Industrial Machinery Anomaly Detection

Train an autoencoder on normal operating data from an industrial machine to predict anomalies.

Industrial Machinery Anomaly Detection

This example applies various anomaly detection approaches to operating data from an industrial machine. Specifically it covers:

• Extracting relevant features from industrial vibration timeseries data using the Diagnostic Feature Designer app
• Anomaly detection using several statistical, machine learning, and deep learning techniques, including:
  • LSTM-based autoencoders
  • One-class SVM
  • Isolation forest
  • Robust covariance and Mahalanobis distance

Setup

This demo is implemented as a MATLAB® project and will require you to open the project to run it. The project will manage all paths and shortcuts you need.

To Run:

1. Open the MATLAB Project AnomalyDetection.prj
2. Open Parts 1-3 on the Project Shortcuts tab

MathWorks® Products (http://www.mathworks.com)

Requires MATLAB® release R2021b or newer and:

• Predictive Maintenance Toolbox™
• Deep Learning Toolbox™
• Statistics and Machine Learning Toolbox™

License

The license for Industrial Machinery Anomaly Detection using an Autoencoder is available in the license.txt file in this GitHub repository.

Community Support

MATLAB Central

Copyright 2021 The MathWorks, Inc.