Computational Intelligence for Sustainable Glass Manufacturing: A Data-Driven Approach for Energy Efficient Conditioning

Abstract

In the glass container manufacturing process, conditioning is a key stage that contributes to energy consumption. The main objective of conditioning is to cool the glass exiting the furnace to a suitable temperature for container forming. Currently, this stage is managed based on the experience of operators, which is functional but not optimized for energy efficiency. While several approaches to minimizing energy consumption based on process control using physical modeling have been proposed in the literature, they do not completely account for all the involved variables. Moreover, none of these studies leverage the power of data to predict energy consumption patterns. In this paper, we introduce a data-driven method to minimize energy consumption during the glass conditioning stage. We applied this methodology to a specific production line and tested it under various scenarios, achieving potential savings of 5% to 45% in energy consumption. Operational validations in two additional real forehearths showed energy reductions of 26.3–89.3 kWh per operating hour, corresponding to relative savings of 8.2%–22.1%, including a same-production A/B test. The implementation of this method has the potential to significantly contribute to the decarbonization goals of the glass manufacturing industry.