Sensitivity Filtering Method Considering Partition Blending Weights

In order to get continuum structures of topology optimization with a crisp boundary, a sensitivity filtering method considering partition blending weights is proposed. The method applies a continuous weighting in the neighborhood to eliminate gray-scale cells. Firstly, the partition parameter is used to divide the cells in the filter radius area into sub-areas inside and outside, and the number of cells in each sub-domain can be controlled by adjusting the parameter. Secondly, the weight factor of cells in sub-region is evaluated by mean weight and exponential function respectively. While the influence of the units near the center on the sensitivity of the target function is strengthened, the averaging effect of the topological boundary caused by low density cells is weakened. The proposed method is combined with the framework of solid isotropic microstructures with penalization model. The optimization results of different numerical examples,which are solved by various methods, are compared and analyzed. Numerical results show that the gray-scale cells are inhibited and do not appear to the phenomenon of instability.