CN110647037B - 一种基于二型模糊神经网络的污水处理过程协同控制方法 - Google Patents

一种基于二型模糊神经网络的污水处理过程协同控制方法 Download PDF

Info

Publication number
CN110647037B
CN110647037B CN201910899611.7A CN201910899611A CN110647037B CN 110647037 B CN110647037 B CN 110647037B CN 201910899611 A CN201910899611 A CN 201910899611A CN 110647037 B CN110647037 B CN 110647037B
Authority
CN
China
Prior art keywords
time
concentration
layer
output
dissolved oxygen
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201910899611.7A
Other languages
English (en)
Other versions
CN110647037A (zh
Inventor
韩红桂
李嘉明
伍小龙
乔俊飞
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing University of Technology
Original Assignee
Beijing University of Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Beijing University of Technology filed Critical Beijing University of Technology
Priority to CN201910899611.7A priority Critical patent/CN110647037B/zh
Priority to US16/694,911 priority patent/US11530139B2/en
Publication of CN110647037A publication Critical patent/CN110647037A/zh
Application granted granted Critical
Publication of CN110647037B publication Critical patent/CN110647037B/zh
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • CCHEMISTRY; METALLURGY
    • C02TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
    • C02FTREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
    • C02F1/00Treatment of water, waste water, or sewage
    • C02F1/20Treatment of water, waste water, or sewage by degassing, i.e. liberation of dissolved gases
    • GPHYSICS
    • G05CONTROLLING; REGULATING
    • G05BCONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B13/00Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
    • G05B13/02Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
    • G05B13/04Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
    • G05B13/042Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators in which a parameter or coefficient is automatically adjusted to optimise the performance
    • CCHEMISTRY; METALLURGY
    • C02TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
    • C02FTREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
    • C02F3/00Biological treatment of water, waste water, or sewage
    • C02F3/006Regulation methods for biological treatment
    • CCHEMISTRY; METALLURGY
    • C02TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
    • C02FTREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
    • C02F1/00Treatment of water, waste water, or sewage
    • C02F1/58Treatment of water, waste water, or sewage by removing specified dissolved compounds
    • C02F1/586Treatment of water, waste water, or sewage by removing specified dissolved compounds by removing ammoniacal nitrogen
    • CCHEMISTRY; METALLURGY
    • C02TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
    • C02FTREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
    • C02F3/00Biological treatment of water, waste water, or sewage
    • C02F3/30Aerobic and anaerobic processes
    • C02F3/302Nitrification and denitrification treatment
    • C02F3/305Nitrification and denitrification treatment characterised by the denitrification
    • GPHYSICS
    • G05CONTROLLING; REGULATING
    • G05BCONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B13/00Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
    • G05B13/02Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
    • G05B13/0265Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric the criterion being a learning criterion
    • G05B13/0285Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric the criterion being a learning criterion using neural networks and fuzzy logic
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10Complex mathematical operations
    • G06F17/16Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10Complex mathematical operations
    • G06F17/18Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00Machine learning
    • G06N20/10Machine learning using kernel methods, e.g. support vector machines [SVM]
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks
    • G06N3/04Architecture, e.g. interconnection topology
    • G06N3/043Architecture, e.g. interconnection topology based on fuzzy logic, fuzzy membership or fuzzy inference, e.g. adaptive neuro-fuzzy inference systems [ANFIS]
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks
    • G06N3/08Learning methods
    • CCHEMISTRY; METALLURGY
    • C02TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
    • C02FTREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
    • C02F2101/00Nature of the contaminant
    • C02F2101/10Inorganic compounds
    • C02F2101/16Nitrogen compounds, e.g. ammonia
    • C02F2101/163Nitrates
    • CCHEMISTRY; METALLURGY
    • C02TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
    • C02FTREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
    • C02F2209/00Controlling or monitoring parameters in water treatment
    • C02F2209/005Processes using a programmable logic controller [PLC]
    • C02F2209/006Processes using a programmable logic controller [PLC] comprising a software program or a logic diagram
    • CCHEMISTRY; METALLURGY
    • C02TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
    • C02FTREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
    • C02F2209/00Controlling or monitoring parameters in water treatment
    • C02F2209/15N03-N
    • CCHEMISTRY; METALLURGY
    • C02TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
    • C02FTREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
    • C02F2209/00Controlling or monitoring parameters in water treatment
    • C02F2209/22O2

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Software Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Health & Medical Sciences (AREA)
  • General Engineering & Computer Science (AREA)
  • Pure & Applied Mathematics (AREA)
  • Mathematical Optimization (AREA)
  • Mathematical Analysis (AREA)
  • Computational Mathematics (AREA)
  • Computing Systems (AREA)
  • Environmental & Geological Engineering (AREA)
  • Chemical & Material Sciences (AREA)
  • Molecular Biology (AREA)
  • Hydrology & Water Resources (AREA)
  • Organic Chemistry (AREA)
  • Water Supply & Treatment (AREA)
  • Automation & Control Theory (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Medical Informatics (AREA)
  • Biophysics (AREA)
  • General Health & Medical Sciences (AREA)
  • Fuzzy Systems (AREA)
  • Computational Linguistics (AREA)
  • Biomedical Technology (AREA)
  • Algebra (AREA)
  • Databases & Information Systems (AREA)
  • Biodiversity & Conservation Biology (AREA)
  • Microbiology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Operations Research (AREA)
  • Probability & Statistics with Applications (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Computational Biology (AREA)

