CFD-based evolutionary algorithm for the realization of target fluid flow distribution among parallel channels

Abstract

In energy and process systems, the bad fluid flow distribution (flow maldistribution) is one of the main causes of performance deterioration, equipment dysfunction, safety defaults, product quality degradation, reduced lifetime and over-cost. Flow equidistribution among numerous parallel channels is a classical goal, but is not always the optimal option according to a given objective and constraints. Target flow distribution (uniform or non-uniform) is far more difficult to achieve because it requires the precise control of flow-rate in given time for every channel. There is no such attempt in the existing literature.In this study, a CFD-based evolutionary algorithm is developed to optimize the topology of an inserted perforated baffle, for the realization of target flow distribution among parallel channels. Several 2D numerical examples with different circuit geometries and different target curves are studied and compared. Results show that the optimized distribution curves obtained by performing the evolutionary algorithm are in good agreement with the target curves, with acceptable increase in pressure drop. This optimality algorithm may be easily applied to different working conditions. In this sense, the proposed algorithm is robust, effective, general and flexible, showing its promising application in various engineering fields dealing with fluid distribution problem.