CONTROL ENGINEERING LABORATORY
MODUL 4: Water Level Control System PCT-100
Laboratorium Teknik Pengaturan Departemen Teknik Elektro Fakultas Teknologi Elektro Institut Teknologi Sepuluh Nopember
Water Level Control System PCT-100 OBJECTIVE To provide controlled process in a learning environment which reflects the control problems experience in industry. Students may know about control system in li quid level control and can carry out detailed analysis of alternative control techniques. To illustrate simply and clearly, the fundamental control techniques of proporsional, integral and derivative control. REFERENCE Ogata K. Modern Control Engineering. 2010 Anonymous. User’s Manual Process Control Technology-100 Ljung L. System Identification, Theory for the user 2 nd edition
EQUIPMENT REQUIRED 1 set Process Rig 1 set Control Module 1 set PC INTRODUCTION Process Control Technology (PCT-100) Process control is a branch of control engineering relating to the operation of the plant in an industries such as petrochemicals, foodstuffs, steel, glass, paper, energy, etc. The main objective is to maintain the stability of all variables in the process. Temperature, level, flow, and pressure are the four most common process variables. This four variables are key to process control because it provide a critical condition for boiling, chemical reaction, distillation, extrusion, vacuuming, and air conditioning. Bad control of this four variables can cause safety, quality, and productivity problems. Therefore, it is highly desirable to keep under control and maintained within its safety limits. Process Control Technology PCT-100 is an instrument used to demonstrate various aspects of process control. This equipment facilitate the process control via computer or Programmable Logic Controller (PLC). The main elements of the PCT-100 are the process rig and control module. Process Rig Process rig is the main elements of PCT-100 for a diagram of the PCT- 100 can be shown in Figure 2. PCT-100 includes the following elements: 1. Process tank 12. Pressure relief valve 2. Sump tank 13. Heater 3. Cooler Unit 14. Level sensor 4. Sump tank Temperature sensor (PRT) 15. Pressure transducer 5. Variable speed pump with filter 16. Float switch 6. 3/2 Diverter valve 17. Overvlow 7. 2/2 Proportional control valve 18. Digital LCD displays 8. Flow rate sensor 19. Indicator Light 9. One way check valve 10. 2/2 Proportional drain valve
2
Laboratorium Teknik Pengaturan Departemen Teknik Elektro Fakultas Teknologi Elektro Institut Teknologi Sepuluh Nopember
11. Needle valve
Figure 1. Process Rig of PCT-100. The sump contains a store of distilled water which may be pumped around the system at flowrates up to about 3.2 liters/minute. Water may be pumped directly from the sump tank to the process tank or be diverted via the cooler. The fluid in the process tank may be drained via the manual or computer controlled valves below the tank, so completing the fluid cycle. The five digital displays are used to show the sump tank and process tank temperatures, flow-rate, pressure and level and the indicator lamps reveal the on/off status of the cooler fan, and computer controlled drain and diverter valves and the heater. Control Module The control module incorporates all of the electronic ci rcuitry required to link the process rig to the controller. The design of the circuits demonstrates the interfacing principles required in many process control situations where a mix of analogue, digital and frequency signals have to be processed. Shown in Figure 3, the front of the Control Module has a schematic of the Process Rig, On/Off indicator, six illuminated fault switches, test points, indicators to show the operational status of the elements on the rig, and a backlight switch to turn on the backlights for the displays on the rig.
Figure 2. Front part of Control Module of PCT-100
3
Laboratorium Teknik Pengaturan Departemen Teknik Elektro Fakultas Teknologi Elektro Institut Teknologi Sepuluh Nopember
All connections to the process rig and power supply unit are made at the rear of the control module. The USB port on the PC is connected to the USB sockets on the control module and there are analog and digital I/O that is connected to the Process Rig as shown in Figure 4.
