Investigation of the Accuracy Requirements for Permeable Surfaces Used in Rotor Noise Prediction

This paper addresses the rotor noise prediction problem using the permeable Ffowcs Williams – Hawkings (FW-H) surface approach. The permeable surface method has become very popular and been widely used since it has been demonstrated that the Kirchhoff method can produce unreliable and erroneous results in normal applications. However, a detailed analysis of the accuracy requirements and error associated with permeable surfaces has not been widely addressed. In this paper, estimates of the spatial and temporal accuracy requirements for the acoustic prediction are provided. Three different type of acoustic data surfaces have been used for the noise prediction; impermeable surface, rotating permeable surface and non-rotating permeable surface. Wind tunnel test data is used as a baseline to compare the acoustic predictions for each of the three surface types. It is shown that impermeable surface did not capture the sound pressure negative peak amplitude caused by the transonic flow around the blade, but the correct amplitude and waveform shape are predicted by the permeable surface prediction. An analysis has been performed to determine which parts of the permeable surface are most responsible for the acoustic signal. Rotating permeable surfaces ensure more accuracy of flow data on the surface since they are positioned close to the actual blade surface while the solution accuracy of flow data on the nonrotating permeable surface is lower since it is located far away from actual blade. However, rotating permeable surfaces experience a Doppler singularity problem when the surface moves with sonic speed. Non-rotating permeable surfaces avoid this singularity because they encloses all blades and only translate with rotor. High-speed forward flight cases have been evaluated to examine the effect of the permeable surface type on the acoustic prediction accuracy. The sample cases show that the nonrotating permeable surface does not captured the negative peak of sound pressure as well as the rotating surface for a microphone on the advancing side of the rotor when the CFD flow solution is of questionable accuracy to begin with. Another potential source of error, associated with the vorticity passing through the permeable surface, is also discussed.