CN106745739A - A kind of method that SBR short distance nitrations are realized based on Neural Network model predictive pH changes - Google Patents
A kind of method that SBR short distance nitrations are realized based on Neural Network model predictive pH changes Download PDFInfo
- Publication number
- CN106745739A CN106745739A CN201611199982.7A CN201611199982A CN106745739A CN 106745739 A CN106745739 A CN 106745739A CN 201611199982 A CN201611199982 A CN 201611199982A CN 106745739 A CN106745739 A CN 106745739A
- Authority
- CN
- China
- Prior art keywords
- neural network
- value
- data
- layer
- sbr
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 55
- 238000003062 neural network model Methods 0.000 title claims abstract description 25
- 238000006396 nitration reaction Methods 0.000 title claims abstract description 24
- 230000008859 change Effects 0.000 claims abstract description 27
- 230000008569 process Effects 0.000 claims abstract description 27
- 238000013528 artificial neural network Methods 0.000 claims abstract description 24
- 238000005273 aeration Methods 0.000 claims abstract description 20
- 238000012544 monitoring process Methods 0.000 claims abstract description 10
- 238000011217 control strategy Methods 0.000 claims abstract description 7
- 230000006641 stabilisation Effects 0.000 claims abstract description 6
- 238000011105 stabilization Methods 0.000 claims abstract description 6
- 238000012549 training Methods 0.000 claims description 9
- 238000005096 rolling process Methods 0.000 claims description 8
- 238000006243 chemical reaction Methods 0.000 claims description 7
- 230000006870 function Effects 0.000 claims description 7
- 230000004913 activation Effects 0.000 claims description 6
- 230000007935 neutral effect Effects 0.000 claims description 6
- 239000011159 matrix material Substances 0.000 claims description 5
- 229910002651 NO3 Inorganic materials 0.000 claims description 3
- NHNBFGGVMKEFGY-UHFFFAOYSA-N Nitrate Chemical compound [O-][N+]([O-])=O NHNBFGGVMKEFGY-UHFFFAOYSA-N 0.000 claims description 3
- IOVCWXUNBOPUCH-UHFFFAOYSA-M Nitrite anion Chemical compound [O-]N=O IOVCWXUNBOPUCH-UHFFFAOYSA-M 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 3
- 238000013461 design Methods 0.000 claims description 3
- 230000000638 stimulation Effects 0.000 claims description 2
- 238000004065 wastewater treatment Methods 0.000 abstract description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 11
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 description 9
- 229910052799 carbon Inorganic materials 0.000 description 9
- LFQSCWFLJHTTHZ-UHFFFAOYSA-N Ethanol Chemical compound CCO LFQSCWFLJHTTHZ-UHFFFAOYSA-N 0.000 description 8
- 239000010865 sewage Substances 0.000 description 7
- IJGRMHOSHXDMSA-UHFFFAOYSA-N Atomic nitrogen Chemical compound N#N IJGRMHOSHXDMSA-UHFFFAOYSA-N 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 6
- 230000003647 oxidation Effects 0.000 description 5
- 238000007254 oxidation reaction Methods 0.000 description 5
- 238000012545 processing Methods 0.000 description 5
- 241000894006 Bacteria Species 0.000 description 3
- 230000008901 benefit Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 230000001546 nitrifying effect Effects 0.000 description 3
- 229910052757 nitrogen Inorganic materials 0.000 description 3
- 238000001556 precipitation Methods 0.000 description 3
- 239000010802 sludge Substances 0.000 description 3
- QGZKDVFQNNGYKY-UHFFFAOYSA-N Ammonia Chemical compound N QGZKDVFQNNGYKY-UHFFFAOYSA-N 0.000 description 2
- MMDJDBSEMBIJBB-UHFFFAOYSA-N [O-][N+]([O-])=O.[O-][N+]([O-])=O.[O-][N+]([O-])=O.[NH6+3] Chemical compound [O-][N+]([O-])=O.[O-][N+]([O-])=O.[O-][N+]([O-])=O.[NH6+3] MMDJDBSEMBIJBB-UHFFFAOYSA-N 0.000 description 2
- 238000007599 discharging Methods 0.000 description 2
- 238000005265 energy consumption Methods 0.000 description 2
- 239000011521 glass Substances 0.000 description 2
- GQPLMRYTRLFLPF-UHFFFAOYSA-N nitrous oxide Inorganic materials [O-][N+]#N GQPLMRYTRLFLPF-UHFFFAOYSA-N 0.000 description 2
- 238000003756 stirring Methods 0.000 description 2
- QGZKDVFQNNGYKY-UHFFFAOYSA-O Ammonium Chemical compound [NH4+] QGZKDVFQNNGYKY-UHFFFAOYSA-O 0.000 description 1
- IOVCWXUNBOPUCH-UHFFFAOYSA-N Nitrous acid Chemical compound ON=O IOVCWXUNBOPUCH-UHFFFAOYSA-N 0.000 description 1
- 229910021529 ammonia Inorganic materials 0.000 description 1
- XKMRRTOUMJRJIA-UHFFFAOYSA-N ammonia nh3 Chemical compound N.N XKMRRTOUMJRJIA-UHFFFAOYSA-N 0.000 description 1
- 230000033228 biological regulation Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000001186 cumulative effect Effects 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 239000010840 domestic wastewater Substances 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 238000004134 energy conservation Methods 0.000 description 1
