Simplified CFD for pressure drop predictions in ducts

Abstract

In the design of a ventilation system, the ductwork is generally composed based on analytical equations in combination with loss coefficients. This widely used approach can lead to large deviations in the prediction of the total pressure drop. Poor performance optimization could on its turn lead to a higher energy demand. While the use of computational fluid dynamics (CFD) could improve this prediction, it is often not feasible in practice as the use of CFD requires a high computational demand and a high-level of expertise. An alternative could be the use of simplified CFD. In this study the application of several conventional and simplified CFD methods and analytical prediction methods in the ventilation system design process was assessed. The simplified CFD methods include the use of coarse-grid CFD and voxel-based CFD simulations. Measurements on a single and double elbow configuration were performed and were used for validation purposes. For the investigated configurations the analytical prediction started to deviate from the measurements with increasing complexity of the system. For the conventional and coarse-grid CFD methods the prediction of the pressure drop was highly sensitive to the applied near-wall treatment and roughness parameters. For the voxel-based CFD method an average percentage difference of only 3% with respect to the measurements was found for the single elbow configuration. However, it overestimated the measured pressure drop by 64% on average for the double elbow configuration. A general sufficiently accurate method for predicting the pressure drop, which would also be feasible for use in practice ventilation system design, was not yet found. As this paper is part of a larger study, the assessment of additional simplified CFD methods on more complex duct configurations is the subject of future work.