Abstract

Basic structural optimization has been available for the past three decades and with the capabilities of additive manufacturing and implicit geometric generation tools, these generative design techniques have expanded the capabilities of structural optimization. There are a few existing efforts to generate optimized geometries specifically for fluid applications, however, these efforts are quite cumbersome and hard to implement early in the design phase. Typically, the CFD simulations are performed late in the design process for Validation & Verification of the design. To enable truly fluid-optimized components the use of CFD needs to be used earlier in the design process. This paper will present a new method to leverage the use of CFD simulation fields at the onset of the design process to generate an optimum geometry for fluid performance. There are some common constraints a design engineer is faced with. One is generating a geometry that equalizes the flow rate at several outlets with a given inlet flow rate. While another is to determine an optimum topology within a given design domain that minimizes pressure drop. Skilled CFD & design practitioners can come up with these topologies but this process requires several iterations and significant domain knowledge. As product complexity increases and multiple design constraints get introduced this task quickly becomes infeasible with our current methodologies. Today, a simple, reusable design process can utilize the three dimensional fields (velocity, pressure, turbulence, etc) from CFD simulation coupled with a powerful implicit geometric modeler to generate optimal geometries. The imported three dimensional field is truncated based on the magnitude of the nodal results to produce an optimized geometry. A new CFD simulation can be performed on the generated geometry and the above process repeated. This process will be demonstrated on a variety of geometries and fluid applications with various constraints. A final CFD simulation validates that the design meets the performance requirements.