Hybrid Gaussian Process Regression with Temporal Feature Extraction for Partially Interpretable Remaining Useful Life Interval Prediction in Aeroengine Prognostics

TL;DR

An adapted Gaussian Process Regression model for RUL interval prediction is introduced, tailored for the complexities of manufacturing process development and provides transparent, interpretable insights into uncertainty, contributing to robust process development and management.

Abstract

The estimation of Remaining Useful Life (RUL) plays a pivotal role in intelligent manufacturing systems and Industry 4.0 technologies. While recent advancements have improved RUL prediction, many models still face interpretability and compelling uncertainty modeling challenges. This paper introduces a modified Gaussian Process Regression (GPR) model for RUL interval prediction, tailored for the complexities of manufacturing process development. The modified GPR predicts confidence intervals by learning from historical data and addresses uncertainty modeling in a more structured way. The approach effectively captures intricate time-series patterns and dynamic behaviors inherent in modern manufacturing systems by coupling GPR with deep adaptive learning-enhanced AI process models. Moreover, the model evaluates feature significance to ensure more transparent decision-making, which is crucial for optimizing manufacturing processes. This comprehensive approach supports more accurate RUL predictions and provides transparent, interpretable insights into uncertainty, contributing to robust process development and management.

Figures (9)

• Figure 1: Dual-line interval RUL prediction framework.
• Figure 2: Illustration of the network in our method HRP.
• Figure 3: Distribution and importance for sub-dataset FD001. Red dashed lines represent the density of testing datasets, and blue lines represent the density of training datasets. The lower figure shows the importance ranking of 14 features.
• Figure 4: RUL prognostic performances of our algorithm for the testing engine units in four sub-datasets. The blue polyline represents the actual RUL. The range between the red curve and the green curve represents the predicted interval. The red segment indicates that the predicted RUL exceeds the actual RUL, misleading engineers into believing the machine can still operate. Conversely, the green segment signifies that the predicted RUL is less than the actual RUL. Engine 46 and 62 belong to FD001. Engine 145 and 185 belong to FD002. Engine 20 and 99 belong to FD003. Engine 32 and 68 belong to FD004.
• Figure 5: Subroutines of the engine simulation.