2.765

2022影响因子

(CJCR)

  • 中文核心
  • EI
  • 中国科技核心
  • Scopus
  • CSCD
  • 英国科学文摘

留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

基于宽度混合森林回归的城市固废焚烧过程二噁英排放软测量

夏恒 汤健 崔璨麟 乔俊飞

夏恒, 汤健, 崔璨麟, 乔俊飞. 基于宽度混合森林回归的城市固废焚烧过程二噁英排放软测量. 自动化学报, 2023, 49(2): 343−365 doi: 10.16383/j.aas.c220012
引用本文: 夏恒, 汤健, 崔璨麟, 乔俊飞. 基于宽度混合森林回归的城市固废焚烧过程二噁英排放软测量. 自动化学报, 2023, 49(2): 343−365 doi: 10.16383/j.aas.c220012
Xia Heng, Tang Jian, Cui Can-Lin, Qiao Jun-Fei. Soft sensing method of dioxin emission in municipal solid waste incineration process based on broad hybrid forest regression. Acta Automatica Sinica, 2023, 49(2): 343−365 doi: 10.16383/j.aas.c220012
Citation: Xia Heng, Tang Jian, Cui Can-Lin, Qiao Jun-Fei. Soft sensing method of dioxin emission in municipal solid waste incineration process based on broad hybrid forest regression. Acta Automatica Sinica, 2023, 49(2): 343−365 doi: 10.16383/j.aas.c220012

基于宽度混合森林回归的城市固废焚烧过程二噁英排放软测量

doi: 10.16383/j.aas.c220012
基金项目: 国家自然科学基金(62073006, 62173120, 62021003), 北京市自然科学基金资助项目(4212032, 4192009), 科技创新2030 — “新一代人工智能”重大项目(2021ZD0112301, 2021ZD0112302)资助
详细信息
    作者简介:

    夏恒:北京工业大学信息学部博士研究生. 主要研究方向为小样本数据建模, 城市固废焚烧过程二噁英排放预测. E-mail: xiaheng@emails.bjut.edu.cn

    汤健:北京工业大学信息学部教授. 主要研究方向为小样本数据建模, 城市固废处理过程智能控制. 本文通信作者. E-mail: freeflytang@bjut.edu.cn

    崔璨麟:北京工业大学信息学部硕士研究生. 主要研究方向为城市固废焚烧过程风险预警. E-mail: cuicanlin @emails.bjut.edu.cn

    乔俊飞:北京工业大学信息学部教授. 主要研究方向为污水处理过程智能控制, 神经网络结构设计与优化. E-mail: junfeiq@bjut.edu.cn

Soft Sensing Method of Dioxin Emission in Municipal Solid Waste Incineration Process Based on Broad Hybrid Forest Regression

Funds: Supported by National Natural Science Foundation of China (62073006, 62173120, 62021003), Beijing Natural Science Foundation (4212032, 4192009), National Key Research and Development Program of China (2021ZD0112301, 2021ZD0112302)
More Information
    Author Bio:

    XIA Heng Ph.D. candidate at the Faculty of Information Technology, Beijing University of Technology. His research interest covers small sample data modeling and dioxin emission prediction of the municipal solid waste incineration process

    TANG Jian Professor at the Faculty of Information Technology, Bei-jing University of Technology. His research interest covers small sample data modeling and intelligent control of municipal solid waste treatment process. Corresponding author of this paper

    CUI Can-Lin Master student at the Faculty of Information Technology, Beijing University of Technology. His main research interest is risk warning of municipal solid waste incineration process

    QIAO Jun-Fei Professor at the Faculty of Information Technology, Beijing University of Technology. His research interest covers intelligent control of waste water treatment process and structure design & optimization of neural networks

  • 摘要: 二噁英是城市固废焚烧过程排放的痕量有机污染物. 受限于相关技术的复杂度和高成本, 二噁英排放浓度检测的大时滞已成为制约城市固废焚烧过程优化控制的关键因素之一. 虽然具有低成本、快响应、高精度等特点的数据驱动软测量模型能够有效解决上述问题, 但二噁英建模方法必须要契合数据的小样本、高维度特性. 对此, 提出了由特征映射层、潜在特征提取层、特征增强层和增量学习层组成的宽度混合森林回归软测量方法. 首先, 构建由随机森林和完全随机森林构成的混合森林组进行高维特征映射; 其次, 依据贡献率对全联接混合矩阵进行潜在特征提取, 采用信息度量准则保证潜在有价值信息的最大化传递和最小化冗余, 降低模型的复杂度和计算消耗; 然后, 基于所提取潜在信息训练特征增强层以增强特征表征能力; 最后, 通过增量式学习策略构建增量学习层后采用Moore-Penrose伪逆获得权重矩阵. 在基准数据集和城市固废焚烧过程二噁英数据集上的实验结果表明了方法的有效性和优越性.
  • 图  1  城市固废焚烧工艺流程图

    Fig.  1  Process flow chart of municipal solid waste incineration process

    图  2  宽度混合森林回归建模策略图

    Fig.  2  Modeling strategy of broad hybrid forest regression

    图  3  RF和CRF的建模过程

    Fig.  3  Modeling process of RF and CRF

    图  4  基准数据集中潜在特征的贡献率曲线

    Fig.  4  Contribution rate curves of latent features of benchmark datasets

    图  5  基准数据集潜在特征与真值的互信息值

    Fig.  5  Mutual information values of latent features and true values of benchmark datasets

    图  6  基准数据集训练误差收敛曲线

    Fig.  6  Training error convergence curve of benchmark datasets

    图  7  NIR数据集的拟合曲线

    Fig.  7  Fitting curves of NIR dataset

    图  8  CT数据集的拟合曲线

    Fig.  8  Fitting curves of CT dataset

    图  9  RB数据集的拟合曲线

    Fig.  9  Fitting curves of RB dataset

    图  10  DXN数据潜在特征的贡献率曲线

    Fig.  10  Contribution rate curve of latent featureof DXN dataset

    图  11  DXN数据潜在特征与真值的互信息值

    Fig.  11  Mutual information value of latent feature and true value of DXN dataset

