A New Explicit Three Dimension Topology Optimization Approach And Length Scale Control Based On MMC Framework

As one of the most powerful tool for structure design,Topology Optimization has been widely applied in many engineering fields.Since three dimension topology optimization(3D Top Opt)method hold theory background for engineering structure design,it has driven much research attention.However,traditional 3D Top Opt approach always has poor computational efficiency for large computational effort.The widely adopted SIMP approach use density of discrete element to get implicit topology description for continuum structure,while level set method use values of function in high dimension on mesh grids.Tremendous description variables are needed in these approaches because both of them see structure topology implicitly.Besides,the optimization model and the analysis model are usually strongly coupled in this kind of approach,thus to get clear picture of result there is no other choice but making denser meshes,and inevitably leading to large computational effort in finite element analysis.Explicit topology optimization method is a novel approach proposed in recent few years.It's distinguish character is using explicit parameters to depict structure topologies.By combination of moving morphable components/voids,an extreme complex structure can be represent.The explicit method has many natural advantages such as less design variables,black-and-white distribution of material,decoupling with analysis model,explicit structure boundaries and direct connecting with CAD system.These merits make it a more potential way to adopt this method to deal with geometry-related and computational effort-related problems in topology optimization.In this thesis,a new three dimension topology optimization approach based on explicit description of topologies is presented.The bottle-neck problem associated with 3D Top Opt will be well addressed with use of the proposed approach.Numerical test show that even with very little design variables detailed and complex result can still be obtained,and the solution efficiency will be enhanced significantly.An explicit and straightforward scheme to control length scale in topology optimization will also be presented in this paper.Inspired by the fact that in the explicit topology optimization framework component already has geometrical characters,we find that maximum/minimum thickness of topology result can be simply restricted by upper/lower bound of thickness variables,with only all components in design domain form ideally connection.