Root-Cause Analysis of Momentum Wheel Fault Based on Counterfactual Reasoning

Current root-cause analysis methods rely on correlation to identify the underlying causes of faults. In contrast, this paper introduces a novel approach for analyzing momentum wheel faults based on counterfactual reasoning. This method distinguishes correlation from causality, aiming to provide more reliable, interpretable, and accurate results when identifying the root causes of faults. The effectiveness of this proposed method is validated through 29 momentum wheel fault experiments, demonstrating a root-cause analysis accuracy of 79.3%, surpassing that of traditional correlation-based methods.