Implementation of an auto-tuning pid controller based on a microcontroller

In this thesis, it is aimed to implement an auto- tuning PID controller based on a microcontroller. For this purpose, at the first step a digital PID controller has been designed and then this has been enhanced by means of Relay Method which is one of the best, but simple auto- tuning method, proposed by Astrom and Hâgglund. In this project, Intel 80C31 microcontroller board, providing both digital PID controller and auto-tuning properties, has been designed. This was a new approach to auto-tuning PID controllers using such advanced Integrated Circuits. On the other hand, it is possible to see some microprocessors application on implementing auto-tuning PID controllers, during development period. However, microcontroller based applications differ from microprocessor based ones. Microprocessor application can be efficient when they carry out various program of users, multiple data processing, calculating sensitive numerical values. On the contrary, microcontroller based system is quite useful in case of particular states which do not require reprogramming and runs unchanged programs. The advantage of such a system is containing the main compo nents of a digital computer in single chip. A microcon troller individually has a Central Processing Unit-CPU, a memory unit and intensive peripherals such as Analog to Digital Converter-ADC and Digital to Analog Converter-DAC. Therefore, it can be suitable for digital control by adding some extra components. The structure of PID controller which is implemented in this study, has been illustrated in Figure 1.e(kT).H1^ +« <& +)+*\ Z \ ' Vu'(kT) T>0- DAC Implemented Boatd MAINS VOLTAGE m n Induction Machine Driver 3®=0=?) Rectifier EDDY Current Break Figure 4 - Speed control of an Induction Machine. In this application the induction machine and its driver are considered as a plant to be controlled. Control signals feeding the plant is produced by a PID algorithm which process reference and feedback signals. After the control signal is converted into a continuous signal between 0-10V, it is applied to the driver of induction machine. The driver produce pulse width modulated voltage as an equivalence of the input signal. In the next some amount of speed data are collected to obtain graphical speed characteristics. Thus, perfor mance of the controller can be figured out. The characteristics in Figure 5 has been obtained for K = 2 » Kj = 2 and Kd " 0. In this case the overshoot is found to be slightly high. This, actually, disturb the motor speed particularly in case of instant load. But motor speed turn back to its reference value when the load is released. VI 11ntt> 1750 1300 1850 10O0 750 son USD o J_ a s\ loo 1.6 260 3.2 ~3O0 1.8 400 6.4 5O0 8.0 Figure 5 - Graphical speed characteristic of induction machine for controller parameters K- m 2 $ Kj = 2 Ka = 0 When auto-tuning process is applied to the system, the reference value is set to desired speed value. A relay with hysteresis characteristics oscillates the system. This is illustrated in Figure 6. n( t ) 1750 1SOO 1250 ? IOOO ? 7S0 SOO 250 ' \ / \ / \ / \ o 100 1.6 200 3.2 300 a.o a do e.A 500 8.0 Figure 6 - Oscillation of the system for the speed reference of 750 rpm. At the end of this routine, j^-ultimate gain and y - ultimate period can be determined from oscillation period. Then controller parameters are obtained as follows, JTp = 2,063, j^- 2,101, JCd- 0,114 ixFor sampling period of T=16ms, these parameters are loaded to controller. The characteristics in Figure 7 has been obtained for Kp = 2,063. K± = 2, 101 and Kd - 0, 114. In this case the overshoot is smoother than it is found in previous example and induction machine reaches the given reference speed. 1730 1300 1230 ? 100O 730 300. 230 ? too l.G 300 4. a "385- 20(1 3.2 400 6.4 Figure 7 - After auto-tuning graphical speed character istic of induction machine for controller parameters jf. 2,063. J^- 2,101. Kd =0,114 ? From collected measurement, it has been shown that microcontroller based auto-tuning PID controller performs efficiently. In case study of speed control induction machine, the transient behavior of speed is found very sufficient ignoring small oscillations and low level of overshoot. In this work, it also shown that how these oscillations can be removed totally, particularly of speed control of an induction machine. smal 1 in case When the implemented controllers is applied to the non-linear systems, the results are given above. If the plant has a non-linear characteristic such as induction machine, it is advisable to use one of the optimization algorithms. If DC motor is used instead of induction machine, it can be easily said that the auto-tuning PID controller provides better results.