A hybrid model for wear prediction of a single revolute joint considering a time-varying lubrication condition

• Published in: Wear 2020-02, Vol.442-443, p.203124, Article 203124
• Main Authors: Zhuang, Xinchen, Saraygord Afshari, Sajad, Yu, Tianxiang, Liang, Xihui
• Format: Article
• Language: English
• Subjects: Bayesian inference; Boundary lubrication; Coefficient of friction; Comparative studies; Computer simulation; Finite element method; Journal bearings; Lubrication; Mathematical models; Revolute joint; Service life; Time-varying wear coefficient; Wear; Wear prediction
• ISSN: 0043-1648; 1873-2577
• DOI: 10.1016/j.wear.2019.203124

Wear of revolute joints critically influences the service life and maintenance of certain types of machinery. As a result, many methods have been developed to investigate the wear of revolute joints based on the assumption of either dry contact or full-film lubrication. However, in real applications, a revolute joint may operate under different lubrication conditions during its lifecycle. In this paper, a hybrid model is proposed for wear prediction of a single revolute joint that may experience different lubrication conditions. Three lubrication conditions are considered: full-film lubrication, boundary lubrication, and dry contact. A time-varying wear coefficient is introduced to represent these different lubrication conditions. In our hybrid model, the Archard wear model is used. An adaptive finite element model is used to perform wear simulation of a single revolute joint by continuously updating the contact nodes. In the finite element model, two friction coefficients are used to take into account the change of lubrication conditions. To achieve more accurate predictions and reduce uncertainties of the wear coefficient, the Bayesian updating method is implemented using experimental wear data. The proposed approach is tested using wear experiments of a revolute joint used in an airplane cabin door. The journal material in this experiment is 0.45C steel and the bearing material is brass H58. A comparative study was also conducted between the proposed approach and two other methods. Results show that the proposed approach can more accurately predict the remaining useful life of a single revolute joint. •A hybrid model is proposed for wear prediction of a single revolute joint.•A time-varying lubrication condition is considered.•A time-varying wear coefficient is introduced in wear modeling.•Bayesian inference is used to reduce uncertainties in wear coefficient.