- 239000007789 gas Substances 0.000 description 1
- JEGUKCSWCFPDGT-UHFFFAOYSA-N h2o hydrate Chemical compound O.O JEGUKCSWCFPDGT-UHFFFAOYSA-N 0.000 description 1
- 238000010438 heat treatment Methods 0.000 description 1
- 238000012886 linear function Methods 0.000 description 1
- 239000007788 liquid Substances 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 230000001537 neural effect Effects 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 238000004886 process control Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 150000003839 salts Chemical class 0.000 description 1
- 238000004062 sedimentation Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Classifications
-
- C—CHEMISTRY; METALLURGY
- C02—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F3/00—Biological treatment of water, waste water, or sewage
- C02F3/30—Aerobic and anaerobic processes
- C02F3/302—Nitrification and denitrification treatment
- C02F3/303—Nitrification and denitrification treatment characterised by the nitrification
-
- C—CHEMISTRY; METALLURGY
- C02—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F3/00—Biological treatment of water, waste water, or sewage
- C02F3/34—Biological treatment of water, waste water, or sewage characterised by the microorganisms used
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
-
- C—CHEMISTRY; METALLURGY
- C02—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F2101/00—Nature of the contaminant
- C02F2101/10—Inorganic compounds
- C02F2101/16—Nitrogen compounds, e.g. ammonia
-
- C—CHEMISTRY; METALLURGY
- C02—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F2209/00—Controlling or monitoring parameters in water treatment
- C02F2209/06—Controlling or monitoring parameters in water treatment pH
-
- C—CHEMISTRY; METALLURGY
- C02—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F2209/00—Controlling or monitoring parameters in water treatment
- C02F2209/22—O2
Landscapes
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Microbiology (AREA)
- Biodiversity & Conservation Biology (AREA)
- Hydrology & Water Resources (AREA)
- Environmental & Geological Engineering (AREA)
- Water Supply & Treatment (AREA)
- Chemical & Material Sciences (AREA)
- Organic Chemistry (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Purification Treatments By Anaerobic Or Anaerobic And Aerobic Bacteria Or Animals (AREA)
Abstract
A kind of method for realizing SBR short distance nitrations based on Neural Network model predictive pH changes, belongs to method of wastewater treatment field.In SBR system, aeration time is controlled using Real-Time Control Strategy, the data based on the SBR data of longtime running stabilization, set up 3 layers of BP neural network forecast model, look-ahead pH change curves, according to the pH change points of prediction, setting in advance stops aeration time, prevents NO2 ‑N is further aoxidized, specific as follows:Data acquisition:Data acquisition is carried out by next step by in-line meter;Data selection is processed:The data that the first step is collected are selected, is pre-processed, enough samples of continuous-stable are therefrom selected, using pH as variable parameter;Model is set up:The data of selection are modeled, BP neural network model is selected, are trained, correct and are tested, after reaching required precision, carry out process sensor monitoring pH and on-line prediction pH.
Description
Technical field
The present invention relates to a kind of sewage disposal technology, especially can be by Neural Network model predictive SBR ammoxidations
Journey pH change points, in advance setting stops aeration time, so that SBR system is done step-by-step short distance nitration, it is adaptable to urban life
The denitrogenation processing of sewage, belongs to method of wastewater treatment field.
Background technology
Traditional biological denitrogenation technology mainly includes two stages of nitration denitrification, and nitration reaction includes two steps, first
It is ammoxidation process to walk, by ammonia oxidation bacteria (AOB) by mineralized nitrogen be cultured water, second step is nitrifying process, by nitrous acid
Cultured water is further converted to nitrate nitrogen by salt oxidation bacteria (NOB).Anti-nitration reaction is present in anoxic and organic carbon source
Under the conditions of, it is N by nitrite or nitrate transformation2Or N2The process of O.In traditional biological denitrogenation technology, nitrifying bacteria community growth
It is relatively slow, it is difficult to maintain biological concentration higher in reactor, cause system hydraulic detention time long, volumetric loading rate is low, capital construction
Investment is big, and the additional organic carbon source of denitrification causes energy waste.
