Skip to Main Content
In this paper, a CMAC (cerebellar model articulation controller) neural network diagnosis system of turbine-generator is proposed. This novel fault diagnosis system contains an input layer, quantization layer, binary coding layer, and fired up memory addresses coding unit. Firstly, we construct the configuration of diagnosis system depending on the fault patterns. Secondly, the known fault patterns were used to train the neural network. Finally, the diagnosis system can be used to diagnose the fault types of turbine-generator system. By using the characteristic of self-learning association and generalization, like the cerebellum of human being, the proposed CMAC neural network fault diagnosis system enables a powerful, straighforward, and efficient fault diagnosis. Furthermore, the following merits are obtained: (1) high learning and diagnosis speed; (2) high noise rejection ability; (3) alleviates the dependency for additional expert expertise; (4) eliminates the weight interference between different fault type patterns; (5) memory size is reduced by new excited address coding technique; and (6) suitable for multiple fault diagnosis.