Adaptive density-based robust topology optimization under uncertain loads using parallel computing

Abstract This work presents an efficient parallel implementation of density-based robust topology optimization (RTO) using adaptive mesh refinement (AMR) schemes permitting us to address the problem with modest computational resources. We use sparse grid stochastic collocation methods (SCMs) for transforming RTO into a weighted multiple-loading deterministic at points. The calculation these problems and functional sensitivity is computationally expensive. combine distributed-memory computing AMR techniques efficiently. former allows exploit resources available, whereas latter permits increase performance significantly. propose incremental contribution maintaining similar memory allocation one used in counterpart. cumulative uses buffers adapt evaluation points exploitation embarrassing parallelism SCMs. evaluate coarse generated each iteration performance. perform regularization design variable update fine obtain equivalent such mesh. proposal two- three-dimensional test its feasibility scalability. also check improvement nodes. Finally, we compare same approach different preconditioners without showing significant improvements.