Validation of computational fluid dynamics simulations for determining pressure loss coefficients of ventilation components

Abstract

Ventilation systems include a variety of components for which necessary pressure loss data is often unavailable. Computational fluid dynamics simulations could substitute for expensive measurements, but validation simulations with suitable data are crucial to assess model uncertainties. Existing CFD validation studies either did not focus specifically on pressure losses, only covered few components, or did not include recent developments in turbulence modelling. In the present work, 33 bends, 4 gates and 2 tees were simulated using a consistent approach. Computational fluid dynamics simulations were validated with published data: rectangular high-edge and wide-edge bends from the experimental dataset of Sprenger, gates and diverging tees from the SMACMA guide. The considered flows cover important basic flow phenomena: deflection, splitting and flow separation. The 39 components were simulated with three turbulence models at 14 Reynolds numbers. The simulations predicted pressure loss coefficients accurately for various components. Cases with strong flow separation regions were most challenging. The model prediction uncertainty was assessed by carrying out simulations with three selected turbulence models. As in the experimental data from Sprenger, the simulations showed a distinct dependence of pressure loss coefficients on the Reynolds number for bends. In contrast, for abrupt deflections and flow separation at sharp edges, the Reynolds number dependency was minor. Technical pressure loss data of ductwork components is needed for the dimensioning, optimisation, and energy assessment of ventilation systems. The present validation study assesses the present state of the art of CFD simulations to determine pressure loss coefficients and the resulting prediction uncertainties.