Abstract: The theory and algorithm of the Least Square Support Vector Machine (LSSVM) are applied in the trend forecasting of gas path parameters of aero engine. First, the protocol and content of the Aircraft Communication Addressing and Reporting System (ACARS) report are analyzed, the parallel subsystem is constructed for realtime decoding and storing the engine condition report, and the samples for online model training are acquired. Then, the Final Prediction Error (FPE) principle is used to optimize the embedding dimension, the model selection is conducted using Pattern Search Method. An online forecasting model based on the LSSVM is established. The data link between aircraft and ground is established to acquire the realtime data of the gas path. The online model and the realtime data are used to predict the trend of the gas path parameters. Comparison with the ANN model shows that both one step forecasting accuracy and multistep forecasting accuracy by the online model are higher than that by ANN model. The proposed online model is suitable for prognosis of gas turbine engine.

Key words: propulsion system of aviation and aerospace, aircraft engine, online prediction, support vector machine, gas path parameters