Figure 3. Rear part of Control Module of PCT-100
Heater Control Since the heating element is capable of heating the water in the process tank to a high temperature or causing damage to the tank, several safety features have been incorporated into the PCT-100 system. The safety requirements are as follows: The heating element must only be energized when the water level in the process tank is at the required level. If any connection between the computer and control module fails or the computer "crashes", the heating element must be switched off automatically. Tank full signal. The software supplied incorporates an inter lock to prevent power being applied to the heater until the signal from the level sensor shows that the process tank is the required level. If students are to write their own control software it is crucially important that they duplicate this software interlock in their program, before they use the heating element. Failures or crashes. If the computer-to-control module communication fails, the circuitry on the control module detects this and turns the heating element off. The heating element contained in the process tank is controlled by the computer using a pulse width modulation (PWM) technique. The required "mark/space ratio" is determined by a software algorithm in the case of computer control. As the mark/space ratio increases the average rate at which electrical energy is dissipated by the heating element increases , increasing the rate at which the water is heated. Pump Control The software initially outputs zero and gradually increases the output to whatever value is necessary to achieve the required flow rate. A PID algorithm utilizing feedback from the flowmeter re-calculates the output value at each sample interval. Switched Faults With the understanding of how systems operate within industrial control systems there is a great need for individuals with fault finding skills. To help develop these skills, PCT-100 incorporates
4
Laboratorium Teknik Pengaturan Departemen Teknik Elektro Fakultas Teknologi Elektro Institut Teknologi Sepuluh Nopember
six individually selectable switched faults located on the front panel of t he control module (see Figure 2. Front Part of Control Module). The switched faults are typical of those found in real industrial applications. Generally an electrical f ault may be due to one or more of the following: 1. Component failure 2. Electrical short circuit to supply. 3. Electrical short circuit to ground. 4. Crossed signal wires either due to a short or incorrect commissioning. 5. Open circuit due to a broken wire, burnt circuit track or bad connection. The student may be given fault finding tests following his observation of the correct operation of the PCT-100. The effect of each fault is shown in Table 1. Using standard test equipment in conjunction with circuit diagrams, the faults may be successf ully diagnosed. Table 1. List of switched faults
PID Controller PID controller has three main components there are the proportional or denoted by P , Integral or I, and a differential or D as shown in Figure 2. A typical structure of a PID control system
is shown in Equation 1, where it can be seen that in a PID controller, the error signal e(t) is used to generate the proportional, integral, and derivative actions, with the resulting signals weighted and summed to form the control signal u(t) applied to the plant model. A mathematical description of the PID controller is t 1 de(t ) u (t ) K p e(t ) e( )d d , dt i 0
(1)
Figure 4. Block Diagram of PID Controller.
5
Laboratorium Teknik Pengaturan Departemen Teknik Elektro Fakultas Teknologi Elektro Institut Teknologi Sepuluh Nopember
Where u(t) is the input signal to the plant model, e(t) is defined as e(t) r(t) − y(t), and r(t) is the reference input signal. =
Proportional Controller Proportional controller providing action that is proportional to the recent error. The controller consists of a constant proportional gain shown in Equation 2. Proportional controller used to speed up the response but the greater the value, the greater the error that occurred and will also cause the system to oscillate. u(t ) K p e(t ) K p (r (t )
(2)
y(t )),
Integral Controller Integral controller providing action is proportional to the integral of the error. This controller can eliminate errors that occurred but system response is slow. The relationship between the error and the integral action can be seen in Equation 3. Integral controller can be use to eliminate the steady state error. t
u (t )
K i e( )d
(3)
0
