您的位置:山东大学 -> 科技期刊社 -> 《山东大学学报(工学版)》

山东大学学报 (工学版) ›› 2020, Vol. 50 ›› Issue (3): 133-142.doi: 10.6040/j.issn.1672-3961.0.2019.009

• 其他 • 上一篇    

变量选择在废水处理过程软测量建模中的应用

刘鸿斌1,2(),吴启悦1,宋留1   

  1. 1. 南京林业大学江苏省林业资源高效加工利用协同创新中心, 江苏 南京 210037
    2. 华南理工大学制浆造纸工程国家重点实验室, 广东 广州 510640
  • 收稿日期:2019-01-07 出版日期:2020-06-20 发布日期:2020-06-16
  • 作者简介:刘鸿斌(1981—),男,山西大同人,博士,副教授,主要研究方向为制浆造纸过程监测与控制研究. E-mail: hongbinliu@njfu.edu.cn
  • 基金资助:
    制浆造纸工程国家重点实验室开放基金资助项目(201813);南京林业大学高层次人才科研启动基金(GXL029)

Application of variable selection in soft sensor modeling of wastewater treatment processes

Hongbin LIU1,2(),Qiyue WU1,Liu SONG1   

  1. 1. Jiangsu Co-Innovation Center of Efficient Processing and Utilization of Forest Resources, Nanjing Forestry University, Nanjing 210037, Jiangsu, China
    2. State Key Laboratory of Pulp and Paper Engineering, South China University of Technology, Guangzhou 510640, Guangdong, China
  • Received:2019-01-07 Online:2020-06-20 Published:2020-06-16
  • Supported by:
    制浆造纸工程国家重点实验室开放基金资助项目(201813);南京林业大学高层次人才科研启动基金(GXL029)

摘要:

化学需氧量与悬浮固形物含量是造纸工业废水排放中需要重点监测的指标,建立有效的废水出水水质预测模型是优化控制废水中污染物排放量的有效方法。由于实际工业废水处理过程的复杂性,可测变量之间存在强相关性,利用偏最小二乘法提取变量的投影重要性信息进行变量选择,将选择后的最优变量子集作为软测量模型的输入,建立出水水质的最优预测模型。以最小二乘支持向量机模型为例,基于变量选择的最小二乘支持向量机模型对出水化学需氧量进行预测时均方根误差降低了15.2%,相关系数提高了14.4%;对于出水悬浮固形物模型,均方根误差降低了20.5%,相关系数提高了16.1%。结果表明在建模时进行变量选择可以降低模型的复杂度和提高模型的泛化能力。

关键词: 废水处理, 出水水质, 变量选择, 变量投影重要性, 偏最小二乘

Abstract:

Chemical oxygen demand and suspended solid were important monitoring indices of effluent discharge in paper-making industry. An effective model of effluent quality of wastewater treatment processes was of key importance to monitoring and controlling pollution emission. Concerning the strong correlations among the input variables and the complicated characteristics of wastewater treatment processes in paper-making industry, partial least squares (PLS) method was applied to extract information of variables importance in projection (VIP) for variable selection (VS). Then the optimal variables were chosen as new input variables for soft sensor models to predict the effluent qualities of a papermaking wastewater treatment process. Compared to the LSSVM model, the root mean square error (RMSE) of VS-based LSSVM model was reduced by 15.2%, and the correlation coefficient (r) was increased by 14.4%. For the effluent SS, the value of RMSE was decreased by 20.5%, and the value of r was increased by 16.1%. The results showed that the proposed method not only reduced the model complexity, but also enhanced the model generalization capacity.

Key words: wastewater treatment, effluent quality, variable selection, variables importance in projection, partial least squares

中图分类号: 

  • X703

图1

辅助变量的VIP得分"

图2

逐步回归系数"

图3

出水SS模型预测结果"

表1

模型的预测结果"

模型出水COD出水SS
RMSEMAPE/%rRMSEMAPE/%r
BPNN训练集5.196 14.370 50.807 90.698 32.415 80.832 3
测试集5.523 66.373 70.660 40.714 82.610 80.715 7
VS-BPNN训练集4.396 44.553 10.860 40.632 62.259 30.864 7
测试集4.404 15.324 30.797 30.649 72.270 30.801 6
LSSVM训练集4.324 84.529 50.883 60.723 42.580 60.829 1
测试集4.948 15.740 40.697 60.758 42.680 70.701 6
VS-LSSVM训练集4.003 94.169 70.890 50.596 62.134 50.889 5
测试集4.194 64.800 00.798 40.603 32.167 60.814 3

