This paper describes an exercise carried out to establish the capability of time efficient CFD based simulations to describe the injection of fuel from a plain jet into a crossflow. The exercise provides a quantitative assessment of the accuracy with which simulation tools can predict spray plume trajectory, dispersion, droplet size and velocity. The effort involved collection of detailed data associated with a spray plume produced by Jet-A injected from a plain jet into a high speed crossflow at 1000 deg K. A document outlining the details of the inlet and boundary conditions needed to set up the CFD solution was provided to seven participants which represented gas turbine manufacturers, CFD vendors, research firms, and component providers. These participants provided at least one simulation of the problem posed. The participants were given 6 weeks to complete the exercise and return the results to ERC which then carried out the detailed comparison of the results with the experimental data. Penetration height, plume width, volume flux, and integrated volume flux were compared at two downstream axial locations. D_(32) and average spray axial velocity are compared amongst the CFD results. The results shown herein emphasize the characteristics of the spray plume (i.e., penetration and plume width) and indicate (1) a wide range in the simulation results and (2) generally marginal agreement with the experimental results. For example, the penetration 1" downstream was overestimated by 53% on average and width overpredicted by 111%. No dominant trend relative to modeling approach and overall agreement with the measured results was found. As a result, it is difficult to derive a "best practice" recommendation relative successful application of the modeling approaches evaluated in this study.
展开▼