Abstract: The milling force is an important process parameter in the milling of thin-walled aluminium alloy parts, with accurate feedback playing a significant role in reducing the deformation of the work pieces. In order to achieve the accurate prediction of milling force, firstly, the milling force data was obtained through simulation tests for the milling of thin-walled aluminum alloy parts. Secondly, to address the drawbacks of the traditional BP neural network, the Lion Swarm algorithm was used to improve it. The milling force data was imported into the improved network for training, to establish the LSO-BP prediction model. Finally, the LSO-BP model, PSO-BP model, and the traditional BP neural network model were used to predict the milling force respectively. The comparison results of evaluation indexes such as root mean square error, average relative error, and correlation coefficient show that the LSO-BP model significantly outperforms both the PSO-BP model and the traditional BP neural network model in predicting the milling force.

Key words: thin-walled parts, milling force, neural network, predictive model