Short distance nitration is in NO by nitrifying process control2 -In-N the stages, then carry out denitrification.Realize the key of short distance nitration
It is in NO by ammoxidation control2 -In-N the stages, prevent NO2 -The further oxidation of-N, then directly carry out denitrification.Realize at present
The conventional control strategy of short distance nitration is to observe pH change points according to on-line sensor, is closed when ammoxidation process is completed and exposed
Gas, but nitrate nitrogen is converted into due to still having part cultured water under anaerobic environment, because NOB must produce nitrous in AOB
Can be grown after hydrochlorate, if it is possible to close aeration in time before the completion of ammoxidation process, just can reduce NOB's
Growth, going down can reach the purpose of elutriation system NOB for a long time.
City domestic sewage ammoxidation process has non-linear, time variation and uncertainty, and this is for setting up accurate number
Learn model and cause certain difficulty, even if setting up, structure is also sufficiently complex, it is difficult to effectively control.Fuzzy control is found based on research
Lack self-learning capability, adaptability is poor, and control accuracy is not high, and practical application is relatively difficult, and is based on the pre- of neutral net
Observing and controlling system is that neutral net and PREDICTIVE CONTROL are combined, and the rolling thought and neutral net of PREDICTIVE CONTROL are accurately described into non-
The characteristic of linear and uncertain dynamic process is organically combined, and can be predicted control to nonlinear system well.Therefore, profit
Realize that SBR short distance nitrations are a kind of accurate effective methods with Neural Network model predictive pH change points.
The content of the invention
The purpose of the present invention is to set up a kind of method for realizing short distance nitration based on neural network model control SBR methods, is led to
Neural Network model predictive pH change points are crossed, setting in advance stops aeration time, reach the quick effect for realizing short distance deep denitrogenation
Really.
The technical scheme is that:
In SBR system, aeration time is controlled using Real-Time Control Strategy, the SBR data with longtime running stabilization are as base
Plinth data, set up 3 layers of BP neural network forecast model (Fig. 1), look-ahead pH change curves, according to the pH change points of prediction,
Setting in advance stops aeration time, prevents NO2 -- N is further aoxidized, and preferably realizes short distance nitration.
Model is set up the data acquisition of process (Fig. 2) Main Basiss, data processing and model and is set up.Using BP neural network
Model, data is trained, corrects and is tested.After reaching required precision, then neural network prediction pH models are applied to SBR
In system.
Concrete operation method is as follows:
(1) data acquisition:Data acquisition is for further processing by in-line meter;
(2) data selection treatment:The data that the first step is collected are selected, is pre-processed, therefrom selected continuous-stable
Enough samples, using pH as variable parameter;
(3) model is set up:The data of selection are modeled, BP neural network model is selected, are trained, correct and are surveyed
Examination (is combined) by on-line monitoring and prediction, after reaching required precision, carries out process sensor monitoring pH and on-line prediction pH.
To predict the pH change curves of ammoxidation process, the input layer of the BP neural network that the present invention sets up includes online prison
The pH value of survey, output layer is lower a moment (time interval 1min) pH value of prediction.Specific method is as follows:
1) to write neural network function using Matlab as follows
Net=newff (X, Y, [3,1], ' tansig''purelin'}, ' traingdx', ' learngdm');
Net.trainParam.epochs=2000;
Net.trainParam.goal=0.0001;
Net.trainParam.show=10;
Net.trainParam.lr=0.05;
Net.trainParam.min_grad=0.00000001;
Net.trainParam.max_fail=1500;
Net=train (net, X, Y);
Activation primitive:Activation primitive for discrimination or convergence rate no matter have significant impact.It is bent high order is approached
During line, sigmoid function ratio of precision linear function is much higher, and the activation primitive of the hidden layer that the present invention is used takes S function:f(x)
=1/ (1+exp (- x)).
2) design of BP neural network:Typical BP neural network is 3-tier architecture:Input mistake, hidden layer and output layer;
Input layer receives the input of algorithm, and input layer is x in this model.The number of input layer should be with the dimension of x
Number correspondence, in this model, input x is scalar, therefore input layer number is 1;
Hidden layer node number can be with sets itself, and this model specification is 3;
Output layer node number should be corresponding with the dimension of output y, and in this model, output y is scalar, therefore output layer knot
Point number is 1;
3) basic ideas of algorithm
In SBR system, one 3 layers of BP neural network model is set up, output layer is estimated using the error after output
Direct preceding conducting shell error, then with this directly before conducting shell estimation error more preceding layer error, so in layer
Anti-pass is gone down, and just obtains the estimation error of every other each layer.Neutral net constantly changes under the stimulation of extraneous input sample
Become the connection weight of network, so that the process that the output of network is constantly constantly adjusted close to desired output, weights, that is,
The learning training process of network.The error that this process is performed until network output is reduced to acceptable degree or proceeds to pre-
Untill the study number of times for first setting.
