A probabilistic neural network with adaptive capability to optimize fault trees for the intelligent fault diagnosis of rapier loom throughout its operating cycle¹.

This article focuses on developing an improved fault diagnosis system for rapier looms, which are characterized by strongly coupled mechanical structures and complex failure modes. It proposes a method combining fault tree analysis with probabilistic neural networks (PNN) to build a comprehensive and intelligent fault diagnosis model covering the entire operation cycle of rapier loom equipment. The fault tree models are constructed for key subsystems—start-up inspection, weaving, and warp feeding and winding—based on expert knowledge and historical data, while the PNN optimizes fault classification by learning from fault sign samples. Simulation and field experiments demonstrate that this integrated approach achieves high accuracy (over 99%) in fault sign recognition, provides effective fault localization, and significantly improves troubleshooting efficiency compared to traditional methods. The study highlights the method’s strong self-learning capability and its potential to enhance maintenance management and operational safety in rapier loom systems.