Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Explainable, Interpretable & Trustworthy AI for Intelligent Digital Twin: Case Study on Remaining Useful Life

Kazuma Kobayashi, Syed Bahauddin Alam

TL;DR

This paper advocates integrating explainable AI (XAI) and interpretable ML (IML) into intelligent digital twins (DTs) for robust remaining useful life (RUL) prognostics in energy and nuclear contexts. It introduces a DT framework with an update module that combines Bayesian filtering and ML, and explores fast surrogate models via operator learning (DeepONet) within a platform-agnostic Python/Flask implementation. The authors compare several interpretable or explainable models (e.g., ReLU-DNN, EBM, FIGS, and decision-tree surrogates) on the PHM08 data, using a suite of local and global XAI techniques (PDP, ALE, LIME, SHAP) to identify cycle operation as the dominant predictor of RUL and to assess trustworthiness through model diagnosis. The work demonstrates how XAI/IML can enhance transparency, trust, and maintenance decision-making in DT-enabled prognostics, while outlining key challenges and future directions for real-time, uncertain, and multi-fidelity DT applications.

Abstract

Artificial intelligence (AI) and Machine learning (ML) are increasingly used in energy and engineering systems, but these models must be fair, unbiased, and explainable. It is critical to have confidence in AI's trustworthiness. ML techniques have been useful in predicting important parameters and in improving model performance. However, for these AI techniques to be useful for making decisions, they need to be audited, accounted for, and easy to understand. Therefore, the use of explainable AI (XAI) and interpretable machine learning (IML) is crucial for the accurate prediction of prognostics, such as remaining useful life (RUL), in a digital twin system, to make it intelligent while ensuring that the AI model is transparent in its decision-making processes and that the predictions it generates can be understood and trusted by users. By using AI that is explainable, interpretable, and trustworthy, intelligent digital twin systems can make more accurate predictions of RUL, leading to better maintenance and repair planning, and ultimately, improved system performance. The objective of this paper is to explain the ideas of XAI and IML and to justify the important role of AI/ML in the digital twin framework and components, which requires XAI to understand the prediction better. This paper explains the importance of XAI and IML in both local and global aspects to ensure the use of trustworthy AI/ML applications for RUL prediction. We used the RUL prediction for the XAI and IML studies and leveraged the integrated Python toolbox for interpretable machine learning~(PiML).

Explainable, Interpretable & Trustworthy AI for Intelligent Digital Twin: Case Study on Remaining Useful Life

TL;DR

This paper advocates integrating explainable AI (XAI) and interpretable ML (IML) into intelligent digital twins (DTs) for robust remaining useful life (RUL) prognostics in energy and nuclear contexts. It introduces a DT framework with an update module that combines Bayesian filtering and ML, and explores fast surrogate models via operator learning (DeepONet) within a platform-agnostic Python/Flask implementation. The authors compare several interpretable or explainable models (e.g., ReLU-DNN, EBM, FIGS, and decision-tree surrogates) on the PHM08 data, using a suite of local and global XAI techniques (PDP, ALE, LIME, SHAP) to identify cycle operation as the dominant predictor of RUL and to assess trustworthiness through model diagnosis. The work demonstrates how XAI/IML can enhance transparency, trust, and maintenance decision-making in DT-enabled prognostics, while outlining key challenges and future directions for real-time, uncertain, and multi-fidelity DT applications.

Abstract

Artificial intelligence (AI) and Machine learning (ML) are increasingly used in energy and engineering systems, but these models must be fair, unbiased, and explainable. It is critical to have confidence in AI's trustworthiness. ML techniques have been useful in predicting important parameters and in improving model performance. However, for these AI techniques to be useful for making decisions, they need to be audited, accounted for, and easy to understand. Therefore, the use of explainable AI (XAI) and interpretable machine learning (IML) is crucial for the accurate prediction of prognostics, such as remaining useful life (RUL), in a digital twin system, to make it intelligent while ensuring that the AI model is transparent in its decision-making processes and that the predictions it generates can be understood and trusted by users. By using AI that is explainable, interpretable, and trustworthy, intelligent digital twin systems can make more accurate predictions of RUL, leading to better maintenance and repair planning, and ultimately, improved system performance. The objective of this paper is to explain the ideas of XAI and IML and to justify the important role of AI/ML in the digital twin framework and components, which requires XAI to understand the prediction better. This paper explains the importance of XAI and IML in both local and global aspects to ensure the use of trustworthy AI/ML applications for RUL prediction. We used the RUL prediction for the XAI and IML studies and leveraged the integrated Python toolbox for interpretable machine learning~(PiML).
Paper Structure (19 sections, 1 equation, 18 figures, 3 tables)

This paper contains 19 sections, 1 equation, 18 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Core Components of Digital Twins: ML Algorithms and the XAI Requirement for DTs
  3. Update Module in a Digital Twin
  4. Operator Learning as a Faster Surrogate for a Digital Twin
  5. Platform-Agnostic and Surrogate-Driven Digital Twin Framework
  6. Explainable and Interpretable AI for Digital Twins
  7. Inherent Interpretable ML Models: Problem Setup, Methodologies, and Data Preparation
  8. XAI for RUL: Results and Discussion
  9. Feature Selection
  10. Global Explainability: Partial Dependence Plot
  11. Global Explainability: Accumulated Local Effects
  12. Local Explainability: Local Interpretable Model Agnostic Explanations
  13. Local Explainability: SHapley Additive Explanation
  14. Interpretability of the Model
  15. Local Interpretability
  16. ...and 4 more sections

Figures (18)

  • Figure 1: Intelligent digital twin framework with explainable AI and interpretable ML module. The diagram shows the ML components (red boxes) exploited in different segments of the digital twin framework for prognostics and justifies the explainability of these ML algorithms to understand RUL/prognostic behaviors kobayashi2023NET.
  • Figure 2: Developed update module with an unscented Kalman filter (UKF) and ML method for an intelligent digital twin framework kobayashi2023NET.
  • Figure 3: Authors leveraged the architecture of Deep Neural Operator (DeepONet) following the approach proposed by Lu lu2021learning. DeepONet comprises two networks: (1) Branch net and (2) Trunk net. The case of a 2D diffusion system kobayashi2023NET.
  • Figure 4: Developed architecture of DT framework following the approach by Bonney bonney2021digitalbonney2022development. The framework is composed of four parts: (1) Python application programming interface (API), (2) third-party software, (3) database, and (4) user interface. Flask serves as the core for providing connectivity to all components.
  • Figure 5: Demonstration of a Flask-based framework developed by the authors. Parameter setting, running Python code, and visualization can be performed on a web browser; following the approach by Bonney bonney2021digitalbonney2022development.
  • ...and 13 more figures