Scholarlink

In the era of Industry 4.0, where smart manufacturing prevails, machinery plays an indispensable role. Rotating machines, comprising a significant portion of these setups, demand meticulous maintenance to curb unplanned shutdowns and prolong their operational lifespan. The adoption of predictive maintenance stands as the cornerstone in the pursuit of smart maintenance strategies, particularly fa-voured by maintenance engineers. Leveraging the advancements in Artificial Intelligence (AI), a da-ta-centric approach to predictive maintenance has emerged, reshaping the way manufacturing is per-formed around the world. The crux of predictive maintenance lies in fault diagnosis, a formidable challenge. This study centers on vibrational data from shafts, extracted via multiple sensors embedded within an experimental setup. Three distinct machine learning models were employed for comprehensive training and testing using this dataset, revealing the superiority of the Support Vector Ma-chine model over its counterparts. The discerning performance of the SVM model in fault diagnosis based on the vibrational data, emphasizes its potential for widespread implementation in predictive maintenance across industrial settings.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Machine Learning Models for Predictive Maintenance of Unbalanced Shafts

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Arpan Gupta,
  • Sarthak Mittal,
  • Gaurav Luniya,
  • Hardik Agrawal,
  • Shuchi Jain

摘要

In the era of Industry 4.0, where smart manufacturing prevails, machinery plays an indispensable role. Rotating machines, comprising a significant portion of these setups, demand meticulous maintenance to curb unplanned shutdowns and prolong their operational lifespan. The adoption of predictive maintenance stands as the cornerstone in the pursuit of smart maintenance strategies, particularly fa-voured by maintenance engineers. Leveraging the advancements in Artificial Intelligence (AI), a da-ta-centric approach to predictive maintenance has emerged, reshaping the way manufacturing is per-formed around the world. The crux of predictive maintenance lies in fault diagnosis, a formidable challenge. This study centers on vibrational data from shafts, extracted via multiple sensors embedded within an experimental setup. Three distinct machine learning models were employed for comprehensive training and testing using this dataset, revealing the superiority of the Support Vector Ma-chine model over its counterparts. The discerning performance of the SVM model in fault diagnosis based on the vibrational data, emphasizes its potential for widespread implementation in predictive maintenance across industrial settings.