Welcome to the documentation of fault-diagnosis¶
MATLAB libraries for fault-diangosis in dynamic, large-scale systems¶
fault-diagnosis is research software in MATLAB, targeted for fault diagnosis in large-scale systems. It employs methodologies of Structural Analysis for qualitative abstraction and efficient extraction of residual-generators.
This software is open-source and its source code can be found in this github repository: https://github.com/Georacer/fault-diagnosis
Overview¶
fault-diangosis uses the methodology of Structural Analysis to perform Fault Diagnosis, which performs a qualitative abstraction of the mathematical model of a system into a bipartite Structural Graph. Graph methodologies are then applied onto that graph for the extraction of residual generators.
Graph representation of the bipartite graph is favoured, against the biadjacency matrix one, to allow more natural handling and processing of graph meta-information.
The codebase is split into several logical parts. There is dedicated documentation for each one:
- Structural Graph
- Graph Interface
- Display engine
- Subgraph Generator
- Matcher
Demos are available in the dedicated section.
System Compatibility¶
This software is developed and tested under Ubuntu Linux
External Software Requirements¶
fault-diagnosis requires the following software to operate:
- Fault Diagnosis Toolbox - Research software on fault diagnosis using structural graphs. Used for the generation of PSO sets and its interface of the Dulmage-Mendelsohn decomposition.
- matlab_networks_routines - Library for graph operations.
- munkres - Fast implementaiton of the Hungarian weighted assignment algorithm.
The above software is distributed alognside fault-diagnosis, according to the copyright set by their respective authors/owners. Licence files for each software are included in the corresponding directories.
Also the following software is required to be installed on the system:
- The dot language compiler, part of the GraphViz package - Used for graph visualization.