CN109928493A - It is a kind of that method is controlled to adjust based on the accurate dissolved oxygen of big data and evolution algorithm - Google Patents
It is a kind of that method is controlled to adjust based on the accurate dissolved oxygen of big data and evolution algorithm Download PDFInfo
- Publication number
- CN109928493A CN109928493A CN201910137709.9A CN201910137709A CN109928493A CN 109928493 A CN109928493 A CN 109928493A CN 201910137709 A CN201910137709 A CN 201910137709A CN 109928493 A CN109928493 A CN 109928493A
- Authority
- CN
- China
- Prior art keywords
- data
- model
- value
- control
- 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.)
- Pending
Links
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02W—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO WASTEWATER TREATMENT OR WASTE MANAGEMENT
- Y02W10/00—Technologies for wastewater treatment
- Y02W10/10—Biological treatment of water, waste water, or sewage
Landscapes
- Feedback Control In General (AREA)
Abstract
Method is controlled to adjust based on the accurate dissolved oxygen of big data and evolution algorithm the present invention relates to a kind of, which includes three subsystems: the pretreatment system of " emulation+model ", carries out Technology Modeling using the historical data of sewage plant；The control system of " feedforward+model+feedback ", operating parameter is obtained by pretreatment system and airflow balance steady-state value is calculated through model, each control zone is accurately controlled according to the tolerance, the control system includes data acquisition module, data processing module, air compressor control module, valve control module；The check system of " big data+evolution algorithm " carries out simulation calculating to system parameter and corrects, improve control stability using operation front and back data.The present invention has many advantages, such as that design is reasonable, easy to operate, reliable performance is stable, can be widely used to promote.
Description
Technical field
The present invention relates to municipal sewage treatment aerobic tank aeration system optimal controls more particularly to accurate aeration to control.
Background technique
The aeration process of activated sludge process is to provide suitable dissolved oxygen for processes such as removal organic carbon, nitrification, suction phosphorus, with
Promote being normally carried out for three kinds of biochemical reactions.The target for being aerated flow control is to form stable dissolved oxygen conditions, is microorganism
Growth establishes a dynamic equilibrium and reliable living environment with contaminant degradation.The essence of this homeostasis process is to make always
Oxygen transfer rate is approximately equal to total oxygen consumption rate.Since the influent quality and water of sewage plant are variations, in the specific time
Its oxygen demand is also variation in section, only makes the oxygen-supplying amount in the period mutually balanced with oxygen demand, just can guarantee processing environment
Stabilization, to guarantee effluent quality.
In order to accurately be controlled dissolved oxygen in biological tank (DO) environment, it is necessary to be carried out to the homeostasis process of DO
Adequately understanding, it includes two processes: first is that the diffusion and dissolution of oxygen.Air is mainly reflected in blast aeration system
After the aeration head release of aeration bottom of pond portion, the oxygen in air is shifted from gas phase into liquid phase.Second is that the utilization of dissolved oxygen and
Consumption.This process synthesis the organic carbon removal process of Aerobic, Ammonia Nitrification, biological phosphate-eliminating etc., are by multiple
The result of process synthesis effect.
Summary of the invention
For accurate aeration in the prior art control it is unstable, it is an object of the invention to: improve accurate aeration system
Stability, stablize effluent quality, save fan energy use, reduce manual operation, have design it is reasonable, easy to operate, can
Depending on changing the advantages that strong, reliable performance is stable, use easy to spread.
In order to achieve the above object, the present invention adopts the following technical scheme:
Accurate aeration control system uses the mentality of designing of " emulation+model+feedforward+feedback+big data+evolution algorithm ",
It is broadly divided into three subsystems:
(1) pretreatment system of " emulation+model ".Divided by history data to sewage plant and analysis data
Analysis processing, determines sewage plant operating parameter；Operating parameter is recycled to establish " model "；Finally by real-time running data, chemical examination
Data and " big data+evolution algorithm " tool verify simulation model, finally obtain reliable operating parameter.By this process,
Dynamic DO setting value, sludge back flow quantity steady-state value, airflow balance steady-state value and these values of sewage plant can be obtained with water inlet
The variation tendency value of load.
(2) control system of " feedforward+model+feedback ".By obtaining operating parameter in design process above and through model
Airflow balance steady-state value is calculated, each control zone is accurately controlled according to the tolerance；Recycle model total to air blower
Tolerance carries out real-time control；Finally by feedback data and " big data+evolution algorithm " tool verification control effect and revise control
Parameter processed reaches stability contorting effect.
(3) checking system of " big data+evolution algorithm ".The strategic importance of big data technology, which is not lain in, grasps huge number
It is believed that breath, and be to carry out specialized process containing significant data to these.Technically, big data and sewage plant optimize
The relationship of operation is inseparable just as the front and back sides of one piece of coin.In accurate aeration control process, big data is through the beginning
Eventually, support is provided for optimal control parameter.
Detailed description of the invention
The single seat biochemistry pool dissolved oxygen control zone Fig. 1 divides
Fig. 2 blower unit non-blind area curve graph
Fig. 3 is the overall flow and structural schematic diagram of accurate aeration of the present invention.
Specific embodiment
Accurate aeration control system of the present invention is described in detail with reference to the accompanying drawing, and combines a specific embodiment pair
Effect is explained.
