Physically Consistent Real‐Time Measurement of Batch‒Cullet Mixing

ABSTRACT Real‐time process metrology and control have long been considered as effective measures in direct response to operational disruptions and sustainability appeals within the glass industry. However, common challenges such as insufficient granularity, complex interdependencies, and harsh operating environments persist. Batch‒cullet mixing quality, a critical foundation of high‐quality glass production, for instance, remains loosely inferred from pre‐weighed bulk materials due to the lack of robust and reliable characterization methods. The current work develops a physically consistent measurement integration framework that enables continuous, in situ quantification of batch–cullet mixing quality using thermographic imaging and complementary sensor inputs with thermodynamic constraints embedded into image segmentation to ensure physical consistency. The resulting spatial heterogeneities are then mapped to effective thermal properties through analysis of statistically reconstructed representative volume elements, addressing their inherent temporal and spatial incompatibilities with process‐level evaluation. As a demonstration of its potential application in future digital twin development, the derived properties, along with other concurrent operational data, are incorporated into a continuously updated glass melting furnace simulation framework. Finally, future extensions are discussed to illustrate how the proposed capability may further support supervisory decision making and process optimization.