School of Energy Science and Engineering, Central South University

基于机器学习的方法,探究了从模拟废水中以鸟粪石的形式回收氮和磷的问题。 利用极限梯度提升算法(XGBoost)和随机森林(RF) 模型对磷回收率和氮回收率进行单目标和多目标预测,明确了 7 种工艺条件对鸟粪石结晶的影响。 XGBoost 在单目标(R2 = 0.91 ~ 0.93) 和多目标(R2 = 0.89)的预测方面表现均优于 RF。 此外,在 P 初始浓度为10 mg / L和 1 000 mg / L 的情况下,通过实验验证了多目标模型的优化解集,得到鸟粪石回收的最佳工艺条件为 N ∶ P 比值为1.2 ∶ 1,Mg ∶ P 为 1 ∶ 1,pH 为 9.5,反应时间为 80 min,反应温度为25 ℃,搅拌速率为 240 r/ min。

The recovery of nitrogen and phosphorus from simulated wastewater in the form of struvitewas investigated through a Machine Learning (ML) -based approach. The Extreme Gradient BoostingAlgorithm (XGBoost) and Random Forest (RF) models were used for single-objective and multi-ob⁃jective prediction of the recovery rates of N and P, respectively. The effects of seven process conditionson struvite crystallization were identified. The results showed that XGBoost outperformed RF in both sin⁃gle-objective (R2 = 0.91 ~ 0.93) and multi-objective (R2 = 0.89) predictions. Furthermore, experi⁃mental validation was conducted with initial phosphorus concentrations of 10 mg / L and 1 000 mg / L todetermine the optimized process conditions for struvite recovery using the multi-objective model. Theoptimal conditions were found to be: N ∶ P ratio of 1.2 ∶ 1, Mg ∶ P ratio of 1 ∶ 1, pH of 9.5, reactiontime of 80 min, reaction temperature of 25 ℃, and stirring rate of 240 r/ min.

国家重点研发计划资助项目(2021YFE0104900)

佟颖, 蒋绍坚, 康冰艳, 等. 废水中鸟粪石回收的机器学习预测和优化[J]. 能源环境保护, 2023, 37(6):79-88.

TONG Ying, JIANG Shaojian, KANG Bingyan, et al. Prediction and optimization of struvite recovery fromwastewater by machine learning[J]. Energy Environmental Protection, 2023, 37(6): 79-88.