The accurate control of aeration is divided into two dimensions of the time and space, spatially carries out the control of dissolved oxygen subregion to biological tank
System, meets the needs of different process section is to aeration quantity；Change dynamic setting DO value with influent load in time, meet it is different into
Aeration demand under water condition.
Designed according to certain sewage plant biological tank and aerating pipelines, by single group biological reaction pool be divided into Π (Π in this example=
4) 2 Π dissolved oxygen control zones are arranged in a dissolved oxygen control zone, biological tank altogether, realize the accurate control to dissolved oxygen.Single seat pond
Dissolved oxygen control zone divides as shown in Figure 1:
One, air compressor control
1. accurate aeration system controls blower system
Accurate aeration system and blower system can communicate between both sides in a looped network.Accurate aeration system
System calculates blower system according to influent quality, the state change of water and biological tank, according to model and big data dynamic
Pressure or flow, then blower system is given by communication modes, the aeration quantity of real-time dynamic regulation blower system is saved and is exposed
Gas energy consumption.
2. accurate aeration system controls blower system and requires
Accurate aeration has following function to blower system control principle:
(1) control function on the spot
Air blower has on-site control LCP (Local Control Panel), LCP can carry out on the spot and remotely into
Row switching: when being switched to regard to ground mode, air blower can carry out start and stop on LCP and adjust air quantity operation；It is long-range when being switched to
When mode, start and stop can be carried out by control cabinet MCP (Master Control Panel) and adjust air quantity.
(2) remote control function
Blower system MCP is the control cabinet of blower system, plays the window role connecting with third party system.MCP
It can receive to set pressure (or flow) and actual pressure (or flow) signal from third party system, by these signals to drum
Blower fan system is effectively adjusted.
3. accurate aeration system is for blower system requirement
(1) blind area is adjusted
Blower system does not adjust blind area, by taking flow is adjusted as an example: there can be no opening, a blower air quantity is inadequate,
Two blower air quantities situation more than needed is opened, otherwise will make air blower frequent start-stop in accurate aeration adjustment process.Drum
Start and stop adjustment curve such as Fig. 2 between blower:
(2) response time
When blower system not automatic start-stop blower operation, reach in 5 minutes accurate aeration system pressure or
Tolerance requirement；When blower system automatic start-stop blower operation, system tolerance is made to reach stable state in 15 minutes.
(3) it communicates with each other
To make accurate aeration system and blower system both sides' accessible communication, it need to guarantee that air blower MCP has Ethernet
Communication function, while guaranteeing that PLC brand should be consistent with live robot control system(RCS) or accurate aeration system PLC system brand.MCP branch
It holds and detects communication heartbeat function mutually with accurate aeration system.
(4) failover
Blower system has failure protection function, when communicating failure, can be immediately switched to local runtime, protects
Blower system.Control of the accurate aeration to air blower is divided into following several situations:
Accurate aeration and air blower MCP are communicated under normal circumstances, be given by accurate aeration air blower setting pressure (or
Flow) and actual pressure (or flow), it automatically adjusts to air blower, is not necessarily to manual intervention；
When accurate aeration and air blower MCP communication abnormal suddenly, MCP automatically switches to local runtime state, keeps drum
Blower standing state is constant, and after communicating recovery, accurate aeration control model can be switched to by central control system；
Accurate aeration system and sewage plant central control system are parallel systems, when a failure occurs, can a key exit and open
With accurate aeration system；
After exiting accurate aeration system, central control system also remote controlled air blower MCP is divided into two kinds of feelings at this time
Condition:
Central control system and air blower MCP are communicated under normal circumstances, are to give air blower setting pressure manually by central control system
Power (or flow) and actual pressure (or flow), are adjusted air blower；
When central control system and air blower MCP communication abnormal suddenly, MCP automatically switches to local runtime state, keeps drum
Blower standing state is constant, and after communicating recovery, middle control MANUAL CONTROL mode can be switched to by central control system；
This hair can be understood and applied the above description of the embodiments is intended to facilitate those skilled in the art
It is bright.Person skilled in the art obviously easily can make various modifications to these embodiments, and described herein
General Principle is applied in other embodiments without having to go through creative labor.Therefore, the present invention is not limited to the realities of this explanation
Example, those skilled in the art's announcement according to the present invention are applied, improvement and modification made without departing from the scope of the present invention all should
Within protection scope of the present invention.
Two, based on big data+evolution algorithm Optimization about control parameter method
Accurate aeration control parameter is mostly manually set by operator in water factory at present, to make water factory run behaviour substantially up to standard
Operating parameter need to repeatedly be adjusted by making personnel, and this method has that time-consuming, adjust coarse, and system robustness is poor, is unable to reach accurate
The shortcomings that control, find a kind of method for capableing of rapid Optimum control parameter be very it is necessary to.Due to practical water factory's history
Data lack aeration and accurately control target, are only capable of providing dissolved oxygen control range, so that the problem, which has, is not suitable for supervised learning
Algorithm, while having the characteristics that hysteresis quality is big.Different from supervised learning, evolution algorithm needs intelligent body to carry out in continuous trial and error
Study, by obtaining reward thus the behavior for instructing intelligent body with interacting for environment.
