Optimization of low-turbulence structures and neural network-driven precise wind speed Control: A synergistic design for novel jet wind tunnel systems

Abstract

Jet wind tunnels are critical equipment for achieving high-precision aerodynamic testing, particularly suited for aerospace applications with extremely high requirements for turbulence intensity and control accuracy. However, traditional wind tunnel design has long been limited by the independent development of aerodynamic structures and control systems, resulting in poor overall system performance and limited flow field control precision. To address this, this paper proposes a synergistic design method centered on a Coaxial Intelligent Flow Converger (CIFC), through deep integration of aerodynamic structural optimization and intelligent algorithms. In terms of structural optimization, the CIFC adopts an innovative main/auxiliary dual-channel design, which suppresses airflow disturbances while enabling precise flow regulation. In terms of intelligent algorithms, a multi-parameter, multi-task CNN-BiLSTM-Attention deep learning architecture is developed to realize real-time coordinated control of the CIFC's dual channels and establish a high-precision nonlinear mapping model among temperature, pressure, channel flow rates, and wind speed/pressure difference coefficients. Experimental validation shows that this synergistic design exhibits excellent performance in flow field uniformity, temperature gradient control, flow stability, and environmental adaptability, providing theoretical support and engineering pathways for the construction of next-generation high-precision aerodynamic testing platforms.