Abstract
Decarbonizing high-temperature glass production requires coordinated decisions on batch composition, energy sourcing, and furnace operation. However, existing research has not sufficiently integrated virgin raw materials, internal and external cullet, multiple hydrogen pathways, electric boosting, inventories, lead times, and operational constraints within a unified short-term production planning and control framework for oxy-fuel glass furnaces. This study develops a multi-period, multi-objective linear programming model to address this gap. The model jointly optimizes batch composition and energy allocation across natural gas, green, blue, and grey hydrogen, and renewable and grid electricity. It links material choices to melting-energy requirements, process emissions, and direct and indirect energy-related emissions, while incorporating storage, electricity availability, ordering decisions, ramping limits, hydrogen-blending limits, and energy-mix constraints. Two conflicting objectives are considered: minimizing operating cost and emissions. Lexicographic optimization generates cost- and emission-priority solutions, while a normalized weighted-sum approach identifies intermediate trade-offs. The model is evaluated using three European industrial cases in container-glass and fiberglass production over a 24-h planning horizon. Results show that higher cullet utilization can reduce emissions by up to 25.6%, energy use by up to 19.3%, and cost by up to 18.9%. Under emission-priority operation, green-hydrogen scale-up reduces emissions by up to 10.7% at a cost increase of approximately 4%. Cost-priority solutions remain dominated by natural gas, whereas lower-emission operation requires greater use of green hydrogen and renewable electricity. Although the model is deterministic and does not capture detailed furnace physics, it provides structured decision support for plant-specific decarbonization planning.