Abstract

本发明提出了一种基于二型模糊神经网络的污水处理过程协同控制方法,针对污水处理过程难以建立精确的数学模型,污水处理过程具有较强的非线性和不确定性,溶解氧DO浓度和硝态氮NO3‑N浓度难以有效控制的特点,实现污水处理过程中溶解氧DO浓度和硝态氮NO3‑N浓度的协同控制。该控制方法利用二型模糊神经网络,建立协同模糊神经控制器,搭建协同模糊神经控制器和控制对象的回路,利用协同模糊神经控制方法对溶解氧DO浓度和硝态氮NO3‑N浓度进行控制,能够在不同运行工况下对溶解氧DO浓度和硝态氮NO3‑N浓度进行快速、准确的控制,提高了污水处理过程在不同工况下的运行性能,实现了令人满意的控制精度。

Description

一种基于二型模糊神经网络的污水处理过程协同控制方法
技术领域
本发明利用基于二型模糊神经网络的协同控制方法实现污水处理过程中溶解氧DO浓度和硝态氮NO3-N浓度协同控制,溶解氧DO浓度和硝态氮NO3-N 浓度是污水生化反应过程中关键控制参数,对污水处理过程、出水水质,能耗都有着重要影响;将基于二型模糊神经网络的协同控制方法应用在污水处理过程中,实现溶解氧DO浓度和硝态氮NO3-N浓度的协同控制;保证污水处理过程在多种运行工况下的稳定运行,既属于水研究领域,又属于智能控制领域;
背景技术
城镇人口的快速增加与国家环保标准的日趋严格使得国家环保部门对城市污水处理厂的出水水质提出了更高要求;我国生活用水总量逐年攀升,至2018 年底已达850亿立方米,这对城市污水处理能力构成极大的挑战;同时,越来越严格的排放标准和多变的进水负荷也对污水处理过程的平稳运行构成威胁;
作为污水处理过程中关键过程变量,溶解氧DO浓度和硝态氮NO3-N浓度对于污水处理过程中的生化反应过程起着直接控制作用;在污水处理厂中,活性污泥法是处理污水最常用的方法,其中涉及的生化反应包括氨化反应、硝化反应、反硝化反应等可以通过降解有机物来去除废水中的污染物;氨化反应可以将有机氮化合物转化为氨氮,硝化反应可将氨氮合成为硝态氮NO3-N;同时,通过反硝化反应可以将硝态氮NO3-N分解为氮气;尽管这些生化反应能有效地去除有机物,但这些反应过程很难得到有效控制;硝化反应是一种需氧反应,在硝化反应中,较高的溶解氧DO浓度会促进氨氮的降解;与此相反,反硝化反应需要缺氧环境,在缺氧条件下反硝化反应可以去除绝大多数硝态氮NO3-N;由于这些相互矛盾的反应条件,当前的传统PID控制方法很难有效地同时控制硝化反应和反硝化反应,并且污水处理过程中存在大量的干扰和不确定性也增加了控制难度;此外,过高的溶解氧DO浓度会抑制硝化反应和反硝化反应,过低的硝态氮 NO3-N浓度则会降低反硝化反应的脱氮效果;因此,对溶解氧DO浓度和硝态氮 NO3-N浓度进行协同控制是很有必要的;二型模糊神经网络通过利用二型模糊规则可以很好的表达具有强非线性和高不确定性的污水处理过程中过程变量的变化,同时具有较强的学习能力和自适应能力;基于二型模糊神经网络的污水处理过程协同控制具有强大的鲁棒性和较高的控制精度,同时通过对二型模糊神经网络的全局参数和局部参数的协同优化,提高了控制响应速度,能够实现对溶解氧 DO浓度和硝态氮NO3-N浓度的协同控制,提高污水处理过程在多种工况下的稳定运行能力,保证了出水水质的实时达标,具有很好的实际应用价值;
本发明设计了一种基于二型模糊神经网络的污水处理过程协同控制方法,主要通过二型模糊神经网络控制器对控制目标进行控制,实现溶解氧DO浓度和硝态氮NO3-N浓度的协同控制;
发明内容
本发明获得了一种基于二型模糊神经网络的污水处理过程协同控制方法,通过基于二型模糊神经网络的污水处理过程协同控制方法求解的溶解氧DO浓度和硝态氮NO3-N浓度的控制量,利用得到的曝气控制量和内回流控制量实现对溶解氧DO浓度和硝态氮NO3-N浓度的协同控制,在多种运行工况下,保证出水水质实时达标,降低能耗和提高污水处理过程运行稳定性;
本发明采用了如下的技术方案及实现步骤:
1.一种基于二型模糊神经网络的污水处理过程协同控制方法,
针对污水处理过程中溶解氧DO浓度和硝态氮NO3-N浓度进行控制,其中,以曝气量和内回流量为控制量,溶解氧DO浓度和硝态氮NO3-N浓度为被控量;
其特征在于,包括以下步骤:
(1)设计用于溶解氧DO浓度和硝态氮NO3-N浓度控制的二型模糊神经网络,二型模糊神经网络分为五层:输入层、隶属函数层、规则层、后件层、输出层;具体为:
①输入层:该层由4个输入神经元组成:
X(t)=[x1(t),x2(t),x3(t),x4(t)]T (1)
其中,X(t)表示二型模糊神经网络的输入向量,x1(t)为t时刻溶解氧DO浓度设定值与实际测量值的误差,x2(t)为t时刻溶解氧DO浓度设定值与实际测量值误差的变化率,x3(t)为t时刻硝态氮NO3-N浓度设定值与实际测量值的误差,x4(t) 为t时刻硝态氮NO3-N浓度设定值与实际测量值误差的变化率,T为矩阵的转置;
②隶属函数层:该层有4×M个神经元,每个神经元代表一个二型隶属函数,表示如下:
Figure RE-GDA0002268425240000021
Figure RE-GDA0002268425240000031
其中,M为规则层神经元的总个数,1<M≤20;m ij(t)表示t时刻关于第i个输入和第j个规则层神经元的隶属函数层神经元的输出值下界,
Figure RE-GDA0002268425240000032