图4

出水SS模型预测结果"

图5

出水COD模型预测结果"

图6

出水COD模型预测结果"

1 OLSSON G , NEWELL B . Wastewater treatment systems, modeling, diagnosis and control[M]. Britain, London: IWA Publishing, 1999.
2 SHEN Wenhao , CHEN Xiaoquan , CORRIOU J P . Application of model predictive control to the BSM1 benchmark of wastewater treatment process[J]. Computers & Chemical Engineering, 2008, 32 (12): 2849- 2856.
3 黄银蓉, 张绍德. MIMO最小二乘支持向量机污水处理在线软测量研究[J]. 自动化与仪器仪表, 2010, (4): 15- 17.
doi: 10.3969/j.issn.1001-9227.2010.04.006
HUANG Yinrong , ZHANG Shaode . Online soft measurement for wastewater treatment based on MIMO least squares support vector machine[J]. Automation & Instrumentation, 2010, (4): 15- 17.
doi: 10.3969/j.issn.1001-9227.2010.04.006
4 王欣, 宋翼颉, 秦斌, 等. 基于LSSVM的污水处理过程建模[J]. 湖南工业大学学报, 2016, 30 (1): 59- 63.
doi: 10.3969/j.issn.1673-9833.2016.01.011
WANG Xin , SONG Yijie , QIN Bin , et al. Modeling of sewage treatment process based on MIMO-LSSVM[J]. Journal of Hunan University of Technology, 2016, 30 (1): 59- 63.
doi: 10.3969/j.issn.1673-9833.2016.01.011
5 汪瑶, 徐亮, 殷文志, 等. 基于ANN和LSSVR的造纸废水处理过程软测量建模[J]. 中国造纸学报, 2017, 32 (1): 50- 54.
WANG Yao , XU Liang , YIN Wenzhi , et al. Soft sensor modeling of papermaking treatment processes based on ANN and LSSVR[J]. Transactions of China Pulp and Paper, 2017, 32 (1): 50- 54.
6 宋留, 杨冲, 张辉, 等. 造纸废水处理过程的高斯过程回归软测量建模[J]. 中国环境科学, 2018, 38 (7): 2564- 2571.
doi: 10.3969/j.issn.1000-6923.2018.07.023
SONG Liu , YANG Chong , ZHANG Hui , et al. Soft-sensor modeling of papermaking wastewater treatment process based on Gaussian process[J]. China Environmental Science, 2018, 38 (7): 2564- 2571.
doi: 10.3969/j.issn.1000-6923.2018.07.023
7 KOHAVI R , JOHN G H . Wrappers for feature subset selection[J]. Artificial Intelligence, 1997, 97 (1-2): 273- 324.
doi: 10.1016/S0004-3702(97)00043-X
8 MALDONADO S , WEBER R , FAMILI F . Feature selection for high-dimensional class-imbalanced data sets using support vector machines[J]. Information Sciences, 2014, 286, 228- 246.
doi: 10.1016/j.ins.2014.07.015
9 MALDONADO S , LÓPEZ J . Dealing with high-dimensional class-imbalanced datasets: embedded feature sele-ction for SVM classification[J]. Applied Soft Com-puting, 2018, 67, 94- 105.
doi: 10.1016/j.asoc.2018.02.051
10 ANDERSEN C M , BRO R . Variable selection in regression: a tutorial[J]. Journal of Chemometrics, 2010, 24 (11/12): 728- 737.
11 SHAHRIARI S , FARIA S , GONÇCALVES A M . Variable selection methods in high-dimensional regre-ssion: a simulation study[J]. Communications in Statistics-Simulation and Computation, 2015, 44 (10): 2548- 2561.
doi: 10.1080/03610918.2013.833231
12 DUPUIS D J , VICTORIA F M . Robust VIF regression with application to variable selection in large data sets[J]. The Annals of Applied Statistics, 2013, 7 (1): 319- 341.
doi: 10.1214/12-AOAS584
13 王树云, 宋云胜. 线性模型下基于AIC准则的Bayes变量选择[J]. 山东大学学报(理学版), 2010, 45 (6): 43- 45.
