查看文章

Today, we are at the beginning of Industry 4.0. Machines are becoming increasingly sensorized and connected to the internet. Streaming data will thus be sent continuously to cloud computing data-centers. Condition monitoring techniques can leverage these huge volumes of available data to increase detection potential and insights in system behavior by long-term trending, anomaly detection and learning approaches. Additionally, the fact that data of similar machines of a fleet is collected allows for exploiting system similarity. This paper illustrates an integrated monitoring approach for the Industry 4.0 context. Our cloud processing approach is shown. Furthermore, we show an integrated failure detection and severity assessment approach. Asset performance in the fleet is used as a proxy for potential failure. Health assessment and fault localization combines state-of-the-art vibration signal processing on high frequency data (> 10kHz) with machine learning models trained on low frequency (1Hz) bearing temperature data to create a health score.