process control week 8
-
Upload
dhruv-motwani -
Category
Documents
-
view
227 -
download
1
description
Transcript of process control week 8
![Page 1: process control week 8](https://reader030.fdocuments.net/reader030/viewer/2022020117/563db8df550346aa9a97c384/html5/thumbnails/1.jpg)
Chapter 7: Feedback Controllers
Lecturer: Mukta Bansal
Department of Chemical & Biomolecular Engineering (CBE)Nanyang Technological University (NTU)
1
![Page 2: process control week 8](https://reader030.fdocuments.net/reader030/viewer/2022020117/563db8df550346aa9a97c384/html5/thumbnails/2.jpg)
2
Feedback Control?
OutputsInputsPlant
Controller
![Page 3: process control week 8](https://reader030.fdocuments.net/reader030/viewer/2022020117/563db8df550346aa9a97c384/html5/thumbnails/3.jpg)
Learning Objectives
√ Control Algorithms√ Basic Control ModesProportional ControlIntegral ControlDerivative Control
√ Features of PID Controllers√ Typical Responses of Feedback Control Systems
3
![Page 4: process control week 8](https://reader030.fdocuments.net/reader030/viewer/2022020117/563db8df550346aa9a97c384/html5/thumbnails/4.jpg)
Recall from Lecture 1
• A controller ensures that the process operates at itsdesign specifications in presence of disturbances
• The most commonly used control structure is feedback,which updates manipulated variables based onmeasurements of controlled or output variables
• Unlike feedforward control, feedback control does notanticipate effect of disturbances on controlledvariables and is less sensitive to modeling errors (dueto direct use of measurements)
4
![Page 5: process control week 8](https://reader030.fdocuments.net/reader030/viewer/2022020117/563db8df550346aa9a97c384/html5/thumbnails/5.jpg)
Revisiting The Continuous Blending Tank
5
Process
Sensor
ControllerControl Valve
![Page 6: process control week 8](https://reader030.fdocuments.net/reader030/viewer/2022020117/563db8df550346aa9a97c384/html5/thumbnails/6.jpg)
Real Life Feedback Controllers
• Feedback Control only in Industries?
6
Driving a car Marathon running
Sensor
Control calculation
Final control element
![Page 7: process control week 8](https://reader030.fdocuments.net/reader030/viewer/2022020117/563db8df550346aa9a97c384/html5/thumbnails/7.jpg)
Simple Example of Feedback Control
7
Controller
Set point ysp
Input Signal ym(from transmitter)
Output Signal p(to Control Valve)
![Page 8: process control week 8](https://reader030.fdocuments.net/reader030/viewer/2022020117/563db8df550346aa9a97c384/html5/thumbnails/8.jpg)
Basic Control Modes
• Proportional/Integral/Derivative
• Proportional Control (P)√ Objective is to reduce the error signal to zero
√ Error Signal, Ysp = set pointYm = measured value of the controlled variable (orequivalent signal from transmitter)
8
(7‐1)
![Page 9: process control week 8](https://reader030.fdocuments.net/reader030/viewer/2022020117/563db8df550346aa9a97c384/html5/thumbnails/9.jpg)
1. Proportional Control
• Proportional Controller√ Output is proportional to the error signal√ √ p(t): Controller Output, Kc: Controller Gain, :steady‐statevalue
• Transfer Function for Proportional Controller
• Laplace Transform
9
(7‐2)
![Page 10: process control week 8](https://reader030.fdocuments.net/reader030/viewer/2022020117/563db8df550346aa9a97c384/html5/thumbnails/10.jpg)
Proportional Control (Cont’d)
10
![Page 11: process control week 8](https://reader030.fdocuments.net/reader030/viewer/2022020117/563db8df550346aa9a97c384/html5/thumbnails/11.jpg)
Proportional Control (Cont’d)
• Proportional control is attractive because of itssimplicity
• Disadvantage√ Steady‐state error (or offset) occurs after set‐point changeor a sustained disturbance
• Example√ Level Control Problem
11
![Page 12: process control week 8](https://reader030.fdocuments.net/reader030/viewer/2022020117/563db8df550346aa9a97c384/html5/thumbnails/12.jpg)
Key Concepts for Proportional Control
• Proportional Band (in %), PB
• ≜ %
• Reverse or Direct Acting Controller√ can be either positive or negative
• Direct‐Acting ( )√ output increases p(t) as input increases Ym(t)
• Reverse‐Acting ( )√ output increases p(t) as input decreases Ym(t)
12
(7‐3)
![Page 13: process control week 8](https://reader030.fdocuments.net/reader030/viewer/2022020117/563db8df550346aa9a97c384/html5/thumbnails/13.jpg)
Example
• Flow Control Loop
• Assume FT is direct‐acting√ Should the flow controller have direct or reverse action?
