Performance of Different Control Strategies for Boiler Pragyan Paramita Kar

Performance of Different Control Strategies for Boiler  Pragyan Paramita Kar

Performance of Different Control Strategies for Boiler

Drum Level Control Using LabVIEW

Pragyan Paramita Kar

Priyam Saikia

Department of Electronics & Communication Engineering

National Institute of Technology, Rourkela

Rourkela-769008, Odisha, India

Performance of Different Control Strategies for Boiler Drum Level Control

Using LabVIEW

A Thesis submitted in partial fulfillment of the requirements for the award of the degree of

Bachelor of Technology

In

Electronics & Instrumentation Engineering

in

May 2013

to the department of

Electronics & Communication Engineering

of

National Institute of Technology, Rourkela

By

Pragyan Paramita Kar &

Priyam Saikia

[Roll no 109EI0327] [Roll no 109EI0343]

Under the guidance of

Prof. Umesh Chandra Pati

Department of Electronics & Communication Engineering

National Institute of Technology, Rourkela

Rourkela-769008, Odisha, India

Department of Electronics & Communication Engineering

National Institute of Technology, Rourkela

Rourkela -769008, Odisha, India

Certificate

This is to certify that the work in the thesis entitled Performance of different

control strategies of Boiler Drum level using LabVIEW by Pragyan Paramita Kar

and Priyam Saikia bearing Roll No. 109EI0327 and 109EI0343 respectively, is a record of an original research work carried out by them under my supervision and guidance in partial fulfillment of the requirement for the award of the degree of

Bachelor of Technology in Electronics & Instrumentation Engineering. Neither this

thesis nor any part of it has been submitted for any degree or academic award elsewhere.

Prof. Umesh Chandra Pati

May 8, 2013 Department of Electronics & Communication Engineering

National Institute of Technology, Rourkela iii

ACKNOWLEDGEMENT

We would like to express our earnest gratitude to our project guide, Prof. Umesh Chandra

Pati for believing in our ability to work and enriching us with knowledge throughout our project work that crowned our efforts with success. His profound insights and working styles have inspired us. The invaluable guidance and support that he has offered us has deeply encouraged us.

We wish to extend our sincere thanks to Mr. Subhransu Padhee, Ph.D Research Scholar for his enriching knowledge about the subject. We would also like to thank all people, faculty and non-teaching staff who have helped and inspired us during our project work at

Department of Electronics & Communication Engineering of National Institute of

Technology, Rourkela for extending their help and support as and when required.

We would conclude with our deepest gratitude to our parents, and all our loved ones. Our full dedication to the work would have not been possible without their blessings and moral support.

Once again, we especially thank Prof. U. C. Pati. It was a great pleasure for us to conduct the project under his supervision.

Pragyan Paramita Kar

Priyam Saikia

B. Tech

National Institute of Technology, Rourkela

Electronics and Instrumentation Engg iv

ABSTRACT

This project work describes the control strategies of boiler drum level system. The design and implementation of this process is done in the LabVIEW software. Three different types of boiler drum level control system are designed in the Circuit Design and Simulation toolkit of

LabVIEW. This work provides insight into the various PID controller design methods such as

Zeigler-Nichol method, Tyreus-Luyben method, Internal Model Control (IMC) as well as

Fuzzy Logic Controller. Comparative study is made on the performances of the PID tunings methods for the different control strategies. The implementation of IMC in PID and its utilization for handling unstable processes with dead time is also done. An insight to fuzzy logic and its simple implementation as a controller is studied. Fuzzy logic controller is designed using LabVIEW Fuzzy System Designer. It is implemented in set point weighting for proportional term in PID controller. The performance evaluation for fixed weight and fuzzy weight set-point are performed and their performance in important parameters like overshoot, rise time and settling time are evaluated. v

CONTENTS

Certificate

ACKNOWLEDGEMENT

ABSTRACT

LIST OF FIGURES

LIST OF TABLES

Chapter 1:

INTRODUCTION

1.1

OVERVIEW

1.2

LITERATURE SURVEY

1.3

NI-LabVIEW

1.3.1 KEY FEATURES OF LABVIEW

Chapter 2:

BOILER DRUM LEVEL CONTROL SYSTEM

2.1

BOILER DRUM LEVEL CONTROL SYSTEM

2.1.1 SINGLE ELEMENT DRUM LEVEL CONTROL

2.1.2 TWO ELEMENT DRUM LEVEL CONTROL

2.1.3 THREE ELEMENT DRUM LEVEL CONTROL

2.2

PID TUNING METHODS

2.2.1 ZEIGLER NICHOLS METHOD

2.2.2 TYREUS-LUYBEN METHOD

2.2.3 INTERNAL MODEL CONTROL

2.2.4 FUZZY LOGIC CONTROL

Chapter 3:

CONTROL STRATEGIES

3.1

ONE ELEMENT DRUM LEVEL CONTROL

3.1.1 ZEIGLER NICHOLS TUNING:

3.1.2 TYREUS LUYBEN TUNING :

3.2

TWO ELEMENT DRUM LEVEL CONTROL:

3.2.1 ZEIGLER NICHOLS TUNING:

3.2.2 TYREUS LUYBEN TUNING:

3.3

SUMMARY

Chapter 4:

INTERNAL MODEL CONTROL

4.1 OVERVIEW OF INTERNAL MODEL CONTROL

4.1.1 THE IMC STRATEGY

4.1.2 CONTROLLER DESIGN PROCEDURE

4.1.3 LabVIEW IMPLEMENTATION FOR INTERNAL MODEL CONTROL

vi

3

3

5

5

5

6

6

7

8

8

9

9

2

2

2

12

12

13

16

17

18

20

21

23

23

23

25

25

iii

iv

v

viii

xii

4.2 IMC BASED PID CONTROL METHOD

4.2.1 THE EQUIVALENT FEEDBACK FORM TO IMC

4.2.2. THE IMC BASED PID CONTROL DESIGN PROCEDURE

4.2.3 RESULT AND SIMULATION OF IMC BASED PID CONTROLLER

4.3

IMC BASED PID TUNING FOR UNSTABLE PROCESSES

4.3.1. IMC BASED PID PROCEDURE

4.3.2 TUNING FOR IMC BASED PID FOR UNSTABLE SYSTEM

Chapter 5:

FUZZY LOGIC CONTROLLER

5.1

OVERVIEW OF FUZZY LOGIC

5.1.1 HOW DOES FUZZY LOGIC WORK?

5.1.2 CONSTRUCTING A FUZZY LOGIC CONTROLLER:

5.2 FUZZY LOGIC BASED CONTROLLER

5.2.1 FUZZY LOGIC CONTROLLER FOR BOILER SYSTEM

5.2.2 LabVIEW IMPLEMENTATION

5.2.3 SIMULATIONS AND RESULTS

5.3 FUZZY LOGIC BASED PID CONTROLLER

5.3.1 LabVIEW IMPLEMENTATION

5.3.2 SIMULATIONS AND RESULTS

5.4 SUMMARY

Chapter 6:

CONCLUSION

PUBLICATIONS

REFERENCES

vii

31

34

34

35

30

30

31

42

45

46

50

51

42

42

42

44

53

55

60

62

63

64

LIST OF FIGURES

FIGURE

Fig. 2.1 Single element boiler drum level control

Fig. 2.2 Two element boiler drum level control

Fig. 2.3 Three element boiler drum level control

Fig. 2.4 Basic design of an IMC Controller

Fig. 2.5 Block Diagram of Fuzzy control system

Fig. 3.1 Block diagram for one element drum level

Fig. 3.2 Ultimate gain response for step input for

( )

Fig. 3.3 Step Response for P Control

Fig. 3.4 Step Response for PI Control

Fig. 3.5 Step Response for PID Control

Fig. 3.6 Ultimate gain response for step input for

( )

Fig. 3.7 Step Response for P Control

Fig. 3.8 Step Response for PI Control

Fig. 3.9 Step Response for PID Control

Fig. 3.10 Step Response for PI Control

Fig. 3.11 Step Response for PID Control

Fig. 3.12 Step Response for PI Control viii

9

10

12

6

7

Pg.