To protect environment, keeps sewage disposal plant effluent up to standard, destructive testing can not be carried out in water factory's actual motion.However
Evolution algorithm needs constantly learn in trial and error, finally provide more excellent control parameter by the interaction with environment.To solve this
Contradiction devises the two-phase method of optimization aeration control parameter: in the first stage, from the historical data analysis water factory, water factory of accumulation
Inherent feature, and establish modeling water factory correlation module；It will be consolidated the first stage in second stage using evolution algorithm
There is feature to be applied to water factory's model, carries out parameter optimization.Steps are as follows for concrete operation:
First stage:
1.1 historical datas based on water factory's accumulation are to water factory's inherent feature (aeration, dissolved oxygen, dissolved oxygen efficiency and temperature
Between relationship) analyzed, determine the dependence between feature.It is exposed with each gallery aeration quantity setting value and corresponding gallery
For tolerance, compared to Pearson correlation coefficient, Spearman's correlation coefficient (Spearman related coefficient) is applied widely, because
This is using the dependence between Spearman correlation coefficient analysis variable.Remember that each gallery aeration quantity setting value is Afsij, aeration
Amount is Afij(wherein i=1,2 ..., Π are numbered, j=1,2 ..., n for gallery.N is number of samples).To the aeration of each gallery
Amount and dissolved oxygen are ranked up according to size order respectively, and result is Afs after note sequenceihk, AfihjWherein h=1,2,3 ..., n, j
It is respectively aeration ordinal number corresponding with dissolved oxygen before sorting with k, then:
dh=j-k (2)
Due to being influenced each other between each variable there are hysteresis quality, when adjusting aeration valve opening, true tolerance value is needed
The regular hour is wanted to can be only achieved the tolerance value of setting.Therefore the reaction relation between each variable of accurate simulation water factory is established just
True model, being determined to become for lag time must condition.However, the increase of sample size will lead to the increase of difference, i.e., generally
For biggish sample will lead to lesser related coefficient.Accurately to find the lagged relationship between correlated variables, using such as lower section
Method remembers nlagTo lag number, then its calculation formula is as follows:
Wherein, m is the number for the variable that each variable is deducted from total sample.M=5, then to aeration quantity setting value AfsijFrom
The 5+1 data starts up to n-th of value, to aeration quantity AfijThen this Pierre is utilized up to the n-th -5 values since the 1st value
Graceful related coefficient, that is, formula (1) calculates ρi5.As m=0, then it represents that the two is without lagged relationship.[] is rounding symbol, as a result
The integer nearest with variable is rounded, [3.2]=3, and [3.6]=4.
To its dependent variable, the calculating such as the dissolved oxygen of each gallery and aeration quantity, temperature and oxygen conversion rate determine system
Number R2, examine whether dissolved oxygen and aeration quantity, temperature and oxygen conversion rate have significant relation using F, if having
Linear relationship then utilize t examine each variable (aeration quantity, temperature and oxygen conversion rate) to Effect of Dissolved Oxygen whether
Significantly.
WhereinFor the coefficient of determination of i-th of gallery,For the regression sum of square of i-th of gallery,It is i-th
The residual sum of squares (RSS) of gallery,I-th of gallery dissolved oxygen sample mean value, DOijFor the dissolved oxygen at i-th of gallery j moment,
Dissolved oxygen for the j moment obtained to i-th of gallery regression forecasting, i=1,2 ..., Π, j=1,2 ..., n.
1.2 establish corresponding model to water factory's necessaries, facility and parameter in accordance with inherent feature respectively, calculate by 1.1
As a result, being each equipment, facility and parameter selection model: if research object has significantly under level of significance α=0.05
Linear relationship then uses linear model, otherwise uses neural network model.It should be noted that needing for characterization relevant device
Model specification threshold value, this is because being protection air volume regulating valve in practical application, only when front and back moment air demand is more than certain
(value is commonly referred to as dead zone to thresholding, is denoted as Aflock, front and back moment dissolved oxygen, which changes, leads to adjust tolerance greater than 200 sides, then
Aflock=200) just air volume regulating valve is adjusted when, is not otherwise adjusted.Simultaneously as valve regulated to final stablize needs
Certain time is wanted, to prevent damage valve, usually sets the guard time of 30s, it may be assumed that valve self-regulation is made to start to next time to adjust
Time interval be set as 30s, the not control valve within this 30s time.It is in the first stage the simulation true operating condition of water factory,
Equally use intrinsic 30s setting value for the regulating time thresholding of control valve in model.However due to being had determined in (1)
Lagged relationship between each variable is remembered that t is valve guard time herein, therefore in second stage, is then used
T=tδ+tlag (8)
Adjusting thresholding as valve, wherein tδThe time needed is completed for valve event, is adjusted 250 side's gas and (is about adjusted
10%), the valve regulated time is 10s (valve variation 1% needs about 1 second), then tδ=10.And tlagIt then indicates that tolerance reaches to set
In the reaction time of definite value, the value is by nlagBeing scaled the time gets.Compared to original 30s fixed value, setting in this way can be protected conscientiously
Relevant device is protected, guard time is only set as 30s such when adjusting the tolerance time plus the reaction time is more than 30 seconds
In the case of just can not play protection valve purpose.Therefore, valve can be obtained using the technology and needs the stable practical time.
