You are on page 1of 13

EXPERIMENT 6 INTRODUCTION TO PID CONTROLLER

SUMMARY
The main objective of this laboratory is to get a basic understanding for how feedback control can be used to modify the behavior of a dynamic system. In particular, we will consider PID-control of a simple process consisting of two water tanks. The letters PID here stand for Proportional, Integral and Derivative control, respectively. PID controllers are by far the most common type of controllers used in industrial systems, mainly because they are relatively simple and still often able to provide good performance. The laboratory experiments will hopefully conrm this. In addition to establish an understanding for the fundamental principles of feedback control, the laboratory will provide some experience of manually tuning PID-controllers. Based on the input supplied from the controller to the system, result will be obtained in form of graph. This graph can be printed later from the computer. There are several conditions that the students will do as to know what the result will be like. Therefore, at the end of this experiment, student will understand on what will happen if some parameters are varied.

OBJECTIVES
Learn the importance of the vital system characteristics in the assessment of control loop efficiency Learn on how to evaluate the PID elements using the PCU computer controlled cycle

THEORY
PID CONTROLLER
A proportionalintegralderivative controller (PID controller) is a generic control loop feedback mechanism (controller) widely used in industrial control systems a PID is the most commonly used feedback controller. A PID controller calculates an "error" value as the difference between a measured process variable and a desired set point. The controller attempts to minimize the error by adjusting the process control inputs. The PID controller calculation (algorithm) involves three separate constant parameters, and is accordingly sometimes called three-term control: the proportional, the integral and derivative values, denoted P, I, and D. Heuristically, these values can be interpreted in terms of time: P depends on the present error, I on the accumulation of past errors, and D is a prediction of future errors, based on current rate of change. The weighted sum of these three actions is used to adjust the process via a control element such as the position of a control valve, or the power supplied to a heating element.

PROPORTIONAL CONTROL MODE

In this mode the output of the controller is proportional to the error between the set point and the measured value. Proportional control may be expressed as either proportional gain or proportional band. Mathematically,

Mp =PG(SP-MV)+C = PG e(t) +C
Where,

Mp = Controller Output PG = Proportional Gain SP = Set point MV = measured value C =Output with zero error e (t) = Error as a function of time.

The error band where the output is between 0% and 100% is called the proportional band (PB), and given by PB = 100/PG. Thus, the higher the gain, the smaller the band.
2

This control mode rarely produce adequate control, where there usually an offset.
INTEGRAL MODE

This mode of control is often used to remove proportional offsets errors. The integral mode determines an output based on the history of error. It is calculated by finding the net area under the error curve versus time and multiplying by a constant called the integral action time (IAT) in seconds. The controller output equation is:

Mi( t )

PG e( t )dt IAT

The integral Action time is defined as the time taken for the integral action to duplicate the proportional action of the controller, if the error remains constant during this period. It is used commonly to remove any steady state errors incurred when using a proportional controller.

DERIVATIVE CONTROL MODE

Derivative control mode is often used to reduce the response time of the system; it is

de( t ) dt based on the time rate of the change of error. The time taken for the proportional action to duplicate the instantaneous output of the derivative element is called derivative action time (DAT). Md PG DAT
The controller output equation is: The derivative control mode is never used alone as there is no controller output corresponding to zero rate of change. So it is commonly used with Proportional controller (PD). However, it can also exaggerate high frequency noise in the system.

SYSTEM RESPONSE

Figure 1 shows the typical system response of a control system. There are three types of response for a second order system, which are over damped, under damped, and critical damped response. The system response depends on the PID gains set in the experiment. The characteristic of the response is shown in Figure 2

Figure 1 Graph of controlled variable versus time

Figure 2 Graph of step input versus time

Some of the important system performance parameters are: Peak overshoot: is often expressed as percent overshoot at the first peak and given by (Peak value- input value)/input value * 100 Settling time: The time taken to settle within 2% of the final value Rise time: The time taken for the system to respond to a fraction of the final value on the initial part. Typically 5-95% or 10-90%. Steady state error: Any error between the set point and the controlled variable once the system has stabilized.

APPARATUS
THE SYSTEM RIG

The System Rig is the hardware for the process, which is to be controlled by the microcomputer. The unit is based around a fluid flow process, where flow and temperature may be controlled. This reflects a typical process control situation such as in the food and drink manufacturing petrochemical industry.

Each feature on the System Rig has a manual or computer control option. Users may select either of the modes allowing a comparison between human and computer control operation to be made. This allows a rapid appreciation of the advantages and disadvantages under both modes of control.

DESCRIPTION

A - Mains switch B - Water pump switch C - Bottom reservoir tank D - Bypass valve E - Return valve F - Water level tank

G - Overflow pipe H - Proportional valve I - Water inlet port J - Water drain port K - Water pump L - Control panel M - Level foot

FEEDBACK

Feedback is an essential requirement for the control of any process. It consists of various transducers measuring the conditions on the rig and feeding this information back to the controlling microcomputer.

On the Process Control Unit the temperature at the sump, flow line and process tank are measured using platinum resistance thermometers. The flow rate is measured by an in-line flow meter. These analogue signals are fed back to the signal conditioners on the Computer Control Module (CCM) from where they are sampled by the microcomputer via an analogue to digital converter (ADC). LED meters are used to display the temperatures and flow rate on the system rig. Indicators are provided for the cooler, tank full sensor and drain/diviner solenoids, giving a status check when the Process Control Unit is in operation.