Air‐to‐open (fail close) valve ==> Flow rate through the valveincreases as the signal to the valve p(t) increases
Air‐to‐close (fail open) valve ==> ?
• Extremely important that Controller action be specified correctly• General Guidelines: The overall product of the gains for all the
components in the feedback control should be positive
13
![Page 14: process control week 8](https://reader030.fdocuments.net/reader030/viewer/2022020117/563db8df550346aa9a97c384/html5/thumbnails/14.jpg)
Qualitative effects of parameters on Performance
• P‐Control: Increasing Kc results in√ Faster response√ Reduced Offset√ Increased Oscillations√ instability
14
![Page 15: process control week 8](https://reader030.fdocuments.net/reader030/viewer/2022020117/563db8df550346aa9a97c384/html5/thumbnails/15.jpg)
2. Integral Control
• The controller output depends on the integral of theerror signals over time
• ∗ ∗
• I reset time (or integral time) – units of time
• Important Advantage: Elimination of offset√ When e ≠ 0, p(t) changes with time until e(t) = 0, where p
reaches a steady state that caused error to be zero
• Disadvantage√ It tends to produce Oscillatory responses and reduces stability
(can be avoided by proper tuning or including derivative action)
15
(7‐7)
![Page 16: process control week 8](https://reader030.fdocuments.net/reader030/viewer/2022020117/563db8df550346aa9a97c384/html5/thumbnails/16.jpg)
Proportional‐Integral (PI) Control
• Integral Controller is seldom used by itself (as it actsslow)
• Integral control action is normally used inconjunction with proportional control as PI
• PI Controller
• ∗ ∗
• Transfer Function for PI
• 1
16
(7‐8)
(7‐9)
![Page 17: process control week 8](https://reader030.fdocuments.net/reader030/viewer/2022020117/563db8df550346aa9a97c384/html5/thumbnails/17.jpg)
PI Controller (Cont’d)
• Response of PI controller to a unit step change in e(t)√ Controller output change instantaneously due to P action√ Integral action causes the ramp increase in p(t)√ Integral Time: time the controller takes to repeat p action
17
![Page 18: process control week 8](https://reader030.fdocuments.net/reader030/viewer/2022020117/563db8df550346aa9a97c384/html5/thumbnails/18.jpg)
Key Concepts for Integral Control
• Reset Windup or Integral windup√ Inherent disadvantage of integral control√ Integral mode causes the controller output to change as long as e(t) ≠
0√ When a sustained error occurs, the integral term becomes quite large
& controller output eventually saturates√ Further build up of the integral term while the controller is saturated
is referred as reset windup or integral windup
18
![Page 19: process control week 8](https://reader030.fdocuments.net/reader030/viewer/2022020117/563db8df550346aa9a97c384/html5/thumbnails/19.jpg)
Summary of PI Controllers
• Two parameters Kc and τI
• I action eliminates offset
• Tends to produce Oscillatory response
• Reduce the stability
19
![Page 20: process control week 8](https://reader030.fdocuments.net/reader030/viewer/2022020117/563db8df550346aa9a97c384/html5/thumbnails/20.jpg)
Qualitative effects of PI parameters on Performance
• Increasing integral time makes the controller moreconservative (sluggish)
• Theoretically, offset will be eliminated for all positive values ofintegral time
20
![Page 21: process control week 8](https://reader030.fdocuments.net/reader030/viewer/2022020117/563db8df550346aa9a97c384/html5/thumbnails/21.jpg)
3. Derivative Control
• Anticipatory Control√ Anticipate the future behavior of the error signal by considering
its rate of change√ Tends to stabilize the controlled process√ Example: A reactor temperature increases by 10 degree in 3
