Docstoc

IN-16-02

Document Sample
IN-16-02 Powered By Docstoc
					Abu Dhabi National Oil Co. ADNOC Technical Institute

INSTRUMENTATION
PROCESS CONTROL FUNDAMENTALS

UNIT 2 TUNING A CONTROLLER

TUNING A CONTROLLER DATE OF ISSUE 8-DEC-09

PAGE 1 OF 9 IN-16-02

ADNOC TECHNICAL INSTITUTE

UNITS IN THIS COURSE

UNIT 1

BASIC CONTROL THEORY

UNIT 2

TUNING A CONTROLLER

UNIT 3

INTRODUCTION TO DCS AND PLC

UNIT 4

HONEYWELL TDC 3000 DCS

UNIT 5

FOXBORO IA DCS

UNIT 6

PRACTICAL TASKS

TUNING A CONTROLLER DATE OF ISSUE 8-DEC-09

PAGE 2 OF 9 IN-16-02

ADNOC TECHNICAL INSTITUTE

TABLE OF CONTENTS
Paragraph 2.0 UNIT OBJECTIVES 2.1 INTRODUCTION 2.2 WHAT IS GOOD CONTROL 2.3 TUNING A LOOP Page 4 5 5 7

TUNING A CONTROLLER DATE OF ISSUE 8-DEC-09

PAGE 3 OF 9 IN-16-02

ADNOC TECHNICAL INSTITUTE

2.0 UNIT OBJECTIVES The student will be able to   Describe good loop control using simple diagrams. Explain how the PID settings of a controller are set to obtain good control.   Describe modern tuning methods. Explain how you can get the best results by co-operating with the panel operator.

TUNING A CONTROLLER DATE OF ISSUE 8-DEC-09

PAGE 4 OF 9 IN-16-02

ADNOC TECHNICAL INSTITUTE

2.1 INTRODUCTION The aim of this unit is to introduce the basics of controller tuning.

2.2 WHAT IS GOOD CONTROL

CU

PROCESS

MV

P

I

D
MV-SP

CONTROLLER

ERROR DETECTOR SET POINT (SP)

100%
MEASURED VALUE

SET POINT CHANGE RESPONSE TIME

O TIME

Figure 2-1 Good Control
Figure 2-1 shows a typical control loop The job of the loop is to keep the measured variable (MV) at the set point. The loop is called a feedback control system. If flow is the process variable, it works as follows.

TUNING A CONTROLLER DATE OF ISSUE 8-DEC-09

PAGE 5 OF 9 IN-16-02

ADNOC TECHNICAL INSTITUTE

If the flow rate increases the error detector sends a signal to the controller. This signal indicates how much the measured value is more than the set point (MV-SP). The controller then adjusts the correcting unit (CU) so that the MV decreases and the flow rate goes down to the set point. If the flow rate decreases to below the set-point, then the error signal is (SP-MV). In this case the controller adjusts the correcting unit (CU) so that the MV increases and the flow rate goes up to the set point. When you are tuning a loop you cannot wait for the variable to change. So, the set point is changed and the MV moves to the new set point. The effect on the control loop is the same. The graph shows the tuning of the three mode controller (PID) to give good control for a step set point change. The response time should be fast and the MV should only go a little over the SP before it stabilises, (small overswing). In some processes it is important to tune the system so that there is no overswing. A slow and smooth response is needed. However, some processes need a fast response and quite big overswings are no problem. The point is you must decide what kind of control is good for each specific loop.

TUNING A CONTROLLER DATE OF ISSUE 8-DEC-09

PAGE 6 OF 9 IN-16-02

ADNOC TECHNICAL INSTITUTE

2.3 TUNING A LOOP If you must tune a loop it is best to tune it when its on automatic (closed loop). It is possible to tune it on manual (open-loop) but this can be dangerous. The operator may not allow any open loop tuning. For any feedback control system, if the loop is closed (the controller is on automatic), you can increase the controller gain. As you do this, the loop will start to swing more and more. As you continue to increase the gain, you will see continuous cycling (oscillation) in the controlled variable. This is the maximum gain at which the system may be operated before it becomes unstable. The period of these continuous oscillations is called the ultimate period (see Figure 2-2).
100%
CURVE A: UNSTABLE CURVE B: CONTINUOUS CYCLING, READ MAXIMUM GAIN & PERIOD CURVE C: STABLE

CURVE A

CURVE B

O U T P U T

CURVE C

PU
ULTIMATE PERIOD

0
TIME

Figure 2-2 Control Loop Response
TUNING A CONTROLLER DATE OF ISSUE 8-DEC-09 PAGE 7 OF 9 IN-16-02

ADNOC TECHNICAL INSTITUTE

To determine the maximum gain and the ultimate period, take the following steps: 1. Tune out all the reset and derivative action from the controller, leaving only the proportional action. This means you should set T1 equal to infinity and Td equal to zero on the controller (or as close to these values as possible). 2. Maintain the controller on automatic i.e., leave the loop closed. 3. Set the gain of the proportional mode of the controller at any value. Then make a disturbance on the process and see what happens. One easy way of making a disturbance is to move the set point for a few seconds and then return it to its original value. 4. If the response curve from step 3 does not damp out, as in curve A (see Figure 2-2), it means the gain is too high (the proportional band setting is too low). The gain should be decreased by increasing the proportional band setting. Then you repeat step 3. 5. If the response curve from step 3 stops swinging, as in Curve C (see Figure 2-2), it means the gain is too low (the proportional band setting is too high). The gain should be increased by decreasing the proportional band setting. Then you repeat step 3. 6. If the response from step 3 cycles continuously, as in Curve B (see Figure 2-2), it means you have best possible gain (optimum proportional band setting). The "ultimate period" of the response curve should be noted. This is the maximum gain at which continuous oscillations are maintained. The ultimate period is written as PU. The values obtained from step 6 are then used to set the PID of the controller. The standard method for setting the values of PID is the Ziegler and Nichols method as follows:

TUNING A CONTROLLER DATE OF ISSUE 8-DEC-09

PAGE 8 OF 9 IN-16-02

ADNOC TECHNICAL INSTITUTE

Proportional only
Set the gain to half the maximum gain (or twice the PB setting)

Proportional plus reset.
Set the gain to 0.45 of the maximum gain (or 2.22 times the PB setting) Set the reset to read 0.83PU

Proportional plus reset plus derivative
Set the gain to 0.6 of the maximum gain (or 1.7 times the PB setting) Set the rest to read 0.5 PU. Set the rate (derivative) to read 0.13 PU If you follow the above procedure it will produce a reasonably tuned loop. However, it may need to be adjusted during operation. The way the loop responds to real changes in the process must be monitored by the operator. required. On older control loops the response to process changes can be seen on the ink/paper recorder. A modern DCS system displays process variables on the VDU, using what are called "trend" displays. You must ask the operator to show these displays when you check the loop. Normally, it is possible to print these "trend" displays as the information is stored on a hard disc (the historian). You can usually get a copy of what the loop has done for the last 24 hrs. It can take hours for a loop to settle down to normal operational requirements. The operator must decide if any fine tuning is

TUNING A CONTROLLER DATE OF ISSUE 8-DEC-09

PAGE 9 OF 9 IN-16-02