Fault Diagnosis for Body-in-White Welding Robot Based on Multi-Layer Belief Rule Base

Fault diagnosis for body-in-white (BIW) welding robots is important ensuring the efficient production of assembly line. As a result complex mechanism robot, its strong correlation components, and many types faults, it difficult to establish complete fault model. Therefore, model BIW-welding robot based on multi-layer belief rule base (BRB) was proposed. This can effectively integrate monitoring data expert knowledge achieve an accurate facilitate traceability. First, according established tree, determined. Second, relationship we BRB Meanwhile, in order improve accuracy parameters, projection covariance matrix adaptive evolutionary strategy (P-CMA-ES) algorithm used optimize update parameters Finally, validity proposed verified by simulation experiment robot.