Prediction of mechanical equipment fault diagnosis based on IPSO-GRU deep learning algorithm

Abstract

Abstract Exploring effective logistics machinery and equipment fault diagnosis and prediction technology to achieve efficient and stable operation of logistics machinery and equipment. In this paper, starting from the logistics machinery and equipment fault diagnosis technology, we optimize the hyperparameters of the recurrent gate unit neural network by using the improved second-order oscillatory particle swarm algorithm and then construct the IPSO-GRU logistics machinery and equipment fault prediction model. The IPSO-GRU model is used to test the prediction effect of the hydraulic lift table and logistics hoist by using the historical data of the hydraulic lift table as training data. The prediction accuracy of the IPSO-GRU model was improved by 6% compared with BP neural network. From the prediction results of the logistic hoist, only 6 out of 250 data samples failed to achieve accurate prediction. This shows that the IPSO-GRU model can effectively achieve the prediction of logistics machinery and equipment fault diagnosis and also provides a proven method for predictive maintenance of logistics equipment.