Pyrolysis of cattle dung: model fitting and artificial neural network validation approach

The pyrolysis of the cattle dung was quantified using the Coats-Redfern method. The thermogravimetric and derivate curves (TG/DTG) divided the decomposition into three stages. Apart from stage I (dehydration), stage II exhibited higher thermal decomposition rate in the temperature range of 220–380 °C whereas, in stage III (390–690 °C), a lower decomposition rate was noticed. Comparative kinetic parameters for solid-state reactions showed that first-order reaction (F1) had the highest value of regression coefficient (R2) in both stages. In stage II, the Power-law (P3/2), reaction order-2 and 3 (F2 and F3), and diffusion models (D1 and D2) produced higher activation energy (Ea) values, while 3-diffusion (D3) produced the lowest Ea value. However, in subsequent stage III, only two reaction mechanisms (F1 and F2) were estimated with significant R2 value and F2 showed higher Ea value. The simulated TG/DTG validated that decomposition of cattle dung was best described by F1 in both stages. In addition to kinetic analysis through Coats-Redfern method, mass change at 20 °C/min was also processed by employing artificial neural network (ANN) and the model was validated with a strong R2 value and lower mean squared error (MSE). In thermodynamic analysis, the increase in the heating rate decreased ∆G and increased ∆S for the whole process with stable ∆H. This study provides the theoretical and practical guideline for the utilization of cattle dung as a potential energy source.