Machining feature process route planning based on a graph convolutional neural network

The machining processes of machining features, as the crucial components of the machining process for the overall part, significantly impact machining quality and production efficiency. However, the existing methods for machining feature process planning primarily focus on the information of individual machining features and lack sufficient consideration of the overall part information. This deficiency results in reduced effectiveness of process design outcomes and limited direct applicability, necessitating significant manual adjustments. To address these limitations, we propose a novel approach for machining feature process planning using graph convolutional neural networks. The proposed method utilizes an attribute graph to efficiently represent the information of a part. In this representation, nodes symbolize machining features, while edges describe their interaction relationships. Subsequently, a graph convolutional neural network is constructed for learning the machining feature process planning model. After training, the proposed model achieved 93.31% accuracy in predicting process routes for machining features. In addition, the experimental results demonstrate the successful resolution of some current limitations in learning-based machining feature process planning. These findings underscore the potential of intelligent automation in this domain. Overall, this research contributes to the progress of intelligent process planning in manufacturing.

Automated summary

This research proposes a novel approach for machining feature process planning using graph convolutional neural networks. By representing part information with attribute graphs and constructing a learning model, the proposed method achieves higher accuracy and resolves current limitations in machining feature process planning.