Rare events prediction in complex technical systems has been very interesting and critical issue for many industrial and commercial fields due to huge increase of sensors and rapid growth of Internet of Things (IoT). To detect anomalies and foresee machine failure during normal operation, various types of Predictive Maintenance (PdM) techniques have been studied. Among these techniques, unsupervised anomaly detection methods for multi-dimensional data set would be of more interest in many practical cases. So, in this demo, I have selected following three typical methods.
1. Htelling's T-square method
2. Gaussian mixture model
3. One-class SVM
To emulate a realistic situation, in this demo, I will use the dataset provided by C-MAPSST (Commercial Modular Aero-Propulsion SystemSimulation) [1, 2].
 A. Saxena, K. Goebel, D. Simon and N. Eklund, "Damage Propagation Modeling for Aircraft Engine Run-to-Failure Simulation," International Conference on Prognostics and Health Management, (2008).
 Turbofan Engine Degradation Simulation Data Set, https://ti.arc.nasa.gov/tech/dash/groups/pcoe/prognostic-data-repository/#turbofan
Akira Agata (2020). Demo Files for Predictive Maintenance (https://www.mathworks.com/matlabcentral/fileexchange/63012-demo-files-for-predictive-maintenance), MATLAB Central File Exchange. Retrieved .
This work is helpful！Thank you！
It's a wonderful job. However, the feature selection seems ignored.
It's there any reason or reference why should we select varName([1:2, 7:9, 12:14, 16:20, 22, 25:26])?
Thank you very much : )
This work is incredible.
Fantastic implementation. I am trying to implement to actual aircraft data. But the functions do not work on my MATLAB versions (2011A). Could you please guide for older MATLAB versions?
good job , but i really didn't anderstand the use of algorithmes ,what are our outputs ,also is there a documentation about this algorithms ,and as i said good work bro ;)
- Updated the link of the Turbofan Engine Degradation Simulation Data Set
Update demo scripts.