実機振動診断カルテに基づいたマルチモーダル・ディープラーニングによる中・小型回転機械の診断効率化
英字タイトル:
An approach to building an efficient diagnosis system for middle range rotational machinery using multimodal deep learning and diagnosis database.
著者:
長野 達朗 （プライア） 沢田 作雄 （沢田テクニカルサービス）
発刊日:
公開日:
Anomaly DetectionMachine LeaningMultimodal Deep LearningVibration Analysis
An approach to building an efficient diagnosis system for middle range rotational machinery using multimodal deep learning and diagnosis database.
著者:
長野 達朗 （プライア） 沢田 作雄 （沢田テクニカルサービス）
発刊日:
公開日:
カテゴリ: 第16回
キーワードタグ:Anomaly DetectionMachine LeaningMultimodal Deep LearningVibration Analysis
概要
For the purpose of detecting anomaly of rotational machinery, the vibration diagnostic method is often used. However, it requires experts to understand insights of cause or sign of abnormal progression. We have developed SA-FRONTIER-PRO, a diagnosis system that assists operators to acquire vibration signals, analyze measured data, and examine machinery through an inquiry approach. To step forward towards further efficiency, we employ multimodal deep learning techniques to mimic decision-making process of experts especially focusing on how to perceive and extract features from measurement data.
