專題演講 主講人：李政崇博士(台灣電力公司)
題 目：Using Machine Learning to Predict the Leakage Current of Insulators
主講人：李政崇博士(台灣電力公司)
時 間：111年4月29日(星期五)上午10:40-11:30
(上午10:20-10:40茶會於交大統計所428室舉行)
地 點：線上演講－使用Google Meet
使用Google Meet線上直播，
演講開始前20分鐘可進入會議，請點選下列連結後按下「要求加入」即可
摘要
From October to next March, the insulators deployed in the western area may be polluted by different factors, such as air pollution, rain, and the change in the climate. Generally, Taipower has to spend considerable money and resources to clean the insulators, to prevent serious consequences, such as partial arc and flashover. The scale of the leakage current of insulators is a good metric to evaluate their status. However, deploying sensors to monitor current values is impractical, and the alternative way is to monitor other influential factors instead and uses machine learning methods to predict the scale, which can be referenced by frontline workers. In this talk, we briefly introduce the motivation, do exploratory data analysis, clean the data, and demonstrate the final results by different models. Future works are also illustrated at the end. Besides, an introduction to Taipower company, Taipower Research Institute, and how to apply for Taipower will also be presented in the talk.