Machine-learning-aided biochar production from aquatic biomass DOI Creative Commons
Zhilong Yuan, Ye Wang, Lingfeng Zhu

et al.

Carbon Research, Journal Year: 2024, Volume and Issue: 3(1)

Published: Nov. 11, 2024

Abstract Modeling hydrothermal carbonization (HTC) and pyrolysis (PLC) for the conversion of biomass into high-quality biochar various applications shows promise. Unlike extensive modeling studies on lignocellulosic biomass, research aquatic (AB) had not been reported until now. In this study, we compiled 586 data points from existing literature trained five tree-based models to predict yields hydrochar pyrochar their properties, including nitrogen recovery degree, energy density, residual sulfur based 10 feedstock process parameters. The random forest regression (RFR) model demonstrated highest predictive accuracy among these models. It achieved R 2 values ranging 0.89 0.98 yield, degree hydrochar, hydrochar. extreme gradient boosting (XGB) also showed exemplary performance, with between 0.84 0.94 density pyrochar. Results feature importance highlighted that, beyond well-documented impact parameters, properties were significantly influenced by elemental compositions, such as contents feedstock. relationship factors was further elucidated using partial dependence plots. Finally, used RFR yield XGB examples, test generalization ability developed new data, explaining application methods. Overall, study provided valuable insights predicting understanding HTC PLC processes AB produce low resources time costs. Besides, presented an iterative learning method where exceptionally high performance data. This is highly versatile can be adopted across directions in field machine learning. Graphical

Language: Английский

Screening biomass by machine learning to construct Fe/Co loaded carbon for organic pollutants degradation via peroxymonosulfate activation DOI
Baoying Wang, Huanyan Xu,

Jingming Lan

et al.

Separation and Purification Technology, Journal Year: 2025, Volume and Issue: unknown, P. 131664 - 131664

Published: Jan. 1, 2025

Language: Английский

Citations

2
Multi-output neural network model for predicting biochar yield and composition DOI
Yifan Wang,

Liang Xu,

Jianen Li

et al.

The Science of The Total Environment, Journal Year: 2024, Volume and Issue: 945, P. 173942 - 173942

Published: June 15, 2024

Language: Английский

Citations

6
Machine learning-driven predictive frameworks for optimizing chemical strategies in Microcystis aeruginosa mitigation DOI

Zobia Khatoon,

Suiliang Huang,

Adeel Ahmed Abbasi

et al.

Journal of Water Process Engineering, Journal Year: 2025, Volume and Issue: 71, P. 107235 - 107235

Published: Feb. 12, 2025

Language: Английский

Citations

0
Deterministic Models for Performance Analysis of Lignocellulosic Biomass Torrefaction DOI Creative Commons
Abbas Azarpour, Sohrab Zendehboudi, Noori M. Cata Saady

et al.

ACS Omega, Journal Year: 2025, Volume and Issue: 10(7), P. 6470 - 6501

Published: Feb. 13, 2025

Energy plays a key role in the socioeconomic development of society, and most its global demand is provided by conventional resources (e.g., fossil fuels). Utilizing renewable energy significantly growing since it can meet while minimizing adverse impacts carbon emissions on climate change. Biomass an appealing option among emerging alternatives wind solar). Torrefaction mild pyrolysis process, this research aims to analyze torrefaction process lignocellulosic biomass. The methodology proposed involves employing hybrid models artificial neural network-particle swarm optimization (ANN-PSO), adaptive neuro-fuzzy inference system (ANFIS), coupled simulated annealing-least-squares support vector machine (CSA-LSSVM). In addition learning algorithms, correlation developed using gene expression programming (GEP) interrelate biomass properties, including moisture content, volatile matter, fixed carbon, ash, sample size, contents oxygen, hydrogen, nitrogen along with operating condition encompassing residence time, temperature, concentration CO2, O2, N2 solid yield as target variable. results reveal that CSA-LSSVM model has highest accuracy, statistical metrics coefficient determination (R2), mean square error (MSE), average absolute relative percentage (AARE%) are 0.98, 0.00082, 2.61%, respectively. parametric sensitivity analysis demonstrates content influential variables, temperature playing crucial findings be used assess similar torrefaction, providing required knowledge for modeling process. Hence, bioenergy industry optimal conditions, cost energy, lessen negative CO2 emission.

