Leveraging Artificial Intelligence of Things and Big Data Analytics for Enhanced Predictive Maintenance

Abstract

This research aims at establishing the role of AIoT and Big Data Analytics in improving PdM techniques by providing failure prognosis and condition-based upkeep. The proposed framework utilizes IoT sensors installed on the industrial assets to always track vital parameters which are as follows; Temperature, Vibration, Pressure and Humidity which inversely reflect the status of equipment. The information gathered from these sensors is preprocessed and analyzed in real time by the means of Edge computing that helps in minimizing the latencies of the large amount of data collected and in turn makes the decision making possible at the edge of the network. Pulse signal data of equipment are used together with the Long Short-Term Memory (LSTM) networks for prediction of degradation, identification of abnormality, and determination of Remaining Useful Life (RUL) that enable prescriptive maintenance. The inclusion of Big Data Analytics in the evaluation of this framework also improves this aspect by providing a way of storing and processing large volumes of data and the review of patterns in the data to refine the model over time. Evaluation metrics such as accuracy, precision, recall together with Mean Absolute Error (MAE) summarize the ability of the proposed AIoT based PdM system to effectively prognosticate maintenance requirements effectively.