Given group sample:I.e. a collection of pH value of stable operation, is trained to model, and detailed process is as follows:
Determine that neural network structure, for 1-3-1 types, selects 0.1 as the margin of error, and provides hidden layer and arrive output layer
Weight matrix, is initialized with a certain small random weight matrix.
1. sequentially inputs sample, the reality output of calculating network
2. calculation error
3. calculates the variable quantity of each layer weights
4. adjusts weights
If 5. errors are in permissible range (such as less than 0.1), turn 2., otherwise turn 6.
If 6. reaches the training time, terminate, otherwise turn the calculating 2. into next round, if directly
Output layer is not set also to reach desired output when number of times (such as 2000 times) for specifying,
Then exit by force, terminator.
4) using the pH data training neural network model of stable operation, pH of lower a moment is predicted by the method for rolling forecast
Change.It is specific as follows:
In 1 one SBR aerobic reaction cycles, preceding 50 points only carry out data acquisition, are not predicted;
2 are predicted since the 51st pH value, and the 51st pH value is predicted with the 1-50 pH value, and the 51st for predicting is worked as
Make known pH value;
3 predict the 52nd pH value with the 2-52 pH value, the 52nd pH value known to that will be predicted;
4 predict the 53rd pH value, by that analogy, rolling forecast with the 3-52 pH value;
5 predicted with the pH of first 50 actual on-line monitorings after 20 pH value.Can at 51st
To show 51-70 pH value of prediction, by that analogy, the pH value of subsequent time is predicted;
5) neural network prediction model is applied to SBR system:In SBR system aeration phase, by on-line checking pH numbers
According to the change point occurred using Neural Network model predictive pH value, the pH change curves according to prediction, at the end of prediction ammoxidation
Between, setting in advance stops aeration time, it is to avoid excess aeration causes nitrite to be further converted to nitrate, it is ensured that AOB fills
Divide propagation, eliminate NOB.
The present invention is a kind of to realize the method for SBR short distance nitrations compared with the life for the treatment of city based on Neural Network model predictive pH
The technology of sewage living is compared, with following advantages:
1) energy conservation and consumption reduction effects are good:Most of expense of sewage treatment plant is used for power consumption and medicine consumes, but short distance nitration is only
25% aeration energy consumption by ammonium oxidation into cultured water, need to be saved, energy resource consumption has been saved;In the anoxic denitrification stage
40% additional carbon has been saved, operating cost has been reduced.
2) stabilization short distance nitration:Accurate neural network prediction pH change points, can make ammonia nitrogen exhaustive oxidation, it is ensured that go out
Water water quality, is the effective ways for stablizing short distance nitration while can also avoid the excess aeration from causing short distance to be destroyed.
3) control strategy simplifies:Choosing pH value can both save data storage as the control parameter of neural network prediction
Space, while the operation times of controller can be reduced, is simplified control strategy.
4) process advantage:Whole technique is completed by process real-time control system, and easy to operate with managing, expense is low, resistance to
The advantages of impact load is strong and is not susceptible to sludge bulking.
5) good stability:The neutral net that neuron is constituted has powerful non-linear mapping capability and learning functionality, phase
Than in other control methods, network response surface method can well realize the control to nonlinear system, improve intelligence
The adaptivity and robustness of energy control system, accurately can effectively predict ammoxidation process, it is ensured that sewage disposal process
Stability.
(6) short distance nitration of stabilization can provide new method and way for prevailing technology (short distance nitration-anaerobic ammoxidation)
Footpath.
Brief description of the drawings
Fig. 1 is 3 layers of BP neural network structure structure chart;
Fig. 2 is Establishment of Neural Model procedure chart;
Fig. 3 is structure of reactor schematic diagram of the invention;
Fig. 4 is a kind of method for realizing SBR short distance nitrations based on Neural Network model predictive pH change points in embodiment
Operation result figure.
Specific implementation method
The present invention is explained with reference to the accompanying drawings and examples, but the present invention is not limited to following examples.
