Aeroengine gas path parameter prediction based on dynamic ensemble algorithm

For the problem that single global modeling methods cannot get satisfactory aeroengine gas path parameter prediction results, ensemble algorithms with a combination method called dynamic weighted kernel density estimation(DWKDE) were proposed.Neighboring samples of the test samples were chosen.The weights of the base learners were dynamically calculated by evaluating the base learners' local performance in the neighboring samples. Based on the calculated weights, the integrated prediction of the time series was realized by the weighted kernel density estimation. The proposed combination method was insensitive to outliers and deviations from normality. Applying this combination method to AdaBoost.RT and AdaBoost.R2, experiments were conducted on gas path parameters of aeroengine, verifying that the proposed algorithms can achieve higher prediction accuracy than the single neural network and the traditional ensemble learning algorithms, for example, the root mean square error(RMSE) can be reduced by at least 27%.