5

13

14

14

14

14

15

15

15

16

16

17

Fig. 3.13 Step Response for PID Control

Fig. 3.14 Block Diagram for two element drum level control

Fig. 3.15 Step Response for P Control

Fig. 3.16 Step Response for PI Control

Fig. 3.17 Step Response for PID Control

Fig. 3.18 Step Response for PID Control with disturbance rejection

Fig. 3.19 Step Response for P Control

Fig. 3.20 Step Response for PI Control

Fig. 3.21 Step Response for PID Control

Fig. 3.22 Step Response for PID Control with disturbance rejection

Fig. 3.23 Step Response for PI Control

Fig. 3.24 Step Response for PID Control

Fig. 3.25 Step Response for PID Control with disturbance rejection

Fig. 3.26 Step Response for PI Control

Fig. 3.27 Step Response for PID Control

Fig. 3.28 Step Response for PID Control with disturbance rejection

Fig.4.1 Schematic of IMC control

Fig. 4.2 Unity feedback for IMC

Fig.4.3 Response for Unity feedback of IMC

Fig. 4.4 Feedback for IMC

Fig. 4.5 Response for feedback of IMC

Fig. 4.6 Response for Unity feedback of IMC ix

27

28

28

29

20

20

20

20

20

19

19

19

20

19

19

19

17

18

18

18

23

27

Fig. 4.7 Response for IMC with feedback

Fig. 4.8 Response for IMC with feedback

Fig. 4.9 Response for Unity feedback of IMC

Fig. 4.10 IMC based PID controller

Fig. 4.11 Block Diagram for IMC based PID using Unity Feedback

Fig. 4.12 Response for IMC based PID with Unity feedback

Fig. 4.13 Response for IMC based PID with Unity feedback

Fig. 4.14 IMC based PID with a set point filter

Fig. 4.15 Block Diagram for IMC based PID (unstable process) Unity feedback

Fig. 4.16 Response for IMC based PID (unstable process)

Fig. 4.17 Response for IMC based PID (unstable process)

Fig. 4.18 Response for IMC based PID (unstable process)

Fig. 5.1 Block diagram of a Fuzzy Interference System

Fig. 5.2 Fuzzy logic control system

Fig. 5.3 A simple fuzzy logic controller

Fig. 5.4 Fuzzy logic controller using two input variables

Fig. 5.5 Membership Function for Error (e(t))

Fig. 5.6 Change in Error (Δe(t)).

Fig. 5.7 Membership Function for output variable, CONTROLLER (u(t))

Fig. 5.8 Block Diagram - Fuzzy logic controller

Fig. 5.9: Step Response for

( )

Fig. 5.10: Step Response for

( ) x

37

38

39

40

43

43

45

46

46

46

47

50

50

32

33

34

35

29

30

30

31

51

Fig. 5.11 PID controller with fixed set point weighting

Fig. 5.12 PID controller with fuzzy based set point weighting

Fig. 5.13 Block Diagram for comparison of fixed set point weighting and fuzzy setpoint weighting for PID controller.

Fig. 5.14:

( )

Fig. 5.15:

( ), ξ=0.2

Fig. 5.16:

( )

( ), T=1, L=0.1

Fig. 5.18:

( ), T=1, L=0.4

Fig. 5.19:

( ), T=1, L=0.8

Fig. 5.20:

( ), T=10, L=0.1

Fig. 5.21:

( ), T=10, L=0.4

Fig. 5.22:

( ), T=10, L=0.8

Fig. 5.23:

( )

Fig. 5.24:

( )

Fig. 5.25

( ), L=0

Fig. 5.26:

( ), L=0.1

52

53

55

56

56

56

55

55

56

56

57

57

57

57

57

57 xi

LIST OF TABLES

TABLE Pg

TABLE 2.1 ZIEGLER NICHOLS TUNING PARMETERS 8

TABLE 2.2 TYREUS LUYBEN TUNING PARMETERS 8

TABLE 3.1 ZIEGLER NICHOLS TUNING PARMETERS for

( )

13

TABLE 3.2 ZIEGLER NICHOLS TUNING PARMETERS for

( )

15

TABLE 3.3 TYREUS LUYBEN TUNING PARMETERS for

( )

TABLE 3.3 TYREUS LUYBEN TUNING PARMETERS for

( )

16

17

TABLE 4.1 PID TUNING PARAMETERS IMC CONTROLLER FOR UNSTABLE 36

SYSTEM

TABLE 5.1 LINGUISTIC VARIABLES FOR FUZZY LOGIC CONTROLLER

TABLE 5.2 IF-THEN RULES FOR FUZZY LOGIC CONTROLLER

46

47

TABLE 5.3 LINGUISTIC VARIABLES FOR FUZZY BASED PID CONTROLLER 53

TABLE 5.4 RULE BASE FOR FUZZY BASED PID CONTROLLER 54

58 TABLE 5.5 VALUE OF OVERSHOOT (%) USING A FIXED WEIGHT b AND A

FUZZIFIED WEIGHT b

TABLE 5.6 VALUE OF RISE TIME USING A FIXED WEIGHT b AND A

FUZZIFIED WEIGHT b

TABLE 5.7 VALUE OF SETTLING TIME USING A FIXED WEIGHT b AND A

FUZZIFIED WEIGHT b

59

59 xii

CHAPTER 1

INTRODUCTION

Overview

Literature Survey

NI LabView

1

INTRODUCTION

In this chapter, the overview of the boiler drum level control system is described.

Literature survey of this work has been discussed. The objective of the project is explained.

1.1 OVERVIEW

Boiler is defined as a closed vessel in which steam is produced from water by the combustion of fuel. Generally, in boilers steam is produced by the interaction of hot flue gases with water pipes which is coming out from the fuel mainly coal or coke. In boilers, chemical energy of stored fuel is converted into the heat energy and this heat energy is absorbed by the water which converts them into a steam. Boilers have many serious injuries and destruction of property. It is critical for the safe operation of the boiler and the steam turbine. Too low a level may overheat boiler tubes and damage them. Too high a level may interfere with separating moisture from steam and transfers moisture into the turbine, which reduces the boiler efficiency. Various controlling mechanism are used to control the boiler system so that it works properly.

1.2 LITERATURE SURVEY

Subhransu Padhee and Yaduvir Singh give an overview of data acquisition, data logging and supervisory control system of a plant consisting of multiple boilers. Data acquisition, data logging and supervisory control are the basic building blocks of plant automation. This paper takes a case study of plant consisting of multiple boilers where multiple process variables of the boilers need to be acquired from the field and monitored. The data of the process variables needs to be logged in a database for further analysis and supervisory control.[1]

Mihai Iacob, Gheorghe-Daniel Andreescu, and Nicolae Muntean present an open loop dispatcher training simulator for boiler-turbine implemented in LabView for COLTERM heating power plant of Timisoara, Romania. The system employs real-time capability, graphical user interface (GUI), uninterrupted operator interaction, having as background a low order boiler-turbine model for dynamic simulation. The operator manually controls the fuel charge on each of the three boilers, the turbine valve position and the steam to consumers, to anticipate parameter evolution on each boiler and the electric power generated by turbine.[2]

2

1.3 NI-LabVIEW

National Instrument’s LabVIEW is a graphical development environment for creating flexible, measurement and control applications rapidly at minimal cost. With LabVIEW, engineers and scientists interface with real-world signals, analyse data for meaningful information and share results. LabVIEW makes development very fast and easy for all users.

1.3.1 KEY FEATURES OF LABVIEW

• Graphical Programming

• Built-in measurement and control function

• Multiple development tools

• Wide array of computing targets

The main programming section of LabVIEW is a Virtual Interface (VI) and a corresponding block diagram. Programming for the VI is done using control palette which contains several controls and indicators. Similarly, the corresponding block diagram is programmed using the function palette.

