An applications of neural network information fusion techniques in fault detection and diagnosis

A neural network information fusion method was designed to the plane steering surface system's fault detection and diagnosis. According to the characteristic of multi-sensor fusion, synthesizing the out-put data of relational sensors, using the Nonlinear approach ability of neural network expand relational information, the fusion diagnosis strategy can ensure the fault sensor and resume the fault signal. Simulation tests show that the structural form can identify and resume the steering Surface' s normal failure. The effect is satisfactory.