A novel intelligent system based on machine learning for hydrochar multi-target prediction from the hydrothermal carbonization of biomass

Abstract

Hydrothermal carbonization (HTC) is a thermochemical conversion technology to produce hydrochar from wet biomass without drying, but it is time-consuming and expensive to experimentally determine the optimal HTC operational conditions of specific biomass to produce desired hydrochar. Therefore, a machine learning (ML) approach was used to predict and optimize hydrochar properties. Specifically, biochemical components (proteins, lipids, and carbohydrates) of biomass were predicted and analyzed first via elementary  composition. Then, accurate single-biomass (no mixture) based ML multi-target models (average R ² = 0.93 and RMSE = 2.36) were built to predict and optimize the hydrochar properties (yield, elemental composition, elemental atomic ratio, and higher heating value). Biomass composition (elemental and biochemical), proximate analyses, and HTC conditions were inputs herein. Interpretation of the model results showed that ash, temperature, and the N and C content of biomass were the most critical factors affecting the hydrochar properties, and that the relative importance of biochemical composition (25%) for the hydrochar was higher than that of operating conditions (19%). Finally, an intelligent system was constructed based on a multi-target model, verified by applying it to predict the atomic ratios (N/C, O/C, and H/C). It could also be extended to optimize hydrochar production from the HTC of single-biomass samples with experimental validation and to predict hydrochar from the co-HTC of mixed biomass samples reported in the literature. This study advances the field by integrating predictive modeling, intelligent systems, and mechanistic insights, offering a holistic approach to the precise control and optimization of hydrochar production through HTC.

Highlights

•	Biochemical components of biomass were first predicted by elemental composition.
•	The multi-target ML model accurately predicted hydrochar with R 2 = 0.93.
•	The ash, T, N, and C were the most critical factors affecting hydrochar properties.
•	An online intelligent system based on optimal models was posted and verified.

Keywords

Biomass / Hydrothermal carbonization / Hydrochar / Machine learning / Intelligent prediction system

References

[1] Buss W, Jansson S, Wurzer C, Mašek O. Synergies between BECCS and biochar—maximizing carbon sequestration potential by recycling wood ash. ACS Sustain Chem Eng, 2019, 7: 4204-4209,

[2] Deng Q, Lin B. Automated machine learning structure-composition-property relationships of perovskite materials for energy conversion and storage. Energy Mater, 2022, 1: 100006,

[3] Duan P, Chang Z, Xu Y, et al.. Hydrothermal processing of duckweed: effect of reaction conditions on product distribution and composition. Bioresour Technol, 2013, 135: 710-719,

[4] Duan PG, Yang SK, Xu YP, et al.. Integration of hydrothermal liquefaction and supercritical water gasification for improvement of energy recovery from algal biomass. Energy, 2018, 155: 734-745,

[5] DuBois M, Gilles KA, Hamilton JK, et al.. Colorimetric method for determination of sugars and related substances. Anal Chem, 1956, 28: 350-356,

[6] Fang Y, Ma L, Yao Z, et al.. Process optimization of biomass gasification with a Monte Carlo approach and random forest algorithm. Energy Convers Manag, 2022, 264,

[7] Gao F, Shen Y, Brett Sallach J, et al.. Predicting crop root concentration factors of organic contaminants with machine learning models. J Hazard Mater, 2022, 424,

[8] He M, Cao Y, Xu Z, et al.. Process water recirculation for catalytic hydrothermal carbonization of anaerobic digestate: water-energy-nutrient nexus. Bioresour Technol, 2022, 361,

[9] He M, Zhu X, Dutta S, et al.. Catalytic co-hydrothermal carbonization of food waste digestate and yard waste for energy application and nutrient recovery. Bioresour Technol, 2022, 344,

[10] Hoekman SK, Broch A, Robbins C. Hydrothermal carbonization (HTC) of lignocellulosic biomass. Energy Fuels, 2011, 25: 1802-1810,

[11] Kim JY, Shin UH, Kim K. Predicting biomass composition and operating conditions in fluidized bed biomass gasifiers: an automated machine learning approach combined with cooperative game theory. Energy, 2023, 280,

[12] Kirchner K, Zec J, Delibašić B. Facilitating data preprocessing by a generic framework: a proposal for clustering. Artif Intell Rev, 2016, 45: 271-297,

