Application of Improved BP Neural Network in Fault Diagnosis for Turbogenerator Set
Yang Ku · Power System Engineering · 2002
Based on standard BP neural networks, the improved BP neural network is given with priority power value. The model structure, BP algorithm, learning algorithm and the method assuring priority power value for the neural network are set forth. Combined with the example of fault diagnosis for turbogenerator set, it is proved that the model has a strong ability of fault recognition from two sides, single fault and multi faults, and the diagnosis results are accord with the facts.