In-Process Tool Condition Forecasting of Drilling CFRP/Ti Stacks Based on ResNet and LSTM Network

Tool condition forecasting (TCF) is a key technology for continuous drilling of CFRP/Ti stacks, as the tool wear always rapid and severe, which may further induce unexpected quality issues. However, cutting spindle power vibration signals change are complex, influenced by many factors due to different materials properties. The TCF stacks remains challenging, sensitive features difficult extract, decide accuracy robustness. Aiming monitor forecast an in-process method based on residual neural network (ResNet) long short-term memory (LSTM) has been proposed in this paper. Using preprocessed method, LSTM with ResNet-based model integrated can tool-wear values next holes. A case study demonstrated effectiveness TCF, where results using raw measured datasets tested comparison. mean absolute error (MAE) 45.01 μm, 2.20 times bigger than that preprocess signals. With data preprocessing improve MAE 20.43μm level, meets demand online TCF.