3

CHAPTER 2

BOILER DRUM LEVEL

CONTROL SYSTEM

Boiler drum level control system

Single element drum level (feedback) control

Two element drum level (feedback and feedforward) control

Three element drum level (cascade) control

PID tuning methods

Ziegler Nichols method

Tyreus Luyben method

Internal Model Control(IMC)

Fuzzy logic controller

4

BOILER DRUM LEVEL CONTROL

SYSTEM

In this chapter, three different types of boiler control system i.e. single element boiler drum level control, two element boiler drum level control and three element boiler drum level control are explained. The different PID and PID based controllers implemented to control the drum level are also explained. It consists of Zeigler Nichols and Tyreus Luyben PID tuning, Internal Model Control (IMC) and Fuzzy Logic control.

2.1 BOILER DRUM LEVEL CONTROL SYSTEM

2.1.1 SINGLE ELEMENT DRUM LEVEL CONTROL

The single element control is the simplest method for boiler drum level control system. It is least effective form of drum level control which requires a measurement of drum water level and feed water control valve. It is mainly recommended for boilers with modest change requirement and relatively constant feed water condition. The process variable coming from the drum level transmitter is compared to a set point and the difference is a deviation value. This signal is given to the controller which generates corrective action output. The output is then passed to the boiler feed water valve, which adjusts the level of feed water flow into the boiler drum.

Fig. 2.1 depicts the control scheme for single-element drum level control. In this configuration, only the water level in the drum is being measured (hence the term "single element")

Fig. 2.1 Single element boiler drum level control

5

2.1.2 TWO ELEMENT DRUM LEVEL CONTROL

The two element drum level control system can be best applied to single element boiler drum level control system where feed water is at a constant pressure. It requires the measurement of drum level, load demand and feed water control valve.

Level Element: This signal is compared to a setpoint and the resultant is a deviation value.

This signal is acted upon by the controller which generates corrective action in the form of a proportional value.

Steam Flow Element: This mass flow rate signal is used to control the feed water flow, giving immediate corrections to feed water demand in response to load changes. Any imbalance between steam mass flow out and feed water mass flow into the drum is corrected by the level controller. This imbalance can arise from blow down variation due to change in the dissolved solids, variations in feed water supply pressure or leaks in steam.

Fig. 2.2 Two element boiler drum level control

Fig. 2.2 depicts the control scheme for two-element drum level control. Two element boiler drum level control system is adequate for a load changes of moderate speed and magnitude. It can be applied to any size of boiler.

2.1.3 THREE ELEMENT DRUM LEVEL CONTROL

This control system is ideally suited where a boiler plant consists of multiple boilers and multiple feed water pumps or feed water valve has variation in pressure or flow. It requires the measurement of drum level, steam flow rate, feed water flow rate and feed water control valve.

By using cascade control mechanism level element act as a primary loop and flow element act as a secondary loop and steam flow element act as a feed forward controller.

Level element and steam flow element mainly correct for unmeasured disturbances within the

6

system such as boiler blow down. Feed water flow element responds rapidly to variations in feed water demand either from the feed water pressure and steam flow rate of feed forward signal.

The performance of the three-element control system during transient conditions makes it very useful for general industrial and utility boiler applications. It handles loads exhibiting wide and rapid rates of change. Fig. 2.3 depicts the control scheme for three-element drum level control.

Fig. 2.3 Three element boiler drum level control

2.2 PID TUNING METHODS

The mnemonic PID refers to the first letters of the names of the individual terms that make up the standard three-term controller. These are P for the proportional term, I for the integral term and D for the derivative term in the controller. Three-term or PID controllers are probably the most widely used industrial controller. Even complex industrial control systems may comprise a control network whose main control building block is a PID control module.

The PID controller separately calculate the three parameters i.e. the proportional, the integral, the derivative values. The proportional value determines the reaction to the current error. The integral value determines the reaction based on the sum of recent errors as past error. The derivative value determines the reaction based on the rate at which the error has been changing as a future error. By tuning these three constants in the PID controller algorithm, the controller can provide control action designed for specific process control requirements.

Here, we will discuss about the different methods used for tuning PID parameters. Zeigler

Nichols and Tyreus Luyben are two of the main methods to tune PID. Apart from that,

Internal Model Control (IMC) and Fuzzy Logic controller can be used along with PID to get desired and improved performances.

7

2.2.1 ZEIGLER NICHOLS METHOD

This method is introduced by John G. Ziegler and Nathaniel B. Nichols. In this method, the

and gains are first set to zero. The gain is increased until it reaches the ultimate gain

, at which the output of the loop starts to oscillate. The ultimate gain, and the oscillation period are used to set the gains as shown in Table 2.1

.

TABLE 2.1

CONTROL TYPE

P

PI

PID

ZIEGLER NICHOLS TUNING PARMETERS

0.5

0.45

0.6

-

1.2

/

2 /

-

-

/ 8

2.2.2 TYREUS-LUYBEN METHOD

This method is introduced by Tyreus-Lubyen. In this method, the and gains are first set to zero. The gain is increased until it reaches the ultimate gain , at which the output of the loop starts to oscillate.

The ultimate gain, and the oscillation period are used to set the gains as shown in Table

2.2.

TABLE 2.2

TYREUS LUYBEN TUNING PARMETERS

CONTROL TYPE

PI

PID

0.3125

0.4545

/

/

-

/ 6.3

8

2.2.3 INTERNAL MODEL CONTROL

In process control applications, model based control systems are often used to track set points and reject low disturbances. The internal model control (IMC) philosophy relies on the internal model principle which states that if any control system contains within it, implicitly or explicitly, some representation of the process to be controlled then a perfect control is easily achieved. Internal Model Control (IMC) is a commonly used technique that provides a transparent mode for the design and tuning of various types of control. In this report, we analyze various concepts of IMC design and IMC based PID controller has been designed for a plant transfer function to incorporate the advantages of PID controller in IMC.

The IMC-PID controller does good set-point tracking but poor disturbance response mainly for the process which has a small time-delay/time-constant ratio. But, for many process control applications, rejection of disturbance for the unstable processes is more important than set point tracking.

The basic block diagram of an IMC controller is shown in Fig. 2.4.

∑ FILTER CONTROLLER

DISTURBANCES

PROCESS

+

+

PROCESS

MODEL

PROCESS

OUTPUT

-

+

Fig. 2.4 Basic design of an IMC Controller

The details about the IMC controller are discussed in chapter 4. Apart from simple IMC controller and IMC based PID controller, we have also studied the tuning of IMC parameters for IMC based PID controller for unstable processes. The processes consisting of dead time are usually unstable and difficult to control and hence traditional PID controller is not preferred. IMC based PID gives appropriate tuning methods to handle such unstable systems.

2.2.4 FUZZY LOGIC CONTROL

Fuzzy logic is a form of logic that is the extension of boolean logic, which incorporates partial values of truth. Instead of sentences being "completely true" or

9

"completely false," they are assigned a value that represents their degree of truth. In fuzzy systems, values are indicated by a number (called a truth value) in the range from 0 to 1, where 0.0 represents absolute false and 1.0 represents absolute truth. Fuzzification is the generalization of any theory from discrete to continuous.

The fuzzy logic controller provides an algorithm, which converts the expert knowledge into an automatic control strategy.The fuzzy control systems are rule-based systems in which a set of fuzzy rules represent a control decision mechanism for adjusting the effects of certain system stimuli. With an effective rule base, the fuzzy control systems can replace a skilled human operator.

A general form for a Fuzzy control system is given in Fig. 2.5.

Rule base

Fuzzification

Normalisation

Sensors

Inference

Process

Defuzzification

Denormalisation

Final Control Element

Fig. 2.5 Block Diagram of Fuzzy control system

We have also designed and studied the performances of set point weighting controller where the proportional term of a PID controller is multiplied using a fixed weight (a constant) and fuzzified weight (based on error and rate of rate of error). Although it is possible to design a fuzzy logic type of PID controller by a simple modification of the conventional ones, via inserting some meaningful fuzzy logic IF- THEN rules into the control system, these approaches in general complicate the overall. However, the fuzzy PID controller used here, is a natural extension of their conventional versions, which preserve the linear structures of the

PID controllers, with simple and conventional analytical formulas as the final results of the design. Thus, they can directly replace the conventional PID controllers in any operating control systems (plants, processes). The main difference is that these fuzzy PID controllers are designed by employing fuzzy logic control principles and techniques, to obtain new controllers that possess analytical formulas very similar to the conventional digital PID controllers.