(1) data acquisition:Data acquisition is for further processing by online pH instrumentations;
(2) data processing:The data that the first step is collected are pre-processed, 217 lot sample sheets are therefrom selected, selection pH is made
It is variable parameter;
(3) model is set up:Data after treatment are modeled, first preference pattern, are trained, correct and test,
After reaching required precision, on-line prediction, process monitoring and Sensor monitoring are carried out.To predict the pH curves of ammoxidation process, this
Inventing the input layer of the BP neural network set up includes the pH value of on-line checking, and output layer is the pH value of prediction.Transported using stabilization
Capable 217 batches of pH data training neural network model, predicts that pH changes by the method for rolling forecast.Specific method is as follows:
1) order is input into Matlab programs, neural network function is as follows
Net=newff (X, Y, [3,1], ' tansig''purelin'}, ' traingdx', ' learngdm');
Net.trainParam.epochs=2000;
Net.trainParam.goal=0.0001;
Net.trainParam.show=10;
Net.trainParam.lr=0.05;
Net.trainParam.min_grad=0.00000001;
Net.trainParam.max_fail=1500;
Net=train (net, X, Y);
Activation primitive:The activation primitive of hidden layer takes s type functions:F (x)=1/ (1+exp (- x))
2) design of BP neural network:Typical BP neural network is 3-tier architecture:Input mistake, hidden layer and output layer.
Input layer receives the input of algorithm, and input layer is x in this model.The number of input layer should be with the dimension of x
Number correspondence, in this model, input x is scalar, therefore input layer number is 1.
Hidden layer node number can be with sets itself, and this model specification is 3.
Output layer node number should be corresponding with the dimension of output y, and in this model, output y is scalar, therefore output layer knot
Point number is 1.
3) basic ideas of algorithm
Given group sample:I.e. a collection of pH value of stable operation, is trained to model, and detailed process is as follows:
Determine that neural network structure, for 1-3-1 types, selects 0.1 as the margin of error, and provides hidden layer and arrive output layer
Weight matrix, is initialized with a certain small random weight matrix.
1. sample, the reality output of calculating network are sequentially input
2. calculation error
3. the variable quantity of each layer weights is calculated
4. weights are adjusted
If 5., turning 2), otherwise to turn 7)
If 6., terminating, otherwise, turn 2) to enter the calculating of next round, if until the number of times for specifying
Output layer is reached desired output when (2000 times), then exit by force,
Terminator.
4) using the pH data training neural network model of stable operation, predict that pH changes by the method for rolling forecast.
It is specific as follows:
1. in a SBR aerobic reaction cycle, preceding 50 points only carry out data acquisition, are not predicted
2. predicted since the 51st pH value, predict the 51st pH value with the 1-50 pH value, the 51st for predicting
As known pH value
3. the 52nd pH value is predicted with the 2-52 pH value, the 52nd for predicting is known to
PH value,
4. the 53rd pH value, by that analogy, rolling forecast are predicted with the 3-52 pH value
5. 20 pH value after being predicted with the pH of first 50 actual on-line monitorings.At 51st
51-70 pH value of prediction is shown, by that analogy, the pH value of subsequent time is predicted.
5) used as experimental subjects, sbr reactor device main body is organic glass to the actual domestic wastewater using certain university's dependents' district discharge
Glass is made, dischargeable capacity 19.5L (1100mm high, internal diameter 150mm), and experimental rig is as shown in Figure 1.Inflow is controlled by the time
System, it is also possible to by liquid level gauge control.Operation a cycle, in whole course of reaction, is controlled by heating rod and temperature control system daily
At 25 DEG C or so, the aerobic stage of course of reaction is aerated temperature by air compressor, and DO is constant in control system, and sludge averagely stops
The time is stayed at 12 days or so;Anoxic phases add ethanol as denitrifying carbon source, and whole course of reaction is stirred to protect always
Hold the homogeney of system and be thoroughly mixed state and run according to Real-Time Control Strategy.React precipitation after terminating, draining and leave unused,
Whole cycle monitors pH value and DO on-line.
Comprise the following steps that:
1) intake:Inlet valve is opened, starts intake pump, it is anti-that by intake pump the sanitary sewage in water tank is added into SBR
Answer in device, intake pump discharge 24L/min, 12L of being intake after 30s close intake pump and inlet valve, and stirring is started while water inlet
Device is stirred.
2) it is aerated:After water inlet end, air intake valve is opened, open air blower, aeration is adjusted by adjusting spinner flowmeter
Amount, is aerated to system, and the compressed air that air blower is provided is entered in system by air inlet pipe, is shown by observing DO on-line checkings
The DO concentration shown, real-time regulation spinner flowmeter, DO constant concentrations (such as 0.5mg/L) in control water, by on-line checking pH numbers
According to ANN Control unit reads pH data, it is contemplated that go out the change point and setting time of pH value.Then change point is defined,
And set point that this change point is predicted with ANN Control unit and setting time are checked.If the change for detecting
Change point within the time interval that set point and setting time are allowed, then this change point can be as real-time control point, in advance
The setting ammoxidation end time, stop aeration, open mud discharging mouth, discharge the sludge of 100ml, after spoil disposal, close mud discharging mouth.
