Health status assessment and fault warning methods for aircraft engines under time varying operating conditions

With the growth of the aviation transportation industry, aircraft engines, as the core components of flight safety, are facing increasingly severe challenges in health status assessment and fault warning technology. To achieve accurate evaluation and fault warning of engine status, this study proposes a new method using improved multi-channel network and hybrid network models. The new method can achieve life prediction and evaluation of engine health status in different time-varying scenarios by improving the multi-channel network. Meanwhile, the method achieves early warning of operational faults by using a hybrid network model for real-time analysis of aircraft engine operation data. The results demonstrated that the new method had root mean square errors of only 12.35 and 12.84 on different datasets, significantly better than other models. The score of the new model has also significantly decreased, with accuracy rates of 91.5% and 93.4% on different datasets, far exceeding other models. Moreover, although the new model had a large number of parameters, it had short training time, low latency, small memory usage, and excellent system performance. The new method can significantly improve the health status assessment and fault warning of engines, which has good guiding significance for achieving stable operation of aircraft engines.