A combined experimental and computational research effort investigated the noise sources of a pneumatic bleed valve used in turbofan engines and developed engineering solutions for attenuating those sources. The experimental effort employed 1/4-scale, rapid-prototyped valve designs which enabled the exploration of a large parameter space. Microphone array systems surveyed the sound and its sources, and Pitot surveys measured the mean velocity downstream of the valve. The numerical code solved the Reynolds-Averaged Navier Stokes (RANS) in both steady and unsteady modes. The research addressed the flow and acoustics of the valve without and with a muffler. The noise of the isolated valve consists of conventional jet noise with strong excess sound originating from the valve exit. The excess sound was traced to vortex shedding by support struts. Streamlining of the struts, aided by computational parametric studies, eliminated this noise component. The noise of the complete valve, with muffler attached, was attenuated by two principal means: using the streamlined struts and adding a honeycomb flow straightener to the muffler cavity. Proper installation of the honeycomb is critical for maximizing noise reduction. The resulting best design achieved a reduction of approximately 8 decibels in perceived noise level.

No abstract

The performance of four different turbulence models in addressing shock wave-boundary layer instabilities is investigated. The problem chosen for this goal is a transonic flow over a 10% thick circular arc airfoil in a channel. The self-excited shock motion over the circular arc airfoil has been investigated before experimentally and those results are used as a benchmark for current study. Unsteady RANS and DES methods in combination with different turbulence models are used. All the methods can successfully predict the overall shock oscillatory behavior. Yet there are minor differences in frequency prediction. Another reason for choosing this problem is to better understand the physics governing the problem. It is found that the shock oscillation frequency strongly depends on mean shock wave location.