Linear model form, the general type of linear model are as follows:
Y=a0+a1x1+a2x2+…+acxc (9)
Wherein y is explained variable (y is true tolerance in valve aeration equation), xθIt (is exposed in valve for explanatory variable
X in gas equationθFor tolerance setting value), a0For the constant term obtained using least square fitting, aθFor regression coefficient, θ=1,
2 ..., c, c are the dimension of the explanatory variable of linear model.
Neural network model form:
Since prediction object is the unitary variants such as dissolved oxygen, but it is polynary that it, which relies on factor, thus using have have more into
Single output has the rear feed neural network (BP neural network) of single hidden layer.Hidden layer excitation function is sigmoid function:
Wherein v is the corresponding input of each neuron of hidden layer, to hidden layer theThe input of a neuron are as follows:
WhereinFor hidden layerThe γ weight of a neuron,It isThe biasing of a neuron,γ=1,2 ..., M, K=10 are hidden neuron number, and M is that neural network inputs number.According to excitation letter
Number is it is found that hidden layer theThe output of a neuron are as follows:
1.3 simulate water factory's corresponding portion (air blower and aeration link) using the model established, and fitting effect is small
In 10-5Then receive the equation；Otherwise, it returns to (1) and re -training is adjusted to model until effect is no more than 10-5.So far,
The first stage of adjusting method terminates.
Second stage:
2.1 improve the fixed threshold in previous stage according to formula (8), and are transferred to 2.2 and optimize；
2.2 establish evolution algorithm model, construct reward function, setting model boundary condition: DOij>0,0<Afij, Afij≤
Afmax,
Wherein DOijAnd DOsijRespectively i-th of gallery j moment dissolved oxygen and dissolved oxygen setting value；AfijAnd AfsijRespectively
For i-th of gallery j moment aeration quantity and aeration quantity setting value, i=1,2 ..., Π and j=1,2 ..., n, AflockFor dead band value,
AfmaxFor maximum tolerance.
2.3 are learnt using evolution algorithm, and until reaching termination condition output optimal control parameter, we are utilized herein
Differential evolution algorithm optimizes variable relevant in objective function, has compared to other evolution algorithm differential evolution algorithms
The advantage that the variable for needing to adjust is less, stability is stronger, convergence is more preferable, the specific steps of which are as follows:
1) algorithm initialization: p individual is randomly generated in search space, and (usual p is optimised problem space dimension
10 times, dead band value, each gallery control parameter that we will be aerated valve to different gallerys herein optimize, therefore p
=10* Π * 4, Π are gallery number), it is denoted as It is the i-th of the initial value composition of parameter to be optimized
Group vector, i=1,2 ..., p, to P0According to formula (18) calculating target function value.
2) carry out mutation operation to Search of Individual: note population algebra is z, three Different Individual difference of selection from z generationL ≠ q ≠ r and l, q, r are integer).It is generated by these three individuals according to formula (13)
The intermediate of z+1 the φ individual of generation
Wherein mut is mutagenic factor, and value interval is [0,2], takes 0.5 herein, right to avoid algorithm from precocious phenomenon occur
Mutation operator increases TSP question algorithm, carries out automatic adjusument according to formula (14):
Mut=mut0*2λ (14)
Wherein lambda definition is formula (15), mut0=0.5, take zmFor maximum number of iterations:
3) crossover operation is carried out to Search of Individual: (each to each genetic fragment of the intermediate in z generation and z+1 generation
Component) it carries out carrying out crossover operation according to formula (16), it may be assumed that
Wherein rrandFor the random number (obey unit interval in Uniformly distributed) in unit section,For the φ individual
The intersection factor, intersect the factorIt is adaptively adjusted according to formula (17).
Wherein QφFor individualFitness, QminAnd QmaxIt is z for the maximum and minimum value of population,It is planted for z generation
The mean value of all individual adaptation degrees in group,WithRespectively intersect the up-and-down boundary value of the factor, value is respectively 0.1 and 0.6.
4) since the needs of all parts in actual production meet some requirements and could operate normally, for guarantee model
These boundary conditions of accuracy are essential.To make model consistent with actual production, by target function value (fitness function)
It is defined as formula (18):
According to above-mentioned two step, it is known that can be obtained according to formula (19) for the optimized parameter in control:
Compared to the setting of general dead band value, the case where being directed to different gallerys using context of methods setting is corresponding
Dead zone (optimization obtains), enables the value to react optimal situation to the greatest extent, is conducive to the accurate control of aerator.
Claims (6)
1. one kind is based on the biological tank accurate aeration system of " emulation+model+feedforward+feedback+big data+evolution algorithm ", special
Sign is to include three subsystems:
(1) pretreatment system of " emulation+model "；It is carried out at analysis by history data to sewage plant and analysis data
Reason, determines sewage plant operating parameter；Operating parameter is recycled to establish " model "；Finally by real-time running data, analysis data
" big data " tool verifies simulation model, finally obtains reliable operating parameter；By this process, the dynamic of sewage plant is obtained
State DO setting value, sludge back flow quantity steady-state value, airflow balance steady-state value and these values with influent load variation tendency value；
(2) control system of " feedforward+model+feedback "；By obtaining operating parameter in design process above and being calculated through model
It obtains airflow balance steady-state value, each control zone is controlled according to the tolerance；Model is recycled to carry out the total tolerance of air blower
Real-time control；Finally by feedback data and " big data " tool verification control effect and control parameter is revised, reaches stable control
Effect processed, control system include data acquisition module, data processing module, air compressor control module, valve control module；
(3) inspection of " big data+evolution algorithm " and parameter calibration system；
In accurate aeration control process, the evolution algorithm model established for the purpose of reducing valve regulated and saving tolerance leads to
The control parameter that the mass data such as the water factory's aeration, dissolved oxygen, water of accumulation are learnt is crossed, raising system control can be reached
The robustness of system, reduction valve frequently switch on to extend valve service life and save the purpose of electric energy.
