DBSCANを用いたContextual Outlier Interpretationによるガスタービン異常検知の説明

カテゴリ: 全国大会

論文No: 3-074

グループ名: 【全国大会】令和6年電気学会全国大会論文集

発行日: 2024/03/01

タイトル(英語): Explanation of Gas Turbine Generator anomaly detection by Contextual Outlier Interpretation with DBSCAN

著者名: 尹家輝（明治大学）,福山良和（明治大学）,島崎祐一（富士電機）,長田悠人（富士電機）,村上賢哉（富士電機）,飯坂達也（富士電機）,松井哲郎（富士電機）

著者名(英語): Jiahui Yin (Meiji University),Yoshikazu Fukuyama (Meiji University),Yuichi Shimasaki (Fuji Electric Co., Ltd.),Yuto Osada (Fuji Electric Co., Ltd.),Kenya Murakami (Fuji Electric Co., Ltd.),Tatsuya Iizaka (Fuji Electric Co., Ltd.),Tetsuro Matsui (Fuji Elec

キーワード: 異常検知|contextual outlier interpretation|ガスタービン|説明可能artificial intelligence|Density-Based Spatial Clustering of Applications with Noise|anomaly detection|contextual outlier interpretation|gas turbine generator|explainable artificial intelligence|Density-Bas

要約(日本語): Real-time anomaly detection in gas turbines and proactive resolution of the issue before it escalates into a breakdown are crucial for gas turbines. In the case of anomaly detection using AI models, it is essential that users agree to use the models. Therefore, it is necessary to explain the black box part of the anomaly detection models. This paper proposes explanation of gas turbine anomaly detection by Contextual Outlier Interpretation with DBSCAN. It can explain gas turbine anomaly detection in real time. Effectiveness of the proposed method is verified using practical gas turbine generator operation data.

本誌掲載ページ: 109-111 p

原稿種別: 日本語

PDFファイルサイズ: 480 Kバイト