Concurrent topology and build orientation optimization based on neural network
3D printing can produce complex structures and has been widely used in various industries such as aviation,industrial manufacturing,and medicine.However,it also has certain limitations.Topology optimization is a method that seeks the best material distribution in space to achieve lighter components with the same performance using less material.In recent years,the combination of 3D printing and topology optimization has attracted widespread attention.It has been shown that the printing direction of a part can significantly affect printing quality and is closely related to its topology structure and required support volume.To address this problem,a concurrent topology and build orientation optimization framework based on the artificial neural network(ANN)is proposed in this paper.The feasibility of the framework is validated on supporting structures.To achieve the objective,a differentiable loss function is constructed based on structural compliance and the required support volume,and then concurrently optimizes structural topology and build orientation.The proposed framework has been tested on various problems in both 2D and 3D domains,which demonstrate that the proposed framework can significantly reduce the required support volume while maintaining a similar level of structural compliance,thus reducing printing costs.
3D printingtopology optimizationbuild orientationneural networksupport struc-ture
万方数据
维普
