Adaptive Multi-Parameter Prediction Model Based on Grey Theory

Abstract

Fault prediction is of great importance to ensuring weapon equipments' safety and reliability. Usually the data for fault detection and prediction of weapon equipments have features like small samples and multi-parameter. Currently the main fault prediction methods have achieved some success in practical applications, but all fall short at some aspects. Based on grey prediction theory and with analysis of disadvantages of GM(1, 1) model, an adaptive prediction model with several characteristic parameters for small samples is put forward. This model modifies initial value and background value, and takes interrelations of the parameters and characteristics of prediction series into account. The model is then used for prediction and analysis with the multi-parameter data of certain aero-engine. The results show that the model has good prediction precision, which in turn validates its availability.