A Novel Semi-Supervised Graph-Guided Approach for Intelligent Health State Diagnosis of a 3-PRR Planar Parallel Manipulator

Abstract

As one of the most important operation and maintenance approaches, health state diagnosis technology plays a crucial role in ensuring the safety and reliability of mechanical equipment. The planar parallel manipulator, as a typical actuator, is widely employed in the field of precision manufacturing due to its advantages of high stiffness, large load support capability, and high precision. However, compared with common key functional components (such as bearings and gearboxes), planar parallel manipulators have more complicated operating mechanisms and failure behaviors. To satisfy the health state diagnosis demands of planar parallel manipulators in the scenario of insufficient label information, a novel intelligent health state diagnosis approach, termed semi-supervised graph-guided network with perception attention (SGN-PA), is developed for a 3-PRR (P and R represent prismatic and revolute pairs, respectively) planar parallel manipulator. Specifically, an improved multiorder graph perception module is constructed to extract multiscale feature information, and achieve feature fusion by combining perceptual attention mechanism, which enables the proposed SGN-PA model to have adaptation adjustment capabilities. Following that, local and nonlocal feature constraint strategies are employed with pseudo-label technology to reduce intraclass differences and maximize interclass differences, and then to fit the demands of health state diagnosis tasks. Eventually, based on the simulation and experimental scenarios of a 3-PRR planar parallel manipulator, the effectiveness and feasibility of the proposed SGN-PA model is extensively confirmed, and the diagnosis results show that it can significantly relax the constraints of label information while maintaining superior performances.