基于Stacking集成学习的有机废物堆肥过程有效磷生成预测及关键影响因素解析

戴旭; 杨晓帆; 屈静; 张勐; 高小中; 程翔; 杨天学; 孙德智

doi:10.13205/j.hjgc.202608006

基于Stacking集成学习的有机废物堆肥过程有效磷生成预测及关键影响因素解析

doi: 10.13205/j.hjgc.202608006

戴旭¹,
杨晓帆¹,
屈静¹,
张勐¹,
高小中¹,
程翔^1, ,,
杨天学²,
孙德智¹

1. 北京林业大学环境科学与工程学院, 北京 100083;
2. 中国环境科学研究院固体废物污染控制技术研究所, 北京 100012

基金项目:

国家重点研发计划项目“湿垃圾多形态磷转化规律及驱动机制”（2024YFC3908901）

详细信息

作者简介:
戴旭（2002—），男，硕士研究生，主要研究方向为固体废物处理处置与资源化。daixu1017@bjfu.edu.cn

通讯作者:
程翔（1981—），男，博士，教授，主要研究方向为低碳水处理与资源回收研究。xcheng@bjfu.edu.cn

计量
- 文章访问数: 2
- HTML全文浏览量: 1
- PDF下载量: 0
- 被引次数: 0
出版历程
- 收稿日期: 2026-03-19

Prediction of available phosphorus formation and analysis of key influencing factors during organic waste composting using stacking ensemble learning

1. College of Environmental Science and Engineering, Beijing Forestry University, Beijing 100083, China;
2. The Institute of Solid Waste Pollution Control Technology, Chinese Research Academy of Environmental Sciences, Beijing 100012, China

摘要

摘要: 有机废物是潜在磷资源库，其资源化利用过程中有效磷（available phosphorus，AP）的变化规律对磷资源高效利用具有重要意义。堆肥作为有机废物资源化利用的重要途径，有机质在降解和腐殖化过程中也伴随着磷素形态的转化。然而，由于堆肥实验周期长、成本高，且体系受多种因素共同影响，传统试验手段难以高效揭示AP的动态演变特征。研究汇编33篇文献数据，构建了包含647组样本的研究数据集，通过迭代插补、独热编码与标准化实现数据预处理，采用Stacking集成学习方法构建了堆肥过程AP生成的预测模型；并利用SHAP（shapley additive explanations，SHAP）和部分依赖图（partial dependence plot，PDP）解析了影响AP生成的关键参数。结果表明：有机废物堆肥过程AP预测的最优集成模型是由XGBoost与SVR作为基模型，ElasticNet作为元模型而构成。该模型在训练集和测试集上的决定系数R²分别为0.954和0.928，预测性能稳定，整体误差较低；SHAP的全局特征重要性分析表明，影响AP含量的关键因素的影响程度顺序为原料类型、调理剂类型、翻堆周期、pH、电导率与C/N。其中，以畜禽粪便作为原料、秸秆类作为调理剂对AP预测值表现为显著正向贡献。进一步结合PDP可知：在全堆肥周期内，较低的pH与C/N水平总体上更有利于AP积累；在堆肥初期，充足的含水率与低电导率水平更利于AP提升；在高温期前段，较高的堆体温度与更高的AP预测水平相对应；在降温腐熟期，保持含水率（<48%）与C/N（<14）在相对较低的水平下，适当延长堆肥天数（>43 d），可促进AP累积。研究通过stacking集成学习方法实现了对有机废物堆肥过程中AP含量的精准预测，并结合SHAP和PDP分析识别了影响AP生成的关键因素。预测模型的建立为堆肥过程有效磷的生成调控和工程应用优化提供了支撑。
- 集成学习 /
- 堆肥 /
- 有效磷 /
- 模型可解释性 /
- SHAP分析
Abstract: Organic waste contains abundant phosphorus （P） and is regarded as a potential P reservoir. The variation in available phosphorus （AP） during its resource utilization is of great importance for phosphorus recovery and utilization. Composting is an important approach for the resource utilization of organic waste， during which P speciation is transformed along with the degradation and humification of organic matter. However， because composting experiments are time-consuming and costly， and the process is influenced by multiple factors， the dynamic behavior of AP is difficult to reveal efficiently using conventional experiments. In this study， a dataset containing 647 samples was compiled from the literature. Data preprocessing was implemented via iterative imputation， one-hot encoding， and standardization. A predictive model for AP formation during composting was developed using a stacking ensemble learning method. Shapley additive explanations （SHAP） and partial dependence plots （PDPs） were further used to identify the key factors affecting AP formation. The results showed that the best ensemble model for predicting AP during organic waste composting consisted of XGBoost and SVR as base learners and ElasticNet as the meta-learner. The model achieved R² values of 0.954 and 0.928 for the training and test sets， respectively， indicating stable predictive performance and low overall error. Global SHAP importance analysis showed that the main factors affecting AP content were feedstock type， bulking agent type， turning interval， pH， electrical conductivity （EC）， and the C/N ratio. In particular， livestock manure as the feedstock and straw-based materials as the bulking agent made significant positive contributions to the predicted AP values. PDP analysis further showed that relatively low pH and C/N ratio generally favored AP accumulation throughout the composting period. During the initial stage of composting， higher moisture content and lower EC were more favorable for AP increase. During the early thermophilic stage， higher pile temperature was associated with higher predicted AP levels. During the cooling and maturation stage， AP accumulation was further promoted when moisture content （< 48%） and C/N ratio （< 14） were kept at relatively low levels and the composting duration was appropriately extended （> 43 d）. In this study， accurate prediction of AP content during organic waste composting was achieved using a stacking ensemble learning method， and the key factors affecting AP formation were identified through SHAP and PDP analyses. The developed model provides support for regulating AP formation and optimizing engineering applications in the composting process.
- ensemble learning /
- composting /
- available phosphorus （AP） /
- model interpretability /
- SHAP analysis