2. system according to claim 1, it is characterised in that: acquired in the pretreatment system of " emulation+data+model "
Real time data includes: Inlet and outlet water COD and ammonia nitrogen, disengaging water flow, aerobic tank MLSS, DO and ORP.
3. accurate aeration system according to claim 1, it is characterised in that: in the control system of " feedforward+model+feedback "
Data acquisition module acquires primary at interval of 1-10s；The data of acquisition have been carried out glide filter processing by data processing module, when
When the data of acquisition have jump, data is carried out and jump are gone to handle；Air compressor control module be according to system-computed go out pressure or
Flow, interval 5-60min control air blower；Valve control module is the flow gone out according to system-computed, interval 1-5min control
Valve.
4. accurate aeration system according to claim 1, it is characterised in that: the real time data and change of 1min acquisition will be spaced
It tests data and correlation is found out by big data analysis tool, and by iterative processing, obtain characteristic parameter variable and characteristic value,
Control stability is improved using these data.
5. the method for application such as accurate aeration system of any of claims 1-4, which is characterized in that this method is specific
The following steps are included:
(1) sewage plant real time data and analysis data are acquired, and is recorded in the form of data sheet；
(2) it is modeled using data, and carries out system emulation test；
(3) simulation result is input in control system, and is regulated and controled；
(4) big data tool and evolution algorithm, combined data are utilized, optimal control parameter improves control stability.
6. according to the method described in claim 5, being divided into two stages: in the first stage, from water factory's historical data analysis of accumulation
Water factory's inherent feature, and establish modeling water factory correlation module；It will be obtained the first stage in second stage using evolution algorithm
It is applied to water factory's model to inherent feature, carries out parameter optimization；It is characterized by:
First stage:
1.1 historical datas based on water factory's accumulation are to it that water factory's inherent feature is aeration, dissolved oxygen, dissolved oxygen efficiency and temperature
Between relationship analyzed, determine the dependence between feature；
Remember that each gallery aeration quantity setting value is Afsij, aeration quantity Afij, wherein i=1,2 ..., Π, are numbered, j=1 for gallery,
2,…,n；N is number of samples；Aeration quantity and dissolved oxygen to each gallery are ranked up according to size order respectively, after note sequence
It as a result is Afsihk, AfihjWherein h=1,2,3 ..., n, j and k are respectively aeration ordinal number corresponding with dissolved oxygen before sorting, then:
dh=j-k (2)
Accurately to find the lagged relationship between correlated variables, with the following method, n is rememberedlagTo lag number, then its calculation formula is such as
Under:
Wherein, m is the number for the variable that each variable is deducted from total sample；M=5, then to aeration quantity setting value AfsijFrom 5+1
A data start up to n-th of value, to aeration quantity AfijThen Spearman phase is utilized up to the n-th -5 values since the 1st value
Relationship number, that is, formula (1) calculates ρi5；As m=0, then it represents that the two is without lagged relationship；[] is to be rounded symbol；
To its dependent variable, these calculating coefficients of determination of the dissolved oxygen and aeration quantity, temperature and oxygen conversion rate of each gallery
R2, examine whether dissolved oxygen and aeration quantity, temperature and oxygen conversion rate have significant relation using F, if having line
Sexual intercourse then using t examine each variable include aeration quantity, temperature and oxygen conversion rate to Effect of Dissolved Oxygen whether
Significantly；
WhereinFor the coefficient of determination of i-th of gallery,For the regression sum of square of i-th of gallery,For i-th gallery
Residual sum of squares (RSS),I-th of gallery dissolved oxygen sample mean value, DOijFor the dissolved oxygen at i-th of gallery j moment,For to
The dissolved oxygen at the j moment that i gallery regression forecasting obtains, i=1,2 ..., Π, j=1,2 ..., n；
1.2 establish corresponding model to water factory's necessaries, facility and parameter in accordance with inherent feature respectively, by 1.1 calculated results,
For each equipment, facility and parameter selection model: if research object has significant linear pass under level of significance α=0.05
System then uses linear model, otherwise uses neural network model；Need for characterize relevant device model specification threshold value, this be by
It is protection air volume regulating valve in practical application, only when front and back moment air demand is more than certain thresholding, the value is commonly referred to as dead
Area is denoted as Aflock, front and back moment dissolved oxygen, which changes, leads to adjust tolerance greater than AflockWhen just air volume regulating valve is adjusted,
