Digital Twin-Driven Adaptive Sim-to-Real Alignment via Reinforcement Learning for Vibration-Based Bearing Health Monitoring Under Data Scarcity

2026-06-23

AI Summary

This study addresses the challenge of bearing health monitoring under limited real-world fault data and domain discrepancies between digital twin simulations and actual sensor signals. To this end, the authors propose a reinforcement learning–based, adaptive, and fault-class-aware feature alignment method. The alignment process is formulated as a Markov decision process, wherein proximal policy optimization (PPO) dynamically learns fault-type-specific affine correction policies that minimize domain shift while preserving class discriminability. Notably, the approach requires only real normal samples and aligned simulated fault data for training, eliminating the need for encoder fine-tuning. Experimental validation on the XJTU-SY and CWRU datasets, along with a custom slewing bearing test rig, demonstrates a cross-device linear probing accuracy of 92.8%, substantially enhancing generalization capability in condition monitoring.

Abstract

Vibration-based health monitoring of rotating machinery requires reliable fault diagnosis under operational data constraints, yet condition assessment remains challenged by structural scarcity of fault events and heterogeneous sim-to-real gaps in digital twin-generated signals. Each fault type generates impulses with distinct periodicity, amplitude modulation, and spectral character, making feature-space discrepancies fundamentally heterogeneous across fault classes. Existing domain adaptation methods apply a class-agnostic global transformation that cannot close all fault-specific gaps without distorting inter-class separability, while uniform source-target mixing introduces distributional noise into the data-abundant Normal class. These limitations stem from treating a sequential, state-dependent alignment problem as a one-shot optimization. Each corrective transformation simultaneously reshapes all class distributions, creating state dependencies that static gradient descent cannot resolve. We formulate feature alignment as a continuous-action Markov decision process solved via Proximal Policy Optimization, where the learned policy issues fault-type-specific affine corrections responsive to the current feature-space configuration, with a dual-objective reward balancing gap minimization against separability preservation. An asymmetry-aware strategy reserves real data for the Normal class while augmenting fault classes with policy-aligned simulated samples. Validation across XJTU-SY, CWRU, and a self-built slewing bearing testbed confirms the dominant gain from reinforcement learning-driven alignment, and cross-equipment linear probing achieves 92.8% without encoder retraining, demonstrating transferable monitoring capability.

Problem

• Digital Twin
• Sim-to-Real Gap
• Data Scarcity
• Bearing Health Monitoring
• Domain Adaptation

Innovation

• Digital Twin
• Reinforcement Learning
• Sim-to-Real Alignment
• Fault-Specific Adaptation
• Vibration-Based Health Monitoring