1. Research on Grinding Force Prediction of Flexible Abrasive Disc Grinding Process of TC17 Titanium Alloy
- Author
-
Jihao Duan, Zhuofan Wu, Jianbo Ren, and Gaochen Zhang
- Subjects
abrasive disc grinding ,normal grinding force ,response surface method ,PSO-BP neural network ,Mechanical engineering and machinery ,TJ1-1570 - Abstract
Abrasive disc grinding is currently a key manufacturing process to achieve better accuracy and high-quality surfaces of TC17 components. Grinding force, which results from the friction and elastic–plastic deformation during the contact and interaction between the abrasive grains and the workpiece, is a critical parameter that represents the grinding accuracy and efficiency. In order to understand the influence factors of grinding force, the characteristics of the flexible abrasive disc grinding process were studied. Considering the contact state between the abrasive tool and the workpiece, the theoretical model of normal grinding force was established in detail, from macro- and micro-perspectives. By conducting single-factor and orthogonal grinding experiments of TC17 components, the influence of different process parameters on the normal grinding force was revealed. The normal grinding force prediction models of the abrasive disc grinding process were developed based on the Box–Behnken design (BBD) and particle swarm optimization–back propagation (PSO-BP) neural networks, respectively. The results showed that the normal grinding force was negatively correlated with the disc rotational speed, and positively correlated with the contact angle, grinding depth, and feed rate, and the interaction of the factor feed rate and grinding depth was the more influential factor. Both the BBD and PSO-BP force models had good reliability and accuracy, and the mean absolute error (MAE) and mean relative error (MRE) of the above two prediction models were 0.22 N and 0.16 N, and 13.3% and 10.9%, respectively.
- Published
- 2024
- Full Text
- View/download PDF