Otherwise it does not adjust；Meanwhile valve self-regulation being made to start to the time interval that next time is adjusted to be set as 30s, within this 30s time not
Control valve；It is in the first stage the simulation true operating condition of water factory, uses intrinsic 30s setting value equally in model to adjust
The regulating time thresholding of valve；However due to having determined that the lagged relationship between each variable in (1), remember that t is valve herein
Door guard time, therefore in second stage, then it uses
T=tδ+tlag (8)
Adjusting thresholding as valve, wherein tδIt is completed for valve event the time needed, and tlagThen indicate that tolerance reaches setting
In the reaction time of value, the value is by nlagBeing scaled the time gets；
Neural network model form:
Since prediction object is the unitary variants such as dissolved oxygen, but it is polynary that it, which relies on factor, therefore is had more using having into single defeated
The rear feed neural network of single hidden layer is provided, hidden layer excitation function is sigmoid function:
Wherein v is the corresponding input of each neuron of hidden layer, to hidden layer theThe input of a neuron are as follows:
WhereinFor hidden layerThe γ weight of a neuron,It isThe biasing of a neuron,γ
=1,2 ..., M, K=10 are hidden neuron number, and M is that neural network inputs number；According to excitation function it is found that hidden layerThe output of a neuron are as follows:
1.3 simulate water factory's corresponding portion using the model of foundation, and error is less than 10-5Then receive the equation；Otherwise, it returns
(1) re -training is adjusted to model until error is no more than 10-5；So far, the first stage terminates；
Second stage:
2.1 improve the fixed threshold in previous stage according to formula (8), and are transferred to 2.2 and optimize；
2.2 establish evolution algorithm model, construct reward function, setting model boundary condition: DOij>0,0<Afij, Afij≤Afmax,
Wherein DOijAnd DOsijRespectively i-th of gallery j moment dissolved oxygen and dissolved oxygen setting value；AfijAnd AfsijRespectively i-th
A gallery j moment aeration quantity and aeration quantity setting value, i=1,2 ..., Π and j=1,2 ..., n, AflockFor dead band value, Afmax
For maximum tolerance；
2.3 are learnt using evolution algorithm, and until reaching termination condition output optimal control parameter, we utilize difference herein
Evolution algorithm optimizes variable relevant in objective function, the specific steps of which are as follows:
1) algorithm initialization: being randomly generated p individual in search space, and p=10* Π * 4, Π are gallery number, is denoted as For i-th group of vector that the initial value of parameter to be optimized is constituted, i=1,2 ..., p, to P0According to formula
(18) calculating target function value；
2) carry out mutation operation to Search of Individual: note population algebra is z, and three Different Individuals of selection distinguish P from z generationz:L, q, r ∈ [1, p], l ≠ q ≠ r and l, q, r are integer；Z+1 is generated according to formula (13) by these three individuals
The intermediate of the φ individual of generation
Wherein mut is mutagenic factor, and value interval is [0,2], to avoid algorithm from precocious phenomenon occur, is increased certainly mutation operator
Adequate variation algorithm carries out automatic adjusument according to formula (14):
Mut=mut0*2λ (14)
Wherein lambda definition is formula (15), mut0=0.5, take zmFor maximum number of iterations:
3) crossover operation is carried out to Search of Individual: to each genetic fragment, that is, each component of the intermediate in z generation and z+1 generation
It carries out carrying out crossover operation according to formula (16), it may be assumed that
Wherein rrandFor the random number in unit section,For the intersection factor of the φ individual, intersect the factorAccording to formula (17)
Adaptive adjustment；
Wherein QφFor individualFitness, QminAnd QmaxIt is z for the maximum and minimum value of population,It is z in population
The mean value of all individual adaptation degrees,WithRespectively intersect the up-and-down boundary value of the factor, value is respectively 0.1 and 0.6；
4) since the needs of all parts in actual production meet some requirements and could operate normally, for the accurate of guarantee model
These boundary conditions of property are essential；To make model consistent with actual production, target function value is defined as formula (18):
According to above-mentioned two step, the optimized parameter in control is obtained according to formula (19):
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910137709.9A CN109928493A (en) | 2019-02-25 | 2019-02-25 | It is a kind of that method is controlled to adjust based on the accurate dissolved oxygen of big data and evolution algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910137709.9A CN109928493A (en) | 2019-02-25 | 2019-02-25 | It is a kind of that method is controlled to adjust based on the accurate dissolved oxygen of big data and evolution algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109928493A true CN109928493A (en) | 2019-06-25 |
Family
ID=66985871