10

CHAPTER 3

CONTROL STRATEGIES

Single element drum level (feedback) control

Ziegler Nichols method

Tyreus Luyben method

Two element drum level (feedback and feedforward) control

Ziegler Nichols method

Tyreus Luyben method

Summary

11

CONTROL STRATEGIES

This chapter describes about the different control strategies used in boiler drum level control. It consists of single element feedback control and two element feedback and feedforward control. In both the control strategies PID controller is implemented and the tuning is done using Zeigler Nichols and Tyreus Luyben tuning methods.

3.1 ONE ELEMENT DRUM LEVEL CONTROL

As already mentioned in Chapter 1, this is the simplest but least effective form of drum level control. This consists of a proportional signal or process variable (PV) coming from the drum level transmitter. This signal is compared to a setpoint and the difference is a deviation value. This signal is acted upon by the controller which generates corrective action in the form of a proportional output. The output is then passed to the boiler feed water valve, which then adjusts the level of feed water flow into the boiler drum.

We consider two set of systems for both Zeigler Nichols and Tyreus Luyben tuning. The transfer functions of the two systems are:

( )

( )

;

Valve function :

( )

( )

( )

;

Valve function :

( )

(3.1)

(3.2)

The block diagram for one element drum level control for the tuning methods is shown in

Fig. 3.1.

Fig. 3.1 Block diagram for one element drum level.

12

3.1.1 ZEIGLER NICHOLS TUNING:

We consider the two systems one by one. We give a step input to evaluate the performances. First, and are set to zero. Then is increased till when the output response starts to oscillate with constant amplitude.

For first system we get,

Ultimate gain,

= 8.3

Time period,

2 sec

Fig. 3.2 Ultimate gain response for step input for

( )

Values of the control parameters are in Table 3.1

CONTROL TYPE

P

PI

PID

TABLE 3.1

ZIEGLER NICHOLS TUNING PARMETERS for

( )

4.15

3.735

4.98

-

1.67

1

-

-

The P, PI and PID controller responses for step input are shown in the figures below.

13

Fig. 3.3: Step Response for P Control Fig. 3.4: Step Response for PI Control

For second system we also get,

Ultimate gain,

= 1

Fig. 3.5: Step Response for PID Control and Time period,

2.8 sec

Fig. 3.6 Ultimate gain response for step input for

( )

14

Values of the control parameters are in Table 3.

TABLE 3.2

ZIEGLER NICHOLS TUNING PARMETERS for

( )

CONTROL TYPE

P

PI

PID

0.5

0.45

0.6

-

1.4

2.33

-

-

0.006

The P, PI and PID controller responses for step input are shown in the figures below

Fig. 3.7: Step Response for P Control Fig. 3.8: Step Response for PI Control

Fig. 3.9: Step Response for PID Control

15

3.1.2 TYREUS LUYBEN TUNING :

For the first system, Ultimate gain,

= 8.3

Values of the control parameters are in Table 3.3 and Time period,

2 sec

TABLE 3.3

TYREUS LUYBEN TUNING PARMETERS for

( )

CONTROL TYPE

PI

PID

2.59

3.77

-

The PI and PID controller responses for step input are shown in the figures below.

Fig. 3.10: Step Response for PI Control Fig. 3.11: Step Response for PID Control

For the second system,

Ultimate gain,

= 1 and Time period,

2.8 sec

Values of the control parameters are in Table 3.4

16

TABLE 3.4

TYREUS LUYBEN TUNING PARMETERS for

( )

CONTROL TYPE

PI

PID

0.3125

0.4545

-

The PI and PID controller responses for step input are shown in the figures below

Fig. 3.12: Step Response for PI Control Fig. 3.13: Step Response for PID Control

3.2 TWO ELEMENT DRUM LEVEL CONTROL:

As mentioned in Chapter 1, the two-element drum level controller can best be applied to a single drum boiler where the feed water is at a constant pressure.

T he two elements are made up of

a Level Element (a proportional signal or process variable (PV) coming from the drum level transmitter)

 a Steam Flow Element (a mass flow rate signal (corrected for density) is used to control the feedwater flow, giving immediate corrections to feedwater demand in response to load changes.)

We consider the two set of systems (eqn. (1) and eqn. (2)) for both Zeigler Nichols and

Tyreus Luyben tuning in this strategy also.

17

The feedback tuning is done using PID (Ziegler –Nichols and Tyreus-Luyben Methods)

However here we also have two additional transfer functions.

Steam flow variable (disturbance),

( )

(3.3)

Feed forward controller,

( )

(3.4)

The feedback tuning is performed using PID (Ziegler –Nichols and Tyreus-Luyben Methods).

The block diagram for the two element boiler drum level control is shown in Fig. 3.13.

Fig. 3.14 Block Diagram for two element drum level control

3.2.1 ZEIGLER NICHOLS TUNING:

The ultimate gain, the time period is already found in one element control for the two systems as well as the control parameters for Zeigler Nichols and Tyreus Luyben methods are also seen in Table 3.1, 3.2, 3.3 and 3.4 respectively. The P, PI and PID controller responses for step input for first system are shown in the figures below.

Fig. 3.15: Step Response for P Control Fig. 3.16: Step Response for PI Control

18

Fig. 3.17: Step Response for PID Control Fig. 3.18: Step Response for PID Control with disturbance rejection.

The P, PI and PID controller responses for step input for second system are shown in the figures below.

Fig. 3.19: Step Response for P Control Fig. 3.20: Step Response for PI Control

Fig. 3.21: Step Response for PID Control Fig. 3.22: Step Response for PID Control with disturbance rejection

19

3.2.2 TYREUS LUYBEN TUNING:

The PI and PID controller responses for step input for first system are shown in the figures below.

Fig. 3.23: Step Response for PI Control Fig. 3.24: Step Response for PID Control

Fig. 3.25: Step Response for PID Control with disturbance rejection

The PI and PID controller responses for step input for second system are shown in the figures below.

Fig. 3.26: Step Response for PI Control Fig. 3.27: Step Response for PID Control

20

Fig. 3.28: Step Response for PID Control with disturbance rejection

3.3 SUMMARY

The enhanced three-element drum level control module incorporates the standard three element level components with the following improvements:

Tighter control through a choice of control schemes. Drum level maintained on failure of steam or feed water flow measurements

This module introduces an additional level control loop Boiler drum level control.

Different PID control method is used i.e. Zeigler-Nichols Method and Tyreus-Luyben

Method to design all the three controller strategies of controlling a boiler drum level.

Comparison between the performances of the control strategies is studied and as a result the response of Ziegler Nichols Method control is more accurate than the other control method.

Also PID Type method gives better stabilization, less oscillation, less overshoot and steady state error as compared to P-type control and PI Type control in both Ziegler Nichols Method as well as Tyreus Luyben Method. So, this controller is selected for the control system of boiler. The response of the PID controller is used, when the plant response is changing with time, or there is uncertainty.

Also we can conclude that Ziegler Nichols method generates classical tuning constants that are aggressive and oscillates. Thus if we require minor oscillation and robustness, then

Tyreus Luyben method for PID tuning is used.

21

CHAPTER 4

INTERNAL MODEL CONTROL

Overview of Internal Model Control

IMC Controller

IMC based PID Controller

IMC based PID tuning for unstable processes

Summary

22

INTERNAL MODEL CONTROL

This chapter describes about Internal Model Control and its application in different controllers. IMC model principle relies on the fact that if the control scheme has been developed based on an exact model of the process, then perfect control can be achieved. In this chapter, the advantages of IMC structure (and controller design procedure), compared to the classical feedback control structure have also been discussed.

