Recently, variable speed refrigeration system (VSRS) has been receiving high attention for energy saving ability according to development of industrial technology. Therefore, it is necessary to design high performance and high precision controller for obtaining precise temperature and energy saving for VSRS. The VSRS is basically to control the capacity and the superheat simultaneously in control scheme. Capacity is controlled by changing compressor speed to cope with partial thermal load condition and superheat is maintained as a constant value to keep maximum COP and to prevent liquid back phenomenon. Hence, precise chamber temperature and energy saving can be achieved at the same time by these two controls.
To design a high-performance and a high-precision controller of the VSRS, an empirical model that represents dynamic characteristics of the system is necessary. The basic refrigeration system mainly consists of two exchangers (evaporator and condenser), an electronic expansion valve (EEV) and a compressor. Typical conventional control method for the VSRS was PID. However, it needs a specific transfer function model and also its modeling process requires tremendous efforts. Even though an empirical model was obtained from number of experiments, the model strongly depends on system parameters. The system parameters inevitably includes uncertainties because of linearization and approximation of real system in modeling process. Thus the PID controller is not enough to get desirable control performance in nonlinear system and the system included big model uncertainty. To solve these problems, the fuzzy logic control (FLC) has been applied in many industrial fields. Actually it is not depending on the mathematical model. Thus, the FLC is better for controlling the nonlinear system like VSRS. So far, most of researches on FLC are only aimed at improving control performance. Therefore, it was not enough for engineers to design the controller systematically.
This paper focuses on analysis about the influence of the main design factors such as an error , an error rate , and sampling time for the fuzzy controller on the control performances. It can help systematical design of fuzzy controller for VSRS even though beginner of the system. In addition, this paper proposes a novel approach to design rule bases easily and quickly using surface diagram (SD) with Matlab function.
Finally, through some computer simulations and experiments for a chamber control system, validity of the analyzed results and effectiveness of the suggested rule base design approach were verified. The results showed that the fuzzy controller is a very useful tool for inherent non-linear system like VSRS. It is expected that the suggested rule base design approach will be very helpful for engineers to build up rule bases with limited experiences for the target system.