Optimization technique for design of automotive air filter housings with improved fluid dynamic performance and filtration

A filter housing design technique has been developed that computes the housing geometry required to provide a user-specified velocity distribution through the filter. Appropriate control of the flow field across the filter can improve filter efficiency. Two-dimensional computational designs have been performed specifying a uniform velocity distribution through the filter. This distribution corresponds to a uniform, constant pressure drop across the filter from the upstream to the downstream side. A computational optimization method is used to minimize the variation in the pressure drop along the filter by changing the geometry of the upper wall. Computational fluid dynamics (CFD) calculations of the viscous laminar flow upstream and downstream of the filter are performed using two-dimensional Navier-Stokes equations. As the upper wall profile is numerically changed, the CFD solution for the computations is repeated and the pressure drop variation is reevaluated until an optimum configuration is achieved. The optimization results have produced a pressure distribution that is very close to the specified uniform distribution. Experimental verifications were performed using a laser Doppler anemometer to measure the velocity distribution just upstream of the filter and a pressure transducer to measure pressure drops across the filter. The velocity distribution in this (optimized) housing is more uniform than other housing models having the same fundamental rectangular geometry.