Study on the method for engine technical state change trend precipitation based on combination forecasting model

The paper constructed a model for forecasting remaining age of tank engines. Firstly, the paper forecast remaining life using AR time model, grey model and artificial neural network model respectively, but the predictive result was not very exact using a single method. So the paper combined the three models linearly, and the weights were calculated using grey correlation degree. The combination predictive model was constructed, which was used to forecast remaining age of tank engines of definite service time. The forecasting result proves that the combination predictive model used to forecast engine remaining age can overcome disadvantages of bad reliability and low precision caused using a signal model to forecast engine remaining life.