Physics-Data Fusion With FBG Sensing for In Situ Solid Lubricant Coating Degradation Monitoring Under Microgravity Emulation

作者：Zhenhui Feng, Lei Qi, Xiaobo Rui, Lina Wang, Qie Yan, Junchi Zhang, Yu Zhang · 发表于：IEEE Sensors Journal · 年份：2026 · DOI：10.1109/JSEN.2026.3698397

The condition state of solid-lubricant coatings is critical for the tribological performance of space gear transmission systems. In microgravity-like environments, condition monitoring is challenged by limited dynamic sensing capability and noisy periodic responses, which complicate robust feature extraction. To address these challenges, this study develops an integrated physics-data fusion pipeline based on fiber Bragg grating (FBG) tooth-root strain sensing for dynamic degradation assessment of gear solid-lubricant coatings. An autocorrelation-driven adaptive window function (AWF) is combined with time-synchronous averaging (TSA) to adaptively determine the window shape and ensure coverage of a complete meshing period, enabling stable extraction of degradation-related features from periodic strain signals. An exponential degradation model is then combined with support vector regression (SVR) to account for nonlinear effects that are difficult to represent in the analytic formulation, and Lasso regression is employed for feature selection to retain interpretability and reduce redundancy. The baseline pipeline is validated on a microgravity emulation test platform, achieving a coefficient of determination of 0.992 and an RMSE of $0.1089~\mu $ m for coating-thickness prediction, and is further evaluated on additional same-specification gear pairs through endpoint validation using final residual coating thickness. The 95% confidence interval width remains within $\pm 0.35~\mu $...