Health forecast of aircraft based on the combination forecast model
Jianguo Cui · 2009
Combing with the advantages of GM(1,1) and MGM(1,n),a new forecast method based on the combination forecast model is presented to solve the veracious forecast to the health information of aircraft.First,use sym4 wavelet of five levels to decompose the original acoustic emission signals of aircraft key parts which are collected by the acoustic emission sensors.Then distill the average of absolute value(AV),standard deviation(ST) and singular value(SV) of the fifth layer wavelet decomposing parameters.Forecast the AV character by using GM(1,1) and MGM(1,n),then make the forecast values of the two models as the inputs of BP neural network and the original values as the outputs so as to make combination forecast.The experiment indicates that,the combination forecast method presented in this paper can realize the veracious forecast to the aircraft fault information.And its forecasting precision is obviously higher than the single forecast model.Also the validity of this method is validated.