Health monitoring of industrial machines using Scene-aware threshold selection

Oral presentation at  "Exploring New Possibilities" 2022 Doctoral College Conference at University of Surrey, UK.

An overview about the current work:

Automated health monitoring of industrial machinery can help in avoiding unplanned downtime, increased productivity and reduce maintenance schedules in large-scale industries. Typically, an operating machine produces sounds that might be useful to identify whether the underlying machine is running normally or not. Sound-based monitoring of machinery provides advantages such as readily available sensors (microphones), non-intrusive sensing and ability for omnidirectional sensing. For instance, portable devices such as smartphones can be used to identify the health of the machines based on their sounds.

This paper presents an autoencoder based, unsupervised approach to identify anomaly in an industrial machine using sounds produced by the machine. The proposed framework is trained using log-mel spectrogram representations of the sound signal. In classification, we hypothesize that the reconstruction error computed for an abnormal machine is larger than that of a normal machine, since only normal machine sounds are used to train the autoencoder. A threshold is chosen to discriminate between normal and abnormal machines.

However, the threshold may change due to variations in the surroundings due to noise, other machine sounds etc. Therefore, the selection of an appropriate threshold is crucial. To select an appropriate threshold irrespective of the surrounding, we propose an acoustic scene classification framework that classifies the underlying surrounding and selects a threshold that adapts to the underlying surrounding. The experiment evaluation is performed on MIMII dataset for industrial machines, namely fan, pump, valve and slide rail. Our experiment analysis shows that utilizing the acoustic scene classification framework, an adaptive threshold can be selected, which significantly improves performance compared to choosing a fixed threshold.