您好,欢迎访问江苏省农业科学院 机构知识库!

Machine learning-driven optimization of metal-modified biochar for phosphorus adsorption and wastewater remediation

文献类型: 外文期刊

作者: Lu, Xinyue 2 ; Han, Yuheng 2 ; Wang, Lisha 5 ; Chen, Haoming 6 ; Feng, Yuanyuan 7 ; Chen, Yudong 1 ; Feng, Yanfang 2 ; Xue, Lihong 2 ;

作者机构: 1.Minist Ecol & Environm, Nanjing Inst Environm Sci, State Environm Protect Key Lab Soil Environm Manag, Nanjing 210042, Peoples R China

2.Jiangsu Acad Agr Sci, Inst Agr Resources & Environm, Nanjing 210014, Peoples R China

3.Jiangsu Acad Agr Sci, Key Lab Agroenvironm Downstream Yangtze Plain, Key Lab Combined Farming & Raising, Minist Agr & Rural Affairs, Nanjing 210014, Peoples R China

4.Nanjing Forestry Univ, Sch Mat Sci & Engn, Nanjing 210037, Peoples R China

5.Nanjing Univ, Sch Environm, State Key Lab Pollut Control & Resource Reuse, 163 Xianlin Ave, Nanjing 210023, Peoples R China

6.Nanjing Univ Sci & Technol, Sch Environm & Bioengn, Nanjing 210094, Peoples R China

7.Nanjing Forestry Univ, Coll Forestry & Grassland, Coll Soil & Water Conservat, Nanjing 210037, Peoples R China

关键词: Phosphorus; Metal-modified biochar; Machine Learning; Random Forest model

期刊名称:JOURNAL OF ENVIRONMENTAL CHEMICAL ENGINEERING ( 影响因子:7.2; 五年影响因子:7.6 )

ISSN: 2213-2929

年卷期: 2025 年 13 卷 5 期

页码:

收录情况: SCI

摘要: This study employed machine learning approaches to systematically analyze pyrolysis parameters, structural characteristics, modification condition and adsorption conditions for identifying critical determinants of phosphorus adsorption capacity in biochar. Four predictive models were evaluated, with Random Forest (RF) demonstrating optimal performance (Test R-2=0.89, RMSE=5.29) through effective overfitting resistance and generalization capability. Key findings revealed that: (1) Structural properties exhibited non-linear interactions affecting adsorption efficiency, with specific surface area showing limited correlation (contrary to conventional assumptions). (2) Adsorption conditions collectively contributed 35.1 % feature importance, dominated by solution pH (3.0-5.7), phosphorus concentration (905-1000 mg/L), and adsorbent dosage (<1 g/L). (3) Metal composition emerged as the most significant factor (31.7 % importance), particularly Mg content demonstrating dose-dependent enhancement effects. Notably, the model revealed synergistic interactions between modification parameters and adsorption variables, suggesting that optimal performance requires coordinated optimization of both material properties and operational conditions. These findings challenge traditional surface-area-centric evaluation methods, emphasizing the necessity for multifactorial analysis in biochar design, while highlighting the need for comprehensive considering environmental risks and economic viability in practical applications.

  • 相关文献
作者其他论文 更多>>