3) additional carbon is added:After stopping aeration, 1ml ethanol is added as carbon source, open the valve that carbon source is added on pipe
Pump is added with ethanol, carbon source is closed after the ethanol 1ml for adding adding the valve that pump and carbon source added on pipe carries out denitrification, and pH is bent
Denitrification terminates after line rapid increase tends towards stability, and closes stirring
4) precipitation and draining:It is 30min according to the sedimentation time for setting after reaction terminates, after precipitation, water decanter starts work
Make, water after treatment is discharged to outside reactor through outlet pipe, set draining 12L, after draining terminates, the valve closed on outlet pipe.
5) leave unused:Standby time as 1h is set, after 1 hour standby time set in advance is reached, system is out of service
Or carry out the operation of next cycle.
Fig. 2 be it is a kind of based on Neural Network model predictive pH change points realize SBR short distance nitrations method operation result show
Show, system nitrous cumulative percentage after 10 cycles reaches more than 95%, total nitrogen is less than 5mg/L in system final outflow water, is far below
Total nitrogen concentration required by national grade one discharge standard.Good short distance deep denitrogenation effect has been reached, short distance nitration has been realized
Quick startup.
Claims (3)
1. it is a kind of to change the method for realizing SBR short distance nitrations based on Neural Network model predictive pH, it is characterised in that in SBR systems
In system, aeration time is controlled using Real-Time Control Strategy, the data based on the SBR data of longtime running stabilization set up 3 layers of BP
Neural network prediction model, look-ahead pH change curves, according to the pH change points of prediction, setting in advance stops aeration time,
Prevent NO2 -- N is further aoxidized, specific as follows:
(1) data acquisition:Data acquisition is carried out by next step by in-line meter;
(2) data selection treatment:The data that the first step is collected are selected, is pre-processed, therefrom selected the enough of continuous-stable
Sample, using pH as variable parameter;
(3) model is set up:The data of selection are modeled, BP neural network model is selected, are trained, correct and are tested,
After reaching required precision, process sensor monitoring pH and on-line prediction pH is carried out.
2., according to a kind of method for realizing SBR short distance nitrations based on Neural Network model predictive pH changes of claim 1, it is special
Levy and be, the input layer of the BP neural network of foundation includes the pH value of on-line monitoring, output layer is the pH value of lower a moment of prediction, tool
Body method is as follows:
1) to write neural network function using Matlab as follows:
The activation primitive of the hidden layer of use takes S function:F (x)=1/ (1+exp (- x));
2) design of BP neural network:Typical BP neural network is 3-tier architecture:Input mistake, hidden layer and output layer;
Input layer receives the input of algorithm, and input layer is x in this model, and the number of input layer is corresponding with the dimension of x, this
In model, input x is scalar, and input layer number is 1;
Hidden layer node number is 3;
Output layer node number is corresponding with the dimension of output y, and in this model, output y is scalar, therefore output layer node number is
1;
3) basic ideas of training
In SBR system, one 3 layers of BP neural network model is set up, the straight of output layer is estimated using the error after output
Connect the error of preceding conducting shell, then with this directly before conducting shell estimation error more preceding layer error, anti-pass so in layer
Go down, just obtain the estimation error of every other each layer;Neutral net constantly changes net under the stimulation of extraneous input sample
The connection weight of network, so that the output of network is constantly close to desired output, in the process that weights are constantly adjusted, that is, net
The learning training process of network;The error that this process is performed until network output is reduced to acceptable degree or proceeds in advance
Untill the study number of times of setting;
4) using the pH data training neural network model of stable operation, predict that pH of lower a moment becomes by the method for rolling forecast
Change, it is specific as follows:
In 1 one SBR aerobic reaction cycles, preceding 50 points only carry out data acquisition, are not predicted;
2 predict that predict the 51st pH value with the 1-50 pH value, the 51st for predicting is as since the 51st pH value
Know pH value;
3 predict the 52nd pH value with the 2-52 pH value, the 52nd pH value known to that will be predicted;
4 predict the 53rd pH value, by that analogy, rolling forecast with the 3-52 pH value;
5 predicted with the pH of first 50 actual on-line monitorings after 20 pH value.51-70 of prediction can be shown at 51st
PH value, by that analogy, predicts the pH value of subsequent time;
5) neural network prediction model is applied to SBR system:In SBR system aeration phase, by on-line checking pH data, profit
Neural Network model predictive pH value is used, the change point that the pH change curves according to prediction occur predicts the ammoxidation end time, carries
Preceding setting stops aeration time, it is to avoid excess aeration causes nitrite to be further converted to nitrate, it is ensured that AOB fully increases
Grow, eliminate NOB.
