Investigation on load distribution among rollers of planetary roller screw mechanism considering machining errors: Analytical calculation and machine learning approach

• Published in: Mechanism and machine theory 2023-07, Vol.185, p.105322, Article 105322
• Main Authors: Hu, Rui, Wei, Peitang, Liu, Huaiju, Du, Xuesong, Zhou, Pengliang, Zhu, Caichao
• Format: Article
• Language: English
• Subjects: Analytical calculation; Load distribution among rollers; Machine learning; Machining error; Planetary roller screw mechanism
• ISSN: 0094-114X; 1873-3999
• DOI: 10.1016/j.mechmachtheory.2023.105322

•A model is established to calculate the load distribution among PRSM rollers.•The importance of machining error on load sharing coefficient is determined.•A load sharing coefficient prediction formula is proposed. Load-carrying capacity is a critical characteristic that determines the performance of the planetary roller screw mechanism (PRSM). Until now, the relationship between the machining error and the load sharing characteristics among PRSM rollers has remained unclear, which is a challenge for the mechanism optimization. A novel analytical model considering multi-coupled machining errors is established to calculate the load distribution among the rollers, which is validated by the finite element method and available literature. Based on the experimental measurements of various machining errors, the load distribution among the rollers is analyzed in detail. By collecting data from the analytical model, a machine learning model is implemented for the first time to predict the load sharing coefficient (LSC) of the PRSM and determine the importance of each machining error on the LSC. The obtained results reveal that the importance values of the roller nominal diameter error, screw eccentricity error, and nut eccentricity error to the LSC are 19%, 34% and 47%, respectively. Accordingly, an LSC prediction formula for the PRSM accounting for various machining errors is proposed.