表示t时刻关于第i个输入和第j个规则层神经元的隶属函数层神经元的输出值上界,
Figure RE-GDA00022684252400000315
e为自然常数,e=2.7183;c ij(t)为t时刻关于第i个输入和第j 个规则层神经元的隶属函数层神经元的中心值下界,
Figure RE-GDA0002268425240000033
为t时刻第i个输入和第 j个规则层神经元的隶属函数层神经元的中心值上界,
Figure RE-GDA0002268425240000034
σij(t)为t 时刻关于第i个输入和第j个规则层神经元的隶属函数层神经元的宽度值;i表示二型模糊神经网络的输入个数,j表示二型模糊神经网络的规则层神经元的个数, i=1,2,3,4;j=1,2,…,M;
③规则层:该层有M个神经元,每个神经元的输出为:
Figure RE-GDA0002268425240000035
Figure RE-GDA0002268425240000036
其中,Fj(t)为t时刻第j个规则层神经元的激活强度,f j(t)为t时刻第j个规则层神经元的激活强度下界,
Figure RE-GDA0002268425240000037
为t时刻第j个规则层神经元的激活强度上界,
Figure RE-GDA0002268425240000038
Figure RE-GDA0002268425240000039
④后件层:该层有4个神经元,每个神经元的输出为:
Figure RE-GDA00022684252400000310
其中,y k(t)为t时刻关于第k个输出层神经元的后件层神经元的输出下界,
Figure RE-GDA00022684252400000311
为t时刻关于第k个输出层神经元的后件层神经元的输出上界;
Figure RE-GDA00022684252400000312
为t时刻关于第j个规则层神经元与第k个输出层神经元的后件权值,
Figure RE-GDA00022684252400000313
为t时刻关于第 i个输入、第j个规则层神经元和第k个输出神经元的权值系数,
Figure RE-GDA00022684252400000314
为t时刻关于第k个输出层神经元和第j个规则层神经元的权值偏置,k=1,2;
⑤输出层:该层有2个神经元,每个输出神经元的输出为:
Figure RE-GDA0002268425240000041
其中,Δuk(t)为t时刻二型模糊神经网络的第k个输出值,qk(t)为t时刻关于第k个输出层神经元的后件层神经元输出下界的比例值,0<qk(t)<1;
(2)训练二型模糊神经网络,具体如下:
①将参数分为全局参数和局部参数;定义全局参数向量和局部参数向量为:
Figure RE-GDA0002268425240000042
其中,Φg(t)为t时刻全局参数向量,Φl(t)为t时刻局部参数向量;定义目标函数为:
Figure RE-GDA0002268425240000043
其中,l(t)为t时刻的综合误差,α(t)为t时刻的误差系数,l1(t)为t时刻溶解氧DO浓度设定值与实际值的误差,l2(t)为t时刻硝态氮NO3-N浓度设定值与实际值的误差,
Figure RE-GDA0002268425240000044
为t时刻溶解氧DO浓度的设定值,y1(t)为t时刻溶解氧DO浓度的实际值,
Figure RE-GDA0002268425240000045
为t时刻硝态氮NO3-N浓度的设定值,y2(t)为t时刻硝态氮NO3-N浓度的实际值;
②利用综合误差和自适应二阶算法以协同优化二型模糊神经网络的全局参数和局部参数;定义参数更新公式为:
Figure RE-GDA0002268425240000046
其中,Φ(t+1)为t+1时刻的参数向量,Φ(t)为t时刻的参数向量,I为单位矩阵,Gl(t)为t时刻的误差梯度向量,θ(t)为t时刻的自适应学习率,θ(t)∈(0,1],J(t) 为t时刻的雅克比向量,
Figure RE-GDA0002268425240000047
为t时刻综合误差关于全局参数的偏导数,
Figure RE-GDA0002268425240000048
为t时刻综合误差关于局部参数的偏导数,
Figure RE-GDA0002268425240000049
为t时刻综合误差关于不确定中心下界的偏导数,
Figure RE-GDA00022684252400000410
为t时刻综合误差关于不确定中心上界的偏导数,
Figure RE-GDA0002268425240000051
为t时刻综合误差关于宽度值的偏导数,
Figure RE-GDA0002268425240000052
为t 时刻综合误差关于第k个输出的权值系数的偏导数,
Figure RE-GDA0002268425240000053
为t时刻综合误差关于第k个输出的权值偏置的偏导数,
Figure RE-GDA0002268425240000054
为t时刻综合误差关于第k个输出的比例值的偏导数;
(3)设计用于污水处理过程中溶解氧DO浓度和硝态氮NO3-N浓度的协同控制方法,具体为:
①根据公式(7)计算二型模糊神经网络的输出;
②判断当前时刻综合误差的大小,如果l(t)>0.01,转到步骤③;如果l(t)≤0.01,转到步骤④;
③根据公式(10)求解各个参数的更新值;
