A standing problem in topology optimization problems is the efficient enforcement of manufacturing constraints such as minimum or maximum feature sizes. A large number of methods have been proposed including filtering techniques, perimeter control and gradient control. In this paper we propose some new filtering schemes that may alleviate the problem of grey transition regions in optimized topologies and ensure a minimum length scale for structural features and holes.
The new filter techniques are based on ideas borrowed from image morphology operators. It is shown that operators like “dilation”, “erosion”, “opening” and “closing” can be used to obtain black and white solutions with minimum feature size control using a continuation scheme. The schemes are implemented based on efficient analytical sensitivity calculations and tested on standard benchmark problems.