Space deployable bistable composite structures with C-cross section based on machine learning and multi-objective optimization

A data-driven computational framework combining machine learning and multi-objective optimization is developed for the design of space deployable bistable composite structures with C-cross section. A C-cross section thin-walled deployable composite structure has bi-stability compared with other deployable structures, which has attracted many attentions thanks to its application prospects in roll-out solar array. In order to get the optimal geometric parameters of subtended angle, thickness, initial radius and longitudinal length, combination methods of finite element method, multi-objective optimization technique and experiment method for bistable composite structures with C-cross section are proposed in this article. The optimal Latin hypercube sampling algorithm is used to obtain the sample points of variables for the design of experiments. The surrogate models of the deployable structure will be obtained by response surface method and the non-dominated sorting genetic algorithm-II is used to obtain Pareto-optimal solution. The mechanical responses of the structure, which are set as optimization objectives are obtained by finite element simulation including the snap-through process and coiling process. Experimental results verify the accuracy and effectiveness of this optimization result. Furthermore, a parametric study of the geometric parameters is performed to determine the effect on the mechanical behavior and fully coiling-up stability. Lastly, the computational strategy proposed can be applied to different design problems in composite structures, which can also guide the design of roll-out solar array.