[13] Leng L, Zhang W, Peng H, et al.. Nitrogen in bio-oil produced from hydrothermal liquefaction of biomass: a review. Chem Eng J, 2020, 401,

[14] Leng L, Zhang W, Chen Q, et al.. Machine learning prediction of nitrogen heterocycles in bio-oil produced from hydrothermal liquefaction of biomass. Bioresour Technol, 2022, 362,

[15] Leng L, Zhang W, Liu T, et al.. Machine learning predicting wastewater properties of the aqueous phase derived from hydrothermal treatment of biomass. Bioresour Technol, 2022, 358,

[16] Leng S, Jiao H, Liu T, et al.. Co-liquefaction of Chlorella and soybean straw for production of bio-crude: effects of reusing aqueous phase as the reaction medium. Sci Total Environ, 2022, 820,

[17] Leng L, Li T, Zhan H, et al.. Machine learning-aided prediction of nitrogen heterocycles in bio-oil from the pyrolysis of biomass. Energy, 2023, 278,

[18] Li Y, Liu H, Xiao K, et al.. Correlations between the physicochemical properties of hydrochar and specific components of waste lettuce: influence of moisture, carbohydrates, proteins and lipids. Bioresour Technol, 2019, 272: 482-488,

[19] Li J, Pan L, Suvarna M, et al.. Fuel properties of hydrochar and pyrochar: prediction and exploration with machine learning. Appl Energy, 2020, 269,

[20] Li L, Flora JRV, Berge ND. Predictions of energy recovery from hydrochar generated from the hydrothermal carbonization of organic wastes. Renew Energy, 2020, 145: 1883-1889,

[21] Li J, Zhang W, Liu T, et al.. Machine learning aided bio-oil production with high energy recovery and low nitrogen content from hydrothermal liquefaction of biomass with experiment verification. Chem Eng J, 2021, 425,

[22] Li J, Zhu X, Li Y, et al.. Multi-task prediction and optimization of hydrochar properties from high-moisture municipal solid waste: application of machine learning on waste-to-resource. J Clean Prod, 2021, 278,

[23] Li H, Chen J, Zhang W, et al.. Machine-learning-aided thermochemical treatment of biomass: a review. Biofuel Res J., 2023, 10: 1786-1809,

[24] Liu A, Su Y, Nie W, Kankanhalli M. Hierarchical clustering multi-task learning for joint human action grouping and recognition. IEEE Trans Pattern Anal Mach Intell, 2017, 39: 102-114,

[25] Liu H, Basar IA, Nzihou A, Eskicioglu C. Hydrochar derived from municipal sludge through hydrothermal processing: a critical review on its formation, characterization, and valorization. Water Res, 2021, 199,

[26] Liu Z, Cui Y, Wang J, et al.. Multi-objective optimization of multi-energy complementary integrated energy systems considering load prediction and renewable energy production uncertainties. Energy, 2022, 254,

[27] Marzbali MH, Kundu S, Halder P, et al.. Wet organic waste treatment via hydrothermal processing: a critical review. Chemosphere, 2021, 279,

[28] Mu L, Wang Z, Wu D, et al.. Prediction and evaluation of fuel properties of hydrochar from waste solid biomass: machine learning algorithm based on proposed PSO–NN model. Fuel, 2022, 318,

[29] Natekin A, Knoll A. Gradient boosting machines, a tutorial. Front Neurorobot, 2013,

[30] Palansooriya KN, Li J, Dissanayake PD, et al.. Prediction of soil heavy metal immobilization by biochar using machine learning. Environ Sci Technol, 2022, 56: 4187-4198,

[31] Putatunda S, Rama K (2018) A Comparative Analysis of Hyperopt as Against Other Approaches for Hyper-Parameter Optimization of XGBoost. In: Proceedings of the 2018 International Conference on Signal Processing and Machine Learning. ACM press, New York, USA, pp 6–10

[32] Qureshi AS, Khan A, Zameer A, Usman A. Wind power prediction using deep neural network based meta regression and transfer learning. Appl Soft Comput, 2017, 58: 742-755,

[33] Ribeiro MT, Singh S, Guestrin C (2016) “Why Should I Trust You?” Explaining the Predictions of Any Classifier. In: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. ACM press, New York, USA, pp 1135–1144