Differential Controller When the controller is proportional giving of action based on the current error and provide integral controller action based on the value of the previous error , then the differential controller provide action based on the prediction error value coming as defined in Equation 4. u (t )
K d
de(t )
(4)
dt
The advantage of the differential controller is have zero at the origin. So that if it is added , the system can become more stable . Differential controllers can’t be stand alone because , if applied would give a zero value if the error that occurs is constant and will increase the control signal is given . As a result, noise at high frequencies will be enlarged, therefore, form a differential controller converted into Equation 5. U ( s ) E ( s )
K d s K d s N
(5) 1
6
Laboratorium Teknik Pengaturan Departemen Teknik Elektro Fakultas Teknologi Elektro Institut Teknologi Sepuluh Nopember
Level Measurement
Figure 5. A magnetising force, H, causes a dimensional change due to the alignment of magnetic domains. Magnetostriction is a property of ferromagnetic materials such as iron, nickel, and cobalt. When placed in a magnetic field, these materials change size and/or shape The physical response of a ferromagnetic material is due to the presence of magnetic moments, and can be understood by considering the material as a collection of tiny permanent magnets, or domains. Each domain consists of many atoms. When a material is not magnetized, the domains arc randomly arranged. When the material is magnetized, the domains are oriented with their axes approximately parallel to one another. Interaction of an external magnetic field with the domains causes the magnetostrictive effect. This effect can be optimized by controlling the ordering of the domains through alloy selection, thermal annealing, cold working, and magnetic field strength. The ferromagnetic materials used in magnetostrictive position sensors are transition metals such as iron, nickel, and cobalt. In these metals, the 3d electron shell is not completely filled, which allows the formation of a magnetic moment. (i.e., the shells closer to the nucleus than the 3d shell arc complete, and they do not contribute to the magnetic moment). As electron spins are rotated by a magnetic field, coupling between the electron spin and electron orbit causes electron energies to change. The crystal then strains so that electrons at the surface can relax to states of lower energy. When a material has positive magnetostriction, it enlarges when placed in a magnetic field; with negative magnetostriction, the material shrinks. The amount of magnetostriction in base elements and simple alloys is small, on the order of 10 6 m/m. Since applying a magnetic field causes stress that changes the physical properties of a magnetostrictive material, it is interesting to note that the reverse is also true: applying stress to a magnetostrictive material changes its magnetic properties (e.g., magnetic permeability). This is called the Villari effect. Normal magnetostriction and the Villari effect are ot use n pro uc ng a magnetostr ct ve pos t on sensor. The Wiedemann effect describes the twisting due to an axial magnetic field applied to a ferromagnetic wire or tube that is carrying an electric current. Figure 6. 7
Laboratorium Teknik Pengaturan Departemen Teknik Elektro Fakultas Teknologi Elektro Institut Teknologi Sepuluh Nopember
An important characteristic of a wire made of a magnetostrictive material is t he Wiedemann effect (see Figure 2). When an axial magnetic field is applied to a magnetostrictive wire, and a current is passed through the wire, a twisting occurs at the location of the axial magnetic field. The twisting is caused by interaction of the axial magnetic fiel d, usually from a permanent magnet, with the magnetic field along the magnetostrictive wire, which is present due to the current in the wire. The current is applied as a short-duration pulse, 1 or 2 µs; the minimum current density is along the centre of the wire and the maximum at the wire surface. This is due to the skin effect. The magnetic field intensity is also greatest at the wire surface. This aids in developing the waveguide twist. Since the current is applied as a pulse, the mechanical twisting travels in the wire as an ultrasonic wave. The magnetostrictive wire is therefore called the waveguide. The wave travels at the speed of sound in the waveguide material, ~ 3000 m/s. The operation of a magnetostrictive position sensor is shown in Figure 3 The interaction of a current pulse with the position magnet generates a strain pulse that travels down the waveguide and is detected by the pickup element.