4.1 OVERVIEW OF INTERNAL MODEL CONTROL

The Internal Model Control (IMC) philosophy relies on the Internal Model Principle, which states that control can be achieved only if the control system encapsulates, either implicitly or

explicitly, some representation of the process to be controlled. In particular, if the control scheme has been developed based on an exact model of the process, then perfect control is theoretically possible. So, if the control architecture has been developed based on the exact model of the process then perfect control is mathematically possible

4.1.1 THE IMC STRATEGY

If one has complete knowledge about the process (as encapsulated in the process model) being controlled, one achieve perfect control. It tells us that the feedback controller is necessary only when knowledge about the process is inaccurate or incomplete.

Set point

R(s)

E(s)

( )

+

+

∑ d(s)

Output Y(s)

U(s) d(s)

PROCESS

( )

PROCESS

MODEL

+

Fig.4.1: Schematic of IMC control

In Fig.4.1, d(s) is an unknown disturbance affecting the system. The manipulated input U(s) is introduced to both the process and its model. The process output(s), is compared with the output of the model, resulting in a signal

̂( ). That is,

23

̂( ) [

(4.1)

If d(s) is zero, then

̂( ) is a measure of the measure of the difference in behavior between the process and its model. If

Thus

̃( ), and can therefore be used to improve control. This can be done by subtracting

̂( ) from the set point

R(s), which is very similar to affecting a set point trim. The resulting control signal is given by,

( ) [ ( ) ̂( )] ( ) = { ( ) [

Thus,

( )

[ ( ) ̂( )] ( )

[ ( ) ( )

Since,

( ) ( ) ( ) ( )

(4.3)

The closed loop transfer function for the IMC scheme is therefore

( )

[ ( ) ( )] ( ) ( )

[

( ) ̃ ( )] ( )

( )

(4.4)

Or

( )

( ) ( ) ( ) [

[

( ) ̃

( ) ̃

( )] ( )

( )] ( )

(4.5)

From this closed loop expression, one can see that if

( )

, and if

( )

̃ ( ), then perfect set point tracking and disturbance rejection is achieved.Theoritically, even if

( ) ̃

( ), perfect disturbance rejection can still be achieved provided

( )

̃( )

.

Additionally, to improve robustness, the effect of process model mismatch should be minimized. Since discrepancies between process and model behavior usually occur at high frequency end of the system‘s frequency response, a low pass filter

( ) is usually added to attenuate the effects of process-model mismatch. Thus, the internal model controller is usually designed as the inverse of the process model in series with a low pass filter, i.e.

( ) ( ) ( ). The order of the filter is usually such that ( ) ( ) is proper, to prevent excessive differential control action. The resulting closed loop then becomes

24

( )

( ) ( ) ( ) [ ( ) ̃

[ ( ) ̃ ( )] ( )

( )] ( )

(4.6)

4.1.2 CONTROLLER DESIGN PROCEDURE

The controller design procedure has been generalized to the following steps.

1. Given a model of the process,

( ), it is factorized into invertible and non-invertible components i.e.

̃ ( ) .

̃ ( )

(4.7)

The non- invertible component

̃ ( ), contains terms which if inverted, will lead to instability and realisability problems, e.g terms containing positive zeros and time delays.

2. The invertible component of the process model is inverted and cascaded with a filter that makes the controller

( ) proper.

The controller is set to

̃ ( )

(4.8) and then

( ) ( )

( ),

(4.9) where

( ), is a low pass filter of appropriate order.

4.1.3 LabVIEW IMPLEMENTATION FOR INTERNAL MODEL CONTROL

CASE-I

The transfer functions involved are

1. Process variable (drum level) is considered to be

( )

2. Valve function is considered to be

( )

25

3. The disturbance function is taken to be

( )

As per the IMC methodology,

( ) ( )

The IMC function is obtained to be

( )

( ) ( )

( )

( )

̃ ( )

( )

The filter function

( ) is taken as

( )

( )

For convenience, the value is considered to be λ=0.2 and n=4.

(4.10)

(4.11)

(4.12)

Substituting all the values in Q(s), the value of Q(s) is found to be

( )

The output is observed for two different cases:

(i) When no H(s) is provided i.e. unity feedback control H(s) = 1

(4.13)

(ii) H(s) i.e. feedback is used.

Here, the feedback is considered to be

( )

The block diagram for unity feedback system is shown in Fig.4.2 and the response is shown in Fig.4.3.

26

Fig 4.2 .Unity feedback for IMC

Fig 4.3. Response for Unity feedback of IMC

The block diagram for feedback system is shown in Fig.4.4 and the response is shown in

Fig.4.5.

27

Fig 4.4. Feedback for IMC

Fig 4.5. Response for feedback of IMC

The above procedure is repeated for a different value of λ. Taking λ =0.1, the disturbance function is taken as the same and response is observed for both unity feedback system as well as system with H(s) as feedback.

Fig 4.6 shows the response for unity feedback system .Fig 4.7 shows the response for a system with feedback.

28

Fig 4.6. Response for Unity feedback of IMC Fig 4.7. Response for IMC with feedback

CASE-II

The transfer functions involved are

1. Process variable (drum level) is considered to be

( )

2. Valve function is considered to be

( )

3. The distribution function is taken to be

( )

The output is observed for system with unity feedback as well as a system with H(s) as feedback.

( )

After substituting the values, the IMC controller function is found to be

( )

29

Fig 4.8 shows the response for unity feedback system .Fig 4.9 shows the response for a system with feedback.

Fig 4.8. Response for Unity feedback of IMC Fig 4.9. Response for IMC with feedback

4.2 IMC BASED PID CONTROL METHOD

The previously discussed IMC method, cannot handle the open loop unstable system. When

IMC structure is rearranged to form a standard feedback control system it can easily handle the situation.

The IMC based PID structure uses a standard feedback which uses a process model in an implicit manner i.e. PID tuning parameters are often adjusted based on the transfer function model but it is not always clear how the process model affects the tuning method.

Also, for open loop unstable processes, it is unnecessary to implement the IMC strategy in the standard feedback form, because IMC suffers from internal stability problems. Though

IMC based PID controller will not give the same performance when there are process time delays because the IMC based PID procedure uses an approximation for dead time. But if the process has no time delay, the input does not hit a constraint, then IMC based PID gives the same performance as IMC.

4.2.1 THE EQUIVALENT FEEDBACK FORM TO IMC

The standard feedback equivalence to IMC can be derived by using block diagram manipulation. Hence the standard feedback controller equivalent to IMC involves both the internal model

̃ ( )and external model controller, ( )

30

( )

̃( ) ( )

Fig. 4.10 IMC BASED PID CONTROLLER

4.2.2. THE IMC BASED PID CONTROL DESIGN PROCEDURE

(1) The IMC controller transfer function = Q(s), which includes a filter (s) to provide the derivative action. For unstable processes or better disturbance rejection, a filter of following form is used:

( )

( )

(2 ) The equivalent standard feedback controller is found

( )

( )

̃ ( ) ( )

(3) To show the above in PID form, are found out.

( ) [

( )

] [ ]

(3.8)

(3.9)

(4) The value of λ is inserted .Initial values will be generally around 1/3 to ½ of the dominant time constant.

4.2.3 RESULT AND SIMULATION OF IMC BASED PID CONTROLLER

CASE I: The transfer functions involved are

1. Process variable (drum level) is considered to be

( )

31

2. Valve function is considered to be

( )

3. The distribution function is taken to be

( )

4. The feedback function is taken to be

( )

5. The filter parameters are considered to be (λ=0.2 , n=4)

After substituting the values, the IMC controller transfer function Q(s) and equivalent standard feedback controller are found to be

( )

( )

The Block Diagram for IMC based PID with Unity feedback is shown in Fig.4.11.The response for IMC based PID with Unity feedback is shown in Fig.4.12

Fig 4.11. Block Diagram for IMC based PID using Unity Feedback

32

Fig 4.12. Response for IMC based PID with Unity feedback

CASE-II: The transfer functions involved are

1. Process variable (drum level) is considered to be

( )

2. Valve function is considered to be

( )

3. The distribution function is taken to be

( )

6. The feedback function is taken to be

( )

7. The filter parameters are considered to be (λ=0.2 , n=4)

After substituting the values, the IMC controller transfer function Q(s) and equivalent standard feedback controller are found to be

( )

33

( )

The response for IMC based PID with Unity feedback is shown in Fig.4.13

Fig 4.13. Response for IMC based PID with Unity feedback

4.3 IMC BASED PID TUNING FOR UNSTABLE PROCESSES

If the process has no time delay and input do not hit a constraint, then the IMC based PID controller will have the same performance as IMC does. But, if there is a dead time, then

IMC based PID controllers will not perform as well as IMC because of Pade’s approximation of time delay. The IMC design method was modified to handle unstable processes. The standard IMC structure cannot handle unstable processes, so the controller for an unstable process is implemented using standard PID feedback system.

