Turbofan Engine Health Assessment Based on Spatial-Temporal Similarity Calculation

Area of Science:

• Aerospace Engineering
• Mechanical Engineering
• Data Science

Background:

• Aviation turbofan engine remaining useful life (RUL) prediction accuracy is often limited by complex operating conditions and sensor noise.
• Accurate RUL prediction is crucial for effective engine operation and maintenance decision-making.

Purpose of the Study:

• To propose a novel residual useful life prediction method for aviation turbofan engines.
• To address the low accuracy issues in RUL prediction caused by complex operational data and sensor noise.
• To enhance the reliability and accuracy of RUL predictions for turbofan engines.

Main Methods:

• A spatial-temporal similarity calculation method is proposed, involving adaptive sequence matching.
• The method constructs spatial-temporal trajectory sequences for optimal sample matching and similarity calculation.
• A weight training module with two allocation algorithms and a life prediction module are utilized to determine the final RUL.

Main Results:

• The proposed method demonstrated improved RUL prediction accuracy on the FD004 dataset, reducing root mean square error (RMSE) by 12.6% and health score (Score) by 14.8%.
• On other datasets, RMSE remained comparable, while the Score index decreased by approximately 10%.
• The method enhances sample similarity through sequence matching, leading to more accurate RUL predictions.

Conclusions:

• The developed spatial-temporal similarity-based method significantly improves the accuracy of remaining useful life prediction for aviation turbofan engines.
• This approach offers a robust solution for operational challenges posed by complex engine data and sensor noise.
• The findings provide valuable support for optimizing the operation and maintenance strategies of turbofan engines.