    图  12  DXN训练误差收敛曲线

    Fig.  12  Convergence curves of DXN training error

    图  13  DXN数据集中的拟合曲线

    Fig.  13  Fit curves of DXN dataset

    图  14  超参数敏感性分析曲线

    Fig.  14  Sensitivity analysis curves of super parameter

    表  1  符号说明

    Table  1  Symbol description

    缩写词中文全称英文全称
    MSWI城市固废焚烧Municipal solid waste incineration
    DXN二噁英Dioxin
    HRGC/HRMS高分辨色谱质谱联用High-resolution chromatography combined with high-resolution mass spectrometry
    NN神经网络Neural network
    PCDDs/PCDFs多氯二苯并二噁英/多氯二苯并呋喃Polychlorinated dibenzo-p-dioxins/Polychlorinated dibenzofurans
    BPNN反向传播神经网络Back-propagation neural network
    SVR支持向量回归Support vector regression
    SEN选择性集成Selective ensemble
    RF随机森林Random forest
    DFR深度森林回归Deep forest regression
    BLS宽度学习系统Broad learning system
    BLS-NN神经网络宽度学习系统Broad learning system neural network
    BHFR宽度混合森林回归Broad hybrid forest regression
    CRF完全随机森林Completely random forest
    MSW城市固废Municipal solid waste
    TEQ毒性当量Toxic equivalent quangtity
    G 1烟气 1Gas 1
    NOx氮氧化物Nitrogen oxides
    HCL氯化氢Hydrogen chloride
    HF氟化氢Hydrogen fluoride
    SO2二氧化硫Sulfur dioxide
    PbPlumbum
    HgMercury
    CdCadmium
    G 2烟气 2Gas 2
    G 3烟气 3Gas 3
    RSM随机子空间法Random subspace method
    PCA主成分分析Principal components analysis
    MI互信息Mutual information
    CTCT 切片在轴轴数据上的相对位置The relative location of CT slices on the axial axis data
    RB住宅建筑数据Residential building data
    NIR橙汁近红外光谱数据The orange juice near infrared spectra data
    RMSE均方根误差Root mean square error
    MAE平均绝对误差Mean absolute error
    R2决定系数Coefficient of determination
    DFR-clfc基于跨层全连接的深度森林回归Deep forest regression based on cross-layer full connection
    下载: 导出CSV

    表  2  实验数据统计结果

    Table  2  Statistical results of experimental datasets

    数据集实际样本量实际样本量特征维数
    总数训练集验证集测试集
    NIR218109552727700
    CT583117592929291
    RB372124623131106
    DXN141141713535116
    下载: 导出CSV

    表  3  NIR数据集实验结果

    Table  3  Experimental results of NIR dataset

    方法数据集RMSEMAE${ {{R} } }^{2}$
    平均值方差平均值方差平均值方差
    RF训练集5.7142$\times {10} ^{0}$2.1347$\times {10} ^{-2}$3.8799$\times {10} ^{0}$9.4469$\times {10} ^{-3}$8.1131$\times {10} ^{-1}$9.2887$\times {10} ^{-5}$
    验证集1.3522$\times {10} ^{1}$1.3660$\times {10} ^{-2}$8.3125$\times {10} ^{0}$7.2473$\times {10} ^{-3}$2.6715$\times {10} ^{-1}$1.6048$\times {10} ^{-4}$
    测试集1.0925$\times {10} ^{1}$3.2313$\times {10} ^{-2}$8.1093$\times {10} ^{0}$1.7493$\times {10} ^{-2}$3.0986$\times {10} ^{-1}$5.1625$\times {10} ^{-4}$
    DFR训练集5.8836$\times {10} ^{0}$1.4331$\times {10} ^{-3}$4.1004$\times \text{10}^{0}$6.1766$\times {10} ^{-4}$8.0007$\times \text{10}^{-1}$6.5971$\times {10} ^{-6}$
    验证集1.3585$\times {10} ^{1}$3.7257$\times {10} ^{-3}$8.3057$\times {10} ^{0}$1.5854$\times {10} ^{-3}$2.6036$\times {10} ^{-1}$4.4251$\times {10} ^{-5}$
    测试集1.0842$\times {10} ^{1}$1.4742$\times {10} ^{-3}$8.0306$\times {10} ^{0}$1.8258$\times {10} ^{-3}$3.2046$\times {10} ^{-1}$2.3176$\times {10} ^{-5}$
    DFR-clfc训练集5.7742$\times {10} ^{0}$1.7160$\times {10} ^{-2}$4.0139$\times {10} ^{0}$9.6271$\times {10} ^{-3}$8.0735$\times {10} ^{-1}$7.6485$\times {10} ^{-5}$
    验证集1.3569$\times {10} ^{1}$3.1520$\times {10} ^{-3}$8.3265$\times {10} ^{0}$1.3592$\times {10} ^{-3}$2.6219$\times {10} ^{-1}$3.7318$\times {10} ^{-5}$
    测试集1.0793$\times {10} ^{1}$2.5668$\times {10} ^{-3}$7.9746$\times {10} ^{0}$3.3925$\times {10} ^{-3}$3.2664$\times {10} ^{-1}$3.9879$\times {10} ^{-5}$
    BLS-NN训练集7.4226$\times {10} ^{-1}$3.7588$\times {10} ^{-1}$4.6530$\times {10} ^{-1}$1.5602$\times {10} ^{-1}$9.9476$\times {10} ^{-1}$3.2486$\times {10} ^{-5}$
    验证集7.8288$\times {10} ^{3}$4.9487$\times {10} ^{7}$2.5450$\times {10} ^{3}$4.6180$\times {10} ^{6}$−4.3402$\times {10} ^{5}$7.2220$\times {10} ^{11}$
    测试集6.4945$\times {10} ^{3}$1.4866$\times {10} ^{7}$2.1689$\times {10} ^{3}$1.5496$\times {10} ^{6}$−3.2544$\times {10} ^{5}$8.9669$\times {10} ^{10}$
    BHFR训练集2.8931$\times {10} ^{0}$5.5126$\times {10} ^{-1}$2.1335$\times {10} ^{0}$3.3004$\times {10} ^{-1}$9.4864$\times {10} ^{-1}$6.0585$\times {10} ^{-4}$
    验证集1.3782$\times {10} ^{1}$1.7263$\times {10} ^{0}$9.2584$\times {10} ^{0}$6.5031$\times {10} ^{-1}$2.3224$\times {10} ^{-1}$2.1525$\times {10} ^{-2}$
    测试集9.9505$\times {10} ^{0}$5.6804$\times {10} ^{-1}$7.3675$\times {10} ^{0}$2.4515$\times {10} ^{-1}$4.2455$\times {10} ^{-1}$7.3749$\times {10} ^{-3}$
    下载: 导出CSV