[34] Rodrigues R, Souza D, Toebe M, Chuquel A. Sample size and Shapiro-Wilk test: AN analysis for soybean grain yield. Eur J Agron, 2023, 142,

[35] Seo MW, Lee SH, Nam H, et al.. Recent advances of thermochemical conversion processes for biorefinery. Bioresour Technol, 2022, 343,

[36] Shafizadeh A, Shahbeik H, Rafiee S, et al.. Machine learning-based characterization of hydrochar from biomass: implications for sustainable energy and material production. Fuel, 2023, 347,

[37] Shapiro SS, Wilk MB. An analysis of variance test for normality (complete samples). Biometrika, 1965, 52: 591-611,

[38] Sheng L, Wang X, Yang X. Prediction model of biocrude yield and nitrogen heterocyclic compounds analysis by hydrothermal liquefaction of microalgae with model compounds. Bioresour Technol, 2018, 247: 14-20,

[39] Shi N, Liu Q, He X, et al.. Molecular structure and formation mechanism of hydrochar from hydrothermal carbonization of carbohydrates. Energy Fuels, 2019, 33: 9904-9915,

[40] Toptas Tag A, Duman G, Yanik J. Influences of feedstock type and process variables on hydrochar properties. Bioresour Technol, 2018, 250: 337-344,

[41] Xiong T, Cui J, Hou Z, et al.. Prediction of arsenic adsorption onto metal organic frameworks and adsorption mechanisms interpretation by machine learning. J Environ Manage, 2023, 347,

[42] Xu D, Lin G, Liu L, et al.. Comprehensive evaluation on product characteristics of fast hydrothermal liquefaction of sewage sludge at different temperatures. Energy, 2018, 159: 686-695,

[43] Xu Z, Ma X, Zhou J, et al.. The influence of key reactions during hydrothermal carbonization of sewage sludge on aqueous phase properties: a review. J Anal Appl Pyrolysis, 2022, 167,

[44] Yap BW, Sim CH. Comparisons of various types of normality tests. J Stat Comput Simul, 2011, 81: 2141-2155,

[45] Yu J, Zhong X, Huang Z, et al.. Mining the synergistic effect in hydrothermal co-liquefaction of real feedstocks through machine learning approaches. Fuel, 2023, 334,

[46] Yuan T-Q, Sun S-N, Xu F, Sun R-C. Characterization of lignin structures and lignin-carbohydrate complex (LCC) linkages by quantitative 13 C and 2D HSQC NMR spectroscopy. J Agric Food Chem, 2011, 59: 10604-10614,

[47] Yuan X, Suvarna M, Low S, et al.. Applied machine learning for prediction of CO 2 adsorption on biomass waste-derived porous carbons. Environ Sci Technol, 2021, 55: 11925-11936,

[48] Zhang W, Li J, Liu T, et al.. Machine learning prediction and optimization of bio-oil production from hydrothermal liquefaction of algae. Bioresour Technol, 2021, 342,

[49] Zhang B, Biswal BK, Zhang J, Balasubramanian R. Hydrothermal treatment of biomass feedstocks for sustainable production of chemicals, fuels, and materials: progress and perspectives. Chem Rev, 2023,

[50] Zhang W, Chen Q, Chen J, et al.. Machine learning for hydrothermal treatment of biomass: a review. Bioresour Technol, 2023, 370,

[51] Zhang X, Liu H, Yang G, et al.. Comprehensive insights into the application strategy of kitchen waste derived hydrochar: Random forest-based modelling. Chem Eng J, 2023, 469,

[52] Zhang S, Luo X, Mai B. Multi-task machine learning models for simultaneous prediction of tissue-to-blood partition coefficients of chemicals in mammals. Environ Res, 2024, 241,

[53] Zhu X, Liu B, Sun L, et al.. Machine learning-assisted exploration for carbon neutrality potential of municipal sludge recycling via hydrothermal carbonization. Bioresour Technol, 2023, 369,

Funding

National Key Research and Development Program of China(2021YFE0104900); the Open Project of Xiangjiang Laboratory(22xj03003); Science and Technology Innovative Research Team in Higher Educational Institutions of Hunan Province(2021GK1210)