Summary

Description

Broadband fan noise ? closely tied to turbulent flow on and around the fan blades ? represents a key challenge to the noise reduction community due to the interaction of a highly turbulent flow field with complex, moving geometries. Prediction and high-fidelity simulation of fan noise demands a fundamental innovation in CFD methods due to moving geometries and accuracy requirements. The objective this work is to develop a flexible approach to handling multiple, overset grids for use in simulations of turbomachinery. In Phase 1 we will develop an innovative computational software tool for efficiently managing multiple, overlapping structured meshes in relative motion. This application will be used concurrently with a compressible Navier-Stokes solver and is an enabling technology in enabling high-fidelity simulations of turbulent flows in complex, moving geometries. Phase 1 will demonstrate software feasibility using a simplified model of the NASA Glenn Source Diagnostic Test (SDT) fan at realistic take-off conditions. We propose a simulation that includes a moving "rotor" blade row adjacent to a static blade row. Tailored post-processing of simulation results will provide information on the turbulent flow ? and implied turbulent noise sources ? including unsteady blade surface pressures, acoustic modes, and overall radiated noise. In Phase 2 we focus primarily on broadband turbulent noise sources of modern turbofan engines. By utilizing a realistic NASA SDT fan geometry and take-off flow conditions, we will use our new tools to simulate real-world systems and commercialize our software product.