A constructive solid geometry-based generative design method for additive manufacturing

Abstract

The unique capability of additive manufacturing (AM) to deal with complex geometries drives the traditional topological optimization and lightweight design to a new paradigm. However, AM constraints are still hard to integrate into the optimization procedure, and the multiple conflict design objectives are difficult to handle when making decisions. In addition, the computation cost is usually high. To solve these problems, this paper proposes a new generative design method with a manufacturing validation so that the designer’s decision-making is more efficient. This method first uses a CSG (constructive solid geometry)-based technique to generate and represent topology geometries with smooth boundaries and parametric control. Then, a genetic algorithm is used to operate the CSG geometries in order to search for optimal solutions. Finally, a set of finite optimal non-dominated design solutions on the Pareto front are located and presented for the designer’s further decision making. The proposed method can generate a large quantity of qualified optimal alternative solutions with smooth geometric boundaries but the computation cost is less. It is promising to design qualified solutions, especially for lightweight structure design, in AM. Several demonstration examples and comparison case studies with existing methods in literature are presented at the end of this paper to show the advantages.