3., according to a kind of method for realizing SBR short distance nitrations based on Neural Network model predictive pH changes of claim 2, it is special
Levy and be, 3) specifically it is trained in:Given group sample:I.e. a collection of pH value of stable operation, is trained, detailed process to model
It is as follows:
Determine that neural network structure, for 1-3-1 types, selects 0.1 as the margin of error, and provide hidden layer to the power square of output layer
Battle array, is initialized with a certain small random weight matrix;
1. sequentially inputs sample, the reality output of calculating network;
2. calculation error;
3. calculates the variable quantity of each layer weights;
4. adjusts weights;
If 5. errors are in permissible range, turn 2., otherwise turn 6.;
If 6. reaches the training time, terminate, otherwise turn the calculating 2. into next round, if not having also when the number of times for specifying
Output layer is reached desired output, then exit by force, terminator.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611199982.7A CN106745739A (en) | 2016-12-22 | 2016-12-22 | A kind of method that SBR short distance nitrations are realized based on Neural Network model predictive pH changes |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611199982.7A CN106745739A (en) | 2016-12-22 | 2016-12-22 | A kind of method that SBR short distance nitrations are realized based on Neural Network model predictive pH changes |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106745739A true CN106745739A (en) | 2017-05-31 |
Family
ID=58897213
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201611199982.7A Pending CN106745739A (en) | 2016-12-22 | 2016-12-22 | A kind of method that SBR short distance nitrations are realized based on Neural Network model predictive pH changes |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106745739A (en) |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108640279A (en) * | 2018-05-25 | 2018-10-12 | 北京工业大学 | A kind of real-time monitoring device and method of continuous flow short distance nitration-anaerobic ammoxidation technique |
CN108640272A (en) * | 2018-06-01 | 2018-10-12 | 杭州师范大学 | The method of qualitative assessment anaerobic ammonia oxidation process impact resistance and recovery capability |
CN109925992A (en) * | 2019-04-23 | 2019-06-25 | 辽宁工业大学 | It is a kind of based on multimode stage by stage continuously stir on-line monitoring method |
CN110672792A (en) * | 2019-10-09 | 2020-01-10 | 中南大学 | Soft measurement method and system for pH value in neutral leaching process of zinc hydrometallurgy |
CN111832693A (en) * | 2019-04-16 | 2020-10-27 | 杭州海康威视数字技术股份有限公司 | Neural network layer operation and model training method, device and equipment |
CN112250166A (en) * | 2020-08-27 | 2021-01-22 | 盐城工学院 | Printing and dyeing wastewater efficient detoxification gradient treatment method based on intelligent SBR electrode process |
CN113003704A (en) * | 2019-12-19 | 2021-06-22 | 同济大学 | Method and device for short-cut nitrification of dirt |
CN115684276A (en) * | 2022-12-28 | 2023-02-03 | 北京华科仪科技股份有限公司 | Desulfurization system pH value prediction method and system based on integrated fusion model |
CN115806344A (en) * | 2022-11-22 | 2023-03-17 | 曲誉环境综合治理有限公司 | ORP control accurate aeration system |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6571228B1 (en) * | 2000-08-09 | 2003-05-27 | Po-Tong Wang | Hybrid neural networks for color identification |
-
2016
- 2016-12-22 CN CN201611199982.7A patent/CN106745739A/en active Pending
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6571228B1 (en) * | 2000-08-09 | 2003-05-27 | Po-Tong Wang | Hybrid neural networks for color identification |
Non-Patent Citations (3)
Title |
---|
张吉强等: "《微生物燃料电池同步脱氮产电研究》", 31 May 2016 * |
管秋: "基于人工神经网络的污水水质指标软测量方法的研究", 《中国优秀博硕士学位论文全文数据库(硕士)信息科技辑》 * |
邵青: "SBR工艺中pH值变化时间序列的BP网络预测模型", 《中国农村水利水电》 * |
Cited By (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108640279A (en) * | 2018-05-25 | 2018-10-12 | 北京工业大学 | A kind of real-time monitoring device and method of continuous flow short distance nitration-anaerobic ammoxidation technique |
CN108640272A (en) * | 2018-06-01 | 2018-10-12 | 杭州师范大学 | The method of qualitative assessment anaerobic ammonia oxidation process impact resistance and recovery capability |
CN108640272B (en) * | 2018-06-01 | 2021-01-19 | 杭州师范大学 | Method for quantitatively evaluating impact resistance and recovery capability of anaerobic ammonia oxidation process |
CN111832693B (en) * | 2019-04-16 | 2024-05-28 | 杭州海康威视数字技术股份有限公司 | Neural network layer operation and model training method, device and equipment |
