C114 予兆診断のためのリモートモニタリング基準設定支援技術(OS6 火力および診断技術)

概要

論文の詳細を見る
Proactive maintenance is important to keep normal operation of plant and equipment systems. A prognosis system for such a system was developed that includes two functions: remote monitoring which detects anomalies by comparing sensor data with a threshold and data-mining which detects anomalies using statistical analysis. This paper proposes an automatic method of setting up criteria for remote monitoring. First, training data is generated based on the result of data-mining. Next, a decision tree is generated by learning the training data, and an if-then rule is extracted. Usefulness of the proposed method was evaluated using 4 data sets including one type of faults. In those data sets, anomalies are recognizable several days before the day of the fault. Two rules extracted using 2 of 4 data sets were different each other but the 2 sensor outputs corresponding with the 2 rules were highly correlated, and the extracted 2 rules can detect anomalies from the other data sets properly.
一般社団法人日本機械学会の論文
2011-06-22