Bayesian Joint Model of Multi-Sensor and Failure Event Data for Multi-Mode Failure Prediction

2025-06-20

AI Summary

For industrial systems subject to multiple failure modes, existing remaining useful life (RUL) prediction methods either decouple failure-mode analysis from lifetime modeling or rely on black-box models lacking statistical interpretability—thus failing to jointly achieve accuracy, interpretability, and rigorous uncertainty quantification. This paper proposes the first hierarchical Bayesian joint modeling framework that unifies three components: a Cox proportional hazards model for time-varying degradation dynamics, a convolutional multi-output Gaussian process to fuse heterogeneous multivariate time-series sensor signals, and a polynomial failure-mode distribution to characterize competing failure types. Efficient posterior inference is enabled via variational inference and Monte Carlo sampling. Evaluated on an aircraft engine dataset, the method achieves significantly improved RUL prediction accuracy and well-calibrated uncertainty estimates, outperforming state-of-the-art independent modeling and black-box approaches in comprehensive performance metrics.

Technology Category

• Reasoning under Uncertainty: Relational Probabilistic Models
• Machine Learning: Calibration & Uncertainty Quantification
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semantics
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

Abstract

Modern industrial systems are often subject to multiple failure modes, and their conditions are monitored by multiple sensors, generating multiple time-series signals. Additionally, time-to-failure data are commonly available. Accurately predicting a system's remaining useful life (RUL) requires effectively leveraging multi-sensor time-series data alongside multi-mode failure event data. In most existing models, failure modes and RUL prediction are performed independently, ignoring the inherent relationship between these two tasks. Some models integrate multiple failure modes and event prediction using black-box machine learning approaches, which lack statistical rigor and cannot characterize the inherent uncertainty in the model and data. This paper introduces a unified approach to jointly model the multi-sensor time-series data and failure time concerning multiple failure modes. This proposed model integrate a Cox proportional hazards model, a Convolved Multi-output Gaussian Process, and multinomial failure mode distributions in a hierarchical Bayesian framework with corresponding priors, enabling accurate prediction with robust uncertainty quantification. Posterior distributions are effectively obtained by Variational Bayes, and prediction is performed with Monte Carlo sampling. The advantages of the proposed model is validated through extensive numerical and case studies with jet-engine dataset.

Problem

• Predict multi-mode failures using multi-sensor time-series data
• Integrate failure modes and RUL prediction in a unified model
• Quantify uncertainty in predictions with Bayesian framework

Innovation

• Bayesian joint modeling of multi-sensor and failure data
• Hierarchical framework integrating Cox model and Gaussian Process
• Variational Bayes and Monte Carlo for uncertainty quantification