    表  4  CT数据集实验结果

    Table  4  Experimental results of CT dataset

    方法数据集RMSEMAE${ {{R} } }^{2}$
    平均值方差平均值方差平均值方差
    RF训练集1.5839$\times {10} ^{0}$1.2389$\times {10} ^{-2}$1.2676$\times {10} ^{0}$6.1900$\times {10} ^{-3}$9.7215$\times {10} ^{-1}$1.4862$\times {10} ^{-5}$
    验证集1.8907$\times {10} ^{0}$2.2323$\times {10} ^{-2}$1.4978$\times {10} ^{0}$1.4923$\times {10} ^{-2}$9.5761$\times {10} ^{-1}$4.2488$\times {10} ^{-5}$
    测试集2.2138$\times {10} ^{0}$3.6173$\times {10} ^{-2}$1.8254$\times {10} ^{0}$2.1898$\times {10} ^{-2}$9.5706$\times {10} ^{-1}$5.3991$\times {10} ^{-5}$
    DFR训练集4.1046$\times {10} ^{0}$1.3520$\times {10} ^{-3}$3.4579$\times {10} ^{0}$9.5891$\times {10} ^{-4}$8.1383$\times {10} ^{-1}$1.1086$\times {10} ^{-5}$
    验证集4.1864$\times {10} ^{0}$1.2940$\times {10} ^{-3}$3.5658$\times {10} ^{0}$1.4003$\times {10} ^{-3}$7.9338$\times {10} ^{-1}$1.2646$\times {10} ^{-5}$
    测试集4.8124$\times {10} ^{0}$2.1197$\times {10} ^{-3}$4.2219$\times {10} ^{0}$1.4647$\times {10} ^{-3}$7.9851$\times {10} ^{-1}$1.4815$\times {10} ^{-5}$
    DFR-clfc训练集3.7411$\times {10} ^{0}$4.1351$\times {10} ^{-2}$3.1392$\times {10} ^{0}$3.2494$\times {10} ^{-2}$8.4491$\times {10} ^{-1}$2.8032$\times {10} ^{-4}$
    验证集3.8251$\times {10} ^{0}$3.9698$\times {10} ^{-2}$3.2193$\times {10} ^{0}$3.5889$\times {10} ^{-2}$8.2708$\times {10} ^{-1}$3.1969$\times {10} ^{-4}$
    测试集4.3811$\times {10} ^{0}$4.9342$\times {10} ^{-2}$3.8122$\times {10} ^{0}$4.5453$\times {10} ^{-2}$8.3262$\times {10} ^{-1}$2.8693$\times {10} ^{-4}$
    BLS-NN训练集1.0447$\times {10} ^{-7}$4.7677$\times {10} ^{-15}$4.1145$\times {10} ^{-8}$5.7001$\times {10} ^{-16}$$1.0000\times {10} ^{0}$7.4604$\times {10} ^{-32}$
    验证集2.5527$\times {10} ^{0}$6.9348$\times {10} ^{-1}$1.7606$\times {10} ^{0}$2.0416$\times {10} ^{-1}$9.1542$\times {10} ^{-1}$3.2479$\times {10} ^{-3}$
    测试集2.9466$\times {10} ^{0}$1.3019$\times {10} ^{0}$1.8882$\times {10} ^{0}$4.4686$\times {10} ^{-1}$9.1371$\times {10} ^{-1}$3.5944$\times {10} ^{-3}$
    BHFR训练集4.4458$\times {10} ^{-1}$8.4237$\times {10} ^{-3}$3.7023$\times {10} ^{-1}$4.5597$\times {10} ^{-3}$9.9773$\times {10} ^{-1}$1.6195$\times {10} ^{-6}$
    验证集9.5728$\times {10} ^{-1}$1.3354$\times {10} ^{-2}$7.4450$\times {10} ^{-1}$6.1644$\times {10} ^{-3}$9.8905$\times {10} ^{-1}$1.0451$\times {10} ^{-5}$
    测试集9.3108$\times {10} ^{-1}$1.2802$\times {10} ^{-2}$6.7034$\times {10} ^{-1}$1.2237$\times {10} ^{-2}$9.9235$\times {10} ^{-1}$5.1917$\times {10} ^{-6}$
    下载: 导出CSV

