비등구간 멤버십함수를 이용한 고성능 냉각챔버 시스템의 퍼지제어기 설계

Abstract

Recently, the performance of chamber cooling system (CCS) focused on high precision of temperature control and strong robust. It is used for offering an accurate and stable environment to support the biotechnology and medical sciences research. In addition, the global warming problem arouse scientist pay more attention on the energy saving of equipment. The variable speed refrigeration cycle is highly focused on an industrial field of chamber cooling system because of its energy saving ability. In the variable speed refrigeration system, the inside temperature is modulated by regulating the motor speed of compressor. And superheat is changed by manipulating the opening angle of electronic expansion valve. For driving the refrigeration cycle to output accurate temperature, there are lots of control manners. Otherwise, the original mathematic model base controllers are difficult to design because the inherent nonlinear characteristics of refrigeration cycle in operation ranges. Thus, the fuzzy logic controller which is not depending on the mathematic model is better for controlling the refrigeration cycle. And the linguistic logic express is easier for engineer to design and modify the controller by empirical. However, the big steady-state error is usually exist in the common fuzzy logic control, it because the characteristic of the fuzzy control. And the noise is happened from inverter which set in the variable speed refrigeration cycle used for regulating the motor rotation speed. The harmonic noises disturbing the fuzzy inference result in the controller output unstable. Fuzzy logic controller adopting unevenly-distributed membership function (UMF) is presented with the purpose of enhancing performance of the temperature control precision, robustness and work efficiency for the CCS. The control precision realized by reducing the chamber temperature steady-state errors, robustness is realized by the noise rejecting. The histogram equalization was applied to modify the error membership function for reducing the steady-state error of chamber temperature and the domain encompasses method was applied to modify the error change membership function for rejecting the noise disturbance to improv the robustness. In the characteristic experiment of CCS, the system achieved highest working efficiency when the superheat around 8℃. The working efficiency is improved by reducing the steady-state error of superheat through the manipulated opening angle of electronic expansion valve by fuzzy logic controller with unevenly-distributed membership function. The comparison results in simulation and real experimental is listed for proving the control effective of fuzzy logic controller with unevenly-distributed membership function is better than the evenly-distributed membership function. In each simulation and real experiment, all of the conditions are trying to keep in same except the membership function of fuzzy logic controllers. The smaller steady-state error of chamber temperature and superheat by unevenly-distributed membership function in the simulation proved the control precision is enhanced successfully. Under the noise disturbance, the stable frequency output by unevenly-distributed membership functions proved the noise disturbance is rejected successfully. The experimental results show that the steady-state error was reduced around 40% and the noise disturbance was rejected successfully even though noise range was more than 60% of the control precision range. The control precision was improved by reducing the steady-state error and the robustness was enhanced by rejecting noise disturbance through the fuzzy logic controller with unevenly-distributed membership function. Moreover, the stable output frequency can contribute to improvement of superheat control performance, thereby keeping the steady-state error of superheat within ±2℃, the system can achieve high efficiency for saving energy cost and the service time of EEV is extend by applying UMF to chamber temperature control process of CCS. This thesis is studied on enhancing performance of the fuzzy logic controller by membership function’s modification. It can be quoted by the engineer to improve the control performance not only base on the experience but also base on the real response of the control plant.