Biochar assessment through integration of heuristic Bayesian with chemical simulator

• The maximum char yield of 67.23 % obtained from continuous stirred tank. • Gas yield of around 20.05 % derived through plug-flow reactor. • Noticeable rise in C stock by 1.61 % to 2.04 % with the increase in temperature. • The least MSE was observed while using gradient boosting technique. This article examines the impact of operating conditions on the quality of biochar, a byproduct of pyrolysis, produced from hardwood chips. The process simulation was performed to generate the biochar at different thermal degradation conditions, and the chemical properties, pH, cation exchange capacity, and polycyclic aromatic hydrocarbons were also compared with the corresponding yield of biochar. The Peng-Robinson model was considered to simulate the behaviour of generated gas during the pyrolysis process. The five different types of reactors, conversion (R1), equilibrium (R2), Gibbs (R3), continuous stirred tank (R4) and plug flow (R5) reactors are used to estimate the product yield derived from the pyrolysis of G-30 wood chip at 500°C, 550°C, 650°C, 850°C and 950°C. For comparison and integration purposes, the PyChem was also added. The Python language was used for programming purpose. Additionally, the RothC model was incorporated to determine the carbon stock in the soil for 10 years. The maximum deviation in biochar yield with temperature was noticed for R2 and R3 reactors. The biochar yield was, respectively, dropped by 0.99 to 4.01 %, 0.14 to 5.79 %, 5.30 to 9.29 %, 0.42 to 1.75 % and 0.03 to 0.97 % for R1, R2, R3, R4 and R5 with the temperature rise. R4 reactor was noticed to have the highest char yield, while the gas and water yields were maximised in R5 and R1, respectively. The overall rise in the average carbon sequestration was 1.61 to 15.72 % with the increasing temperature. However, the inert carbon pool fraction showed that the higher temperature biochar would have a higher turnover rate than those derived at a lower temperature regime. The mean squared error encountered while measuring pH value and Cation exchange capacity from the Gradient boosting technique was 0.11 and 0.01, respectively. Bimodality in the posterior distribution functions of biochar parameters was seen.