Prediction of Exhaust Gas Temperature Margin Based on LSSVR

Exhaust gas temperature is one of the important performance characterization parameters of the state of engine. The prediction analysis of the Exhaust Gas Temperature Margin (EGTM) series is helpful to estimate engine's performance, which can offer theory support for the fault detection and diagnose. Aiming at the non-linear and non-stationary features of EGTM data, a prediction method based on Empirical Mode Decomposition (EMD) and Least Squares Support Vector Regression (LS-SVR) is proposed. The EMD method is used to decompose the EGTM data to reduce the complexity of the time series. The EGTM data were decomposed into Intrinsic Mode Function (IMF) and the residual series by EMD. Finally, the prediction model were build the different LS-SVR for predicting each IMF and the residual series. Each prediction results of the series were combined to obtain EGTM forecast results. The results show that compared with the traditional prediction method, RMSE and MAE are reduced to 2.3178 and 1.8388, which improves the prediction accuracy effectively.