Stress-based topology optimization using bi-directional evolutionary structural optimization method

This work proposes an evolutionary topology optimization method for stress minimization design using the bi-directional evolutionary structural optimization (BESO) method. The discrete nature of the BESO method avoids naturally the well-known “singularity” problem in density-based methods with degenerated materials. The p-norm stress aggregation scheme is adopted for the measure of global stress level. A computationally efficient sensitivity number formulation is derived from the adjoint sensitivity of the global stress measure. With regard to the highly nonlinear stress behavior, both sensitivity numbers and topology variables are filtered to stabilize the optimization procedure; meanwhile, the filtered sensitivity numbers are further stabilized with their historical information. The method has been shown efficient, practical and easy-to-implement through a series of 2D and 3D benchmark designs. •An extended BESO method for stress-based topology optimization is proposed.•The discrete nature of the method avoids naturally the “singularity” problem.•The proposed method is efficient, practical and easy-to-implement.•The method has been validated through both 2D and 3D benchmark tests.