    表  5  RB数据集实验结果

    Table  5  Experimental results of RB dataset

    方法数据集RMSEMAE${ {{R} } }^{2}$
    平均值方差平均值方差平均值方差
    RF训练集5.3243$\times {10} ^{1}$3.7463$\times {10} ^{0}$4.2919$\times {10} ^{1}$2.4224$\times {10} ^{0}$8.4698$\times {10} ^{-1}$1.2156$\times {10} ^{-4}$
    验证集1.3211$\times {10} ^{2}$6.6035$\times {10} ^{0}$8.1855$\times {10} ^{1}$2.9950$\times {10} ^{0}$5.3824$\times {10} ^{-1}$3.2140$\times {10} ^{-4}$
    测试集7.5697$\times {10} ^{1}$4.3622$\times {10} ^{0}$6.0986$\times {10} ^{1}$2.7466$\times {10} ^{0}$7.0126$\times {10} ^{-1}$2.7457$\times {10} ^{-4}$
    DFR训练集4.8752$\times {10} ^{1}$1.4944$\times {10} ^{-1}$3.9200$\times {10} ^{1}$8.4396$\times {10} ^{-2}$8.7185$\times {10} ^{-1}$4.1370$\times {10} ^{-6}$
    验证集1.2600$\times {10} ^{2}$1.6487$\times {10} ^{-1}$7.8890$\times {10} ^{1}$1.1939$\times {10} ^{-1}$5.8012$\times {10} ^{-1}$7.3315$\times {10} ^{-6}$
    测试集7.4221$\times {10} ^{1}$3.1170$\times {10} ^{-1}$6.0387$\times {10} ^{1}$2.4493$\times {10} ^{-1}$7.1299$\times {10} ^{-1}$1.8723$\times {10} ^{-5}$
    DFR-clfc训练集3.0978$\times {10} ^{1}$2.6657$\times {10} ^{1}$2.4856$\times {10} ^{1}$1.7056$\times {10} ^{1}$9.4690$\times {10} ^{-1}$3.4386$\times {10} ^{-4}$
    验证集1.1830$\times {10} ^{2}$4.9405$\times {10} ^{0}$7.2901$\times {10} ^{1}$2.7676$\times {10} ^{0}$6.2977$\times {10} ^{-1}$1.9820$\times {10} ^{-4}$
    测试集6.9427$\times {10} ^{1}$1.7460$\times {10} ^{0}$5.5570$\times {10} ^{1}$2.1741$\times {10} ^{0}$7.4879$\times {10} ^{-1}$9.2839$\times {10} ^{-5}$
    BLS-NN训练集9.4877$\times {10} ^{-4}$4.8003$\times {10} ^{-6}$4.3563$\times {10} ^{-4}$7.7161$\times {10} ^{-7}$1.0000$\times {10} ^{0}$1.1561$\times {10} ^{-18}$
    验证集5.0098$\times {10} ^{2}$1.3099$\times {10} ^{5}$2.6163$\times {10} ^{2}$2.6950$\times {10} ^{4}$−8.9285$\times {10} ^{0}$1.6631$\times {10} ^{2}$
    测试集5.2093$\times {10} ^{2}$1.5354$\times {10} ^{5}$2.8934$\times {10} ^{2}$3.2539$\times {10} ^{4}$−2.0737$\times {10} ^{1}$9.3226$\times {10} ^{2}$
    BHFR训练集8.4893$\times {10} ^{0}$6.4125$\times {10} ^{-1}$6.4970$\times {10} ^{0}$3.1874$\times {10} ^{-1}$9.9608$\times {10} ^{-1}$5.6404$\times {10} ^{-7}$
    验证集9.9275$\times {10} ^{1}$2.2922$\times {10} ^{0}$5.4880$\times {10} ^{1}$1.0407$\times {10} ^{0}$7.3930$\times {10} ^{-1}$6.3926$\times {10} ^{-5}$
    测试集4.5645$\times {10} ^{1}$1.7188$\times {10} ^{0}$3.2284$\times {10} ^{1}$9.0105$\times {10} ^{-1}$8.9137$\times {10} ^{-1}$3.8975$\times {10} ^{-5}$
    下载: 导出CSV

    表  6  DXN数据集实验结果

    Table  6  Experimental results of DXN dataset

    方法数据集RMSEMAE${ {{R} } }^{2}$
    平均值方差平均值方差平均值方差
    RF训练集1.1159$\times {10} ^{-2}$5.7497$\times {10} ^{-8}$9.0221$\times {10} ^{-3}$4.0684$\times {10} ^{-8}$8.5346$\times {10} ^{-1}$3.9360$\times {10} ^{-5}$
    验证集2.0051$\times {10} ^{-2}$1.8026$\times {10} ^{-7}$1.4677$\times {10} ^{-2}$8.2255$\times {10} ^{-8}$5.0196$\times {10} ^{-1}$4.3515$\times {10} ^{-4}$
    测试集1.6922$\times {10} ^{-2}$1.6150$\times {10} ^{-7}$1.3548$\times {10} ^{-2}$8.9520$\times {10} ^{-8}$5.9001$\times {10} ^{-1}$3.7817$\times {10} ^{-4}$
    DFR训练集1.1493$\times {10} ^{-2}$8.7413$\times {10} ^{-9}$9.4568$\times {10} ^{-3}$4.6626$\times {10} ^{-9}$8.4463$\times {10} ^{-1}$6.3663$\times {10} ^{-6}$
    验证集2.0735$\times {10} ^{-2}$9.7835$\times {10} ^{-9}$1.5780$\times {10} ^{-2}$1.1121$\times {10} ^{-8}$4.6759$\times {10} ^{-1}$2.5813$\times {10} ^{-5}$
    测试集1.7791$\times {10} ^{-2}$1.7308$\times {10} ^{-8}$1.4608$\times {10} ^{-2}$1.5235$\times {10} ^{-8}$5.4701$\times {10} ^{-1}$4.5066$\times {10} ^{-5}$
    DFR-clfc训练集8.0852$\times {10} ^{-3}$2.9078$\times {10} ^{-6}$6.6040$\times {10} ^{-3}$2.0819$\times {10} ^{-6}$9.1986$\times {10} ^{-1}$1.1887$\times {10} ^{-3}$
    验证集2.0187$\times {10} ^{-2}$1.4562$\times {10} ^{-7}$1.5626$\times {10} ^{-2}$2.3355$\times {10} ^{-8}$4.9520$\times {10} ^{-1}$3.6404$\times {10} ^{-4}$
    测试集1.7025$\times {10} ^{-2}$1.5755$\times {10} ^{-7}$1.4068$\times {10} ^{-2}$6.0233$\times {10} ^{-8}$5.8501$\times {10} ^{-1}$3.7843$\times {10} ^{-4}$
    BLS-NN训练集1.2924$\times {10} ^{-9}$1.5756$\times {10} ^{-18}$9.5358$\times {10} ^{-10}$7.2150$\times {10} ^{-19}$1.0000$\times {10} ^{0}$8.2358$\times {10} ^{-29}$
    验证集6.8845$\times {10} ^{-2}$7.0040$\times {10} ^{-4}$5.3153$\times {10} ^{-2}$3.3474$\times {10} ^{-4}$−5.6928$\times {10} ^{0}$3.7799$\times {10} ^{1}$
    测试集7.8396$\times {10} ^{-2}$6.7692$\times {10} ^{-4}$6.0922$\times {10} ^{-2}$4.1785$\times {10} ^{-4}$−8.7153$\times {10} ^{0}$4.7630$\times {10} ^{1}$
    BHFR训练集6.0665$\times {10} ^{-3}$1.6330$\times {10} ^{-8}$3.9665$\times {10} ^{-3}$8.4708$\times {10} ^{-9}$9.5669$\times {10} ^{-1}$3.3481$\times {10} ^{-6}$
    验证集2.1551$\times {10} ^{-2}$3.5181$\times {10} ^{-8}$1.2384$\times {10} ^{-2}$3.5083$\times {10} ^{-8}$4.2484$\times {10} ^{-1}$9.8731$\times {10} ^{-5}$
    测试集1.6189$\times {10} ^{-2}$2.2474$\times {10} ^{-8}$1.1226$\times {10} ^{-2}$1.0102$\times {10} ^{-8}$6.2491$\times {10} ^{-1}$4.8607$\times {10} ^{-5}$
    下载: 导出CSV

