Modelling Methodology for the Full-Field Fuel Distribution in a Scramjet Combustor with Adjustable Strut/Cavity

Abstract

The variable-geometry scramjet combustor represents a pivotal technology for wide-range and high-maneuverability aerospace vehicles. This study proposes an innovative adjustable strut/cavity configuration to achieve efficient fuel regulation. The fuel control characteristics are systematically analyzed using planar laser shadowgraphy experiments. A predictive model integrating proper orthogonal decomposition (POD) with deep multi-task learning (MTL) is developed for prediction of the full-field fuel distribution. The combustor operates under inflow conditions of the Mach number 2.0, the total temperature 300 K, and the momentum ratio 12, with geometric variations covering: the strut length (0–20%), the strut height (0–10%), the cavity length (0–30%), and the cavity depth (0–20%). The results show that (1) geometric adjustments of strut/cavity effectively modulate the global fuel distribution patterns; (2) the POD-MTL framework accurately establishes correlations between the geometric parameters and the fuel distribution, achieving the prediction accuracy with a relative error of 10%. This methodology provides theoretical foundations for real-time combustion optimization in hypersonic propulsion systems.