CN111832693A (en) * | 2019-04-16 | 2020-10-27 | 杭州海康威视数字技术股份有限公司 | Neural network layer operation and model training method, device and equipment |
CN109925992B (en) * | 2019-04-23 | 2021-05-25 | 辽宁工业大学 | Multi-module staged continuous stirring on-line monitoring method |
CN109925992A (en) * | 2019-04-23 | 2019-06-25 | 辽宁工业大学 | It is a kind of based on multimode stage by stage continuously stir on-line monitoring method |
CN110672792A (en) * | 2019-10-09 | 2020-01-10 | 中南大学 | Soft measurement method and system for pH value in neutral leaching process of zinc hydrometallurgy |
CN110672792B (en) * | 2019-10-09 | 2020-08-28 | 中南大学 | Soft measurement method and system for pH value in neutral leaching process of zinc hydrometallurgy |
CN113003704A (en) * | 2019-12-19 | 2021-06-22 | 同济大学 | Method and device for short-cut nitrification of dirt |
CN113003704B (en) * | 2019-12-19 | 2022-07-26 | 同济大学 | Method and device for dirt short-cut nitrification |
CN112250166A (en) * | 2020-08-27 | 2021-01-22 | 盐城工学院 | Printing and dyeing wastewater efficient detoxification gradient treatment method based on intelligent SBR electrode process |
CN115806344A (en) * | 2022-11-22 | 2023-03-17 | 曲誉环境综合治理有限公司 | ORP control accurate aeration system |
CN115806344B (en) * | 2022-11-22 | 2023-08-04 | 曲誉环境综合治理有限公司 | ORP (integrated pressure and pressure) control precise aeration system |
CN115684276A (en) * | 2022-12-28 | 2023-02-03 | 北京华科仪科技股份有限公司 | Desulfurization system pH value prediction method and system based on integrated fusion model |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106745739A (en) | A kind of method that SBR short distance nitrations are realized based on Neural Network model predictive pH changes | |
Borzooei et al. | Optimization of the wastewater treatment plant: From energy saving to environmental impact mitigation | |
AU2021101438A4 (en) | Adaptive control method and system for aeration process | |
Lee et al. | Neural network modeling for on-line estimation of nutrient dynamics in a sequentially-operated batch reactor | |
CN106651032B (en) | Method for intelligently controlling neural network in ammoxidation process by SBR (sequencing batch reactor) method | |
CN105439285B (en) | A kind of regulation method of sewage treatment | |
CN106277330A (en) | A kind of Sewage Plant intelligence control system based on nitrogen balance and control method | |
CN108640276A (en) | A kind of sewage treatment plant AAO process optimization operation methods based on WEST models | |
CN107032506A (en) | It is segmented the apparatus and method that water outlet short distance nitration Anammox/ denitrifications handle sanitary sewage | |
JP2004025160A (en) | System and method for artificial intelligence control of sewage waste water treatment by neural network and inversion wave algorithm | |
CN104182794A (en) | Method for soft measurement of effluent total phosphorus in sewage disposal process based on neural network | |
CN102643742A (en) | Autotrophic bacteria kinetic parameter measurement device and method | |
CN113104961A (en) | Real-time aeration accurate control method based on activated sludge treatment sewage process | |
CN109809560A (en) | A kind of accurate dosing control device of carbon source and method of multipoint water feeding multistage A/O technique | |
KR102311657B1 (en) | Smart management system for wastewater treatment | |
CN112782975A (en) | Sewage treatment aeration intelligent control method and system based on deep learning | |
CN208008554U (en) | A kind of detection device of microbial activity of activated sludge | |
KR101016394B1 (en) | Real-time wastewater composition analyzer using a rapid microbial respiration detector, ss and ec combined sensing system and its measuring method | |
CN105548474A (en) | Determination method of biochemical characteristic map of microorganism carrier for sewage treatment | |
Drewnowski et al. | Computer simulation in predicting biochemical processes and energy balance at WWTPs | |
Watts et al. | On line respirometry: A powerful tool for activated sludge plant operation and design | |
CN107010723B (en) | Method and device for controlling deep denitrification of landfill leachate through two-stage anaerobic ammonia oxidation | |
CN103145243A (en) | N2O gas collecting device and method in reaction process of sequencing batch reactor (SBR) method based on process control | |
CN203319768U (en) | Process control-based device for collecting N2O gas in SBR (Sequencing Batch Reactor) reaction process | |
Muloiwa et al. | Modelling and optimization of energy consumption in the activated sludge biological aeration unit |
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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170531 |