Input — Output Classification Mapping for the fault detection, identification and accommodation

Early detection and isolation of faults can help avoid major system breakdowns. This paper presents a non parametric fault diagnosis method to detect and isolate faults in industrial environment. The proposed Input Output Classification Mapping (IOCM) algorithm is based on mapping the input parameters from Gaussian hidden layer functions of an RBF neural network to an output layer. In result the input data of each situation is clustered in a specific surface so every machine state, whether it is normal or faulty it will be represented with its own output layer. Parameters are extracted from input signals or input sub-signals after applying wavelet decomposition and then classified using IOCM. The two techniques before and after decomposition are applied on mechanical system and Tennessee Eastman Challenge Process (TECP) chemical reactor to detect and identify the faults.