Integrated Group Method of Data Handing Framework for Remaining Useful Life Prediction

Abstract

Considering the shortcomings of a single Group Method of Data Handling (GMDH) network that is easy to fall into local optimum, this paper proposes an integrated GMDH framework for Remaining Useful Life (RUL) prediction. The framework generates three GMDH networks through different division of training data, and integrates the results of the three GMDH networks with a three-layer back propagation (BP) neural network. The NASA C-MAPSS dataset is used to evaluate the effectiveness of the proposed methodˈ by comparison with the prediction results of a single GMDH network and Long Short-Term Memory (LSTM) network. The results show that the proposed method can effectively improve the generalization ability of the GMDH network and is superior to the LSTM in terms of root mean squared error (RMSE).