Nonlinear prediction of condition parameter degradation trend for hydropower unit based on radial basis function interpolation and wavelet transform

Abstract

A nonlinear prediction model of condition parameter degradation trend of hydropower unit is proposed. This model is based on radial basis function interpolation, wavelet transform, largest Lyapunov exponent prediction method, and grey prediction model (GM(1, 1) method). The condition parameter degradation trend model of hydropower unit is built by using RBF interpolation regression method. In this model, the effect of active power and working head is taken into consideration. The degradation trend time series is decomposed into several high-frequency parts and one low-frequency part. For high-frequency parts, their chaotic characteristics are identified. The largest Lyapunov exponent prediction method or GM(1, 1) method is selected to predict each frequency part according to their different properties. For low-frequency part, the GM(1, 1) method is used to predict it. Finally, the predicted results of high-frequency parts and low-frequency part are reconstructed by wavelet theory. The predicted results of the original condition parameter degradation trend time series are obtained. The results show that the proposed method has a high prediction precision.