Smart manufacturing for snacks: Integrating SMED, predictive maintenance, line balancing, IoT, and machine learning to improve efficiency

This study proposes an integrated smart manufacturing model aimed at im-proving Production Efficiency (PE) in the snack industry, focusing on a Peruvian fried corn chip (FCC) production line. It addresses critical problems caused by extended changeover times, unplanned downtimes, and idle times by workload imbalances. The methodology combines three industrial engineering tools: Single Minute Exchange of Die (SMED), predictive maintenance, and line balancing, with Machine Learning (ML) and Internet of Things (IoT) technologies. A simulation-based implementation using ARENA software evaluates the conceptual smart manufacturing model's impact across three bottleneck stations: washer, fryer, and bagger. Results demonstrate a significant PE increase from 79.80% to 82.55%, surpassing the industry benchmark. Washer changeover time was reduced by 30.1%, Fryer downtime dropped by 44.9%, and Bagger idle time nearly matched the optimal level. Although residual troubles remain in Washer and Fryer stations, the improvements validate the model's effectiveness in reducing nonvalue-Added time. This study offers a novel, data driven framework that integrates traditional manufacturing tools with digital technologies to support industry 4.0 adoption in food processing environments. From a managerial perspective, the model serves as a practical guide for companies seeking to enhance PE, reduce costs and optimize resource use. Its value lies in addressing previously underexplored applications of smart manufacturing in the snack sector, bridging theoretical research with actionable outcomes.