ShuffleNet v2.3-StackedBiLSTM-Based Tool Wear Recognition Model for Turbine Disc Fir-Tree Slot Broaching

At present, deep learning technology shows great market potential in broaching tool wear state recognition based on vibration signals. However, traditional single neural network structure is difficult to extract a variety of different features simultaneously and has low robustness, so the accuracy status not high. In view above problems, model ShuffleNet v2.3-StackedBiLSTM proposed this paper. The integrates v2.3, which been channel shuffling, StackedBiLSTM, long short-term memory network, effectively spatial temporal for recognition. Based innovative model, turbine disc fir-tree slot experiment designed, performance index system confusion matrix adopted. experimental research results show that outstanding accuracy, precision, recall, F1 value, rate reaches 99.37%, significantly better than v2.3 StackedBiLSTM models. speed sample was improved 8.67 ms, 90.32% less model.