Low-Order, Data-Driven Method for Analysis of Impact of Fan Design on Fan Stage Broadband Interaction Noise

Fan blade design plays a crucial role in both fan stage aerodynamic performance and noise. Fast methods for assessing the broadband noise associated with a fan stage can facilitate the design process. This paper presents a two-part study of fan stage broadband interaction noise using a multi-geometry dataset with 59 fan geometries at various operating points. First, a RANS-informed machine learning model for predicting the fan wake flow quantities that are critical to the low-order acoustic model is assessed. Second, correlations that could be used in lieu of even a low-order noise calculation are considered. The downstream fan-stage broadband interaction noise associated with the fans in the dataset is computed using the low-order acoustic model with RANS generated fan wake flow quantities. The results are first viewed as a function of mass flow and total temperature rise as per existing correlations. Next the noise is associated to efficiency through a 5th power of the Mach number downstream of the fan and the 2nd power of the static pressure coefficient across the fan.