Optimization of a Micromixer with Automatic Differentiation

As micromixers offer the cheap and simple mixing of fluids suspensions, they have become a key device in microfluidics. Their performance can be significantly increased by periodically varying inlet pressure, which leads to non-static flow improved process. In this work, micromixer with T-junction meandering channel is considered. A periodic pulse function for pressure numerically optimized regard frequency, amplitude shape. Thereunto, fluid adsorptive concentration are simulated three-dimensionally lattice Boltzmann method (LBM) OpenLB. Its implementation then combined forward automatic differentiation (AD), allows generic application fast gradient-based optimization schemes. The quality shown 21.4% comparison static, passive regime. Methodically, results confirm suitability combination LBM AD solve process-scale problems accuracy over difference quotient approaches context.