Adjoint-based topology optimization of filter structures for gas–particle systems

Conventional filters for aerosol particle deposition consist of one or more filter layers, which are either woven or composed of tangled fibers. The quality of the separation results almost exclusively from the density of the fiber arrangement. Due to the manufacturing process, compromises between separation efficiency and pressure loss, which are in the opposite relationship to each other, are inevitable. The objective of this work is to develop a method for topology optimization to optimize filter structures for both higher filtration efficiency and lower pressure drop simultaneously using the adjoint method based on computation fluid dynamic simulations. The key to topology-optimized “bionic” filters is to find suitable cost functions controlling the optimization. These cost functions should take into account different separation mechanisms and pressure loss. The force coefficients for pressure and shear and the surface integrals of pressure and wall shear stress were evaluated for their contribution to the deposition as part of a combined cost function. In this work, a simple algorithm is devised to combine two opposing cost functions. First, promising results are obtained by considering solid particle separation from gas. For example, it was possible to increase the total filtration efficiency by 2% and reduce the pressure drop by 3.6% in one single deformation step.