Paper Contents
Abstract
Unplanned outages in industry due to machine failures can result in considerable production losses and higher maintenance expenses. Predictive maintenance approaches leverage data collected from IoT-enabled sensors deployed in operational machinery to detect potential flaws and prevent catastrophic breakdowns. This work presents a predictive maintenance system based on machine learning techniques, specifically AdaBoost, that can classify different sorts of machine pauses in real time and is applicable to knitting machines. Machine speeds and steps were collected and pre-processed before being supplied into the machine learning model, which classified six types of machine stops: gate stop, feeder stop, needle stop, finished roll stop, idle stop, and lycra stop. The model is trained and improved using a combination of hyperparameter tuning and cross-validation techniques, resulting in a 92% test set accuracy. The results show that the suggested system has the potential to accurately detect machine pauses and enable prompt maintenance measures, thereby enhancing the textile industry's overall efficiency and production.
Copyright
Copyright © 2025 VENNILA P. This is an open access article distributed under the Creative Commons Attribution License.