mins√ P Control? I control?√ Controller output proportional to the rate of change of the error
signal√ √ In presence of noise (high frequency random fluctuations),
derivative of the measured variable will change wildly andderivative action will amplify the noise
21
(7‐10)
![Page 22: process control week 8](https://reader030.fdocuments.net/reader030/viewer/2022020117/563db8df550346aa9a97c384/html5/thumbnails/22.jpg)
Proportional‐Integral‐Derivative Control
• Combination of the proportional, integral, andderivative control modes as a PID controller
• Many variations of PID control are used in practice• We consider the three most common forms• 1. Parallel Form of PID Control
• Corresponding Transfer Function
22
(7‐13)
(7‐14)
![Page 23: process control week 8](https://reader030.fdocuments.net/reader030/viewer/2022020117/563db8df550346aa9a97c384/html5/thumbnails/23.jpg)
1. Parallel Form of PID Control
23
• What if e(t) becomes constant?• P action?• D action?• I action?
![Page 24: process control week 8](https://reader030.fdocuments.net/reader030/viewer/2022020117/563db8df550346aa9a97c384/html5/thumbnails/24.jpg)
2. Series Form of PID Control
• PI and PD operated in series
• Commercial versions of the series‐form controllerhave a derivative filter
24
(7‐15)
![Page 25: process control week 8](https://reader030.fdocuments.net/reader030/viewer/2022020117/563db8df550346aa9a97c384/html5/thumbnails/25.jpg)
3. Expanded Form of PID Control
• Expanded form
• Three Gains Kc, KI, KD
• Expanded form of PID is used in MATLAB
25
(7‐16)
![Page 26: process control week 8](https://reader030.fdocuments.net/reader030/viewer/2022020117/563db8df550346aa9a97c384/html5/thumbnails/26.jpg)
PID Controller
• Three tuning Parameters: Kc, τI, τD
• Better response than PI controller
• P, I, and D mode respectively depends on√ The present error√ Accumulation of past error√ Prediction of future errors
26
![Page 27: process control week 8](https://reader030.fdocuments.net/reader030/viewer/2022020117/563db8df550346aa9a97c384/html5/thumbnails/27.jpg)
Qualitative effects of Derivative time on Performance
• Difficult to generalize the effects of derivative time τD√ For small values, increasing τD tends to improve theresponse by reducing the maximum deviation, responsetime and degree of oscillation
√ However, if τD is too large measurement noise is amplifiedand response may become oscillatory
√ Intermediate value of τD is desirable
27
![Page 28: process control week 8](https://reader030.fdocuments.net/reader030/viewer/2022020117/563db8df550346aa9a97c384/html5/thumbnails/28.jpg)
Typical response of a Feedback Control Systems
• A step change in a disturbance variable occurs
• Actual Behavior would depends on the type ofcontroller, controller settings and process dynamics
28
![Page 29: process control week 8](https://reader030.fdocuments.net/reader030/viewer/2022020117/563db8df550346aa9a97c384/html5/thumbnails/29.jpg)
Summary of the Characteristics of the Most Commonly Used Controller Modes
• Proportional√ Simple.√ Inherently stable when properly tuned.√ Easy to tune.√ Experiences offset at steady state
• Proportional plus integral√ No offset√ Better dynamic response than reset alone.√ Possibilities exist for instability due to lag introduced
29
![Page 30: process control week 8](https://reader030.fdocuments.net/reader030/viewer/2022020117/563db8df550346aa9a97c384/html5/thumbnails/30.jpg)
Summary (Cont’d)
• Proportional plus derivative√ Stable
√ Less offset than proportional alone (use of higher gain possible)
√ Reduces lags, i.e., more rapid response
• Proportional plus integral plus derivative√ Most complex
√ Rapid response
√ No offset
√ Best control if properly tuned
30