FLOW MEASUREMENT

Example of flow meter 7

The flow rate of the fluid is measured by means of a flow meter of the impeller type. The fluid flows through the meter rotating the impeller, which has six blades. Mounted either side of the impeller is an infra red transmitter and receiver producing an infra red beam which is broken by the rotating impeller. Six pulses are therefore produced for one revolution of the rotor, thus producing a frequency output 'which is proportional to the flow rate. The approximate full-scale frequency is 570Hz (pulses/sec) which is converted to a voltage by the signal conditioning circuit. This voltage is used to drive the flowrate LED display on the rig and also converted into a digital word by the Data Acquisition circuit.

Figure 5 process flow on the flow measurement system

PUMP

Example of centrifugal pump

The pump used is a centrifugal type. It is not a positive displacement type and thus its output is not necessarily linearly proportional to speed, though variation in speed will, of course, vary the output flow rate.
8

Activating Voltage : 12V D.C Maximum Continuous Current: 6 Amps

PROCEDURE
SOFTWARE OPERATION

a) Turn on both the computer system and the process control unit. b) Once you get into Windows click start, program, shut down, go to DOS mode. c) Type cd\pcu4 Enter. Then type pcu d) Once youre in the program this to familiarize yourself try these commands: F1 (to on/off), Yes, F1-F9 (to toggle and control each unit), then try out other commands too.
SECTION 1 : ASSESSMENT OF SYSTEM PERFORMANCE

a) By operating the Process Control Unit using computer controlled, the vital characteristics can be easily demonstrated by varying the tuning of PID controller. b) Click F2 (flow control), Yes (for control using PID), F2 (computer control) c) Set the Set point to 1.0 instead of 1.6 and Time length to 50 seconds instead of 100 seconds. d) Set the PID controller using the given values and your own values. Print out your results and observe the graphs. Label the graphs.

DATA, OBSERVATIONS AND RESULTS

Under Controller Setting Peak Overshoot (liter/min) Proportional Gain Integral Derivative Setting Time (s) Rise Time (s) Steady state error damped/ over damped/ Critical

R1

0.1

6.5

Over damped Under damped Critical

R2

3.5

0.01

7.8

40

5.9 6.8 5.5 6.8

R3

0.05

0.5

6.75

40

10

R4

14

0.025

0.5

7.5

30

10

4.8 6.7 1.0 3.4

Critical

R5

0.5

0.1

5.5

10

Critical

Table 1 Result obtained in this experiment

10

DISCUSSION

1- In this experiment, there are 5 graphs that were drawn. Namely R1, R2, R3, R4, and R5. All these graphs have different value of proportional gain, integral time and derivative time. For details about it, it can be explained here: a- Graph R1 It is in overdamped response. This is because, the system does not reach the set point through in 50 seconds. Plus, the graph also does not stop oscillating with a steady state value. b- Graph R2 This graph is in underdamped response. The reason behind this is because, it requires short time to reach the set point. Peak overshoot occurs furthermore it takes longer time to reach the steady state because of high oscillation occurs on high underdamped system. c- Graph R3, R4, R5 These three graphs are in critically damped response. This happens because the time in all these 3 graphs takes to reach steady state is fast due to the fact that they have linear oscillation. In ideal cases, there should be no oscillation at steady state value.

2- For case 1 Based on the graph, the proportional gain, P is changing for every 20 seconds and it becomes the only manipulated variable. It means, for every increase in gain, the response will also changes. On the graph, it can be seen that the most stable gain is at 7.2

11

3- For case 2 Based on this case 2 graph, we can simplify that the flow rate is increasing and closely match the steady state value. On the graph, at 30 seconds, the pass value is opened, the flow rate drop significantly and gradually increasing again until it reaches 1.3 lpm and becomes stable at 40 seconds of time until the end of process (maintained at 1.3 lpm). 4- For case 3 Based on this case 3 graph, the fixed value on the proportional gain is used as a reference point which is 7.2. The procedure conducted in this case 3 is the same as in case 1. Based on the graph, the optimum value at the integral time is 0.05 as the graph must stable within that interval. After that, it can be seen that the flow rate becomes stable showing that the optimum values at I and P are already determined. 5- For case 4 In this case, it is actually almost the same as with case 2 but the only thing different is, the value of P and I are fixed using the value that have been determined earlier. Those values are (P=7.2, I=0.05). Therefore, when the experiment is conducted, the flow rate amount almost reaches the steady state and becomes constant. After 30 seconds, disturbances was introduces again. 6- For case 5 Based on the graph obtained in this case, the value increases from 200 and becomes stable at first 20 seconds time. After that, the graph started to fluctuate until the end of the experiment. The optimum value achieved for D is 0 and the final value for others are P=7.2, I=0.05 and D=3. 7- For case 6 From the graph obtained, it shows the same result as what have been achieved in case 4. Some disturbances were introduced into the system. The flow rate drops and becomes stable again and maintained its stability until the end of the experiment.

12

CONCLUSION
At the end of this experiment, we managed to understand the importance at the vital system characteristics in this experiment of control loop efficiency and we have successfully learnt on how to evaluate the PID Control Circuit using the computer control flow cycle. We can conclude that the optimum PID controller can be determined through test and experiment. Therefore, we can say that we have fulfilled the objective of this experiment.

13