Unmeasurable Aero-engine Gas Path Parameters Reconstruction based on Generation Diffusion Model

As a complex aerodynamic and thermodynamic system, the deployment of onboard sensors in aero-engines is constrained by harsh operating environments, cost, and design limitations. These restrictions pose challenges for the comprehensive measurement of critical gas path parameters. Traditional data-driven methods struggle to accurately reconstruct unmeasured parameters under complex flight conditions, limiting their effectiveness in onboard engineering applications. To address this issue, this paper proposes an improved generative diffusion model for high-precision reconstruction of unmeasured gas path parameters. The proposed method establishes a reconstruction framework based on onboard sensor data by learning the mapping relationships between measurable and unmeasurable parameters. Specifically, a component-level model is first pre-trained using historical sensor data, and then real-time measurable parameters are incorporated to infer the unmeasured gas path parameters. Experimental results demonstrate that under limited sensor measurement conditions and complex flight scenarios, the proposed method can accurately reconstruct unmeasured parameters. Compared to traditional data-driven methods, the reconstruction error is reduced from 6.35% to 0.42%, significantly improving prediction accuracy. In conclusion, the proposed method provides high-quality data support for aero-engine health management systems.