Bayesian Physics Informed Neural Networks for Reliable Transformer Prognostics

2025-09-19

AI Summary

Transformer aging prediction faces challenges including incomplete physical models, weak uncertainty quantification, and difficulty embedding partial differential equations (PDEs) into data-driven frameworks. Method: This paper proposes a Bayesian physics-informed neural network (Bayesian PINN) framework. It encodes the thermal-stress-driven insulation degradation mechanism—formulated as a heat diffusion PDE—into a prior distribution and employs variational inference for full Bayesian uncertainty quantification over both parameters and predictions. The model jointly constrains the PDE residual, finite-element simulation data, and sparse real-world measurements from photovoltaic power plants. Contribution/Results: Experiments demonstrate that the proposed method significantly outperforms the Dropout-PINN baseline in both predictive accuracy and uncertainty calibration. It enhances generalizability and interpretability under data scarcity, enabling probabilistic lifetime assessment and supporting intelligent operation and maintenance decisions for critical power equipment.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Relational Probabilistic Models
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Social Networks and Social Media: Influence propagation, information diffusion, and the prediction on networks

Abstract

Scientific Machine Learning (SciML) integrates physics and data into the learning process, offering improved generalization compared with purely data-driven models. Despite its potential, applications of SciML in prognostics remain limited, partly due to the complexity of incorporating partial differential equations (PDEs) for ageing physics and the scarcity of robust uncertainty quantification methods. This work introduces a Bayesian Physics-Informed Neural Network (B-PINN) framework for probabilistic prognostics estimation. By embedding Bayesian Neural Networks into the PINN architecture, the proposed approach produces principled, uncertainty-aware predictions. The method is applied to a transformer ageing case study, where insulation degradation is primarily driven by thermal stress. The heat diffusion PDE is used as the physical residual, and different prior distributions are investigated to examine their impact on predictive posterior distributions and their ability to encode a priori physical knowledge. The framework is validated against a finite element model developed and tested with real measurements from a solar power plant. Results, benchmarked against a dropout-PINN baseline, show that the proposed B-PINN delivers more reliable prognostic predictions by accurately quantifying predictive uncertainty. This capability is crucial for supporting robust and informed maintenance decision-making in critical power assets.

Problem

• Bayesian Physics-Informed Neural Networks for transformer prognostics
• Incorporating heat diffusion PDEs for insulation degradation modeling
• Quantifying predictive uncertainty for maintenance decision-making

Innovation

• Bayesian Physics-Informed Neural Networks framework
• Embedding Bayesian Neural Networks into PINN architecture
• Heat diffusion PDE as physical residual component