Research on Fault Prediction Method of Printing Equipment Based on GRU Neural Network
HU Jie-ping, Shulin Yang · 2021 IEEE 3rd International Conference on Frontiers Technology of Information and Computer (ICFTIC) · 2021
Aiming at the problem of low prediction accuracy caused by poor timing of existing printing equipment fault prediction methods, this paper proposes a fault prediction method based on the gated recurrent unit (GRU) neural network. Firstly, the data were preprocessed, the sample average value was used to fill the missing value of the sample, and the data was standardized. Then, the GRU neural network was built based on Keras deep learning framework, and the input data were trained and tested. At the same time, the model parameters were optimized by KerasClassifier and GridsearchCV of SkLearn. Finally, the proposed method is compared with RNN, LSTM, SVM and other methods, and the recognition accuracy and network convergence speed are regarded as quantitative characteristics. The experimental results show that the proposed method has smaller error and higher prediction accuracy.