Predicting Key Performance Indicators (KPI) in smart manufacturing through Machine Learning (ML)

Misra P.K.; Melgarejo-Bolivar R.P.; Mendoza-Mollocondo C.I.; Torres-Cruz F.; Lourens M.E.; Shukla S.K. · AIP Conference Proceedings (2024) · DOI: 10.1063/5.0218192

Resumen

With the support of Information and Communication Technology (ICT), industrial sectors have transitioned from mechanization to digitalization. This shift has ushered in the era of Industry 4.0 or smart manufacturing, made possible by technological advancements and the Internet of Things (IoT). Industry 4.0 encompasses a wide array of transformative technologies such as big data, IoT, cloud computing, robotics, blockchain, artificial intelligence (AI), business software, autonomous vehicles, and others, all of which are reshaping the landscape of the manufacturing industry. However, Industry 4.0 is a relatively recent concept, and it suffers from a considerable degree of confusion, a lack of understanding, and limited research in the domain of performance evaluation and quality management. Meanwhile, manufacturing companies are still grappling with the abundance of Industry 4.0 technologies. The current study examines and evaluates several Machine Learning (ML) strategies for applying Key Performance Indicators (KPIs) in smart manufacturing firms in order to anticipate highly efficient production systems for reducing production costs as well as to improve product quality. The Neuro-Fuzzy based Subtractive Clustering (NFSC) approach and the Artificial Neural Network (ANN) technique is utilized for forecasting manufacturing process for KPIs. Models generated using the NFSC approach for six KPIs has lower Mean Absolute Percentage Error (MAPE) than those established utilizing ANN technique. In this investigation, the results suggest that the NFSC approach has greater prediction ability.

Instituciones: Universidad Nacional del Altiplano