4.3.1. IMC BASED PID PROCEDURE

(1)The IMC controller transfer function was found out,

( ) , which includes a filter

( ).The controller transfer function ( ), may be proper or improper giving rise to the resulting PID controller derivative action.

If the process has a time delay, then Pade’s approximation is used. For, good set point tracking, filter with the following form is generally used.

( )

( )

34

For improved disturbance rejection or integrating and open loop unstable process, a filter of following form is used

:

( )

( )

( )

For disturbance rejection,

is selected so that the filter requirement is satisfied ( ).

(2) The IMC based PID controller is found out using transformation.

( )

( )

̃( ) ( )

The above equation is written in the form of ratio of two polynomials

(3) are found out. The procedure results in an ideal PID controller cascaded with a first order filter with a filter time constant ( )

( ) [

( )

] [ ]

(4)The closed loop simulation for both perfect model case and case with model mismatch are performed. The value of λ is performed as a tradeoff between performance and robustness.

Another filter called set point filter is also used for unstable processes tuning for better set point tracking. It is described as,

( )

( )

(4.14)

4.3.2 TUNING FOR IMC BASED PID FOR UNSTABLE SYSTEM

r(s) +

( )

( ) ( )

_

Fig 4.14: IMC based PID with a set point filter

Y(s)

35

TABLE 4.1

PID TUNING PARAMETERS IMC CONTROLLER FOR UNSTABLE SYSTEM

A

B

C

( )

( )( )

(

(

)

)( )

( )

( )

( )

( )

( )

( )

( )

( )

TYPE A

Consider a process variable of type A .

( ) [

]

The process consists of a time delay in the form

. Pade’s approximation for time delay can be used for such a process. The first order Pade’s approximation is described as:

(

(

)

)

(4.15)

Implementing the first order Pade’s approximation in process, the process transfer function will be:

( )

CASE I:

The value of λ is considered to be 0.1

36

From

( ), and

Hence,

( )

(( ) ) and

[ ]

From the table the value following values were obtained:

= [ ]=0.19 and

(

)

Hence substituting the values in the formula of

( ) from the table provides the value

( )

The Block Diagram of IMC based PID for unstable process is shown in Fig.4.15.The response is shown in Fig.4.16

Fig 4.15. Block Diagram for IMC based PID (unstable process) Unity feedback

37

Fig 4.16. Response for IMC based PID (unstable process)

CASE II:

The value of λ is considered to be 0.5

From

( ), and

Hence,

( )

(( ) ) and

[ ]

From the table the value following values were obtained:

= [ ]=0.75 and

(

)

Hence substituting the values in the formula of

( ) from the table provides the value

( )

( )

The response for IMC based PID tuning method of unstable processes is shown in Fig.4.17

38

Fig 4.17. Response for IMC based PID (unstable process)

TYPE B:

Consider a process variable of type B .

( ) [

]

The process consists of a time delay in the form

. Pade’s approximation for time delay can be used for such a process. The first order Pade’s approximation is described as:

(

(

)

)

Implementing the first order Pade’s approximation in process, the process transfer function will be:

( )

, ,

From the table the value following values were obtained:

= [ ]=1.96 and

( )

39

Hence substituting the values in the formula of

( ) from the table provides the value

( )

The response for IMC based PID tuning method of unstable processes is shown in Fig.4.18

Fig 4.18. Response for IMC based PID (unstable process)

40

CHAPTER 5

FUZZY LOGIC CONTROLLER

Overview of Fuzzy Logic

Fuzzy Logic based controller

LabVIEW Implementation

Simulations and Results

Fuzzy logic based PID controller

LabVIEW Implementation

Simulations and Results

Summary

41

FUZZY LOGIC CONTROLLER

This chapter describes about Fuzzy Logic and its application in different controllers.

Fuzzy logic controller poses the ability to mimic the human mind to effectively employ modes of reasoning that are approximate rather than exact. In this chapter, Fuzzy logic is first used as a simple controller where the error gets fuzzified. Secondly, fuzzy logic is used to fuzzify the input to proportional term of PID controller.

5.1 OVERVIEW OF FUZZY LOGIC

Fuzzy logic is a form of many-valued logic or probabilistic logic; it deals with reasoning that is approximate rather than fixed and exact. Unlike traditional binary sets

(where variables may take either true or false values) fuzzy logic variables have an extent of truth value that ranges in degree between 0 and 1. Fuzzy logic has been extended to handle the concept of partial truth, where the truth value may range between completely true and completely false. Moreover, when linguistic variables are used, these degrees may be managed by specific functions. Fuzzy logic incorporates a simple, rule-based IF A AND B

THEN C approach to a problem rather than attempting to model a system mathematically.

The fuzzy logic is based on an operator’s experience.

5.1.1 HOW DOES FUZZY LOGIC WORK?

Fuzzy logic requires some numerical parameters in order to operate like significant error and significant rate-of-change-of-error, but exact values of these numbers are usually not critical unless very responsive performance is required in which case empirical tuning would determine them.

Linguistic variable: Linguistic variable is a representation of a parameter in words or language rather than numerical values. By a linguistic variable we mean a variable whose values are words or sentences in a natural or artificial language. For example, Temperature is a linguistic variable if its values are linguistic rather than numerical, i.e., hot, not hot, very hot, quite hot, cold, not very cold and not very hot, etc., rather than 20, 21, 22, 23 degree etc.

5.1.2 CONSTRUCTING A FUZZY LOGIC CONTROLLER:

The steps for constructing a simple Fuzzy Logic Controller are as follows:

42

1. Fuzzification (Creating the membership functions): The membership function is a graphical representation of the amount of participation of each input. A fuzzy set is an extension of a crisp set. Crisp sets allow either full membership or no membership at all, whereas fuzzy sets allow partial membership. We first create linguistic variables and then decide, on the basis of an expert’s opinion or experienced operator, which values of the parameter, belongs to or is a member of which fuzzy set and how much of it is present or absent in that set.

2. Creating the rule base table: Fuzzy inference systems consist of if–then rules that specify a relationship between the input and output fuzzy sets. Fuzzy relations present a degree of presence or absence of association or interaction between the elements of two or more sets. Linguistic rules describing the control system consist of two parts; an antecedent block (between the IF and THEN) and a consequent block (following

THEN). Evaluation is usually done by an experienced operator, fewer rules can be evaluated, thus simplifying the processing logic and perhaps even improving the fuzzy logic system performance. The inputs are combined logically using the AND operator.

3. Determine your procedure for defuzzifying the result: The fuzzy outputs for all rules are finally aggregated to one fuzzy set. To obtain a crisp decision from this fuzzy output, we have to defuzzify the fuzzy set. Therefore, we have to choose one representative value as the final output. There are several defuzzification methods, one of them is e.g. to take the center of gravity of the fuzzy set which is widely used for fuzzy sets. Fuzzy logic operators, fuzzy if–then rules, aggregation of output sets, and defuzzification.

