本研究提出一個以即時監控資訊為基之油浸式大型變壓器智慧維護保養平台,其目的為協助大型且昂貴之工程資產管理者能快速彙整即時監控設備狀態資訊,提出設備維護診斷建議及設備剩餘使用壽命之評估,降低設備故障突發機率,確保工程資產全壽期價值之提升。本平台採用國際電子電機工程師學會(IEEE)修正之Doernenburg及Rogers診斷法,和國際電工委員會(IEC)之Duval Triangle診斷法等三種油中氣體分析模式,診斷變壓器故障,並提供對應之維護對策。又針對壽期預測提出一以Group Method of Data Handling(GMDH)資料群集處理技術為基之創新壽期減損評估方法,以油中氣體及糠醛生成量為變數,建立一全新壽命評估模型,並與IEC壽期減損評估進行對照,使油浸式變壓器之壽命評估更具參考價值。
Transformers are critical assets that require continuous monitoring in the generation of electrical power. The management of engineering assets requires real-time diagnosis and preventive maintenance in order to avoid unexpected and catastrophic equipment shut-downs. In addition, the health status and remaining life of the transformer are critical knowledge for proper maintenance and management. This research focuses on oil-immersed transformers as a case study to help managers compile real-time monitoring and sampling data for information systems. The deterioration of the transformation's insulating paper was measured to derive the transformer’s remaining life and to project the optimal replacement time. The proposed prediction method uses the Group Method of Data Handling (GMDH) to estimate the remaining life of any given transformer in use. This algorithm uses the dissolved gas-in-oil and furfural formation as the input variables to form the transformer life assessment model. Finally, the derived model is compared to models proposed by the International Electrotechnical Commission (IEC). The results show improved performance over the currently used IEC models.