Figure 7. The axial magnetic field is provided by a position magnet. The position magnet is attached to the machine tool, hydraulic cylinder, or whatever is being measured. The waveguide wire is enclosed within a protective cover and is attached to the stationary part of the machine, hydraulic cylinder, etc. The location of the position magnet is determined by first applying a current pulse to the waveguide. At the same time, a timer is started. The current pulse causes a sonic wave to be generated at the location of the position magnet Wiedemann effect. The sonic wave travels along the waveguide until it is detected by the pickup. This stops the timer. The elapsed time indicated by the timer then represents the distance between the position magnet and the pickup. The sonic wave also travels in the direction away from the pickup. In order to avoid an interfering signal from waves travelling in this direction, their energy is absorbed by a damping device (called the damp). The pickup makes use of the Villari effect. A small piece of magnetostrictive material, called the tape, is welded to the waveguide near one end of the waveguide. This tape passes through a coil and is magnetized by a small permanent magnet 8
Laboratorium Teknik Pengaturan Departemen Teknik Elektro Fakultas Teknologi Elektro Institut Teknologi Sepuluh Nopember
called the bias magnet. When a sonic wave propagates down the waveguide and then down the tape, the stress induced by the wave causes a wave of changed permeability (Villari effect) in the tape. This in turn causes a change in the tape magnetic flux density, and thus a voltage output pulse is produced from the coil (Faraday Effect). The voltage pulse is detected by the electronic circuitry and conditioned into the desired output. EXPERIMENT Experiment 1. Proporsional Control It is sometimes necessary to perform the identification experiment under output feedback, i.e., in closes loop. The reason may be that the plant is unstable or that it has to be controller for production, economic, or safety reasons, or that it contains inherent feedback mechanisms. Operational Procedure 1. Open software PCT-100, and go to fluid level window. 2. Set SP value of level by 20% 3. Set PG value by 2 4. Change the value of PG between 2 and 10 in accordance with the Table 2 below and write the steady state value in Table 1. Record the response 5. Set the sample time by 0.1 6. Start the simulation and record the response by File → Save. Experimental Data Steady state value each changing PG and SP can be write in Table 2. Draw the response from data in step 4 changing PG and SP.
Table 2. Experiment data of proportional control SP (%)
PG
20 20
2 4
20 20 20
6 8 10
Steady State
Analysis and Experiment Task 1. What is the effect increasing and decreasing PG value? (prove it) 2. What conclusions about the nature of proportional controller may be drawn from your observations? 3. What is the characteristic P controller? 4. Design proportional controller and calculate the value of error steady state to the system above, by design specification ∗ = x ([3 last digits NRP]/10. Given K=[3 last digits NRP]/100, =[3 last digits NRP]. +
Experiment 2. Proportional + Integral Control
9
Laboratorium Teknik Pengaturan Departemen Teknik Elektro Fakultas Teknologi Elektro Institut Teknologi Sepuluh Nopember
Integral controllers are often used to eliminate the offset caused by proportional control. Short setting controller Integral Action Time (IAT) resulting integral effect greater action on the controller output. Correction is performed on the output by the integral action relating to the PG setting. PG great value means a major correction on the output due to the integral action. Integral Action Time (IAT) is defined as the time taken for the integral action to duplicate the proportional action of the controller, if the error were to remain constant during the period. Operational Procedure 1. Run a set of fluid level control experiments beginning with an empty process tank 2. Ensure that the Auto Drain feature is turned off in each case 3. Set PG value by 5 and the SP value of level 20% 4. Set the value of I between 999 and 1 in accordance with the table 3 below and write the steady state value in Table 3. 5. Start the simulation and record the response by File → Save
SP
Table 3. Experiment data of proportional and integral control PG Steady State Observations I
30%
5
999
30%
5
600
30%
5
300
30%
5
100
30%
5
10
30%
5
1
Analysis and Experiment Task 1. What conclusions about the nature of these fluid level experiments may be drawn from your observations? 2. Why is the characteristic system level tank to be better when given controller Integral? (Prove it use to performance characteristic of transient response and steady state response analysis) 3. Design proportional integral (PI) controller, by design specification ∗ = x ([3 last digits NRP]/ 10). Given K=[3 last digits NRP]/100, =[3 last digits NRP]. +
10