④计算当前时刻控制器的输出值uk(t)
uk(t)=uk(t-1)+Δuk(t) (11)
其中,uk(t)为t时刻控制器的第k个输出,u1(t)为t时刻控制器的曝气量输出,u2(t) 为t时刻控制器的内回流量输出,uk(t-1)为t-1时刻控制器的第k个输出;
⑤uk(t)为当前时刻溶解氧DO浓度和硝态氮NO3-N浓度协同控制器的实际输出;转到步骤①;
(4)利用求解出的t时刻控制器的输出值u1(t)和u2(t)对溶解氧DO浓度和硝态氮NO3-N浓度进行控制,u1(t)即t时刻曝气量的控制输入量,u2(t)即t时刻内回流量的控制输入量,控制系统的输出为实际溶解氧DO浓度值和实际硝态氮NO- 3-N浓度值。
本发明的创造性主要体现在:
(1)本发明针对污水处理过程是一个强非线性的复杂工业过程,且具有多种干扰和不确定性的特点;同时,对污水处理过程难以建立精确的数学模型,利用人工神经网络强大的学习能力和任意精度逼近的特点,采用基于二型模糊神经网络的协同控制方法对溶解氧DO浓度和硝态氮NO3-N浓度进行控制,具有控制精度高、控制稳定性强的特点;
(2)本发明采用了基于二型模糊神经网络的污水处理过程协同控制方法对污水处理过程的溶解氧DO浓度和硝态氮NO3-N浓度进行控制,该控制方法充分利用了二型模糊神经网络的鲁棒性和参数的协同优化过程,针对污水处理过程中不同的运行条件均可实现良好的控制效果;解决了污水处理过程不同运行条件下难以实现稳定控制的问题;
特别要注意:本发明只是为了描述方便,采用的是对污水处理过程溶解氧 DO浓度和硝态氮NO3-N浓度进行协同控制,同样该发明也可适用污水处理过程脱氮加药和除磷加药的控制等,只要采用了本发明的原理进行控制都应该属于本发明的范围。
附图说明
图1是本发明的控制结构图
图2是本发明的二型模糊神经网络结构图
图3是本发明的溶解氧DO浓度控制结果图
图4是本发明的溶解氧DO浓度控制结果误差图
图5是本发明的硝态氮NO3-N浓度控制结果图
图6是本发明的硝态氮NO3-N浓度控制结果误差图
具体实施方式
本发明获得了一种基于二型模糊神经网络的污水处理过程协同控制方法,通过基于二型模糊神经网络的污水处理过程协同控制方法求解的溶解氧DO浓度和硝态氮NO3-N浓度的控制量,利用得到的曝气控制量和内回流控制量实现对溶解氧DO浓度和硝态氮NO3-N浓度的协同控制,在多种运行工况下,保证出水水质实时达标,降低能耗和提高污水处理过程运行稳定性;
1.一种基于二型模糊神经网络的污水处理过程协同控制方法,
针对污水处理过程中溶解氧DO浓度和硝态氮NO3-N浓度进行控制,其中,以曝气量和内回流量为控制量,溶解氧DO浓度和硝态氮NO3-N浓度为被控量,控制结构如图1;
其特征在于,包括以下步骤:
(1)设计用于溶解氧DO浓度和硝态氮NO3-N浓度控制的二型模糊神经网络,二型模糊神经网络分为五层:输入层、隶属函数层、规则层、后件层、输出层,二型模糊神经网络结构如图2,具体为:
①输入层:该层由4个输入神经元组成:
X(t)=[x1(t),x2(t),x3(t),x4(t)]T (1)
其中,X(t)表示二型模糊神经网络的输入向量,x1(t)为t时刻溶解氧DO浓度设定值与实际测量值的误差,x2(t)为t时刻溶解氧DO浓度设定值与实际测量值误差的变化率,x3(t)为t时刻硝态氮NO3-N浓度设定值与实际测量值的误差,x4(t) 为t时刻硝态氮NO3-N浓度设定值与实际测量值误差的变化率,T为矩阵的转置;
②隶属函数层:该层有4×M个神经元,每个神经元代表一个二型隶属函数,表示如下:
Figure RE-GDA0002268425240000071
Figure RE-GDA0002268425240000072
其中,M为规则层神经元的总个数,M=10;m ij(t)表示t时刻关于第i个输入和第 j个规则层神经元的隶属函数层神经元的输出值下界,
Figure RE-GDA0002268425240000073
表示t时刻关于第i 个输入和第j个规则层神经元的隶属函数层神经元的输出值上界,
Figure RE-GDA0002268425240000074
e为自然常数,e=2.7183;c ij(t)为t时刻关于第i个输入和第j 个规则层神经元的隶属函数层神经元的中心值下界,
Figure RE-GDA0002268425240000075
为t时刻第i个输入和第 j个规则层神经元的隶属函数层神经元的中心值上界,
Figure RE-GDA0002268425240000076
σij(t)为t 时刻关于第i个输入和第j个规则层神经元的隶属函数层神经元的宽度值;i表示二型模糊神经网络的输入个数,j表示二型模糊神经网络的规则层神经元的个数, i=1,2,3,4;j=1,2,…,M;
③规则层:该层有10个神经元,每个神经元的输出为:
Figure RE-GDA0002268425240000077
Figure RE-GDA0002268425240000078
其中,Fj(t)为t时刻第j个规则层神经元的激活强度,f j(t)为t时刻第j个规则层神经元的激活强度下界,
Figure RE-GDA0002268425240000079
为t时刻第j个规则层神经元的激活强度上界,
Figure RE-GDA00022684252400000710
Figure RE-GDA00022684252400000711
