Advanced Controller Design For Chilled Water Based Air Conditioning Systems

With the rapid improvement of people’s life quality, the indoor environment requirements are getting more demanding. Conventional chilled water based air conditioner uses low temperature chilled water to handle the latent and sensible load, the chilled water temperature is lower than the dew point temperature. Thus reheating devices are usually needed to warm the overcooled air and lead to energy waste. To enhance the indoor air temperature and humidity control performance and reduce energy use, the coupling, multi-variable, nonlinear feature of the chilled water based air conditioning system has to be handled. By taking advantages of advanced control technique, the system dynamic characteristics have been well modeled and controller developed exhibited better performance than conventional controller.To handle the simultaneous control of indoor air temperature and humidity, this thesis has accomplished the following work:1. Analyzed the heat transfer characteristics between the chiller, chilled water loop and the indoor air loop, a linear parameter varying model has been developed to model the dynamic characteristics of chilled water temperature in the outlet of evaporator. The system nonlinear feature was represented by integration of linear models at typical operating conditions with proper weights. A predictive controller based on the linear parameter varying model has been developed to control the chilled water temperature in the outlet of the evaporator.2. Analyzed the heat transfer behavior between the chilled water and the indoor air loop in the air handling unit. Variables which influence the dynamic characteristics of the air dry bulb temperature have been obtained through mechanism of heat transfer process between chilled water and air. To handle the nonlinear, multi-variable and uncertain features of the process, an adaptive neural network controller has been developed to control the indoor air dry bulb temperature.3. The thesis tried to simultaneously control the indoor air temperature and humidity of the chilled water based air conditioner. By exciting the system with proper signals adding to the compressor in the chiller and supply air fan speed in the air handling unit, the system dynamic response has been obtained. A neural network was used to link the outputs variations with the excite signals to simulate the dynamic characteristics of the system. Based on the developed neural network model, a predictive controller was used to control the indoor air temperature and humidity simultaneously.4. To handle the limited control range of the controller developed in the above point, a neural network inverse model controller was proposed with an adaptive algorithm. The controller can capture the system dynamics by learning from the new operating data information and can offer a satisfied control performance.The innovation of this thesis is the development of simultaneous controller of indoor air temperature and humidity via varying the compressor speed and supply air fan speed by taking the advantages of advanced control technology.