam valve.

▪ Condensate comes through the bottom.
▪ TC will receive temperature input from the sensor TL.
▪ Any changes in the steam will be picked up by TL.
▪ For the temperature to be monitored at TZ
▪ Because the pipeline is long with distance, any changes to the steam valve the effects will be felt after a period
of time this is called Transportation Lag Time.
▪ The changes on the steam valve will now be picked up.
1.3 CLOSED LOOP CONTROL (ANTICIPATORY CONTROL)

A B
➢ Works on feedback
➢ The controller manipulates some variable like the valve.
➢ The controller measures the controlled variable (c), compare the measurement to some desired target value (r –
setpoint)
➢ If there is an error (c is not = r), change its output signal (m) and change the manipulated variable m in a way that
will reduce the deviation from the set point.

EXPECTED SYSTEM RESPONSES (PAGE 176)

1. Unstable Loop Run-Away Oscillation
➢ When a system has a loop gain greater than 1, it gives an Oscillating Unstable Output.

▪ LC has a loop gain or the whole process has a high loop gain.
▪ Any changes in the valve will cause the level of the liquid to increase, then drop making an Oscillation.
▪ Then it will reach a point of Instability because of the overflow.
▪ The dynamic phase lag is a Variable Phase Lag.

2. Stage of Marginal Stability

▪ Has a loop gain of one and dynamic phase lag off -180°
▪ It is a loop gain oscillating around a set point.
▪ It oscillates at fixed phase or period.
➢ Oscillation and Phase Shift When a closed loop is in sustained oscillation, the period and amplitude of oscillation
are constant, and therefore, the total phase shift in the loop must be 360 degrees. This state is called Sustained
Oscillation or Marginal Stability.
3. Stable Loop Tuned for 1/4 Amplitude Damping
▪ Loop gain between 0.5 and 0.8.
▪ It is a Stable System, with the expected response.
▪ There will be changes on the output but will eventually reach the set point.

1.4 CONTROL MODES: PID AND ON/OFF CONTROL (SECTION 2.2)

➢ The controller compares its measurement to its set point (r) and based on the difference between them (e =
error), generates a correction signal to the final control element (e.g control valve).
➢ The controller behaves in different ways which include ON/OFF, Floating, Proportional (P), Integral (I), Differential
(D).

ON/OFF CONTROL
Disadvantage - Oscillate around the set point.
- Has wear and tear.
Advantage – It is very simple to implement.
➢ A perfect ON/OFF Controller is “ON” when the measurement is below the set point. Under such conditions the
manipulated variable is at its maximum value.
➢ When the measured variable is above the set point, the controller is “OFF”, and the manipulated variable is at its
minimum value.
➢ The output is switched ON and OFF
➢ if there is an error greater than 0, and output to the valve will be maximum.
➢ If the error is less than 0, the output to the valve will be minimum.
▪ Assuming there is no liquid, e will be greater than 0 then the valve will fully open.
▪ The level in the tank will increase, after some time e will be less than 0 (reached the set point).
▪ The valve will close, then open after the Differential Gap.
▪ Differential Gap - prevents the valve from switching on and off very quickly, it prevents wear and tear of the valve.
▪ The output switches much faster if the differential gap is absent and this can damage the valve.
▪ If error > 0, output will be ON, if error < 0 the output will be OFF.
▪ Solenoid valve can be used on ON/OFF control.

PID CONTROL
➢ PID control is simple flexible algorithm, and it can be applied in many different loops.
Proportional Control (Page 185-187 vol 2)
➢ Control output is proportional to the error.
➢ The Proportional (P) term in a PID controller is responsible for providing an immediate response to the error
signal, which is the difference between the desired setpoint and the actual process variable.
➢ The Proportional (P) mode alone is the simplest linear control algorithm. It is characterized by a constant
relationship between the controller input and output.
➢ The adjustable parameter of the proportional mode, Kc, is called the Proportional Gain. It is frequently expressed
in terms of percent Proportional Band, PB.

