Thermal error correction method for CNC machine tools based on improved GWO algorithm

Abstract In the contemporary manufacturing industry characterized by the pervasive utilization of computer numerical control tools, machine tool errors have become a key factor affecting machining accuracy. To effectively correct the thermal earrors of computer numerical control tools, a method based on an improved grey wolf optimization algorithm is developed. Firstly, the study combines fuzzy clustering algorithm and average influence value method to effectively collect and select features for temperature data of machine tools to reduce redundant data and improve the quality of model input features. Subsequently, a generalized regression neural network is used to build a thermal error prediction model, and the model parameters are optimized by improving the grey wolf optimization algorithm to enhance prediction accuracy and robustness. The experiment outcomes denoted that the model performed the best in ablation testing, achieving accuracy, recall, and F1 values of 0.95, 0.97, and 0.96, respectively. In the error test, the mean square error value of the model was as low as 0.06. In addition, in practical applications, the model achieves a prediction accuracy of 94.0% and an error correction rate of 89.2% in the robot arm precision processing. It can be seen that this model can provide an efficient and reliable solution for correcting thermal errors in computer numerical control tools, and has strong practical application value.