A Machine Learning Based Body Force Model for Analysis of Fan-Airframe Aerodynamic Interactions

Abstract

Body force modeling is a numerical strategy that allows an accurate representation of the aerodynamics of turbomachinery blade rows at a reduced computational cost, making it suitable for predicting fan-airframe aerodynamic interactions in boundary layer ingestion (BLI) propulsive architectures. This paper focuses on a new approach for building the body force representation using a machine learning technique, rather than analytically modeling the effects of the blades in the flow. This methodology is developed and assessed in a distorted inflow case representative of a BLI configuration and compared to a full annulus unsteady computation.