An assessment on the utilization of machine learning for AI-driven fault identification in the industrial internet of things
V. Narasimha, Manoj Kumar, S Suma, Bathula Mounika · 2025
Any industrial environment must include industrial inspection for defects. Potentially fatal risks can arise in any industrial setting from electrical, mechanical, chemical, biological, ergonomic, and other sources. Whenever these flaws go unnoticed, it can seriously harm people’s lives and their possessions. Early detection is necessary to maintain safety in order to avert any accidents and take necessary action whenever something goes poorly. An AI-Enabled fault recognition conduct is presented in this dissertation for the identification of problems in Industrial Internet of Things (IIoT) systems. This technique concentrated on a mixed approach (vote classifier) that improves the fault identification procedure at an AI-Enabled cloud server by stacking the support vector machine (SVM), decision tree (DT), and logistic regression (LR) approaches. The suggested fault identification technique facilitates the defect identification procedure through gathering sensor data through several IoT devices installed in different machine parts and utilizing supervised machine learning (ML) techniques to analyze this data. The suggested approach is contrasted against a number of supervised machine learning techniques, including DT, SVM, and KNN. Metrics for performance including accuracy, precision, recall, and F1 score are used to assess how well these algorithms work. When weighed against the other ways, the suggested hybrid method is capable to execute significantly. Jupyter notebook and Python 2.7 are used for this study.