Investigating the reduction thermo-kinetics of Ca/Mn-doped perovskite in chemical looping reverse water gas shift with artificial neural network validation

Reduction in chemical looping-reverse water gas shift (CL-RWGS) process is crucial for achieving high syngas selectivity. However, detailed understanding of its kinetics remains limited for scale-up. In this study, a comprehensive thermo-kinetic analysis of Ca/Mn-doped LaNiO₃ was performed under isothermal and non-isothermal reduction. Isothermal kinetics indicated nucleation and growth driven mechanism (E_a = 10 kJ/mol). However, non-isothermal reduction showed more non-linear behaviour at 5 °C/min, with first-order chemical reaction model in closest agreement with experimental values at 10 and 15 °C/min. Furthermore, artificial neural network (ANN) modelling accurately predicted experimental conversion under both conditions with root mean square error under 0.031. To successfully predict the isothermal reduction, ANN model required 6 neurons. However, for non-isothermal data, optimal neuron counts for 700 – 800 °C and 800 – 900 °C were 7 and 9, respectively, indicating greater non-linearity. Hence, ANN-based modelling effectively provides a reliable approach for predicting reduction behaviour in CL-RWGS.