Integrating the Belief Rule Base and Fuzzy Comprehensive Evaluation Health Assessment Method to Enhance the Operational Efficiency of Rapier Loom Maintenance

Abstract Comprehensive fault diagnosis and condition monitoring of rapier looms are of greater practical significance in industrial production than the analysis of individual components. Due to the complex mechanical structure and strong coupling characteristics of rapier looms, existing methods have difficulty in accurately assessing the overall health status of the equipment. To address this challenge, this paper proposes a comprehensive health status evaluation method for rapier looms based on an improved Belief Rule Base (BRB) and a Fuzzy Comprehensive Evaluation (FCE) framework. In the proposed method, a rough set-based confidence rule extraction technique is first employed to construct a confidence rule library for inferring subsystem states, thereby integrating quantitative data with qualitative expert knowledge. Subsequently, hierarchical sampling and Evidential Reasoning (ER) secondary fusion methods are applied according to the data distribution characteristics. The Whale Optimization Algorithm (WOA) is utilized to optimize BRB parameters, ensuring robust performance of the BRB evaluation system even when premise attribute information is incomplete. Finally, a hybrid evaluation method combining the confidence rule library with FCE is proposed. The confidence outputs from the subsystem-level confidence rule library serve as the membership matrix for the construction of the comprehensive evaluation model. This approach jointly integrates the entropy weight method and the analytic hierarchy process to determine the weights of each subsystem, thereby aggregating subsystem states to effectively evaluate the overall system health condition. The proposed improved BRB-FCE fusion method satisfies the health assessment requirements of rapier looms. Compared with single evaluation methods, this fusion approach significantly enhances health assessment accuracy. Experimental results demonstrate an overall health assessment accuracy of 97.25% for rapier looms.