Language: Английский

Citations

0
Thermal behavior and conversion of agriculture biomass residues by torrefaction and pyrolysis DOI Creative Commons
Mihai Brebu, Daniela Ioniţă, Elena Stoleru

et al.

Scientific Reports, Journal Year: 2025, Volume and Issue: 15(1)

Published: April 3, 2025

Language: Английский

Citations

0
Lignocellulose‐Derived Energy Materials and Chemicals: A Review on Synthesis Pathways and Machine Learning Applications DOI
Luyao Wang, Shuling Liu, Sehrish Mehdi

et al.

Small Methods, Journal Year: 2025, Volume and Issue: unknown

Published: April 23, 2025

Abstract Lignocellulose biomass, Earth's most abundant renewable resource, is crucial for sustainable production of high–value chemicals and bioengineered materials, especially energy storage. Efficient pretreatment vital to boost lignocellulose conversion bioenergy biomaterials, cut costs, broaden its energy–sector applications. Machine learning (ML) has become a key tool in this field, optimizing processes, improving decision‐making, driving innovation valorization This review explores main strategies – physical, chemical, physicochemical, biological, integrated methods evaluating their pros cons It also stresses ML's role refining these supported by case studies showing effectiveness. The examines challenges opportunities integrating ML into storage, underlining pretreatment's importance unlocking lignocellulose's full potential. By blending process knowledge with advanced computational techniques, work aims spur progress toward sustainable, circular bioeconomy, particularly storage solutions.

Language: Английский

Citations

0
Machine learning modeling of thermally assisted biodrying process for municipal sludge DOI
Kaiqiang Zhang,

Ningfung Wang

Waste Management, Journal Year: 2024, Volume and Issue: 188, P. 95 - 106

Published: Aug. 10, 2024

Language: Английский

Citations

3
Critical insights into ensemble learning with decision trees for the prediction of biochar yield and higher heating value from pyrolysis of biomass DOI

Saurav Kandpal,

Ankita Tagade,

Ashish N. Sawarkar

et al.

Bioresource Technology, Journal Year: 2024, Volume and Issue: 411, P. 131321 - 131321

Published: Aug. 22, 2024

Language: Английский

Citations

3
Co-pyrolysis of coal-derived sludge and low-rank coal: Thermal behaviour and char yield prediction DOI Creative Commons
Tianli Zhang, Chenxu Zhang, Hai‐Peng Ren

et al.

Fuel Processing Technology, Journal Year: 2024, Volume and Issue: 267, P. 108165 - 108165

Published: Nov. 29, 2024

Language: Английский

Citations

3
Machine Learning-Based Prediction of the Adsorption Characteristics of Biochar from Waste Wood by Chemical Activation DOI Open Access
Jinman Chang,

Jai-Young Lee

Materials, Journal Year: 2024, Volume and Issue: 17(21), P. 5359 - 5359

Published: Nov. 1, 2024

This study employs machine learning models to predict the adsorption characteristics of biochar-activated carbon derived from waste wood. Activated is a high-performance adsorbent utilized in various fields such as air purification, water treatment, energy production, and storage. However, its vary depending on activation conditions or raw materials, making explaining predicting them challenging using physicochemical mathematical methods. Therefore, techniques determine activated advance will provide economic time benefits for production. Datasets, consisting 108 points, were used The input variables conditions, iodine number was output variable. datasets randomly split into 75% training 25% model validation normalized by min-max function. Four models, including artificial neural networks, random forests, extreme gradient boosting, support vector machines, properties carbon. After optimization, network identified best model, with highest coefficient determination (0.96) lowest mean squared error (0.004017). As result SHAP analysis, most crucial variable influencing properties. precisely predicts can optimize production process.

Language: Английский

Citations

2