HTML全文

参考文献(47)

[1]	CORDELL D，DRANGERT J O，WHITE S. The story of phosphorus：global food security and food for thought[J]. Global Environmental Change，2009，19（2）：292-305.
[2]	ELSER J，BENNETT E. A broken biogeochemical cycle[J]. Nature，2011，478（7367）：29-31.
[3]	GUO X，LIANG S，ZOU Z，et al. Enhanced phosphorus bioavailability of biochar derived from sewage sludge co-pyrolyzed with K，Ca-rich biomass ash[J]. Water Research，2025，271：122901.
[4]	WANG J，WANG Y，YU F，et al. Efficient reclamation of phosphorus from wetland biomass waste via liquid-recirculated hydrothermal carbonization and precipitation[J]. Water Research，2024，265：122278.
[5]	SHAO M. Analysis of total organic waste and potential for replacing chemical fertilizers in China[J]. Polish Journal of Environmental Studies，2024，33（3）：3397-3405.
[6]	MANEA E E，BUMBAC C，DINU L R，et al. Composting as a sustainable solution for organic solid waste management：current practices and potential improvements[J]. Sustainability，2024，16（15）：6329.
[7]	XU P，SHU L，LI Y，et al. Pretreatment and composting technology of agricultural organic waste for sustainable agricultural development[J]. Heliyon，2023，9（5）：e16311.
[8]	TENG H H，HU C Z，NING J B，et al. Speciation transformation of phosphorus during sludge composting[C]// Proceedings of Annual Conference of Chinese Society for Environmental Sciences，2014：6922-6926. 滕洪辉，胡财政，宁军博，等. 污泥堆肥过程中磷元素的形态变化[C]// 中国环境科学学会学术年会论文集，2014：6922-6926.
[9]	DEBICKA M，JAMROZ E，BEKIER J，et al. The influence of municipal solid waste compost on the transformations of phosphorus forms in soil[J]. Agronomy，2023，13（5）：1234.
[10]	LANNO M，KRIIPSALU M，SHANSKIY M，et al. Distribution of phosphorus forms depends on compost source material[J]. Resources，2021，10（10）：102.
[11]	ONWOSI C O，IGBOKWE V C，ODIMBA J N，et al. Composting technology in waste stabilization：on the methods，challenges and future prospects[J]. Journal of Environmental Management，2017，190：140-157.
[12]	WAN F，BIAN C，WENG E，et al. TECO-CNP Sv1.0：a coupled carbon-nitrogen-phosphorus model with data assimilation for subtropical forests[J]. Geoscientific Model Development，2025，18（20）：7545-7573.
[13]	LAKE R W，SHAERI S，SENEVIRATHNA S. Review of the limitations and potential empirical improvements of the parametric group method of data handling for rainfall modelling[J]. Environmental Science and Pollution Research，2022，30（44）：98907-98921.
[14]	ZHU J J，YANG M，REN Z J. Machine learning in environmental research：common pitfalls and best practices[J]. Environmental Science& Technology，2023，57（46）：17671-17689.
[15]	AGHAABBASI M，CHALERMPONG S. Machine learning techniques for evaluating the nonlinear link between built-environment characteristics and travel behaviors：a systematic review[J]. Travel Behaviour and Society，2023，33：100640.
[16]	GUO H N，WU S B，TIAN Y J，et al. Application of machine learning methods for the prediction of organic solid waste treatment and recycling processes：a review[J]. Bioresource Technology，2021，319：124114.
[17]	MONCKS P C S，CORRÊA É K，GUIDONI L C，et al. Moisture content monitoring in industrial-scale composting systems using low-cost sensor-based machine learning techniques[J]. Bioresource Technology，2022，359：127456.
