Abstract: Traditional engine overheating detection technology is unable to overcome the engine temperature fluctuations caused by changes in the signal run nonlinear distortion, poor fault detection. The paper presents a diagnostic network integration algorithm fighter engine overheating fault mining models, using wavelet transform method, collecting dynamic fighter engine overheating fault signal characteristics, taking signal characteristics to overheating as the basis, through transmitting the data which is the integration of diagnostics in the hidden layer network data mining process to overheating signal, gets overheated temperature information to achieve the depth of digging according fighter engine overheating signal. Experimental results indicate that the use of the proposed model can improve the fighter engine overheating fault mining accuracy, and conducive quick fix faults.

Key words: engine, overheating, signal detection, wavelet transform, fusion diagnosis network