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910137709.9A Pending CN109928493A (en) | 2019-02-25 | 2019-02-25 | It is a kind of that method is controlled to adjust based on the accurate dissolved oxygen of big data and evolution algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109928493A (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110378533A (en) * | 2019-07-22 | 2019-10-25 | 中展环能(北京)技术有限公司 | A kind of intelligence aeration management method based on big data analysis |
| CN111273706A (en) * | 2020-03-02 | 2020-06-12 | 软通动力信息技术（集团）有限公司 | Aeration tank dissolved oxygen control method, device, computer equipment and medium |
| CN113845205A (en) * | 2021-10-18 | 2021-12-28 | 江苏科易达环保科技有限公司 | High-salinity high-nitrogen sewage intelligent integrated desalting and denitrification control system |
| CN114132980A (en) * | 2022-01-27 | 2022-03-04 | 广东省广业环保产业集团有限公司 | Short-range intelligent accurate aeration control method, equipment and system for sewage treatment |
| CN114291911A (en) * | 2021-12-20 | 2022-04-08 | 安徽泛湖生态科技股份有限公司 | Sewage aeration control method based on oxygen transfer efficiency |
| CN114380378A (en) * | 2022-03-23 | 2022-04-22 | 济安永蓝(北京)工程技术开发有限公司 | Intelligent phosphorus-control drug feeding method and device and storage medium |
| CN115124153A (en) * | 2022-07-06 | 2022-09-30 | 安吉国源水务集团有限公司 | Accurate aeration control system for sewage plant |
| CN115385445A (en) * | 2022-09-21 | 2022-11-25 | 南京晓庄学院 | Soft measurement management system for sewage biochemical treatment based on big data |
| CN116969582A (en) * | 2023-09-22 | 2023-10-31 | 深圳市友健科技有限公司 | Intelligent regulation and control method and system for sewage treatment |
| CN117142630A (en) * | 2023-10-26 | 2023-12-01 | 北京华科仪科技股份有限公司 | Automatic control method and system of rotary disc type biological reaction device |
| CN117509928A (en) * | 2023-11-03 | 2024-02-06 | 安徽泛湖生态科技股份有限公司 | OTE real-time monitoring-based aeration optimization energy-saving control system and method |
| CN118131605A (en) * | 2024-05-07 | 2024-06-04 | 浙江信合阀门有限公司 | Intelligent ball valve control system |
| CN118348811A (en) * | 2024-06-14 | 2024-07-16 | 江西交科交通工程有限公司 | Treatment facility scheduling method and system based on wharf initial rainwater collection |
| WO2024198227A1 (en) * | 2023-03-24 | 2024-10-03 | 青岛思普润水处理股份有限公司 | Intelligent control method for two-stage wastewater treatment system |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5852817A (en) * | 1991-08-14 | 1998-12-22 | Kabushiki Kaisha Toshiba | Intelligent control apparatus |
| EP1376276A1 (en) * | 2002-06-21 | 2004-01-02 | H2L Co., Ltd | An AI based control system and method for treating sewage/waste water by means of a neural network and a back-propagation algorithm |
| CN105573115A (en) * | 2015-12-09 | 2016-05-11 | 中山大学 | Sewage treatment process energy-saving optimization control method based on quantum genetic algorithm |
| CN206109149U (en) * | 2016-09-21 | 2017-04-19 | 昆明滇池水务股份有限公司 | Sewage treatment system that one -level strengthening treatment and biochemical treatment combined rainy season |
| CN107555590A (en) * | 2017-10-27 | 2018-01-09 | 中冶华天南京电气工程技术有限公司 | A kind of multi-point combination formula accurate aeration control method |
-
2019
- 2019-02-25 CN CN201910137709.9A patent/CN109928493A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5852817A (en) * | 1991-08-14 | 1998-12-22 | Kabushiki Kaisha Toshiba | Intelligent control apparatus |
| EP1376276A1 (en) * | 2002-06-21 | 2004-01-02 | H2L Co., Ltd | An AI based control system and method for treating sewage/waste water by means of a neural network and a back-propagation algorithm |
| CN105573115A (en) * | 2015-12-09 | 2016-05-11 | 中山大学 | Sewage treatment process energy-saving optimization control method based on quantum genetic algorithm |
| CN206109149U (en) * | 2016-09-21 | 2017-04-19 | 昆明滇池水务股份有限公司 | Sewage treatment system that one -level strengthening treatment and biochemical treatment combined rainy season |
| CN107555590A (en) * | 2017-10-27 | 2018-01-09 | 中冶华天南京电气工程技术有限公司 | A kind of multi-point combination formula accurate aeration control method |
Non-Patent Citations (2)
| Title |
|---|
| 余云龙等: "南方某污水处理厂精确曝气技术示范及评估分析", 《给水排水》 * |
| 刘载文等: "《污水处理过程优化控制系统》", 30 April 2014, 中国轻工业出版社 * |
Cited By (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110378533B (en) * | 2019-07-22 | 2023-04-18 | 中展环能(北京)技术有限公司 | Intelligent aeration management method based on big data analysis |
| CN110378533A (en) * | 2019-07-22 | 2019-10-25 | 中展环能(北京)技术有限公司 | A kind of intelligence aeration management method based on big data analysis |
| CN111273706A (en) * | 2020-03-02 | 2020-06-12 | 软通动力信息技术（集团）有限公司 | Aeration tank dissolved oxygen control method, device, computer equipment and medium |
| CN111273706B (en) * | 2020-03-02 | 2024-02-06 | 软通动力信息技术(集团)股份有限公司 | Dissolved oxygen control method and device for aeration tank, computer equipment and medium |
| CN113845205A (en) * | 2021-10-18 | 2021-12-28 | 江苏科易达环保科技有限公司 | High-salinity high-nitrogen sewage intelligent integrated desalting and denitrification control system |
| CN114291911A (en) * | 2021-12-20 | 2022-04-08 | 安徽泛湖生态科技股份有限公司 | Sewage aeration control method based on oxygen transfer efficiency |
| CN114132980A (en) * | 2022-01-27 | 2022-03-04 | 广东省广业环保产业集团有限公司 | Short-range intelligent accurate aeration control method, equipment and system for sewage treatment |