WANG Shuyun , SONG Yunsheng . Bayesian variable selection based on AIC criteria in linear models[J]. Journal of Shandong University (Natural Science), 2010, 45 (6): 43- 45.
14 杨慧中, 章军, 陶洪峰. 基于互信息的软测量变量选择[J]. 控制工程, 2012, 19 (4): 562- 565.
doi: 10.3969/j.issn.1671-7848.2012.04.004
YANG Huizhong , ZHANG Jun , TAO Hongfeng . A variable selection for soft sensor based on mutual information[J]. Control Engineering of China, 2012, 19 (4): 562- 565.
doi: 10.3969/j.issn.1671-7848.2012.04.004
15 朱群雄, 郎娜. 工业软测量模型结构与输入变量选择的研究[J]. 控制工程, 2011, 18 (3): 388- 392.
doi: 10.3969/j.issn.1671-7848.2011.03.017
ZHU Qunxiong , LANG Na . Research on model structure and input variable selection for industry soft sensing[J]. Control Engineering of China, 2011, 18 (3): 388- 392.
doi: 10.3969/j.issn.1671-7848.2011.03.017
16 WAN J , HUANG M , MA Y , et al. Prediction of effluent quality of a paper mill wastewater treatment using an adaptive network-based fuzzy inference system[J]. Applied Soft Computing, 2011, 11 (3): 3238- 3246.
doi: 10.1016/j.asoc.2010.12.026
17 CHONG I-G , JUN C-H . Performance of some variable selection methods when multicollinearity is present[J]. Chemometrics and Intelligent Laboratory Systems, 2005, 78 (1/2): 103- 112.
18 FARRÉ M , PLATIKANOV S , TSAKOVSKI S , et al. Comparison of the variable importance in projection (VIP) and of the selectivity ratio (SR) methods for variable selection and interpretation[J]. Journal of Chemometrics, 2015, 29 (10): 528- 536.
19 韩放, 吴晶辰, 徐江峰, 等. 利用PLS-VIP方法筛选差异表达基因[J]. 北京大学学报(自然科学版), 2009, 45 (1): 1- 5.
doi: 10.3321/j.issn:0479-8023.2009.01.001
HAN Fang , WU Jingchen , XU Jiangfeng , et al. Searching for differentially expressed genes by PLS-VIP method[J]. Acta Scientiarum Naturalium Universitatis Pekinensis, 2009, 45 (1): 1- 5.
doi: 10.3321/j.issn:0479-8023.2009.01.001
20 刘国海, 程锦翔, 丁煜函, 等. 青霉素发酵过程中基于PLS-VIP的神经网络逆软测量方法[J]. 计算机与应用化学, 2013, 30 (11): 1294- 1298.
doi: 10.3969/j.issn.1001-4160.2013.11.012
LIU Guohai , CHEN Jinxiang , DING Yuhan , et al. Neural network inverse soft-sensing method based on PLS-VIP in penicillin fermentation process[J]. Computers and Applied Chemistry, 2013, 30 (11): 1294- 1298.
doi: 10.3969/j.issn.1001-4160.2013.11.012
21 阎威武, 邵惠鹤. 支持向量机和最小二乘支持向量机的比较及应用研究[J]. 控制与决策, 2003, 18 (3): 358- 360.
doi: 10.3321/j.issn:1001-0920.2003.03.025
YAN Weiwu , SHAO Huihe . Application of support vector machines and least squares support vector machines to heart disease diagnoses[J]. Control and Decision, 2003, 18 (3): 358- 360.
doi: 10.3321/j.issn:1001-0920.2003.03.025
22 丁兰, 张文阳, 张良均, 等. 基于人工神经网络的居民生活垃圾可燃成分热值预测[J]. 环境工程学报, 2016, 10 (2): 899- 905.
DING Lan , ZHANG Wenyang , ZHANG Liangjun , et al. Prediction of household waste combustible component calorific value based on artificial neural network[J]. Chinese Journal of Environmental Engineering, 2016, 10 (2): 899- 905.
23 严文峰, 李晓东, 高智花, 等. 污泥厌氧消化的人工神经网络模型[J]. 环境工程学报, 2015, 9 (5): 2425- 2429.
YAN Wenfeng , LI Xiaodong , GAO Zhihua , et al. Artificial neural network model of sludge anaerobic digestion[J]. Chinese Journal of Environmental Engineering, 2015, 9 (5): 2425- 2429.
24 张恒德, 张庭玉, 李涛, 等. 基于BP神经网络的污染物浓度多模式集成预报[J]. 中国环境科学, 2018, 38 (4): 1243- 1256.
ZHANG Hengde , ZHANG Tingyu , LI Tao , et al. Forecast of air quality pollutants' concentrations based on BP neural network multi-model ensemble method[J]. China Environmental Science, 2018, 38 (4): 1243- 1256.
25 薛美盛, 冀若阳, 王旭. 基于多元线性回归与滚动窗的NOx排放量软测量[J]. 化工自动化及仪表, 2017, 44 (8): 721- 724.
doi: 10.3969/j.issn.1000-3932.2017.08.002
XUE Meisheng , JI Ruoyang , WANG Xu . Soft sensor for NOx emissions based on multiple linear regression and sliding window[J]. Control and Instruments in Chemical Industry, 2017, 44 (8): 721- 724.
doi: 10.3969/j.issn.1000-3932.2017.08.002
26 HWANG J S , HU T H . A stepwise regression algorithm for high-dimensional variable selection[J]. Journal of Statistical Computation & Simulation, 2015, 85 (9): 1793- 1806.
27 MA Mingda , KO J W , WANG S J , et al. Development of adaptive soft sensor based on statistical identification of key variables[J]. Control Engineering Practice, 2008, 17 (9): 1026- 1034.
[1] 刘鸿斌,宋留. 废水处理过程的典型相关分析建模方法研究[J]. 山东大学学报 (工学版), 2020, 50(1): 101-108.
[2] 杨冬璐,马逍天,洪静兰. 基于生命周期评价的造纸废水的水足迹[J]. 山东大学学报 (工学版), 2019, 49(3): 114-119, 128.
[3] 向润,陈素芬,曾雪强. 基于多重多元回归的人脸年龄估计[J]. 山东大学学报 (工学版), 2019, 49(2): 54-60.
[4] 叶晓丰, 王培良, 杨泽宇. 基于混合MPLS的多阶段过程质量预报方法[J]. 山东大学学报(工学版), 2017, 47(5): 246-253.
[5] 方丽英,李爽,王普,陈培煜. 变系数模型在医学纵向数据研究中的应用[J]. 山东大学学报(工学版), 2013, 43(6): 21-26.
[6] 饶敏1,李燕红1,陈伟华1,宋建波2,陈广大1,刘晓作1. 二硝基重氮酚工业废水处理研究[J]. 山东大学学报(工学版), 2012, 42(6): 100-106.
[7] 曾雪强1,李国正2. 基于偏最小二乘降维的分类模型比较[J]. 山东大学学报(工学版), 2010, 40(5): 41-47.
[8] 余锋俊,施来顺*. 二氧化锰催化二氧化氯氧化酸性铬蓝K模拟废水[J]. 山东大学学报(工学版), 2010, 40(2): 88-94.
[9] 董增寿 刘明君. 锅炉氮氧化物排放量检测中PLS的应用改进[J]. 山东大学学报(工学版), 2010, 40(1): 126-128.
[10] 王静,李玉江,张晓瑾, 毕研俊,陈位锁 . 粉煤灰去除水中活性紫KN-B[J]. 山东大学学报(工学版), 2006, 36(6): 100-103 .
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
[1] 张永花,王安玲,刘福平 . 低频非均匀电磁波在导电界面的反射相角[J]. 山东大学学报(工学版), 2006, 36(2): 22 -25 .
[2] 李梁,罗奇鸣,陈恩红. 对象级搜索中基于图的对象排序模型(英文)[J]. 山东大学学报(工学版), 2009, 39(1): 15 -21 .
[3] 王波,王宁生 . 机电装配体拆卸序列的自动生成及组合优化[J]. 山东大学学报(工学版), 2006, 36(2): 52 -57 .
[4] 秦通,孙丰荣*,王丽梅,王庆浩,李新彩. 基于极大圆盘引导的形状插值实现三维表面重建[J]. 山东大学学报(工学版), 2010, 40(3): 1 -5 .
[5] 刘文亮,朱维红,陈涤,张泓泉. 基于雷达图像的运动目标形态检测及跟踪技术[J]. 山东大学学报(工学版), 2010, 40(3): 31 -36 .
[6] 岳远征. 远离平衡态玻璃的弛豫[J]. 山东大学学报(工学版), 2009, 39(5): 1 -20 .
[7] 王,张艳宁,申家振,刘俊成 . 基于信息测度和支持向量机的图像边缘检测[J]. 山东大学学报(工学版), 2006, 36(3): 95 -99 .
[8] 李芳佳, 高尚策, 唐政, 石井雅博, 山下和也. 基于元胞自动化模型的三维雪花晶体近似模式的产生(英文)[J]. 山东大学学报(工学版), 2009, 39(1): 102 -105 .
[9] 程代展,李志强. 非线性系统线性化综述(英文)[J]. 山东大学学报(工学版), 2009, 39(2): 26 -36 .
[10] 曲延鹏,陈颂英,李春峰,王小鹏,滕书格 . 低压大流量自激脉冲清洗喷嘴内部气液两相流数值模拟[J]. 山东大学学报(工学版), 2006, 36(4): 16 -20 .