④后件层:该层有4个神经元,每个神经元的输出为:
Figure RE-GDA0002268425240000081
其中,y k(t)为t时刻关于第k个输出层神经元的后件层神经元的输出下界,
Figure RE-GDA0002268425240000082
为t时刻关于第k个输出层神经元的后件层神经元的输出上界;
Figure RE-GDA0002268425240000083
为t时刻关于第j个规则层神经元与第k个输出层神经元的后件权值,
Figure RE-GDA0002268425240000084
为t时刻关于第 i个输入、第j个规则层神经元和第k个输出神经元的权值系数,
Figure RE-GDA0002268425240000085
为t时刻关于第k个输出层神经元和第j个规则层神经元的权值偏置,k=1,2;
⑤输出层:该层有2个神经元,每个输出神经元的输出为:
Figure RE-GDA0002268425240000086
其中,Δuk(t)为t时刻二型模糊神经网络的第k个输出值,qk(t)为t时刻关于第k个输出层神经元的后件层神经元输出下界的比例值,0<qk(t)<1;
(2)训练二型模糊神经网络,具体如下:
①将参数分为全局参数和局部参数;定义全局参数向量和局部参数向量为:
Figure RE-GDA0002268425240000087
其中,Φg(t)为t时刻全局参数向量,Φl(t)为t时刻局部参数向量;定义目标函数为:
Figure RE-GDA0002268425240000088
其中,l(t)为t时刻的综合误差,α(t)为t时刻的误差系数,l1(t)为t时刻溶解氧DO浓度设定值与实际值的误差,l2(t)为t时刻硝态氮NO3-N浓度设定值与实际值的误差,
Figure RE-GDA0002268425240000089
为t时刻溶解氧DO浓度的设定值,y1(t)为t时刻溶解氧DO浓度的实际值,
Figure RE-GDA00022684252400000810
为t时刻硝态氮NO3-N浓度的设定值,y2(t)为t时刻硝态氮NO3-N浓度的实际值;
②利用综合误差和自适应二阶算法以协同优化二型模糊神经网络的全局参数和局部参数;定义参数更新公式为:
Figure RE-GDA0002268425240000091
其中,Φ(t+1)为t+1时刻的参数向量,Φ(t)为t时刻的参数向量,I为单位矩阵,Gl(t)为t时刻的误差梯度向量,θ(t)为t时刻的自适应学习率,θ(t)∈(0,1],J(t) 为t时刻的雅克比向量,
Figure RE-GDA0002268425240000092
为t时刻综合误差关于全局参数的偏导数,
Figure RE-GDA0002268425240000093
为t时刻综合误差关于局部参数的偏导数,
Figure RE-GDA0002268425240000094
为t时刻综合误差关于不确定中心下界的偏导数,
Figure RE-GDA0002268425240000095
为t时刻综合误差关于不确定中心上界的偏导数,
Figure RE-GDA0002268425240000096
为t时刻综合误差关于宽度值的偏导数,
Figure RE-GDA0002268425240000097
为t 时刻综合误差关于第k个输出的权值系数的偏导数,
Figure RE-GDA0002268425240000098
为t时刻综合误差关于第k个输出的权值偏置的偏导数,
Figure RE-GDA0002268425240000099
为t时刻综合误差关于第k个输出的比例值的偏导数;
(3)设计用于污水处理过程中溶解氧DO浓度和硝态氮NO3-N浓度的协同控制方法,具体为:
①根据公式(7)计算二型模糊神经网络的输出;
②判断当前时刻综合误差的大小,如果l(t)>0.01,转到步骤③;如果l(t)≤0.01,转到步骤④;
③根据公式(10)求解各个参数的更新值;
④计算当前时刻控制器的输出值uk(t)
uk(t)=uk(t-1)+Δuk(t) (11)
其中,uk(t)为t时刻控制器的第k个输出,u1(t)为t时刻控制器的曝气量输出,u2(t) 为t时刻控制器的内回流量输出,uk(t-1)为t-1时刻控制器的第k个输出;
⑤uk(t)为当前时刻溶解氧DO浓度和硝态氮NO3-N浓度协同控制器的实际输出;转到步骤①;
(4)利用求解出的t时刻控制器的输出值u1(t)和u2(t)对溶解氧DO浓度和硝态氮NO3-N浓度进行控制,u1(t)即t时刻曝气量的控制输入量,u2(t)即t时刻内回流量的控制输入量,控制系统的输出为实际溶解氧DO浓度值和硝态氮NO3-N 浓度值;图3显示污水处理过程的溶解氧DO浓度值,X轴:时间,单位是天,Y轴:溶解氧DO浓度,单位是毫克/升,黑色实线为溶解氧DO浓度设定值,黑色虚线是溶解氧DO浓度实际值;溶解氧DO浓度设定值与溶解氧DO浓度实际值的误差如图4,X轴:时间,单位是天,Y轴:溶解氧DO浓度误差值,单位是毫克/升;图5显示污水处理过程的硝态氮NO3-N浓度值,X轴:时间,单位是天,Y轴:硝态氮NO3-N浓度,单位是毫克/升,黑色实线为硝态氮NO3-N 浓度设定值,黑色虚线是硝态氮NO3-N浓度实际值;硝态氮NO3-N浓度设定值与硝态氮NO3-N浓度实际值的误差如图6,X轴:时间,单位是天,Y轴:硝态氮NO3-N浓度误差值,单位是毫克/升;实验结果证明该方法的有效性。