    表  7  BHFR超参数及其区间设置

    Table  7  Super parameters and interval setting of BHFR

    模型超参数符号NIRCTRBDXN
    决策树叶
    节点最小
    样本数
    Nsamples[1, 55][1, 59][1, 62][1, 71]
    RSM特征
    选择数量
    MRSM[1, 700][1, 291][1, 103][1, 116]
    决策树数量Mtree[5, 100][5, 100][5, 100][5, 100]
    混合森林组
    数量
    NForest[5, 100][5, 100][5, 100][5, 100]
    潜在特征
    贡献率阈值
    $\eta $[0.7, 1][0.4, 1][0.71, 1][0.3, 1]
    互信息阈值$\zeta $[0.65, 0.75][0.75, 0.85][0.72, 0.8][0.68, 0.76]
    正则化参数$\lambda$[0.1−10, 1][0.1−10, 1][0.1−10, 1][0.1−10, 1]
    下载: 导出CSV

    A1  DXN数据集的过程变量信息

    A1  The process variable information for DXN datasets

    序号过程变量单位
    1燃烧室左侧温度 1
    2燃烧室左侧温度 2
    3燃烧室右侧温度 3
    4燃烧室左侧温度4
    5燃烧室右侧温度 5
    6燃烧室右侧温度 6
    7燃烬段炉排顶端气温左
    8燃烬段炉排顶端气温右
    9干燥炉排温度左入口
    10干燥炉排温度左出口
    11干燥炉排温度右入口
    12干燥炉排温度右出口
    13干燥段与燃烧段炉排的炉墙左侧内温度
    14干燥段与燃烧段炉排的炉墙左侧外温度
    15干燥段与燃烧段炉排的炉墙右侧内温度
    16干燥段与燃烧段炉排的炉墙右侧外温度
    17燃烧炉排 1-1 左内侧温度
    18燃烧炉排 1-1左外侧温度
    19燃烧炉排 1-1 右内侧温度
    20燃烧炉排 1-1 右外侧温度
    21燃烧炉排 1-2 左内侧温度
    22燃烧炉排 1-2 左外侧温度
    23燃烧炉排1-2右内侧温度
    24燃烧炉排1-2右外侧温度
    25燃烧炉排 2-1 左内侧温度
    26燃烧炉排 2-1 左外侧温度
    27燃烧炉排 2-1 右内侧温度
    28燃烧炉排 2-1 右外侧温度
    29燃烧炉排 2-2 左内侧温度
    30燃烧炉排 2-2 左外侧温度
    31燃烧炉排 2-2 右内侧温度
    32燃烧炉排 2-2 右外侧温度
    33二次燃烧室出口温度左
    34二次燃烧室出口温度右
    35一次空预器出口空气温度
    36燃烧炉排入口空气温度
    37干燥炉排入口空气温度
    38二次空预器出口空气温度
    39炉墙左侧冷却风出口温度
    40炉墙右侧冷却风出口温度
    41炉排左侧冷却风出口温度
    42炉排右侧冷却风出口温度
    43一级过热器出口蒸汽温度
    44二级过热器出口蒸汽温度
    45三级过热器出口蒸汽温度
    46省煤器出口给水温度
    47反应器入口温度
    48布袋 A 入口温度
    49布袋 B 入口温度
    50流化风机出口温度
    51FGD 出口烟气温度 A
    52FGD 出口烟气温度 B
    53烟道入口烟气温度
    54干燥炉排左 1 空气流量km3N/h
    55干燥炉排右 1 空气流量km3N/h
    56干燥炉排左 2 空气流量km3N/h
    57干燥炉排右 2 空气流量km3N/h
    58燃烧炉排左 1-1 空气流量km3N/h
    59燃烧炉排右 1-1空气流量km3N/h
    60燃烧炉排左 1-2 空气流量km3N/h
    61燃烧炉排右 1-2 空气流量km3N/h
    62燃烧炉排左 2-1 空气流量km3N/h
    63燃烧炉排右 2-1 空气流量km3N/h
    64燃烧炉排左 2-2 空气流量km3N/h
    65燃烧炉排右 2-2 空气流量km3N/h
    66燃烬炉排左空气流量km3N/h
    67燃烬炉排右空气流量km3N/h
    68二次风流量km3N/h
    69省煤器 No.2 给水流量t/h
    70省煤器 No.1 给水流量t/h
    71一级过热器冷却水流量t/h
    72二级过热器冷却水流量t/h
    73锅炉出口主蒸汽流量t/h
    74烟道入口烟气流量km3N/h
    75混合器水流量 Akg/h
    76混合器水流量 Bkg/h
    77尿素溶剂供应流量L/h
    78石灰储仓给料量kg/h
    79活性炭储仓给料量kg/h
    80一次风机出口空气压力kPa
    81二次风机出口空气压力kPa
    82省煤器出口压力kPa
    83烟道入口烟气压力Pa
    84汽包压力MPa
    85三级过热器出口蒸汽压力MPa
    86FGD 出口烟气压力Pa
    87布袋压差 AkPa
    88布袋压差 BkPa
    89进料器左内侧速度%
    90进料器左外侧速度%
    91进料器右内侧速度%
    92进料器右外侧速度%
    93干燥炉排左内侧速度%
    94干燥炉排左外侧速度%
    95干燥炉排右内侧速度%
    96干燥炉排右外侧速度%
    97燃烧炉排 1 左内侧速度%
    98燃烧炉排 1 左外侧速度%
    99燃烧炉排 1 右内侧速度%
    100燃烧炉排 1 右外侧速度%
    101燃烧炉排 2 左内侧速度%
    102燃烧炉排 2 左外侧速度%
    103燃烧炉排 2 右内侧速度%
    104燃烧炉排 2 右外侧速度%
    105燃烬炉排左侧速度%
    106燃烬炉排右侧速度%
    107汽包水位mm
    108FGD 出口烟气压力-O2%
    109出口烟气分析-O2%
    110出口烟气分析−灰尘mg/m3N
    111出口烟气分析-NOmg/m3N
    112出口烟气分析-SO2mg/m3N
    113出口烟气分析-HCLmg/m3N
    114出口烟气分析-COmg/m3N
    115出口烟气分析-CO2%
    116出口烟气分析-HFmg/m3N
    下载: 导出CSV
  • [1] Kammen D M, Sunter D A. City-integrated renewable energy for urban sustainability. Science, 2016, 352(6288): 922-928 doi: 10.1126/science.aad9302
    [2] Tobias W, Ludwig K L, Robert B, Esther E, Luca F, Bodo H, et al. Persistence of engineered nanoparticles in a municipal solid-waste incineration plant. Nature Nanotechnology, 2012, 7(8): 520-524 doi: 10.1038/nnano.2012.64
    [3] 乔俊飞, 郭子豪, 汤健. 面向城市固废焚烧过程的二噁英排放浓度检测方法综述. 自动化学报, 2020, 46(6): 1063-1089 doi: 10.16383/j.aas.c190005