Disadvantages
o There will always be an offset.
o Cannot keep the controlled variable on set point.
Example Float System
Equal Impact
▪ The liquid come into the tank, there is a process load.
▪ Float System gives feedback in terms of level.
▪ The Set Point is h, which is 3.5 inches. Large Impact on the Valve.
▪ The float system is connected to the valve. Kc is large.
▪ It has a Lever System or Fulcrum System.
Small Kc, Small Impact
▪ When the level goes down, the valve will Open.
on the valve
▪ When the level rises, the float will go up causing the valve to Close.
▪ Kc is the controller gain which is how much the valve will open or close.
▪ Proportional Band is the range of the controller output over which final control element moves.
▪ PID requires the valve to move from 0% to 100%, closed is 0% then increases as it opens until 100%.
▪ Kc Is given by the fulcrum position any small change on the float will cause a change on the valve.
▪ Proportional gain can be tuned to suit that process.
Disadvantage – is has an offset or undershoot.

Integral Control (Page 185 vol 2)

➢ The Integral (I) term in a PID controller addresses the cumulative error over time by integrating the error signal.
➢ It is adding integral to the proportional control.
➢ Integral control cannot exist on its own.
➢ It is a predictive control tries to predict the error.
➢ In integral control we find the area under the graph to find the new term or error affecting the head/m.
➢ Integral control integrates the error.
Advantage – Reduces offset.

➢ Integral Wind-Up is a problem, which occurs when the plant is offline or when the error persists longer than
normal.
➢ If the error persists, the integral wind up goes higher and higher.
➢ Integral Anti-Wind-up, limits the output to avoid the wind up, because if the integral action gets too large the
controller will not respond correctly if there is a change on a set point or a new set point when the controller is
offline again.
➢ With anti wind up the output can be limited or clamped there will be positive or negative limit.
Comparison Of Control Mode Responses.

➢ Proportional control has an offset.

➢ Integral control receives a set point quicker and respond to changes much faster.
➢ If it is left uncontrolled it goes to the state of instability.

Derivative Control
➢ The Derivative (D) term in a PID controller anticipates future changes in the error signal by measuring its rate of
change.
➢ It is a derivative of the error.
➢ Derivative Of the Error is the rate of change of the error.
Advantage - provides a rapid response to consistently changing errors.
Disadvantages
o It is susceptible or prone to noise, process noise coming from the sensor.
o The error will have noise artefacts imposed to it.
o Filters can be used low pass filter to Filters the noise.
1.5 FEEDBACK AND FEEDFORWARD STRATEGIES (SECTION 2.9)
FEEDBACK CONTROL (PAGE 244 VOL 2)
Disadvantage - if the disturbance is often and substantial the controller will back track.

➢ Only one sensor is required.

➢ No process model is needed.
➢ Reactive control.

Same Operation principle or Procedure as THIS

➢ The PID is tuned to tweak the controller to get the expected response from the plant.
➢ If there is no disturbance, feedback control can be used.
➢ If there are many disturbances or if the set point change occur often feedforward can be used.
➢ Set point r come in, generated error e is from the feedback from the process output c.
➢ The controller reacts to the error.
➢ Then sends a signal out to a final control element which then affects the plant.
➢ It is a continuous process.
➢ No process model, no need to know tank dimensions or the pipe diameter.
➢ The PID is tuned, the gains KC, TI, Td are tuned to get the expected response from the plant.
➢ If there is no disturbance feedback control can be used

FEEDFORWARD CONTROL
➢ Reduces disturbances before they affect the process.
➢ May require more than one sensor.
➢ Process model is required.
➢ Anticipatory control.
➢ Feedforward looks at the effects of the disturbance in the control loop.
➢ Measures all the disturbances in the control loop.
➢ Measure all the disturbances on the input and feed it back.

➢ No measurement taken at the output.

➢ The goal is to measure the disturbance and keep the temperature set point.
▪ Cool water come in, if the flow rate increase, if there is a quick demand.
▪ The valve must open to let more steam in.
▪ Feedback is moved to the input.
▪ If there are other disturbances, like pressure variation, steam is not constant, when the valve open it would create
an effect on hot water temperature pipeline.
▪ Any disturbance that may be present in the process loop are measured.

