A framework for predictive kinetic modeling of materials for thermochemical energy storage using algorithmic optimization

The effective design and optimization of Thermochemical Energy Storage (TCES) systems are often hindered by the lack of predictive kinetic models for materials with complex, multi-step reaction behavior. This study addresses this challenge for sodium sulfide (Image 1), a promising TCES material whose multi-step reaction pathways have complicated kinetic characterization. A robust and accessible framework is presented, using global optimization algorithms to directly calibrate reaction models from standard thermal analysis data. Hysteresis observed in the equilibrium properties of Image 2 is first quantified through Simultaneous Thermal Analysis (STA). Subsequently, eight variations of reaction kinetic models are formulated and calibrated using the Shuffled Complex Evolution (SCE) algorithm on time-series STA data. The resulting models are validated, with the best-performing model enabling accurate prediction of reaction rates under varying operating conditions. Predictive accuracy is reduced by a factor of 16.1 outside the calibration temperature range for the best-performing model, highlighting the value of a framework for rapid, application-specific model generation. Sensitivity analysis reveals that model performance is most dependent on activation energy and equilibrium conditions, with an average absolute sensitivity index of 38.6 and 12.4, respectively. Through this novel integration of algorithmic optimization with standard thermal analysis, this study provides a streamlined framework for developing kinetic models, addressing a critical gap that has limited the effective design of TCES systems.