No abstract

Environmental testing apparatuses are widely used for testing endurance of industrial products to both temperature and relative humidity (hereinafter called humidity). It means that both of these physical values inside of environmental testing apparatus are changed over a wide range to evaluate the durability of the industrial products. These two physical values are changed automatically, and may be changed separately or simultaneously. The goal of this study is to control these two physical values effectively. In general, commercially available environmental testing apparatuses have pre-installed controllers which are traditional PID control systems to control the two physical values. However, PID controllers cannot control two physical values effectively because it is impossible to design for MIMO (Multiple-Input and Multiple-Output) systems. This means that the PID controller for temperature control disrupts the other PID controller for humidity control and vice versa. In this study, an MPC (Model Predictive Control) was introduced because it can be designed for MIMO systems and has recently become widely used in industrial manufacturing systems because of the simplicity of its design as a PID (1) . However, the environmental testing apparatus has two difficulties in controlling the temperature and humidity. One of which is that the refrigerator is always working to cool down and dehumidify, which affects the disturbance factor of the whole control system. The other is that the humidifier works not only to humidify but also as another heater. An accurate model is required to design an MPC with good performance despite these complex difficulties. In this study, an accurate method of modeling the controlled system is proposed. The MPC based on these models is designed using MATLAB ® / Simulink ® and implemented with a new type of PLC (Programmable Logic Controller) (2) . The experimental results show good tracking performance to the changing set-points of the temperature and humidity.