Heat and Mass Balance on a Curing Oven

Abstract

Glass wool manufacturing requires simultaneous drying and curing of an impregnated glass wool mat in a multi-zone conveyor oven. This work develops a lumped-capacity model that couples heat and mass transfer mechanisms with curing kinetics to predict the temperature profile and conversion level throughout the process. The model uses a Number of Transfer Units formulation to model the cross-flow drying process and implements an isoconversional kinetic model using Differential Scanning Calorimetry data. Heat transfer coefficients were computed from two years of production data and show discrepancies that reflect the physical configuration of the oven. Simulations on 60 production runs showed that the model captures the global temperature profile of the product by identifying three main periods: preheating, constant-rate drying, and falling-rate periods where curing predominantly occurs. Additionally, a sensitivity analysis of the most important parameters was performed to assess and understand their impact on the drying process and the temperature profile of the mat. When the outlet gas temperature is fixed from process data, the model achieves good agreement with experimental observations. The conversion profile also shows physically consistent behaviour as the final conversion increases with residence time under favourable temperature conditions. Several limitations were encountered, including the estimation of initial and critical moisture content, the inability to capture internal diffusion that leads to thermal and material gradients along the mat thickness, and uncertainty regarding the accuracy of the computed heat transfer coefficients. To tackle one of these limitations, the critical moisture content was defined to produce a transition between the constant drying rate period and the falling rate period that matches the measured temperature profile. Future work should focus on experimental determination of characteristic drying curves and development of appropriate correlations that will enable accurate computation of transport properties. Validation experiments in controlled pilot-scale equipment would also strengthen model predictions. Finally, the lumped assumption can be relaxed by taking into account both heat and mass transfer limitations along the mat thickness.