Structural Health Monitoring of Aircraft Engines Using Self Attention Multi-Head Machine Learning Technique

One of the basic tools in prognosis and health care is data based life-expectancy estimation prediction. In this paper, We provide a model to enhance the precision of her RUL forecasts for aviation engines, which is purely based on attentional mechanisms. This attention models include two subtypes: timing enhanced attention models and multi headed intra-attention models. Scaled dot product attention is used in multi-headed self-attention models to extract temporal dependencies, while to improve and expedite the feature selection process, timing-improved attention models are employed. In this research, data from the Commercial Modular Air Propulsion System Simulation (C-MAPSS) turbo fan engine simulation developed at NASA Ames is used to compare the suggested technique with existing models. With an RMSE value of 8.08 for FD001 and 8.84 for FD003 in the C-MAPSS data set, the suggested model was able to perform better. The model’s superiority in terms of methodology is supported by experimental findings.