    Qiao Jun-Fei, Guo Zi-Hao, Tang Jian. Dioxin emission concentration measurement approaches for municipal solid wastes incineration process: A survey. Acta Automatica Sinica, 2020, 46(6): 1063-1089 doi: 10.16383/j.aas.c190005
    [4] Bin Wang, Peilong Wang, Xie Lin-Hua, Lin Rui-Biao, Lv Jie, Li Jian-Rong, et al. A stable zirconium based metal-organic framework for specific recognition of representative polychlorinated dibenzo-p-dioxin molecules. Nature Communications, 2019, 10(1): 1-8 doi: 10.1038/s41467-018-07882-8
    [5] Wei J X, Li H, Liu J G. Fate of dioxins in a municipal solid waste incinerator with state-of-the-art air pollution control devices in China. Environmental Pollution, 2021, 289: 1-10 doi: 10.1016/j.envpol.2021.117798
    [6] 代伟, 陆文捷, 付俊, 马小平. 工业过程多速率分层运行优化控制. 自动化学报, 2019, 45(10): 1946-1959

    Dai Wei, Lu Wen-Jie, Fu Jun, Ma Xiao-Ping. Multi-rate layered optimal operational control of industrial processes. Acta Automatica Sinica, 2019, 45(10): 1946-1959
    [7] Xue W Q, Fan J L, Victor G L, Li J N, Jiang Y, Chai T Y, et al. New methods for optimal operational control of industrial processes using reinforcement learning on two time scales. IEEE Transactions on Industrial Informatics, 2020, 16(5): 3085-3099 doi: 10.1109/TII.2019.2912018
    [8] 国家标准. HJ 77.2-2008. 环境空气和废气. 二噁英类的测定. 同位素稀释高分辨气相色谱−高分辨质谱法. 2008

    National standard. HJ 77.2-2008. Ambient air and flue gas. Determination of polychlorinated dibenzo-p-dioxions (PCDDs) and ploychlorinated dibenzofurans (PCDFs) Isotope dilution HRCC-HRMS. 2008
    [9] 李阿丹, 洪伟, 王晶. 激光解吸 /激光电离-质谱法二恶英及其关联物的在线检测. 燕山大学学报, 2015, 39(6): 511-515 doi: 10.3969/j.issn.1007-791X.2015.06.007

    Li A-Dan, Hong Wei, Wang Jing. Online detection of dioxins and their related substances by laser desorption/laser ionization mass spectrometry. Journal of Yanshan University, 2015, 39(6): 511-515 doi: 10.3969/j.issn.1007-791X.2015.06.007
    [10] Nakui H, Koyama H, Takakura A, Watanabe N. Online measurements of low-volatile organic chlorine for dioxin monitoring at municipal waste incinerators. Chemosphere, 2011, 85(2): 151-155 doi: 10.1016/j.chemosphere.2011.06.042
    [11] Zhang H J, Ni Y W, Chen J P, Zhang Q. Influence of variation in the operating conditions on pcdd/f distribution in a full-scale msw incinerator. Chemosphere, 2008, 70(4): 721-730 doi: 10.1016/j.chemosphere.2007.06.054
    [12] 王通, 段泽文. 基于模糊评估自适应更新的油井动液面软测量建模. 化工学报, 2019, 70(12): 4760-4769

    Wang Tong, Duan Ze-Wen. Soft sensor modeling for dynamic liquid level of oil well based on fuzzy inference adaptive updating. Ciesc Journal, 2019, 70(12): 4760-4769
    [13] 秦美华, 朱红求, 李勇刚, 陈俊名, 张凤雪, 李文婷, 等. 基于STA-K均值聚类的电化学废水处理过程离子浓度软测量. 化工学报, 2019, 70(9): 3458-3464

    Qin Mei-Hua, Zhu Hong-Qiu, Li Yong-Gang, Chen Jun-Ming, Zhang Feng-Xue, Li Wen-ting, et al. Soft-sensor method for ion concentration of electrochemical wastewater treatment based on sta-k-means clustering. Ciesc Journal, 2019, 70(9): 3458-3464
    [14] 刘卓, 汤健, 柴天佑, 余文. 基于多模态特征子集选择性集成建模的磨机负荷参数预测方法. 自动化学报, 2021, 47(8): 1921-1931

