Aero-engine Degradation Evaluation Based on First-Order Transition Function and Gaussian Mapping Networks

This paper presents significant advancements in the domain of aero-engine degradation evaluation by introducing an amalgamation of state transition functions, Gaussian mapping networks, and Kalman filtering. Conventional thermodynamic models, when applied to military turbofan engines, often encounter issues related to convergence and computational efficiency. To overcome these challenges, we introduce a first-order state transition function designed specifically for degradation assessment, taking degradation rate into consideration. Furthermore, we propose a pioneering Gaussian mapping network-based measurement function to address convergence challenges and enhance computational efficiency. To seamlessly integrate a degradation prior with measurement regression, we introduce an innovative Kalman filtering iteration method. This paper provides a comprehensive exposition of the mathematical models, derivation processes, and practical implementation of these techniques. Our approach demonstrates precise degradation evaluation on a simulated degradation dataset, capitalizing on the precision of first-order transition functions and Kalman iterative process and the computational efficiency achieved through the combination of Gaussian mapping networks. Additionally, this innovative framework exhibits substantial potential for real-world applications in aero-engine performance evaluation, as exemplified by the actual flight dataset test case.