Prediction of crater tool wear using artificial intelligence models in 7075 Al alloy machining

This paper reports how artificial intelligence (AI) tools are used to predict the evolution of crater tool wear. The choice of using AI methods for predicting tool wear in this study is driven by the inherent capabilities of AI algorithms to handle complex and non-linear relationships within the data. Traditional methods may have limitations in capturing the intricate patterns and interactions between various machining parameters and tool wear. Tests were executed with tungsten carbide cutting tools to machining Aluminum 7075 Alloy and performed with a CNC lathe. Corner radius, feed rate, cutting speeds, and cut depth were studied in response to tools crater wear. Thirty experiments were performed: twenty-four were used in model training and six in tests, and another experiment was carried out with different cutting conditions to approve the chosen models. The novelty of this article lies in its effective prediction of tool wear. Additionally, this study is the first to explore 10 independent AI models in the context of tool wear prediction. Through hyperparameter search and careful tuning, the optimal learning rate for each model was determined to ensure effective convergence. The paper contends that the Gradient Boosting Model has been proven the best according to performance indices (R 2 = 0.9085, MAE = 0.05425, RMSE = 0.06635, RAE = 0.24265 and RSE = 0.09115) with deviations, among predicted and actual crater tool wear, has an average deviation of 8.27%.