Claims (1)

1.一种基于二型模糊神经网络的污水处理过程协同控制方法,针对污水处理过程中溶解氧DO浓度和硝态氮NO3-N浓度进行控制,其中,以曝气量和内回流量为控制量,溶解氧DO浓度和硝态氮NO3-N浓度为被控量;
其特征在于,包括以下步骤:
(1)设计用于溶解氧DO浓度和硝态氮NO3-N浓度控制的二型模糊神经网络,二型模糊神经网络分为五层:输入层、隶属函数层、规则层、后件层、输出层;具体为:
①输入层:该层由4个输入神经元组成:
X(t)=[x1(t),x2(t),x3(t),x4(t)]T (1)
其中,X(t)表示二型模糊神经网络的输入向量,x1(t)为t时刻溶解氧DO浓度设定值与实际测量值的误差,x2(t)为t时刻溶解氧DO浓度设定值与实际测量值误差的变化率,x3(t)为t时刻硝态氮NO3-N浓度设定值与实际测量值的误差,x4(t)为t时刻硝态氮NO3-N浓度设定值与实际测量值误差的变化率,T为矩阵的转置;
②隶属函数层:该层有4×M个神经元,每个神经元代表一个二型隶属函数,表示如下:
Figure FDA0002211413800000011
Figure FDA0002211413800000012
其中,M为规则层神经元的总个数,1<M≤20;mij(t)表示t时刻关于第i个输入和第j个规则层神经元的隶属函数层神经元的输出值下界,
Figure FDA0002211413800000013
表示t时刻关于第i个输入和第j个规则层神经元的隶属函数层神经元的输出值上界,
Figure FDA0002211413800000014
e为自然常数,e=2.7183;c ij(t)为t时刻关于第i个输入和第j个规则层神经元的隶属函数层神经元的中心值下界,
Figure FDA0002211413800000015
为t时刻第i个输入和第j个规则层神经元的隶属函数层神经元的中心值上界,
Figure FDA0002211413800000016
σij(t)为t时刻关于第i个输入和第j个规则层神经元的隶属函数层神经元的宽度值;i表示二型模糊神经网络的输入个数,j表示二型模糊神经网络的规则层神经元的个数,i=1,2,3,4;j=1,2,…,M;
③规则层:该层有M个神经元,每个神经元的输出为:
Figure FDA0002211413800000021
Figure FDA0002211413800000022
其中,Fj(t)为t时刻第j个规则层神经元的激活强度,f j(t)为t时刻第j个规则层神经元的激活强度下界,
Figure FDA0002211413800000023
为t时刻第j个规则层神经元的激活强度上界,
Figure FDA0002211413800000024
④后件层:该层有4个神经元,每个神经元的输出为:
Figure FDA0002211413800000025
其中,y k(t)为t时刻关于第k个输出层神经元的后件层神经元的输出下界,
Figure FDA0002211413800000026
为t时刻关于第k个输出层神经元的后件层神经元的输出上界;
Figure FDA0002211413800000027
为t时刻关于第j个规则层神经元与第k个输出层神经元的后件权值,
Figure FDA0002211413800000028
为t时刻关于第i个输入、第j个规则层神经元和第k个输出神经元的权值系数,
Figure FDA0002211413800000029
为t时刻关于第k个输出层神经元和第j个规则层神经元的权值偏置,k=1,2;
⑤输出层:该层有2个神经元,每个输出神经元的输出为:
Figure FDA00022114138000000210
其中,Δuk(t)为t时刻二型模糊神经网络的第k个输出值,qk(t)为t时刻关于第k个输出层神经元的后件层神经元输出下界的比例值,0<qk(t)<1;
(2)训练二型模糊神经网络,具体如下:
①将参数分为全局参数和局部参数;定义全局参数向量和局部参数向量为:
Figure FDA00022114138000000211
