Analysis of equipment anomalies using multiple sensing data.

Abstract

There are many types of sensors and data for analyzing the behavior of measurement objects in equipment diagnosis. This paper proposes an anomaly detection method that can capture equipment anomalies from multiple angles based on the evaluation of relationships between 3-axis acceleration data, acoustic data, and temperature data obtained from sensors installed on compressors. The main focus here is on the transition of measurements, intercomparisons, and comparisons based on time-frequency analysis, and the usefulness of the wavelet transform, which is a time-frequency analysis, is asserted.