engleski

Application of expert systems, fuzzy algorithms and neural networks to fault diagnosis in control systems

An overview of modern artificial intelligence-based techniques for processing heuristic knowledge is presented. Basic definitions and comparison with classical mathematical tools are given. A qualitative symptom observer is suggested as a replacement for an analytical symptom generator in some particular situations (complex system, unknown process model etc.). All that led to the idea of an integrated fault diagnosis system dealing with as many information sources as possible: measured signals, observed parameters, quantitative and qualitative models and analytical data.

fault diagnosis; fault detection; integrated systems; fuzzy logic; neural networks; expert systems

nije evidentirano