Prediction Of Heavy Metal Adsorption Efficiency Based On Biochar Properties Analysis And Machine Learning Model

Heavy metals in water have posed a serious threat to aquatic systems and human health,adsorption method is one of the effective ways for heavy metal wastewater treatment.Biochar,which has porous structure and abundant surface functional groups,is a cheap and easily available high-efficiency absorbent.In order to realize the efficient utilization of biomass resources in Tibet,barley straw and yak dung were selected as raw materials,and their composition and pyrolysis characteristics were analyzed.Biochar was prepared by oxygen-limited pyrolysis under different environmental conditions by mixing the two materials in a certain mass ratio.Elemental analysis,SEM-EDS,FTIR,BET,TGA and other technologies were used to characterize and analyze the physical and chemical properties of biochar,and explored the effect of adsorption conditions on biochar adsorption efficiency of heavy metal Cd.Combined with this study and related literature reports,based on machine learning theory,DT model,GBDT model,XGBoost model and Stacking model were used to predict the adsorption efficiency of biochar for various heavy metal ions according to the physical and chemical properties of biochar and adsorption conditions.The results were as follows:?1?The content of carbon,volatile organic compounds and fixed carbon in barley straw was significantly higher than that in yak dung,and its cellulose content was 6.5 times higher than that in yak dung,which contained more lignin.The pyrolysis characteristics of barley straw and yak dung were analyzed by thermogravimetry,the suitable temperature of biomass thermal conversion was between 360?590?.?2?biochar from straw had higher carbon content and smaller specific surface area,biochar from manure had higher cation exchange capacity,higher ash content and richer surface functional groups.Biochar prepared at higher temperature had richer pore structure,larger specific surface area,higher degree of carbonization and aromatization.Environmental pressure had little effect on the characteristics of biochar.?3?Co-pyrolytic biochar had more obvious main carbon structure and more abundant pore structure.At 500?,the biochar prepared by mixing barley straw and yak dung in 1:1 had the highest specific surface area of 27.41 m²/g,the highest cation exchange capacity of 62.03cmol/Kg,and the highest O/C and?O+N?/C ratio of 0.154 and 0.177,respectively.The larger the value of these indexes were,the more favorable the adsorption process was.This proved that co-pyrolysis was beneficial to optimize the structure of biochar and improve the adsorption performance of biochar.?4?When the concentration of Cd²⁺solution was 1.0 mmol/L,the maximum adsorption capacity of biochar for Cd²⁺was 16.40 mg/g.when the solution p H was 8,the maximum adsorption capacity was 18.78 mg/g and when the amount of biochar was 0.80 g,the maximum adsorption capacity was 25.34 mg/g.The adsorption of biochar on Cd conformed to the Freundlich model,which belonged to multi-layer adsorption.?5?The prediction results of the machine learning models showed that the Stacking model had the best prediction effect,with RMSE of 0.0288,R² of 0.9816,and MAE of 0.0185.The fitting effect was the best and the root mean square error was the smallest.The influence of solution temperature,initial concentration ratio,p H value,carbon content of biochar and cation exchange capacity on the prediction results of adsorption efficiency was significant.