Wear prediction of a mechanism with multiple joints based on ANFIS. (January 2021)

• Record Type: Journal Article
• Authors: Zhuang, Xinchen; Yu, Tianxiang; Sun, Zhongchao; Song, Kunling
• Is Part Of: Engineering failure analysis. Volume 119(2021)
• Publication Date: 2021-01
• Subjects: Wear prediction -- Mechanism -- Clearance joints -- ANFIS -- Bayesian updating; System failures (Engineering) -- Periodicals; Fracture mechanics -- Periodicals; Reliability (Engineering) -- Periodicals; Pannes -- Périodiques; Rupture, Mécanique de la -- Périodiques; Fiabilité -- Périodiques; Fracture mechanics; Reliability (Engineering); System failures (Engineering); Periodicals; Electronic journals; 620.112
• DOI: 10.1016/j.engfailanal.2020.104958
• Languages: English
• ISSNs: 1350-6307

Abstract

Highlights: An ANFIS-based joints clearance size prognostic method is developed based on motion outputs. A framework for wear prediction based on multi-body dynamics theory is proposed. Bayesian inference is used to reduce uncertainties in wear coefficient. Experiment of a mechanism with multiple joints is designed to validate our proposed method. Abstract: Condition monitoring data of joints in a mechanism contains enormous useful information, and can comprehensively improve the wear prediction accuracy. However, condition data of the joints is sometimes hard to obtain due to technical reasons or cost reasons, especially for some complicated mechanical systems. To obtain the real-time wear data of joints in a mechanism with multiple joints, an ANFIS-based (adaptive-network-based fuzzy inference system) joints clearance size prognostic method is developed based on monitored motion outputs of the mechanism. Then, a framework for wear prediction based on multi-body dynamics theory is proposed to predict joints wear more accurately. In the framework, the Archard's wear model is used. To reduce the uncertainty in the wear coefficient, wear coefficient is treated as a random variable, then a Bayesian updating process is implemented according to the wear data from the ANFIS-based method. The proposed framework is validated using wear experiments of a lock mechanism with three joints in a cabin door. The results show the prediction error is within 3%.