An Fuzzy Interference System (FIS) with multiple outputs can be considered as a collection of independent multi-input, single-output systems. A general model of a fuzzy inference system is shown in Fig. 5.1

Input

Fuzzifier Deffuzifier

Output

Interference

Engine

Engine

Rule base

Fig. 5.1: Block diagram of a Fuzzy Interference System

43

5.2 FUZZY LOGIC BASED CONTROLLER

It provides an algorithm which provides the expert knowledge into automatic control strategy. Fuzzy control systems are rule-based systems in which set of fuzzy rules represent a control decision mechanism for adapting certain stimuli. Fuzzy controllers consist of an input stage, processing stage and an output stage. A general model of a fuzzy logic controller is shown in Fig. 5.2

Process

Sensors

Analog (crisp) to fuzzy interface

(Fuzzification)

Interference

Mechanism

Actuators

Fuzzy to Analog

(crisp) interface

(De-Fuzzification)

Fuzzy Rule base

Fig. 5.2 Fuzzy logic control system

The controller consists of:

1. Input Fuzzifier: It converts the crisp or analog input to a linguistic variable using the membership functions stored in the fuzzy system. Input data are fuzzified by converting them into membership functions. The membership function usually plays a key role in fuzzy computing. The most common shape for membership function graphs are triangular, although trapezoid and gaussian curves can also be used.

2. Rule Base: It holds a set of if then rules as already mentioned earlier, which are quantified through appropriate fuzzy sets.

If M = the number of fuzzy sets,

44

N = the number of system variables.

Then number of rules are .

At a time one or more rules may be fired and the according output are taken into consideration.

3. Fuzzy Interference System: It decides which rules are relevant for a particular system and applies action indicated by the rules. Its functions involve:

Determining the extent to which each rule is relevant to the current situation.

Drawing conclusions using current input and rule base.

4. Output De-fuzzifier: It is the procedure to convert a fuzzified term into deterministic crisp value. This is called de-fuzzification. There are various methods of defuzzification like Centre of gravity (COG) method, maxima method etc.

5.2.1 FUZZY LOGIC CONTROLLER FOR BOILER SYSTEM

In our case we consider a two input variable system. The two inputs are:

Error

Change in error

A simple fuzzy logic controller is shown in Fig. 5.3

I/P

+

-

Δe(t) e(t)

Fuzzy Controller

Fig. 5.3 A simple fuzzy logic controller

( )

O/P

We use a mamdani based fuzzy interference system for the controller. The steps involved are:

We take the input values (error and change in error) and find where they intersect their sets.

The intersection creates a cuts-off line known as the alpha-cut.

We fire our rules to find the corresponding output rule.

The rule is then cut off by the alpha-cut, giving us several trapezoidal shapes.

These shapes are added together to find their total center of gravity.

45

5.2.2 LabVIEW IMPLEMENTATION

We create the membership functions for both the inputs (error as Error and change in error as rate_Err) in LabVIEW environment.

e(t)

Fuzzy

Interference

System

u(t)

Δe(t)

Fig. 5.4 Fuzzy logic controller using two input variables.

The linguistic variables for the membership functions are described below.

TABLE 5.1

LINGUISTIC VARIABLES for Fuzzy logic Controller

NB

NM

NS

ZO

PS

PM

PB

Negative Big

Negative Medium

Negative Small

Zero

Positive Small

Positive Medium

Positive Big

For input variable Error ( e(t) ), Change in Error ( Δe(t) ) and controller ( u(t) ) ,the range is normalized between -1 to 1.

Fig.5.5 Membership Function for Error (e(t)) Fig. 5.6 Change in Error (Δe(t)).

46

Fig. 5.7 Membership Function for output variable, CONTROLLER (u(t)).

Rule base: Fuzzy logic uses linguistic representation of engineering knowledge to implement a control strategy. The rules used by an operator or expert to maneuver a process are described by IF-THEN rules in fuzzy logic.

TABLE 5.2

IF-THEN rules for Fuzzy Interference System for Fuzzy logic Controller

u(t)

e(t)

NB NM NS ZO PS PM PB

NB

NB NB NB NB NM NS ZO

NM

NB NB NB NM NS ZO PS

NS

NB NB NM NS ZO PS PS

Δe(t)

ZO

NB NM NS ZO PS PM PM

PS

NM NS ZO PS PS PB PB

PM

NS ZO PS PM PB PB PB

PB

ZO PS PM PB PB PB PB

The Block diagram and the fuzzy interference system is implemented using LabVIEW and

Fuzzy System Designer.

Here is an instance of the HTML report on our LabVIEW PROJECT.fs file:

47

48

49

Block Diagram used for the Fuzzy logic controller is shown in Fig. 5.8

Fig. 5.8 Block Diagram - Fuzzy logic controller

5.2.3 SIMULATIONS AND RESULTS

We have used two transfer functions here (

( ) ).

 ( )

 ( )

The step responses for the above mentioned transfer functions are shown below in Fig. 5.6 and Fig 5.7

Fig. 5.9: Step Response for

( )

50

Fig. 5.10: Step Response for

( )

5.3 FUZZY LOGIC BASED PID CONTROLLER

Proportional-Integral-derivative (PID) controllers are most adapted techniques in practical cases due to their simple structure, easy to tune quality and they provide acceptable performances for a large range of processes. However, real systems bring problems like higher order, non-linearities, dead-time which are affected easily by noise, load disturbances and environmental conditions. For such cases, tuning is done. We have already discussed

Zeigler Nichols and Tyreus Luyben tuning method in Chapter 3.

Here, we will implement a method to improvise Zeigler Nichols method of PID tuning.

The idea of multiplying the set-point value for the proportional action by a constant parameter less than one is effective in reducing overshoot.

We know, for PID,

( ) ( ( )

( )

∫ ( ) )

Where,

( ) = system error;

( ) control variable;

(5.1)

proportional gain;

derivative time constant;

integral time constant.

51

Using the Ziegler–Nichols formula generally results in good load disturbance attenuation but also in a large overshoot and settling time for a step response that might not be acceptable for a number of processes. Increasing the analog gain generally highlights these two aspects. An effective way to cope with this problem is to weight the set-point for the proportional action by means of a constant (b<1), so that we get the following modified equality,

( ) (

( )

( )

∫ ( ) ) where,

( ) ( ) ( )

(5.2)

Such that,

( )

( )

and,

( )

( )

The modified PID controller block diagram is in Fig. 5.11

(5.3)

(5.4)

( )

( )

Process

-1

Fig. 5.11 PID controller with fixed set point weighting

However, the use of set-point weighting generally leads to an increase in the rise time since the effectiveness of the proportional action is somewhat reduced. To achieve both the aims of reducing the overshoot and decreasing the rise time, a fuzzy module can be used to modify the weight b. The significant drawback can be avoided by using a fuzzy inference system to determine the value of the weight (b(t)) depending on the current value of the system error

(e(t)) and its time derivative (

̇( ))

For the sake of simplicity, the methodology is implemented in such a way that the output of the fuzzy module is added to a constant parameter resulting in a coefficient that multiplies the set-point. The overall control scheme is shown in Fig. 5.12

52

+

-

e(t)

Δe(t)

Fuzzy

Controller

*

+

-

P

I

D

Fig. 5.12 PID controller with fuzzy based set point weighting

Process

5.3.1 LabVIEW IMPLEMENTATION

The block diagrams for both the controllers are implemented using LabVIEW and the Fuzzy

System Designer.

The meaning of the linguistic variables is explained in Table 5.3.

TABLE 5.3

LINGUISTIC VARIABLES for Fuzzy based PID Controller

NVB

NB

NM

NS

Z

PS

PM

PB

PVB

Negative Very Big

53

Negative Big

Negative Medium

Negative Small

Zero

Positive Small

Positive Medium

Positive Big

Positive Very Big y

TABLE 5.4

IF-THEN rules for Fuzzy Interference System for Fuzzy based PID Controller

u(t)

e(t)

NB NS Z PS PB

NB NVB NB NM NS Z

NS

NB NM NS Z PS

Δe(t)

ZO

NM NS Z PS PM

PS

NS Z PS PM PB

PB

Z PS PM PB PVB

We consider six groups of systems, with different values for the parameters, to test the effectiveness of the methodology.

