Scholarlink

Predicting the failures and faults in industrial machinery and equipment which often leads to the complete breakdown of the plant is the focus of this proposed work. The accuracy in prediction was achieved by using various machine learning algorithms. A total of two datasets with data of 10,000 and 1000 have been used in this study to cover a wide range of data and 6 models were trained on each dataset to find out the most accurate ones. Both the datasets are binary classified. The performance of the models was evaluated for both the datasets. Prominent evaluation metrics have been used to exhibit the efficiency of the work. These models were already used in numerous research, but the accuracy has always been differed. This paper also reasons the difference in the accuracy of all the six training models by comparing their statistics. In conclusion, with the results obtained from this study, it was seen that KNN (K-Nearest neighbor) is highly efficient in predicting faults and failures. The results validate the feasibility of the approach proposed.

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

Failure Prediction in Industrial Machinery Using Machine Learning Algorithms

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • V. Jaswanth,
  • R. Ehthikash,
  • M. S. Harish,
  • Gerardine Immaculate Mary,
  • Anitha Julian,
  • Mayur Rele

摘要

Predicting the failures and faults in industrial machinery and equipment which often leads to the complete breakdown of the plant is the focus of this proposed work. The accuracy in prediction was achieved by using various machine learning algorithms. A total of two datasets with data of 10,000 and 1000 have been used in this study to cover a wide range of data and 6 models were trained on each dataset to find out the most accurate ones. Both the datasets are binary classified. The performance of the models was evaluated for both the datasets. Prominent evaluation metrics have been used to exhibit the efficiency of the work. These models were already used in numerous research, but the accuracy has always been differed. This paper also reasons the difference in the accuracy of all the six training models by comparing their statistics. In conclusion, with the results obtained from this study, it was seen that KNN (K-Nearest neighbor) is highly efficient in predicting faults and failures. The results validate the feasibility of the approach proposed.