Proceedings on Engineering Sciences (Mar 2024)

FAULT DETECTION AND CATEGORIZATION USING AN ADVANCED MACHINE LEARNING TECHNIQUE FOR INDUSTRIAL ROTATIONAL MACHINERY

  • Divya Paikaray ,
  • Naveen Kumar Rajendran ,
  • Vaishali Singh Maharishi ,
  • Pulkit Srivastava

DOI
https://doi.org/10.24874/PES.SI.24.02.008
Journal volume & issue
Vol. 6, no. 1
pp. 251 – 260

Abstract

Read online

The difficulty of fault identification as well as categorization in industrial rotating machinery is fixed by this study, which introduces a revolutionary Dandelion Optimized CatBoost (DO-CB) technique. The suggested framework makes use of the CB algorithm, which is enhanced by the DO method. The first step in the suggested DO-CB approach is gathering sensor data from rotating gear to record different operational settings. To ensure robustness, the recommended approach is developed on identified data and includes a variety of fault scenarios. Additionally, the Python tool used for identifying faults and classification is the basis for the implementation of the DO-CB approach. The experimental findings show how well the suggested method works to precisely identify and classify problems in industrial rotating gear. In comparison to benchmark defect detection techniques, the suggested DO-CB approach performs better, demonstrating its capacity to manage intricate patterns and fluctuations in the data.

Keywords