Working Principle Of Feedforward Control

▪ Setpoint, r, come into the system - the controller will have a set point.
▪ It goes into the Feedforward calculation.
▪ Feedforward looks set point and all the inputs from the load (load- changing flow rate of the cold water).
▪ It generates an output, the output is generated to control the flow loop and the steam valve.
▪ The output generated is c.
▪ The Load Disturbance is added to the output c.
▪ Then it is sent to the Feedforward control system to compensate for that addition.

FEEDBACK PLUS FEEDFORWARD CONTROL

➢ A combination of feedback and feedforward control.
➢ Feed forward input requires a disturbance signal, and this set point to be fed to it to determine the output to the
system.
➢ May require more than one measurement disturbance.

Feedforward Control Example – Heat Exchanger (Page 247, Fig. 2.9d, Vol 2)

Hand Indicating Controller Feedforward Calculation Station

MP T4 - T3 Feedback

T4
T4
T3

CH
MH
Output T1

CP
Outlet T2
MP

➢ Liquid comes in, the goal is to maintain temperature TT or T4 which is the temperature of the Output, the Setpoint.
➢ There is steam entering then exits the heat exchanger through the Outlet to compensate a change in variable flow
of the liquid because the liquid demand could be changing.
➢ If liquid demand changes, the amount of steam going in has to change to compensate for that disturbance.
➢ T1 is the temperature of the steam entering the exchanger, T2 is the temperature of the Steam in the Outlet.
➢ Energy In = Energy Out [HHMH (𝑇1− 𝑇2) = CPMP(T4 – T3)].
➢ To control 𝑇4 steam has to be manipulated, by making Steam Mass Flow Rate (MH) subject of the formula.
➢ Setpoint is set on HIC, T4 – T3 is the same as T2set – T1, this is multiplied (X) by the Mass Flow Rate of the Liquid
(MP) in FY.
➢ This computation is sent to the valve or use feedback to control Mass Flow Rate of the Steam
➢ This uses a combination of Feedforward and Feedback for better control.

− 𝑀𝑃 is the steam mass.

− FB (Feedback) controls the steam flow rate.

− PID is calculated using Proportional, Integral and Derivative
time constants.
CASCADE CONTROL (SECTION 2.6) (Page 220, Vol 2)
➢ Uses a combination of 2 PID controllers in series.
➢ Has two feedback loops, Inner Loop and Outer Loop
Advantages
o Disturbances are corrected in the secondary loop first before they affect primary loop.
o Achieves better control.

Working Principle Of A Cascade Control

▪ The set point comes in, main set point is in the Primary Loop.
▪ The Outer Controller is a pH analytical controller.
▪ Saturation Point which is based on the Secondary Set Point.
▪ Inner Controller which is the flow rate FC.
▪ Process is the flow affected by internal disturbances, it has flow loop and pH process loop.
▪ There are two sensors, flow sensor (Sensor2) and pH sensor (Sensor1).
▪ This results in a much improved control.

Example – To control pH in a tank of water.

Primary Loop

Secondary Loop

▪ The pH is the measured variable, it goes to an analytical controller AC, set point is the pH.
▪ Valve controls the acid coming in.
▪ The output of AC goes the input of the Secondary Feedback Loop.
▪ Flow Controller gets input from Flow Transmitter FT.
▪ The Primary Control Loop is a pH control (AT & AC).
▪ The Secondary Control Loop is the flow of the acid (FC & FT).
▪ The output of AC goes to FC, which is another controller and that is where the Cascade comes in.
▪ The Cascade is done to achieve a faster and better control.
▪ If there are variations in the acid, the valve can compensate for that.
▪ The flow control loop is fast.

Room Temperature Control Using Cascade Strategy

T2

25 °C
Setp
Set
int
T1
15 °C

▪ Hot Air comes in the room, Steam heats up the room.