[18]	YILMAZ E C，TEMEL F A，YOLCU O C，et al. Modeling and optimization of process parameters in co-composting of tea waste and food waste：radial basis function neural networks and genetic algorithm[J]. Bioresource Technology，2022，363：127910.
[19]	BAYINDIR Y，YOLCU O C，TEMEL F A，et al. Evaluation of a cascade artificial neural network for modeling and optimization of process parameters in co-composting of cattle manure and municipal solid waste[J]. Journal of Environmental Management，2022，318：115496.
[20]	KABAK E T，YOLCU O C，TEMEL F A，et al. Prediction and optimization of nitrogen losses in co-composting process by using a hybrid cascaded prediction model and genetic algorithm[J]. Chemical Engineering Journal，2022，437：135499.
[21]	GUO H N，LIU H T，WU S. Simulation，prediction and optimization of typical heavy metals immobilization in swine manure composting by using machine learning models and genetic algorithm[J]. Journal of Environmental Management，2022，323：116266.
[22]	WU M，DOU S，LIN N，et al. Estimation and mapping of soil organic matter content using a stacking ensemble learning model based on hyperspectral images[J]. Remote Sensing，2023，15（19）：4713.
[23]	NODA R，ICHIKAWA D，SHIBAGAKI Y. Machine learning-based diagnostic prediction of IgA nephropathy：model development and validation study[J]. Scientific Reports，2024，14：12426.
[24]	EMMANUEL T，MAUPONG T，MPOELENG D，et al. A survey on missing data in machine learning[J]. Journal of Big Data，2021，8（1）：140.
[25]	RAMIREZ S G，WILLIAMS G P，JONES N L，et al. Improving groundwater imputation through iterative refinement using spatial and temporal correlations from in situ data with machine learning[J]. Water，2023，15（6）：1236.
[26]	DING S，HUANG W，XU W，et al. Improving kitchen waste composting maturity by optimizing the processing parameters based on machine learning model[J]. Bioresource Technology，2022，360：127606.
[27]	ZHANG X，KONG Y，YANG Y，et al. Using tree-based machine learning models to predict diverse compost maturity via one-hot encoding：model deployment，experimental validation，and practical application[J]. Waste Management，2025，205：114981.
[28]	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]. Journal of Cleaner Production，2021，278：123928.
[29]	UDDIN M G，RAHMAN A，TAGHIKHAH F R，et al. Data-driven evolution of water quality models：an in-depth investigation of innovative outlier detection approaches-a case study of Irish Water Quality Index（IEWQI）model[J]. Water Research，2024，255：121499.
[30]	FERNÁNDEZ DEL CASTILLO A，YEBRA-MONTES C，GARIBAY M V，et al. Simple prediction of an ecosystem-specific water quality index and the water quality classification of a highly polluted river through supervised machine learning[J]. Water，2022，14（8）：1235.
[31]	PAVLYSHENKO B. Using stacking approaches for machine learning models[C]// 2018 IEEE Second International Conference on Data Stream Mining& Processing（DSMP）. Lviv：IEEE，2018：255-258.
[32]	HASROD T，NUAPIA Y B，TUTU H. Comparison of individual and ensemble machine learning models for prediction of sulphate levels in untreated and treated acid mine drainage[J]. Environmental Monitoring and Assessment，2024，196（4）：332.