The following transfer functions with the indicated values of the parameters have been considered:

( )

;

( )

( )

( )

( )

; T = 1, 10; L = 0.1, 0.4, 0.8

( )

( )

;

( )

( ) ( ) ( ) ( )

; L = 0, 0.1

( )

( )

( )

;

The LabVIEW block diagram implementation in shown below in Fig. 5.13

54

Fig. 5.13 Block Diagram for comparison of fixed set point weighting and fuzzy set point weighting for PID controller.

5.3.2 SIMULATIONS AND RESULTS

The unit step response was simulated with LabVIEW and Fuzzy System Designer for all the processes. A comparison between the responses obtained with the classical Ziegler–Nichols method, a fixed value of the weight b and a fuzzified b was made by evaluating the overshoot percentage ( ), rise time ( ) and settling time (

). The step responses of the different processes for the two controllers are shown below.

The repesentations are:

Fuzzy b

Fixed b

Fig. 5.14 :

( )

Fig. 5.15 :

( )

55

Fig. 5.16 :

( )

( ), T=1, L=0.1

Fig. 5.18 :

( ), T=1, L=0.4

Fig. 5.19 :

( ), T=1, L=0.8

Fig. 5.20 :

( ), T=10, L=0.1

56

Fig. 5.21 :

( ), T=10, L=0.4

Fig. 5.22 :

( ), T=10, L=0.8

Fig. 5.23:

( )

Fig. 5.24:

( )

Fig. 5.25

( )

, L=0

57

Fig. 5.26:

( ), L=0.1

The performance of fixed b and fuzzy b incorporated controllers are evaluated using the considered groups of system and their comparative step responses are shown in Fig.14 to

Fig. 5.26 respectively.

From the graphs above we can calculate the overshoot and rise time along with settling time and steady state error. We have calculated the overshoot, rise time and settling time mainly and analyzed if the desired and expected results are seen.

The overshoot (%), rise time (s) and settling time (s) are reported in Table 5.5, 5.6 and

5.7 for the two controllers, respectively.

TABLE 5.5

Value of overshoot (%) using a fixed weight b and a fuzzified weight b.

Transfer function

( ), ξ=0.8

( ), ξ=0.2

( )

( ), T=1, L=0.1

( ), T=1, L=0.4

( ), T=1, L=0.8

( ), T=10, L=0.1

( ), T=10, L=0.4

( ), T=10, L=0.8

( )

( ), L=0

( ), L=0.1

( )

7

3

1

15

6

3

14

(In %)

Fixed b Fuzzy b

10 1

0

10

0

2

0

8

1

2

3

0.5

0

2

0

5

0.5

0.5

2.5

We can see that the rise time decreases whilst the overshoot does not significantly increase with respect to the fixed b

58

TABLE 5.6

Value of rise time using a fixed weight b and a fuzzified weight b.

Transfer function

( ), ξ=0.8

( ), ξ=0.2

( )

( ), T=1, L=0.1

( ), T=1, L=0.4

( ), T=1, L=0.8

( ), T=10, L=0.1

( ), T=10, L=0.4

( ), T=10, L=0.8

( )

( ), L=0

( ), L=0.1

( )

Fixed b

0.2

1.4

0.5

2.5

0.6

1

0.5

1.5

1.4

1

0.8

1

1

0.1

1.3

0.7

2.6

1.5

1

1.5

Fuzzy b

0.6

0.9

0.5

0.6

2

1.2

TABLE 5.7

Value of settling time using a fixed weight b and a fuzzified weight b.

Transfer function

( ), ξ=0.8

( ), ξ=0.2

( )

( ), T=1, L=0.1

( ), T=1, L=0.4

( ), T=1, L=0.8

( ), T=10, L=0.1

( ), T=10, L=0.4

( ), T=10, L=0.8

( )

( ), L=0

( ), L=0.1

( )

5

3.7

8.4

4

4

6.2

Fixed b

5.5

5.6

3.4

3.8

5.7

6.5

4

3

2.7

5

3.7

3.7

6.8

Fuzzy b

2.5

3.4

2

2.6

7

4.1

3.4

59

5.4 SUMMARY

The Fuzzy Logic Controller is a powerful tool in controlling instruments and equipment. Using Fuzzy Logic can quickly lead to more efficient, precise and accurate controls. The benefits of Fuzzy Logic Controllers are simple; they can be more accurate and more precise than their PID counterparts. The Simple Fuzzy Logic Controller cut down the effective error by nearly half compared to the PID control. Further, there is better control on the overshoot as well as rise time, thus creating a far more precise control.

The use of the fuzzy set-point weighting, in conjunction with the Ziegler–Nichols method for the tuning of the proportional gain and the integral and derivative time constants, leads to a significant improvement in the step response and preserves the good performances in the attenuation of the load disturbance assured by the Ziegler–Nichols formula. It can be shown that the approach is also robust with respect to the controller parameter variations. The approach has been shown to be very effective in the set point following for a large number of processes, whilst the load disturbance attenuation performances obtained by the use of the

Ziegler–Nichols formula are preserved or improved.

We can see that the rise time decreases whilst the overshoot does not significantly increase with respect to the fixed b. This can be deduced by examining Tables 5.5, Table 5.6 and

Table 5.7 where the values of the overshoot (%), rise time (s) and settling time (s) are reported for the two controllers, respectively. The devised control structure seems to be particularly appropriate to be adopted in industrial settings, since it requires a small computational effort, it is easily tuned and it is compatible with a classical PID controller; that is, it consists of a module that can be added or excluded without modifying the parameters of the existing PID.

60

CHAPTER 6

CONCLUSION

61

CONCLUSION

The project work basically evaluates the performance of the different control strategies of a boiler drum level control system and also the implementation of different controllers. The performance of a single element drum level control and a two element drum level control are studied using Zeigler Nichols and Tyreus Luyben PID controller tuning methods. It can be observed that two element control, being a feedback plus feedforward control shows better results as compared to single element drum level control which is a feedback control only. The overshoot, steady state error as well as the settling and rise time are found to be better.

Apart from Zeigler Nichols and Tyreus Luyben tuning methods, we have also implemented

IMC controller for the drum level control. The IMC controller is found to get even better and smoother results than simple PID controller. However due to the fact that most industries used PID controller, since it is easy to tune and has a simple structure, replacing PID with

IMC is an expensive option. Hence we modified PID with IMC based PID and analyzed the results. For unstable processes with dead time involved, we have done tuning=g of the IMC based PID controller. The results are observed and discussed.

Another novel approach for a better controller was using fuzzy logic to control the drum level. A simple fuzzy logic controller with a 7x& rule base was implemented and the corresponding results were studied. Also for the same reason as in case of IMC, fuzzy based

PID controller is studied. In our project work, we have used fuzzy logic to fuzzify a weight given to the proportional term of a PID controller. The results of the modified PID controller, using six group of different processes, are compared to a constant weighted PID controller.

The decrease in overshoot was observed as well as the proportional characteristic of decrease in rise time was also observed to be preserved.

62

PUBLICATIONS

1. Pragyan P. Kar, Priyam Saikia and Umesh C. Pati , “Performance of Different

Control Strategies for Boiler Drum Level Control using LabVIEW” , communicated for publishing to Instrument society of India.

63

REFERENCES

[1] Subhransu Padhee, Yaduvir Singh “Data Logging and Supervisory Control of Process

Using LabVIEW”, Proc. of the IEEE Students Tech. Symp., Jan 2011, pp. 329-334.

[2] Mihai Iacob, Gheorghe-Daniel Andreescu, and Nicolae Muntean “Boiler-Turbine

Simulator with Real-Time Capability for Dispatcher Training Using LabVIEW”, 12th

International Conference on Optimization of Electrical and Electronic Equipment, OPTIM

2010, pp. 864-868.

[3] Antonio Visioli, “Fuzzy Logic Based Set-Point Weight Tuning of PID Controllers”,

IEEE transactions on systems, man, and cybernetics—Part A: systems and humans, Nov

1999, pp. 587-592

[4]

National Instruments, “LabVIEW Control Design Toolkit User Manual”, www.ni.com/pdf/manuals/371057d.pdf, Sept 2004

64

Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement