基于SOM-RBF神经网络的COD软测量方法

doi:10.11949/0438-1157.20190122

摘要/Abstract

摘要：

污水处理是一个复杂的非线性过程，化学需氧量（chemical oxygen demand，COD）是评价污水处理效果的关键指标之一。COD的传统测量方法耗时长、成本高，基于传统神经网络的软测量方法提高了COD参数的测量速度但精度较差。针对这些问题，设计一种结合自组织特征映射 (self-organizing map, SOM)和径向基函数(radial basis function, RBF)神经网络的COD参数软测量方法。该方法利用SOM网络聚类数据样本，根据所得聚类结果确定RBF网络的隐层节点数及节点的数据中心，综合提高RBF网络的收敛速度和拟合精度。利用污水处理厂部分水样数据建立COD软测量模型，模型仿真和硬件在线测试结果表明，相对于传统的BP、RBF等网络，基于SOM-RBF神经网络的COD软测量方法测量时间短、预测精度较高，具有较为广阔的应用前景。

关键词: 化学需氧量, 软测量, 自组织特征映射, 径向基函数网络, 神经网络, 模型, 预测

Abstract:

Sewage treatment is a complex nonlinear process, and chemical oxygen demand (COD) is one of the key indicators for evaluating the effectiveness of wastewater treatment. It is costly and time-consuming to get COD by traditional chemical approaches. By neural networks, it is faster, but it is not accurate enough. To address them, a soft sensor approach, which is based on the combination of self-organizing map (SOM) and radial basis function (RBF) neural network, is designed. SOM is taken to cluster data samples. The number of hidden layer nodes and the center vector of the nodes are determined by clustering results. By such disposal, the rate of convergence and fitting precision have been improved. Part data of water samples from a sewage treatment plant are taken to establish the soft sensor model of COD. Test results provided by numerical model and hardware show that, compared with the traditional BP, RBF and other networks, the soft sensor model of COD designed in this paper has short measurement time and relatively high prediction accuracy. It may be a promising soft sensor approach in applications.

Key words: chemical oxygen demand, soft sensor, self-organizing map, radial basis function, neural network, model, prediction

中图分类号:

TP 183

廉小亲, 王俐伟, 安飒, 魏伟, 刘载文. 基于SOM-RBF神经网络的COD软测量方法[J]. 化工学报, 2019, 70(9): 3465-3472.

Xiaoqin LIAN, Liwei WANG, Sa AN, Wei WEI, Zaiwen LIU. On soft sensor of chemical oxygen demand by SOM-RBF neural network[J]. CIESC Journal, 2019, 70(9): 3465-3472.

图/表 11

图1 RBF网络结构拓扑

Fig.1 RBF network structure topology

图2 SOM神经网络结构

Fig.2 SOM neural network structure

表1 部分实测数据

Table 1 Part of measured data

模型输入				模型输出
氧化还原电位（ORP）/mV	浊度（SS）/ NTU	流量(Q _in)/ (m³·h)	溶解氧（DO）/ (mg·L^-1)	出水化学需氧量（COD）/(mg·L^-1)
-68.00	65.00	1544.00	3.64	80.50
-56.00	105.44	2001.00	3.67	63.40
-56.00	115.00	2423.00	3.43	59.80
-90.00	120.40	2358.00	3.86	56.10
-71.00	65.00	1742.00	3.83	63.40
-81.00	105.44	1883.00	4.79	43.90
-93.00	115.00	2293.00	3.32	43.90
-91.00	120.39	1915.00	3.40	87.00
-84.00	137.22	1907.00	3.85	34.20
-59.00	138.00	1914.00	3.43	53.70

图3 COD训练样本预期输出与实际输出

Fig.3 Expected output and actual output of COD training sample

图4 COD训练样本相对误差曲线

Fig.4 Relative error curve of COD training sample

图5 COD测试样本预期输出与实际输出

Fig.5 Expected output and actual output of COD test sample

图6 COD测试样本相对误差曲线

Fig.6 Relative error curve of COD test sample

表2 四种模型预测结果对比

Table 2 Comparison of prediction results of four models

算法	训练时间/s	训练误差绝对值平均/(mg·L^-1)	测试误差绝对值平均/(mg·L^-1)	均方误差	训练相对误差/%
BP	93.32	1.15	1.23	2.10	12
RBF	8.62	0.78	0.89	1.97	10
K-means-RBF	5.30	—	0.52	0.55	—
SOM-RBF	5.35	0.27	0.36	0.23	3

图7 英伟达Jetson TX2

Fig.7 NVIDIA Jetson TX2

图8 软测量仪表运行界面

Fig.8 Soft measuring instrument running interface

图9 软测量仪表COD预测值与实际值

Fig.9 Predicted value of soft measurement instrument and actual value of COD

参考文献 30

1	穆瑞, 乐高杨, 杨慧中 . 基于O₃/UV法在线COD检测的气体溶解量估计方法[J]. 化工学报, 2019, 70(2): 730-735.
	Mu R , Le G Y , Yang H Z . Estimation method of gas dissolved amount based on O₃/UV method for online COD detection [J]. CIESC Journal, 2019, 70(2): 730-735.
2	宋建军, 赵凌 . 紫外吸收光谱法检测化学需氧量的方法[J]. 传感器与微系统, 2018, 37(5): 30-33.
	Song J J , Zhao L . Methods for determination of chemical oxygen demand by ultraviolet absorption spectroscopy[J]. Transducer and Microsystems, 2018, 37(5): 30-33.
3	李名升, 张建辉, 罗海江, 等 . “十一五”期间中国化学需氧量减排与水环境质量变化关联分析[J]. 生态环境学报, 2011, 20(3): 463-467.
	Li M S , Zhang J H , Luo H J , et al . Relation between water quality and COD discharge in China in the Eleventh Five-Year Plan[J]. Ecology and Environmental Sciences, 2011, 20(3): 463-467.
4	Hanbay D , Turkoglu I , Demir Y . Prediction of chemical oxygen demand (COD) based on wavelet decomposition and neural networks[J]. CLEAN - Soil, Air, Water, 2010, 35(3): 250-254.
5	Li Q , Peng J , Tang D , et al . A study on the UKFNN-based online detection of effluent COD in water sewage treatment[C]// Industrial Electronics & Applications. IEEE, 2015.
6	李雪梅, 朱燕 . 快速消解分光光度法测定污水化学需氧量[J]. 化学分析计量, 2018, 27(3): 36-39.
	Li X M , Zhu Y . Determination of chemical oxygen demand in sewage by rapid digestion spectrophotometry[J]. Chemical Analysis & Metering, 2018, 27(3): 36-39.
7	孙立岩, 姚志鹏, 张军, 等 . 地表水中TOC与COD换算关系研究[J]. 中国环境监测, 2013, 29(2): 125-130.
	Sun L Y , Yao Z P , Zhang J , et al . Study on the relationship between TOC and COD conversion in surface water [J]. Environmental Monitoring in China, 2013, 29(2): 125-130.
8	Zhao H , Jiang D , Zhang S , et al . Development of a direct photoelectrochemical method for determination of chemical oxygen demand[J]. Analytical Chemistry, 2015, 76(1): 155-160.
9	肖峻, 杨孟孟, 缪震华, 等 . 嵌入式污水COD在线监测仪的研制与应用[J]. 环境工程学报, 2017, 11(5): 3312-3319.
	Xiao J , Yang M M , Miu Z H , et al . Development and application of embedded COD online monitoring instrument for sewage [J]. Journal of Environmental Engineering, 2017, 11(5): 3312-3319.
10	Ye Y , Zhuang Y . The COD predictive technique based on neural network[C]// International Conference on Mechatronic Sciences. IEEE, 2014.
11	曹为阳 . 污水处理出水COD软测量预测建模方法的研究[D]. 马鞍山: 安徽工业大学, 2017.
	Cao W Y . Research on COD soft measurement prediction modeling method for wastewater treatment effluent [D]. Ma’anshan: Anhui University of Technology, 2017.
12	Azadi S , Amiri H , Rakhshandehroo G R . Evaluating the ability of artificial neural network and PCA-M5P models in predicting leachate COD load in landfills[J]. Waste Management, 2016, 55: 220-230.
13	Qiao J , Li W , Han H . Soft computing of biochemical oxygen demand using an improved T–S fuzzy neural network[J]. Chinese Journal of Chemical Engineering, 2014, 22(11/12): 1254-1259.
14	Li Y H , Shi Y T , Wang K , et al . Design of online monitoring device for COD parameter in industrial sewage based on soft measurement method[C]// Automation. IEEE, 2017: 959-964.
15	乔俊飞, 郭楠, 韩红桂 . 基于神经网络的BOD参数软测量仪表的设计[J]. 计算机与应用化学, 2013, 30(10): 1219-1222.
	Qiao J F , Guo N , Han H G . Design of soft measuring instrument for BOD parameters based on neural network[J]. Computer and Applied Chemistry, 2013, 30(10): 1219-1222.
16	任东红 . 基于集成神经网络的污水处理出水水质软测量模型[D]. 北京: 北京工业大学, 2013.
	Ren D H . Soft sensor model for wastewater treatment effluent based on integrated neural network [D]. Beijing: Beijing University of Technology, 2013.
17	郭楠 . 基于神经网络的BOD软测量仪表的研究[D]. 北京: 北京工业大学, 2014.
	Guo N . Research on BOD soft measuring instrument based on neural network [D]. Beijing: Beijing University of Technology, 2014.
18	乔俊飞, 安茹, 韩红桂 . 基于RBF神经网络的出水氨氮预测研究[J]. 控制工程, 2016, 23(9): 1301-1305.
	Qiao J F , An R , Han H G . Prediction of effluent ammonia nitrogen based on RBF neural network[J]. Control Engineering, 2016, 23(9): 1301-1305.
19	崔雪梅 . 基于灰色GA-LM-BP模型的COD_(Mn)预测[J]. 水利水电科技进展, 2013, 33(5): 38-41.
	Cui X M . Prediction of COD_(Mn) based on grey GA-LM-BP model[J]. Advances in Science and Technology of Water Resources, 2013, 33(5): 38-41.
20	高峰, 冯民权, 滕素芬 . 基于PSO优化BP神经网络的水质预测研究[J]. 安全与环境学报, 2015, 15(4): 338-341.
	Gao F , Feng M Q , Teng S F . Water quality prediction based on BP neural network optimized by PSO[J]. Journal of Safety and Environment, 2015, 15(4): 338-341.
21	Ye G Q , Li W G , Wan H . Study of RBF neural network based on PSO algorithm in nonlinear system identification[C]// International Conference on Intelligent Computation Technology & Automation. IEEE, 2016.
22	乔俊飞, 孙玉庆, 韩红桂 . 改进K-means算法优化RBF神经网络的出水氨氮预测[J]. 控制工程, 2018, 25(3): 375-379.
	Qiao J F , Sun Y Q , Han H G . Improved K-means algorithm for optimization of effluent ammonia nitrogen prediction in RBF neural network[J]. Control Engineering, 2018, 25(3): 375-379.
23	刘博, 肖长来, 梁秀娟 . SOM-RBF神经网络模型在地下水位预测中的应用[J]. 吉林大学学报(地球科学版), 2015, 45(1): 225-231.
	Liu B , Xiao C L , Liang X J . Application of SOM-RBF neural network model in groundwater level prediction[J]. Journal of Jilin University (Earth Science Edition), 2015, 45(1): 225-231.
24	乔俊飞, 马士杰, 杨翠丽 . 基于ROLS算法的递归RBF神经网络结构设计[J]. 化工学报, 2018, 69(3): 1191-1199.
	Qiao J F , Ma S J , Yang C L . Structural design of recursive RBF neural network based on ROLS algorithm[J]. CIESC Journal, 2018, 69(3): 1191-1199.
25	李文静, 李萌, 乔俊飞 . 基于互信息和自组织RBF神经网络的出水BOD软测量方法[J]. 化工学报, 2019, 70(2): 687-695.
	Li W J , Li M , Qiao J F . Soft measurement method of effluent BOD based on mutual information and self-organizing RBF neural network [J]. CIESC Journal, 2019, 70(2): 687-695.
26	李晋贤 . 基于HS-RBF的污水处理出水COD软测量研究[D]. 绵阳: 西南科技大学, 2017.
	Li J X . Soft-measurement of COD based on HS-RBF for wastewater treatment [D]. Mianyang: Southwest University of Science and Technology, 2017.
27	韩红桂, 林征来, 乔俊飞 . 一种基于混合梯度下降算法的模糊神经网络设计及应用[J]. 控制与决策, 2017, 32(9): 1635-1641.
	Han H G , Lin Z L , Qiao J F . Design and application of a fuzzy neural network based on hybrid gradient descent algorithm [J]. Control and Decision-making, 2017, 32(9): 1635-1641.
28	孙娅楠 . 梯度下降法在机器学习中的应用[D]. 成都: 西南交通大学, 2018.
	Sun Y N . Application of gradient descent method in machine learning [D]. Chengdu: Southwest Jiaotong University, 2018.
29	周萌, 王振华, 王昶, 等 . Lipschitz非线性系统的H_/L _∞故障检测观测器设计[J]. 控制理论与应用, 2018, 35(6): 778-785.
	Zhou M , Wang Z H , Wang C , et al . H_/L _∞ fault detection observer design for Lipschitz nonlinear systems [J]. Control Theory and Applications, 2008, 35(6): 778-785.
30	Jakovetić D , Xavier J , Moura J M F . Convergence rate analysis of distributed gradient methods for smooth optimization[C]// Telecommunications Forum. IEEE, 2013: 867-870.

编辑推荐

Metrics

阅读次数

全文

336

HTML			PDF

最新录用	在线预览	正式出版	最新录用	在线预览	正式出版
0	0	5	8	0	323

来源	本网站	其他网站

次数	161	175
比例	48%	52%

摘要

[1]	宋嘉豪, 王文. 斯特林发动机与高温热管耦合运行特性研究[J]. 化工学报, 2023, 74(S1): 287-294.
[2]	连梦雅, 谈莹莹, 王林, 陈枫, 曹艺飞. 地下水预热新风一体化热泵空调系统制热性能研究[J]. 化工学报, 2023, 74(S1): 311-319.
[3]	金正浩, 封立杰, 李舒宏. 氨水溶液交叉型再吸收式热泵的能量及分析[J]. 化工学报, 2023, 74(S1): 53-63.
[4]	温凯杰, 郭力, 夏诏杰, 陈建华. 一种耦合CFD与深度学习的气固快速模拟方法[J]. 化工学报, 2023, 74(9): 3775-3785.
[5]	李科, 文键, 忻碧平. 耦合蒸气冷却屏的真空多层绝热结构对液氢储罐自增压过程的影响机制研究[J]. 化工学报, 2023, 74(9): 3786-3796.
[6]	王浩, 王振雷. 基于自适应谱方法的裂解炉烧焦模型化简策略[J]. 化工学报, 2023, 74(9): 3855-3864.
[7]	尹刚, 李伊惠, 何飞, 曹文琦, 王民, 颜非亚, 向禹, 卢剑, 罗斌, 卢润廷. 基于KPCA和SVM的铝电解槽漏槽事故预警方法[J]. 化工学报, 2023, 74(8): 3419-3428.
[8]	李锦潼, 邱顺, 孙文寿. 煤浆法烟气脱硫中草酸和紫外线强化煤砷浸出过程[J]. 化工学报, 2023, 74(8): 3522-3532.
[9]	于旭东, 李琪, 陈念粗, 杜理, 任思颖, 曾英. 三元体系KCl + CaCl₂ + H₂O 298.2、323.2及348.2 K相平衡研究及计算[J]. 化工学报, 2023, 74(8): 3256-3265.
[10]	诸程瑛, 王振雷. 基于改进深度强化学习的乙烯裂解炉操作优化[J]. 化工学报, 2023, 74(8): 3429-3437.
[11]	闫琳琦, 王振雷. 基于STA-BiLSTM-LightGBM组合模型的多步预测软测量建模[J]. 化工学报, 2023, 74(8): 3407-3418.
[12]	徐野, 黄文君, 米俊芃, 申川川, 金建祥. 多源信息融合的离心式压缩机喘振诊断方法[J]. 化工学报, 2023, 74(7): 2979-2987.
[13]	郭雨莹, 敬加强, 黄婉妮, 张平, 孙杰, 朱宇, 冯君炫, 陆洪江. 稠油管道水润滑减阻及压降预测模型修正[J]. 化工学报, 2023, 74(7): 2898-2907.
[14]	刘春雨, 周桓宇, 马跃, 岳长涛. CaO调质含油污泥干燥特性及数学模型[J]. 化工学报, 2023, 74(7): 3018-3027.
[15]	邵伟明, 韩文学, 宋伟, 杨勇, 陈灿, 赵东亚. 基于分布式贝叶斯隐马尔可夫回归的动态软测量建模方法[J]. 化工学报, 2023, 74(6): 2495-2502.