[33]	XIA H，HAN J，MILISAVLJEVIC-SYED J. Predictive modeling for the quantity of recycled end-of-life products using optimized ensemble learners[J]. Resources，Conservation and Recycling，2023，197：107073.
[34]	LIM S J，SON M，KI S J，et al. Opportunities and challenges of machine learning in bioprocesses：categorization from different perspectives and future direction[J]. Bioresource Technology，2023，370：128518.
[35]	ZHANG Y，FENG Y，REN Z，et al. Tree-based machine learning model for visualizing complex relationships between biochar properties and anaerobic digestion[J]. Bioresource Technology，2023，374：128746.
[36]	RESHEF D N，RESHEF Y A，FINUCANE H K，et al. Detecting novel associations in large data sets[J]. Science，2011，334（6062）：1518-1524.
[37]	ZHAN Y B，CHANG Y，ZHOU K Y，et al. Phosphorus transformation characteristics during composting of organic wastes from different sources[J]. Environmental Engineering，2023，41（S2）：1098-1102. 詹亚斌，常远，周凯云，等. 不同来源有机废弃物堆肥磷素转化特征[J]. 环境工程，2023，41（增刊2）：1098-1102.
[38]	WEI Y，ZHAO Y，XI B，et al. Changes in phosphorus fractions during organic wastes composting from different sources[J]. Bioresource Technology，2015，189：349-356.
[39]	SARR P S，TIBIRI E B，FUKUDA M，et al. Phosphate-solubilizing fungi and alkaline phosphatase trigger the P solubilization during the co-composting of sorghum straw residues with Burkina Faso phosphate rock[J]. Frontiers in Environmental Science，2020（8）：559195.
[40]	CUI H，OU Y，WANG L，et al. Additive grain-size：An innovative perspective to investigate the transformation among heavy metal and phosphorus fractions during aerobic composting[J]. Journal of Environmental Management，2021，292：112768.
[41]	MA H，LIU W，LI J，et al. Comparison of phosphorus transformation and microbial functional potential in aerobic and facultative composting of livestock and poultry manure[J]. Environmental Research，2025，285：122425.
[42]	MA T，ZHAN Y，CHEN W，et al. Impact of aeration rate on phosphorus conversion and bacterial community dynamics in phosphorus-enriched composting[J]. Bioresource Technology，2022，364：128016.
[43]	RACLAVSKÁ H，RŮŽIČKOVÁ J，RACLAVSKÝ K，et al. Effect of biochar addition on the improvement of the quality parameters of compost used for land reclamation[J]. Environmental Science and Pollution Research，2021，30（4）：8563-8581.
[44]	WANG Y，LI N N，ZHU D W，et al. Effects of rice straw pretreatment on thermophilic composting of pig manure[J]. Journal of Agro-Environment Science，2018，37（9）：2021-2028. 王砚，李念念，朱端卫，等. 水稻秸秆预处理对猪粪高温堆肥过程的影响[J]. 农业环境科学学报，2018，37（9）：2021-2028.
[45]	YANG C. Effect of biochar on phosphorus speciation transformation and its microbial mechanism during pig manure composting[D]. Xi'an：Xi'an University of Architecture and Technology，2022. 杨超. 生物炭对猪粪堆肥中磷形态转化的影响及其微生物学机制[D]. 西安：西安建筑科技大学，2022.
[46]	MARKLEIN A R，HOULTON B Z. Nitrogen inputs accelerate phosphorus cycling rates across a wide variety of terrestrial ecosystems[J]. New Phytologist，2012，193（3）：696-704.
[47]	ZHANG T，LI H，YAN T，et al. Organic matter stabilization and phosphorus activation during vegetable waste composting：Multivariate and multiscale investigation[J]. Science of the Total Environment，2023，891：164608.