Experimental and Numerical Investigation of Effusion Cooling in an S-Bend Duct with Diffuser: Evaluation of Novel RANS Modelling Approach

Abstract

This study investigates both experimental and computational fluid dynamics analyses of a rectangular S-bend duct with diffusion, incorporating effusion cooling at three different locations along the duct. The goal was to explore an alternative CFD simulation method for effusion-cooled surfaces and to predict flow properties and backpressures at the duct inlet. A novel modelling approach, known as the slot method, was examined and compared against experimental data. In this method, effusion holes are replaced with a specific number of slots to simulate the impact of effusion air mass flow on the main duct flow. The first phase focused on investigating the slot method with Reynolds Averaged Navier-Stokes (RANS) models over a range of Reynolds numbers from 2.5×10^5 to 4.5×10^5 and a constant effusion coolant mass flow rate, with validation against experimental tests. The slot method produced acceptable results for the outlet axial velocities, wall static pressures along the duct, and other general characteristics of the flow, such as vortices and flow separation. However, for the diffuser’s pressure recovery performance, different numbers of slots led to varying results, with differences ranging from 10 to 50 percent across all three Reynolds numbers. Experiments were performed on the S-bend duct with effusion cooling at three distinct locations: the first and second convex sections, and the full surface of the left wall. Both cold and hot flow experiments were conducted, with hot flow temperatures ranging from 120 °C to 180 °C. Two effusion coolant mass flow rates were examined at each location: 0.024 kg/s (minimum) and 0.083 kg/s (maximum). The highest coolant mass flow rate resulted in severe back pressure penalties, while moderate penalties were observed at the lowest rate. Severe flow separation was observed for the second convex case even at the lowest mass flow rates. In the third phase, CFD simulations using the slot approach captured flow characteristics like outlet temperatures and axial velocities at the lowest coolant mass flow rates. However, it struggled with pressure recovery performance and wall static pressure predictions at higher mass flow rates because changing the number of slots led to a change in the plenum pressure of up to 30 percent.