Sensitivity Analysis of Optimal Control Schemes of a Refrigeration System

• Main Authors: CHENG-HUNG LIU, 劉政宏
• Other Authors: none
• Format: Others
• Language: zh-TW
• Published: 2005
• Online Access: http://ndltd.ncl.edu.tw/handle/42180525262725289727

碩士 === 逢甲大學 === 自動控制工程所 === 93 === The currently used control schemes of industrial refrigeration systems have the drawback of more energy-consuming. More advanced control schemes were designed by formulating the control of refrigeration systems as a model predictive control problem. The objective function is a modified time-averaged COP. Because of the complexity of the optimization problem, search algorithms were developed for searching the approximate optimal control decisions. The control schemes based on “branch and bound” and “off-line learning neuro-dynamic programming” search algorithms have been developed in previous studies. The simulation results showed that the proposed control schemes achieved higher time-averaged COP and lower power consumption. In this thesis, an “on-line learning neuro-dynamic programming” control scheme is developed. Furthermore, the difference between the predictive model and the refrigeration plant is considered. The performance of the control schemes may be degraded by model uncertainty. The robustness of the performance of control schemes to model uncertainty is investigated by sensitivity analysis。