Manufacturing Waste Reduction Through Data-Driven Process Optimization: Evidence from Smart Production Systems

Manufacturing waste continues to undermine industrial sustainability efforts despite the increasing digitalization of production systems. While smart manufacturing technologies generate vast amounts of operational data, their potential to deliver measurable sustainability benefits remains underexplored. This study develops and empirically validates a data-driven process optimization framework that integrates high-frequency production data, predictive analytics, and multi-objective optimization to reduce material waste and energy inefficiency in smart production systems. Using large-scale, real-time data collected from digitally enabled manufacturing lines, machine learning models were employed to anticipate waste-generating process states and energy-intensive operating conditions. These predictive insights are embedded within a Pareto-based optimization and adaptive process control architecture that dynamically adjusts production parameters. Empirical results reveal an average reduction of 17.6% in material waste, 12.4% in energy consumption per unit, and 10.7% in carbon intensity, alongside an 8.9% improvement in overall equipment effectiveness. Importantly, these sustainability gains are achieved without compromising throughput or product quality, which demonstrates that environmental and operational objectives can be mutually reinforcing. Through the provision of robust, data-driven evidence from real production systems, this study advances the operationalization of sustainability within Industry 4.0 and offers a scalable methodological pathway for manufacturers who seek low-waste, low-carbon production. The findings contribute to manufacturing sustainability research by shifting the focus from digital adoption to outcomeoriented optimization with quantifiable environmental and economic benefits.