Numerous challenges are being faced by implementing the Industrial Internet of Things (IIoT), which enables anomaly detection and optimal maintenance management of industrial production systems and their critical machines. Industries have started adopting this revolutionary technology along with other allied technologies to reap the full benefits out of Industry 4.0 environment. The objectives of this research work were to use the IIoT to develop a continuous monitoring system of the behavior of bottleneck facilities in a production system, to predict and avoid all possible failures, and to improve the overall productivity of the manufacturing system. The proposed prognostic health monitoring system employs IIoT sensors to measure the current values of the operating parameters of a machine, using built-in intelligent decision support mechanism to compare with optimal ranges of values, and to message the appropriate alarming signals as per the severity of the deviation. The system developed was tested with a prototype model comprising the Internet of Things, internet communication technology, and a machine learning algorithm, MEMS with standard input, output, storage and display of components, which was developed in a laboratory but implemented in a case study in a real industrial production plant. After the successful implementation of the developed system, the performance of the critical machine was evaluated in terms of metrics such as the average number of failures, average downtime and average service time spent. It was found that after the implementation, the downtime has decreased by nearly 22% and for the performance in terms of its output, the flow rate exhibited a steady increase with the passage of time.
Keywords: industrial internet of things; machine learning algorithm; anomaly detection; optimal decision-making process
*Corresponding author: Tel.: (+91) 8344349940
E-mail: velmurugan2601@gmail.com
