本研究主要是探討儲冰空調系統滷水主機故障診斷，提出結合徑向基類神經網路及穩健參數設計的演算法－增強型徑向基類神經網路演算法，推導一套儲冰空調系統滷水主機故障診斷的模式，以追求快速及精確地執行故障預測及故障診斷，以提高空調系統效率與避免設備無預警毀損。增強徑向基類神經網路，整合了穩健參數設計和徑向基類神經網路，將用於儲冰空調系統滷水主機的故障診斷，以確保系統效能。

The purpose of this study proposes the fault diagnosis for a brine chiller in an ice-storage air-conditioning system to ensure better capacity of energy conservation, contacts and reduce maintenance cost. Combining the Radial Basis Function Network with Robust Quality Design, Enhanced Radial Basis Function Network is proposed to solve this problem. In this project, ERBFN is developed to accurately predict and diagnose the faults of the chiller in an ice-storage air-conditioning system.