D2P Toolbox
Software toolbox for fault detection, diagnosis and prognosis in rotating systems.
Asset Description
The asset provides a fault detection, diagnosis and prognosis software toolbox for rotating systems. It is a modular and extendible toolbox that allows robust and adaptive diagnostics (fault detection) and prognostics (lifetime prediction) of rotating systems such as bearings and gears.
Asset Details
Technical Information
The fault detection and lifetime prediction toolbox consists of the following data processing steps:
- Preprocessing
- Feature extraction
- Anomaly/fault detection
- Feature fusion
- State estimation and remaining useful life prediction
These processing steps are reflected in the following IP blocks:
- Bearing fault detection and diagnostic toolbox
- Bearing fault feature library
- Feature fusion method: Autoencoder based
- Feature fusion method: Logistic-Regression based
- Hybrid Remaining Useful Life prognosis tool for bearings
- Shock Pulse Method (SPM) imitator
- Thresholding method: generalized extreme values based
- Wavelet scattering network features
- Way of Working - Monitoring Reusable Pipeline and GitHub
Usage
Modules in the toolbox are available through GitHub and have a well-defined I/O interface, allowing modules to be swapped and connected to the pipeline in a “plug-and-play” manner. The toolbox is suitable for both online and offline purposes, for validation as well as demonstration purposes.
Maturity
The toolbox remains actively maintained, with implementations of new state-of-the-art processing techniques being added when available.
Licence
Proprietary
Resources
- The D2P toolbox was created by Flanders Make (Belgium). For further information, please contact Kerem Eryilmaz at Kerem.Eryilmaz@flandersmake.be
- Provided by Flanders Make
Acknowledgement
The D2P toolbox was created in the frame of several projects (Smartsensor, Gearvibdiag, Vibmon, DGTwinPrediction, AI-initiative), with the support of VLAIO (Flanders Innovation & Entrepreneurship) in Flanders, Belgium