Potential Noise Source Identification in a Single Stage Rotor-Stator Compressor
2.1 Noise Generation Mechanisms
As the demand for the control of domestic environments continues to grow, state-led legislation has driven the requirement for more efficient and quieter domestic air moving technologies. Typical domestic Dyson air moving machines are driven by low speed low pressure rise compressors with Reynolds numbers below 100,000. A single stage rotor-stator research compressor was used in this work to evaluate the key aero-acoustic performance characteristics across the compressor map. A combination of low fidelity full annulus RANS CFD simulations and higher fidelity DES CFD simulations were used to investigate a range of operating points. All of the CFD simulations were validated using ducted compressor aero-acoustic measurements as well as high fidelity phase-locked hotwire measurements at various inter-stage locations.
The RANS data was used to explore the onset of stall and the bulk inter-stage flow parameters that could lead to increased noise generation. The results demonstrated that steady state RANS CFD has significant difficulty in predicting the stall margin, and this was heavily dependent on the RANS setup; the turbulence modelling applied as well as solver being used predicted notably different stall flow rates. The length of the simulated upstream and downstream duct was also assessed, and the impact of the Greitzer ‘b’ parameter was trialled in the simulations in relation to the measured data.
A detailed investigation was completed looking at noise increases toward choke and stall, and the corresponding flow phenomena that potentially caused this. The spectral content of the flow was assessed at the extremities of the compressor map to deliver further understanding of the secondary flow features generating noise. The DES simulations were used to further interrogate the flow physics at the various compressor operating points, with keen attention being paid to the impeller tip leakage flows, impeller suction side separation, and finally wake interaction with the stator. Finally the hot-wire data was linked to the DES simulation data and used to further describe the unsteady flow features causing the increase and decrease of acoustics across the compressor operating range.