機械学習を用いた電界算出新手法の開発

カテゴリ: 部門大会

論文No: 47

グループ名: 【B】平成29年電気学会電力・エネルギー部門大会

発行日: 2017/09/05

タイトル(英語): New Method for Electric Field Analysis by Application of Machine Learning

著者名: 加藤 克巳(新居浜工業高等専門学校),加藤 茂(新居浜工業高等専門学校),松友 真哉(新居浜工業高等専門学校)

著者名(英語): Katsumi Kato|Shigeru Kato|Shinya Matsutomo

キーワード: 電界解析|絶縁設計|機械学習|ニューラルネットワーク|人工知能,Electric field analysis,Insulation design,Machine learning,Neural networks,Artificial Intelligence

要約(日本語): Electric field analysis is one of the most important tools for the insulation design of high voltage power apparatus. The efficient method of electric field analysis is required for the insulation design and its optimization processes. In this paper, we developed a new method for electric field analysis with the aid of a machine learning technique. A lot of data by electric field analysis are used for the machine learning process. The machine learning is carried out before the analyzed configuration are determined. As a result, we obtained the electric field analysis results with less calculation time. The calculation efficiency and effectiveness are confirmed by the calculation examples.

PDFファイルサイズ: 455 Kバイト