    Liu Zhuo, Tang Jian, Chai Tian-You, Yu Wen. Selective ensemble modeling approach for mill load parameter forecasting based on multi-modal feature sub-sets. Acta Automatica Sinica, 2021, 47(8): 1921-1931
    [15] 乔景慧, 柴天佑. 数据与模型驱动的水泥生料分解率软测量模型. 自动化学报, 2019, 45(8): 1564-1578 doi: 10.16383/j.aas.c180734

    Qiao Jing-Hui, Chai Tian-You. Data and model-based soft measurement model of cement raw meal decomposition ratio. Acta Automatica Sinica, 2019, 45(8): 1564-1578 doi: 10.16383/j.aas.c180734
    [16] Chang N B, Chen W C. Prediction of pcdds/pcdfs emissions from municipal incinerators by genetic programming and neural network modeling. Waste Management & Research, 2000, 18(4): 341-351
    [17] Bunsan S, Chen W Y, Chen H W, Chuang Y H, Grisdanurak N. Modeling the dioxin emission of a municipal solid waste incinerator using neural networks. Chemosphere, 2013, 92(3): 258-264 doi: 10.1016/j.chemosphere.2013.01.083
    [18] 肖晓东, 卢加伟, 海景, 廖利. 垃圾焚烧烟气中二噁英类浓度的支持向量回归预测. 可再生能源, 2017, 35(8): 1107-1114

    Xiao Xiao-Dong, Lu Jia-Wei, Hai Jing, Liao Li. Prediction of dioxin emissions in flue gas from waste incineration based on support vector regression. Renewable Energy Resources, 2017, 35(8): 1107-1114
    [19] Wu J, Yang H. Linear regression-based efficient svm learning for large-scale classification. IEEE Transactions on Neural Networks and Learning Systems, 2015, 26(10): 2357-2369 doi: 10.1109/TNNLS.2014.2382123
    [20] 乔俊飞, 郭子豪, 汤健. 基于多层特征选择的固废焚烧过程二噁英排放浓度软测量. 信息与控制, 2021, 50(1): 75-87 doi: 10.13976/j.cnki.xk.2021.9663

    Qiao Jun-Fei, Guo Zi-Hao, Tang Jian. Soft sensing of dioxin emission concentration in solid waste incineration process based on multi-layer feature selection. Information and Control, 2021, 50(1): 75-87 doi: 10.13976/j.cnki.xk.2021.9663
    [21] 汤健, 乔俊飞. 基于选择性集成核学习算法的固废焚烧过程二噁英排放浓度软测量. 化工学报, 2019, 70(2): 696-706 doi: 10.11949/j.issn.0438-1157.20181354

    Tang Jian, Qiao Jun-Fei. Dioxin emission concentration soft measuring approach of municipal solid waste incineration based on selective ensemble kernel learning algorithm. Ciesc Journal, 2019, 70(2): 696-706 doi: 10.11949/j.issn.0438-1157.20181354
    [22] Chang N B, Huang S H. Statistical modelling for the prediction and control of pcdds and pcdfs emissions from municipal solid waste incinerators. Waste Management & Research, 1995, 13(4): 379-400.
    [23] 汤健, 乔俊飞, 徐喆, 郭子豪. 基于特征约简与选择性集成算法的城市固废焚烧过程二噁英排放浓度软测量. 控制理论与应用, 2021, 38(1): 110-120 doi: 10.7641/CTA.2020.90874

    Tang Jian, Qiao Jun-Fei, Xu Zhe, Guo Zi-Hao. Soft measuring approach of dioxin emission concentration in municipal solid waste incineration process based on feature reduction and selective ensemble algorithm. Control Theory & Applications, 2021, 38(1): 110-120 doi: 10.7641/CTA.2020.90874
    [24] Xia H, Tang J, Qiao J F, Yan A J, Guo Z H. Soft measuring method of dioxin emission concentration for mswi process based on rf and gbdt. In: Proceedings of the Chinese Control and Decision Conference. Hefei, China: IEEE, 2020. 2173−2178
    [25] 段艳杰, 吕宜生, 张杰, 赵学亮, 王飞跃. 深度学习在控制领域的研究现状与展望. 自动化学报, 2016, 42(5): 643-654

    Duan Yan-Jie, Lv Yi-Sheng, Zhang Jie, Zhao Xue-Liang, Wang Fei-Yue. Deep learning for control: The state of the art and prospects. Acta Automatica Sinica, 2016, 42(5): 643-654
    [26] Zhou Z H, Ji F. Deep forest. National Science Review, 2019, 6: 74-86. doi: 10.1093/nsr/nwy108
    [27] 汤健, 夏恒, 乔俊飞, 郭子豪. 深度集成森林回归建模方法及应用. 北京工业大学学报, 2021, 47(11): 1219-1229