▪ The Primary Loop is the room temperature Master, and the Secondary Loop is the Slave.
▪ Master and Slave are normal PID controllers.
▪ T1 is the Room temperature, T2 is the Steam temperature.
▪ If the Setpoint is 25 ℃, and the room temperature is 15 °C
▪ The Master sends output to the Slave which is the change in Setpoint.
▪ Using T2 feedback, Slave Controller adjusts the Steam Valve such that T1 goes up to the desired temperature of
25 °C.
▪ Uses two loops to achieve faster control, because if Master sends Output directly to the valve there would be lags
▪ If the output of the master went to valve there would lags, the system would not respond quickly enough to the
changes because there is one temperature measurement.

SPLIT RAGE CONTROL (Page 221, Fig. 2.6d, Vol 2)

➢ Uses two valves on the same output.
➢ If there is a temperature control on the reactor, the reactor has a cascade control.
➢ There is a master set point which is the primary setpoint for the temperature inside the reactor.
➢ Has a cascade control.
➢ There are two valves because of the cold water coming into the jacket and steam coming into the jacket as well.
➢ Split Range Splits the output to two separate valves controlling the steam and cold water.
➢ The two valves improve the system.
 Jacket

▪
▪ Liquid is pumped or Injected around the reactor- Jacket.
▪ Temperature of the Cooling Jacket is measured by TT, and that is the Secondary Source of control.
▪ The Master Controller TRC-1 send a set point to the Slave Controller TRC-2 which is the Secondary Loop.
▪ The Output of Slave TRC-2 has to go to the Final Control Element which are the 2 Valves because Cold Water is
coming into the jacket or Steam coming into the jacket is controlled.
▪ The output of Slave goes through a calculation.
▪ These valves are pneumatic valves and will respond differently to the different changes in the output from
Secondary Controller.
▪ If the Output from Slave 1 goes to the two valves.
▪ From 0% to 50% Cold the water valve.
▪ If slave is at 0% the valve is at 100% position.
▪ When the Slave Output increases, Cold Water Valve closes.
COLD WATER STEAM ▪ Reactor needs hot water, Steam Valve opens if slaves goes to 50% to
100%.
▪ Split range splits output to two separate values.
▪ The valves are controlling cold Walter and steam.

RATIO CONTROL (Page 220, Vol 2)

➢ Looks at 2 inputs and decides on the user maintenance ratio.
Advantages - the controlled variable— flow ratio— can be kept constant and can be directly recorded to verify control
performance.
Disadvantages - Analog transmitters in ratio control is that it has no memory.
 ▪ The set point for the Flow Fraction Controller (FFC-2) is calculated
by an adjustable-gain device, which is known as the Ratio Station.
If the ratio station is outside the control loop, it does not interfere
with the secondary loop’s response
▪ The second method is to calculate the ratio R from the individual
measurements of flows A and B and use this calculated ratio as
the measurement input (controlled variable) into a manually set
ratio controller

➢ The main aim is to control the flow rate of two liquids coming to a pipeline.
➢ Uses normal feedback control to manipulate the liquid or gas.
➢ Two pipelines A and B, flow rate of the Wild Stream is measured.
➢ Control Stream is manipulated by using a value.
➢ The valve is manipulated such that it is in ratio with the Wild Stream.
➢ The controller will measure the actual wild stream and do a computation in order to double the flow rate necessary
for the controlled stream.
➢ Can be applied on a boiler system where fuel is being burned.

Ratio Station

A ▪ The Primary Flow Signal 𝐹𝑃, in percent of

scale, is multiplied by the gain setting to
produce a set point for The Secondary Flow
Signal (𝐹𝑆), in percent of scale.
▪ The true flow ratio must consider the scales
of the two flowmeters.
B

➢ Both are A & B controlled to Wild Stream

➢ The Primary Flow FP is measured and compensated for.
➢ It will have its own PID controller FIC controlling the flow rate.
➢ The output of FT is sent to the calculation FY, which is the ratio block.
➢ The flow stream B is controlled.
❖ The setting of the ratio station is related to the true flow ratio by:

(𝑇𝑟𝑢𝑒 𝑓𝑙𝑜𝑤 𝑟𝑎𝑡𝑖𝑜)(𝑆𝑐𝑎𝑙𝑒 𝑜𝑓 𝐹𝑃)

𝑅=
𝑆𝑐𝑎𝑙𝑒 𝑜𝑓 𝐹𝑆