基于组合采样和LightGBM的工业故障预测方法

Industrial equipment fault prediction plays an important role in improving industrial production efficiency. However, the currently collected industrial equipment operation data is unbalanced, which makes equipment fault prediction difficult. To address this problem, this paper proposes an ensemble learning model based on combined sampling and LightGBM for industrial equipment fault prediction. First, ADASYN performs adaptive synthesis of a few fault samples to increase the number of fault samples. Second, NearMiss is used for undersampling to solve the problem of local uniformity of data and overlap between classes. Finally, the LightGBM integrated learning algorithm is used for training to reduce overfitting and improve the fault prediction accuracy. Through the experimental comparison withRUSBoost, SMOTEBoost, and EasyEnsemble algorithms on the wind turbine blade fault dataset, the experimental results show that the fault prediction model proposed in this paper has better performance and has certain effectiveness and feasibility for equipment fault prediction.