    Tang Jian, Xia Heng, Qiao Jun-Fei, Guo Zi-Hao. Modeling method of deep ensemble forest regression with its application. Journal of Beijing University of Technology, 2021, 47(11): 1219-1229
    [28] Tang J, Xia H, Zhang J, Qiao J F, Yu W. Deep forest regression based on cross-layer full connection. Neural Computing and Applications, 2021, 33(15): 9307-9328 doi: 10.1007/s00521-021-05691-7
    [29] Xu W, Tang J, Xia H, Sun Z J. Prediction method of dioxin emission concentration based on PCA and deep forest regression. In: Proceedings of the 40th Chinese Control Conference. Shanghai, China: IEEE, 2021. 1212−1217
    [30] Graybill F A, Meyer C D, Painter R J. Note on the computation of the generalized inverse of a matrix. Siam Review, 1966, 8(4): 522-524 doi: 10.1137/1008108
    [31] Chen C L P, Liu Z. Broad learning system: An effective and efficient incremental learning system without the need for deep architecture. IEEE Transactions on Neural Networks and Learning Systems, 2018, 29(1): 10−24
    [32] Chen C L P, Liu Z, Feng S. Universal approximation capability of broad learning system and its structural variations. IEEE Transactions on Neural Networks and Learning Systems, 2019, 30(4): 1191-1204 doi: 10.1109/TNNLS.2018.2866622
    [33] Fan J C, Wang X, Wang X X, Zhao J H, Liu X. Incremental wishart broad learning system for fast polsar image classification. IEEE Geoscience and Remote Sensing Letters, 2019, 16(12): 1854-1858 doi: 10.1109/LGRS.2019.2913999
    [34] Ye H L, Li H, Chen C L P. Adaptive deep cascade broad learning system and its application in image denoising. IEEE Transactions on Cybernetics, 2021, 51(9): 4450-4463 doi: 10.1109/TCYB.2020.2978500
    [35] Chu Y H, Lin H F, Liang Y, Zhang D Y, Diao Y F, Fan X C, et al. Hyperspectral image classification based on discriminative locality preserving broad learning system. Knowledge-Based Systems, 2020, 206: 1-17
    [36] Cheng C, Wang W J, Chen H T, Zhang B C, Shao J J, Teng W X. Enhanced fault diagnosis using broad learning for traction systems in high-speed trains. IEEE Transactions on Power Electronics, 2020, 36(7): 7461-7469
    [37] Pu X K, Li C G. Online semisupervised broad learning system for industrial fault diagnosis. IEEE Transactions on Industrial Informatics, 2021, 17(10): 6644-6654 doi: 10.1109/TII.2020.3048990
    [38] Yu W K, Zhao C H. Broad convolutional neural network based industrial process fault diagnosis with incremental learning capability. IEEE Transactions on Industrial Electronics, 2020, 67(6): 5081-5091 doi: 10.1109/TIE.2019.2931255
    [39] Sui S, Chen P L C, Tong S C, Feng S. Finite-time adaptive quantized control of stochastic nonlinear systems with input quantization: A broad learning system based identification method. IEEE Transactions on Industrial Electronics, 2020, 67(10): 8555-8565 doi: 10.1109/TIE.2019.2947844
    [40] Feng J, Yao Y, S, Lu S X, Liu Y. Domain knowledge-based deep-broad learning framework for fault diagnosis. IEEE Transactions on Industrial Electronics, 2021, 68(4): 3454-3464 doi: 10.1109/TIE.2020.2982085
    [41] Han H G, Yang F F, Yang Y H, Wu X L. Type-2 fuzzy broad learning controller for wastewater treatment process. Neurocomputing, 2021, 459(4): 188-200
    [42] Tang H, Dong P, Shi Y. A construction of robust representations for small data sets using broad learning system. IEEE Transactions on Systems, Man, and Cybernetics: Systems, 2021, 51(10): 6074-6084 doi: 10.1109/TSMC.2019.2957818
    [43] Qi G J, Luo J. Small data challenges in big data era: A survey of recent progress on unsupervised and semi-supervised methods. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2022, 44(4): 2168-2187 doi: 10.1109/TPAMI.2020.3031898
    [44] 应雨轩, 林晓青, 吴昂键, 李晓东. 生活垃圾智慧焚烧的研究现状及展望. 化工学报, 2021, 72(2): 886-900

    Ying Yu-Xuan, Lin Xiao-Qing, Wu Ang-Jian, Li Xiao-Dong. Review and outlook on municipal solid waste smart incineration. CIESC Journal, 2021, 72(2): 886-900
    [45] Xia H, Tang J, Aljerf L, Wang T Z, Qiao J F, Xu Q D, Wang Q, Ukaogo P. Investigation on Dioxins Emission Characteristic during Complete Maintenance Operating Period of Municipal Solid Waste Incineration, Environmental Pollution, 2023, 318: 120949
    [46] Deng D Y, Qiao J Q, Liu M Q, Dorota K, Zhang M W, Dionysios D D F, et al. Detoxification of municipal solid waste incinerator fly ash by single-mode microwave irradiation: Addition of urea on the degradation of dioxin and mechanism. Journal of Hazardous Materials, 2019, 369: 279-289 doi: 10.1016/j.jhazmat.2019.01.001
    [47] Stanmore B R. The formation of dioxins in combustion systems. Combustion and Flame, 2004, 136(3): 398-427 doi: 10.1016/j.combustflame.2003.11.004
    [48] Zhang S, Chen Z L, Lin X Q, Wang F, Yan J H. Kinetics and fusion characteristics of municipal solid waste incineration fly ash during thermal treatment. Fuel, 2020, 279: 1-13
    [49] Zhang M M, Alfons G B. De novo synthesis of dioxins: A review. International Journal of Environment and Pollution, 2016, 60: 63-110 doi: 10.1504/IJEP.2016.082115
    [50] Zhou H, Meng A H, Long Y Q, Li Q H, Zhang Y G. A review of dioxin-related substances during municipal solid waste incineration. Waste Management, 2015, 36: 106-118 doi: 10.1016/j.wasman.2014.11.011
    [51] Peng Y Q, Lu S Y, Li X D, Yan J H, Cen K F. Formation, measurement, and control of dioxins from the incineration of municipal solid wastes: Recent advances and perspectives. Energy & Fuels, 2020, 34: 13247-13267
    [52] , Influence factors and mass balance of memory effect on PCDD/F emissions from the full-scale municipal solid waste incineration in China. Chemosphere, 2020, 239: 1-8Chemosphere, 2001, 45(8): 1151-1157
    [53] Breiman L. Bagging predictors. Mach Learn, 1996, 24: 123-140
    [54] Ho T K. The random subspace method for constructing decision forests. IEEE Transactions on Pattern Analysis & Machine Intelligence, 1998, 20(8): 832-844
    [55] Blake C L, Merz C J. UCI repository of machine learning databases [Online], available: http://www.archive.ics.uci.edu/ml/k datasets.html, January 1, 2022
  • 加载中
图(14) / 表(8)
计量
  • 文章访问数:  668
  • HTML全文浏览量:  174
  • PDF下载量:  190
  • 被引次数: 0
出版历程
  • 收稿日期:  2022-01-04
  • 录用日期:  2022-03-04
  • 网络出版日期:  2022-08-09
  • 刊出日期:  2023-02-20

目录

    /

    返回文章
    返回