其中,Φg(t)为t时刻全局参数向量,Φl(t)为t时刻局部参数向量;定义目标函数为:
Figure FDA0002211413800000031
其中,l(t)为t时刻的综合误差,α(t)为t时刻的误差系数,l1(t)为t时刻溶解氧DO浓度设定值与实际值的误差,l2(t)为t时刻硝态氮NO3-N浓度设定值与实际值的误差,
Figure FDA0002211413800000032
为t时刻溶解氧DO浓度的设定值,y1(t)为t时刻溶解氧DO浓度的实际值,
Figure FDA0002211413800000033
为t时刻硝态氮NO3-N浓度的设定值,y2(t)为t时刻硝态氮NO3-N浓度的实际值;
②利用综合误差和自适应二阶算法以协同优化二型模糊神经网络的全局参数和局部参数;定义参数更新公式为:
Figure FDA0002211413800000034
其中,Φ(t+1)为t+1时刻的参数向量,Φ(t)为t时刻的参数向量,I为单位矩阵,Gl(t)为t时刻的误差梯度向量,θ(t)为t时刻的自适应学习率,θ(t)∈(0,1],J(t)为t时刻的雅克比向量,
Figure FDA0002211413800000035
为t时刻综合误差关于全局参数的偏导数,
Figure FDA0002211413800000036
为t时刻综合误差关于局部参数的偏导数,
Figure FDA0002211413800000037
为t时刻综合误差关于不确定中心下界的偏导数,
Figure FDA0002211413800000038
为t时刻综合误差关于不确定中心上界的偏导数,
Figure FDA0002211413800000039
为t时刻综合误差关于宽度值的偏导数,
Figure FDA00022114138000000310
为t时刻综合误差关于第k个输出的权值系数的偏导数,
Figure FDA00022114138000000311
为t时刻综合误差关于第k个输出的权值偏置的偏导数,
Figure FDA00022114138000000312
为t时刻综合误差关于第k个输出的比例值的偏导数;
(3)设计用于污水处理过程中溶解氧DO浓度和硝态氮NO3-N浓度的协同控制方法,具体为:
①根据公式(7)计算二型模糊神经网络的输出;
②判断当前时刻综合误差的大小,如果l(t)>0.01,转到步骤③;如果l(t)≤0.01,转到步骤④;
③根据公式(10)求解各个参数的更新值;
④计算当前时刻控制器的输出值uk(t)
uk(t)=uk(t-1)+Δuk(t) (11)
其中,uk(t)为t时刻控制器的第k个输出,u1(t)为t时刻控制器的曝气量输出,u2(t)为t时刻控制器的内回流量输出,uk(t-1)为t-1时刻控制器的第k个输出;
⑤uk(t)为当前时刻溶解氧DO浓度和硝态氮NO3-N浓度协同控制器的实际输出;转到步骤①;
(4)利用求解出的t时刻控制器的输出值u1(t)和u2(t)对溶解氧DO浓度和硝态氮NO3-N浓度进行控制,u1(t)即t时刻曝气量的控制输入量,u2(t)即t时刻内回流量的控制输入量,控制系统的输出为实际溶解氧DO浓度值和实际硝态氮NO3-N浓度值。
CN201910899611.7A 2019-09-23 2019-09-23 一种基于二型模糊神经网络的污水处理过程协同控制方法 Active CN110647037B (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN201910899611.7A CN110647037B (zh) 2019-09-23 2019-09-23 一种基于二型模糊神经网络的污水处理过程协同控制方法
US16/694,911 US11530139B2 (en) 2019-09-23 2019-11-25 Type-2 fuzzy neural network-based cooperative control method for wastewater treatment process

