Research on acceleration methods for performance simulation of adaptive variable cycle engines based on principle-data fusion drive

The adaptive cycle engine (ACE) achieves comprehensive performance advantages for various flight missions through the collaborative regulation of variable mechanisms. However, the numerous variable mechanisms and intense changes in the operating states of ACE results in significant shortcomings in the convergence and computational efficiency of the existing performance simulation model. These shortcomings make it difficult to meet the demands for large-scale performance optimization design of ACE in the aircraft-engine co-design. A performance simulation acceleration method driven by both principles and data is proposed. This method constructs a general global initial guess prediction model based on principle analysis and a data-driven self-expanding initial guess variable fidelity surrogate model, forming a principle-data joint-driven engine initial guess prediction framework. This framework addresses the problems of poor convergence and low computational efficiency caused by the difficulty in reasonably selecting initial guesses for the performance simulation models under various configurations and design schemes. Numerical simulation verification is conducted to optimize the ACE control laws at equal-inlet flow throttle state, and the results demonstrate that the proposed method achieves more than a 38% improvement in convergence capability and a 45% increase in simulation computation speed compared to traditional methods. This method effectively supports large-scale simulation optimization of ACE design parameters and control laws in aircraft-engine co-design, holding significant engineering application value.