| CN114132980B (en) * | 2022-01-27 | 2022-04-29 | 广东省广业环保产业集团有限公司 | Short-range intelligent accurate aeration control method, equipment and system for sewage treatment |
| CN114380378A (en) * | 2022-03-23 | 2022-04-22 | 济安永蓝(北京)工程技术开发有限公司 | Intelligent phosphorus-control drug feeding method and device and storage medium |
| CN114380378B (en) * | 2022-03-23 | 2022-06-28 | 济安永蓝(北京)工程技术开发有限公司 | Intelligent phosphorus control drug feeding method and device and storage medium |
| CN115124153A (en) * | 2022-07-06 | 2022-09-30 | 安吉国源水务集团有限公司 | Accurate aeration control system for sewage plant |
| CN115385445A (en) * | 2022-09-21 | 2022-11-25 | 南京晓庄学院 | Soft measurement management system for sewage biochemical treatment based on big data |
| WO2024198227A1 (en) * | 2023-03-24 | 2024-10-03 | 青岛思普润水处理股份有限公司 | Intelligent control method for two-stage wastewater treatment system |
| CN116969582A (en) * | 2023-09-22 | 2023-10-31 | 深圳市友健科技有限公司 | Intelligent regulation and control method and system for sewage treatment |
| CN116969582B (en) * | 2023-09-22 | 2023-12-08 | 深圳市友健科技有限公司 | Intelligent regulation and control method and system for sewage treatment |
| CN117142630A (en) * | 2023-10-26 | 2023-12-01 | 北京华科仪科技股份有限公司 | Automatic control method and system of rotary disc type biological reaction device |
| CN117142630B (en) * | 2023-10-26 | 2024-02-20 | 北京华科仪科技股份有限公司 | Automatic control method and system of rotary disc type biological reaction device |
| CN117509928A (en) * | 2023-11-03 | 2024-02-06 | 安徽泛湖生态科技股份有限公司 | OTE real-time monitoring-based aeration optimization energy-saving control system and method |
| CN118131605A (en) * | 2024-05-07 | 2024-06-04 | 浙江信合阀门有限公司 | Intelligent ball valve control system |
| CN118131605B (en) * | 2024-05-07 | 2024-07-23 | 浙江信合阀门有限公司 | Intelligent ball valve control system |
| CN118348811A (en) * | 2024-06-14 | 2024-07-16 | 江西交科交通工程有限公司 | Treatment facility scheduling method and system based on wharf initial rainwater collection |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109928493A (en) | It is a kind of that method is controlled to adjust based on the accurate dissolved oxygen of big data and evolution algorithm | |
| Jeppsson et al. | Benchmark simulation model no 2: general protocol and exploratory case studies | |
| Fernández et al. | Reduction of aeration costs by tuning a multi-set point on/off controller: A case study | |
| Han et al. | Model predictive control of dissolved oxygen concentration based on a self-organizing RBF neural network | |
| CN105404151B (en) | Sewage disposal process dynamic multi-objective optimization control method | |
| Han et al. | Nonlinear multiobjective model-predictive control scheme for wastewater treatment process | |
| Lee et al. | Neural network modeling for on-line estimation of nutrient dynamics in a sequentially-operated batch reactor | |
| CN106873379A (en) | A kind of sewage disposal method for optimally controlling based on iteration ADP algorithms | |
| CN112099354B (en) | Intelligent multi-objective optimization control method for sewage treatment process | |
| bin Mohd Zain et al. | Optimization of fed-batch fermentation processes using the Backtracking Search Algorithm | |
| CN102411308A (en) | Self-adaptive control method of dissolved oxygen based on recurrent neural network model | |
| CN113189881A (en) | Multi-objective optimization control method and system for sewage treatment | |
| CN109657790A (en) | A kind of Recurrent RBF Neural Networks water outlet BOD prediction technique based on PSO | |
| CN117390549A (en) | Aeration control method for town sewage plant | |
| CN111798134B (en) | Method for improving methane yield of sewage treatment plant based on data-driven model | |
| Civelekoglu et al. | Modelling of COD removal in a biological wastewater treatment plant using adaptive neuro-fuzzy inference system and artificial neural network | |
| Salles et al. | Dynamic setpoint optimization using metaheuristic algorithms for wastewater treatment plants | |
| Baeza et al. | In-line fast OUR (oxygen uptake rate) measurements for monitoring and control of WWTP | |
| CN114488822B (en) | Intelligent sewage treatment process optimization method based on high-dimensional multi-objective evolutionary algorithm | |
| Francisco et al. | A comparative study of deterministic and stochastic optimization methods for integrated design of processes | |
| CN111484124A (en) | Intelligent control and biochemical process intensive method and system for sewage treatment | |
| CN110412878B (en) | PID-RBF steady state estimation and LSSVM soft measurement sewage BOD control method | |
| CN101794117B (en) | Method for optimizing and controlling operation of membrane bioreactor based on mechanism model | |
| KR20230086851A (en) | A hybrid machine learningebased multiobjective supervisory control strategy of a fullscale wastewater treatment for costeffective and sustainable operation under varying infuent conditions | |
| CN116520690A (en) | Sewage treatment control system and method for large-scale coal chemical engineering project multiple SBR (sequencing batch reactor) tanks |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190625 |
|
| WD01 | Invention patent application deemed withdrawn after publication |