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910899611.7A CN110647037B (zh) 2019-09-23 2019-09-23 一种基于二型模糊神经网络的污水处理过程协同控制方法

Publications (2)

Publication Number Publication Date
CN110647037A CN110647037A (zh) 2020-01-03
CN110647037B true CN110647037B (zh) 2022-03-15

Family

ID=69011052

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910899611.7A Active CN110647037B (zh) 2019-09-23 2019-09-23 一种基于二型模糊神经网络的污水处理过程协同控制方法

Country Status (2)

Country Link
US (1) US11530139B2 (zh)
CN (1) CN110647037B (zh)

Families Citing this family (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113031445B (zh) * 2021-03-12 2022-09-09 北京工业大学 基于机理模型的污水脱氮过程鲁棒多变量控制方法
CN113283481A (zh) * 2021-05-14 2021-08-20 群智未来人工智能科技研究院(无锡)有限公司 一种基于知识二型模糊的膜污染智能决策方法
CN113759720B (zh) * 2021-08-27 2024-08-20 北京工业大学 一种知识和数据驱动的污水处理生化反应过程控制方法
CN114527656B (zh) * 2022-02-05 2024-07-16 北京工业大学 一种基于时滞补偿策略的硝态氮浓度控制方法
CN114879487B (zh) * 2022-03-02 2024-07-26 北京工业大学 一种基于区间二型模糊神经网络的污水处理过程鲁棒模型预测控制方法
CN114859711B (zh) * 2022-04-15 2024-07-16 北京工业大学 一种基于模糊二阶滑模的城市污水处理曝气过程控制方法
CN114912580B (zh) * 2022-05-07 2024-05-28 北京工业大学 一种基于自适应判别策略的区间二型模糊神经网络出水总磷智能检测方法
CN115981153B (zh) * 2022-12-30 2023-08-04 浙江问源环保科技股份有限公司 基于深度学习的a2o工艺间歇性低碳曝气方法及控制系统
CN117637063B (zh) * 2024-01-10 2024-05-31 广东工业大学 一种水质测量方法、装置、设备和存储介质

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101796928A (zh) * 2009-07-14 2010-08-11 大连水产学院 养殖水体水质参数对养殖生物生长状态影响的预测方法
CN103064290A (zh) * 2013-01-01 2013-04-24 北京工业大学 基于自组织径向基神经网络的溶解氧模型预测控制方法
CN106406094A (zh) * 2016-10-16 2017-02-15 北京工业大学 一种基于区间二型模糊神经网络的污水处理溶解氧浓度跟踪控制方法

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7413654B2 (en) * 2003-12-23 2008-08-19 Siemens Water Technologies Holding Corp. Wastewater treatment control

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101796928A (zh) * 2009-07-14 2010-08-11 大连水产学院 养殖水体水质参数对养殖生物生长状态影响的预测方法
CN103064290A (zh) * 2013-01-01 2013-04-24 北京工业大学 基于自组织径向基神经网络的溶解氧模型预测控制方法
CN106406094A (zh) * 2016-10-16 2017-02-15 北京工业大学 一种基于区间二型模糊神经网络的污水处理溶解氧浓度跟踪控制方法

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
"基于污水处理控制系统智能化技术应用";刘芬芬 等;《轻工标准与质量》;20190430(第4期);第102-103页 *

Also Published As

Publication number Publication date
US20210087074A1 (en) 2021-03-25
CN110647037A (zh) 2020-01-03
US11530139B2 (en) 2022-12-20

Similar Documents

Publication Publication Date Title
CN110647037B (zh) 一种基于二型模糊神经网络的污水处理过程协同控制方法
Manu et al. Artificial intelligence models for predicting the performance of biological wastewater treatment plant in the removal of Kjeldahl Nitrogen from wastewater
CN106873379B (zh) 一种基于迭代adp算法的污水处理最优控制方法
CN106406094B (zh) 一种基于区间二型模糊神经网络的污水处理溶解氧浓度跟踪控制方法
CN109669352A (zh) 基于自适应多目标粒子群的含油污水处理过程优化控制方法
Holenda et al. Aeration optimization of a wastewater treatment plant using genetic algorithm
Qiao et al. Neural network on‐line modeling and controlling method for multi‐variable control of wastewater treatment processes
Do et al. A design of higher-level control based genetic algorithms for wastewater treatment plants
Shiek et al. Design of control strategies for nutrient removal in a biological wastewater treatment process
Yong et al. Intelligent control aeration and external carbon addition for improving nitrogen removal
Sheik et al. Integrated supervisory and override control strategies for effective biological phosphorus removal and reduced operational costs in wastewater treatment processes
Gajewska et al. Kinetics of nitrogen removal processes in constructed wetlands
Saravana Kumar et al. A supervisory fuzzy logic control scheme to improve effluent quality of a wastewater treatment plant
CN114275889B (zh) 一种反硝化工艺有机氮与无机氮的协同优化控制方法
Qiao et al. Recurrent neural network-based control for wastewater treatment process
CN114879487B (zh) 一种基于区间二型模糊神经网络的污水处理过程鲁棒模型预测控制方法
Shen et al. Nitrate control strategies in an activated sludge wastewater treatment process
Aathithya et al. Control of Dissolved Oxygen Concentration in Waste-Water Treatment Plants Using Fuzzy Logic Control
CN115259357A (zh) 基于人工智能的微生物降解控制方法及系统
Sheik et al. Fuzzy Logic Control of Active Sludge–Based Wastewater Treatment Plants
He et al. Intelligent control system based on artificial neural network applied research in autotrophic nitrogen removal process
Vivekanandan et al. Enhanced control of activated sludge process using predetermined set-points
Nascu et al. Multilevel predictive control system for an activated sludge wastewater treatment process
Mihály et al. Optimization of the Wastewater Treatment Plant Aeration Using Artificial Neural Networks Models
CN112487697B (zh) 一种基于改进粒子群算法的污水处理优化控制方法

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant