CN107402586A - Dissolved Oxygen concentration Control method and system based on deep neural network - Google Patents
Dissolved Oxygen concentration Control method and system based on deep neural network Download PDFInfo
- Publication number
- CN107402586A CN107402586A CN201710756466.8A CN201710756466A CN107402586A CN 107402586 A CN107402586 A CN 107402586A CN 201710756466 A CN201710756466 A CN 201710756466A CN 107402586 A CN107402586 A CN 107402586A
- Authority
- CN
- China
- Prior art keywords
- neural network
- deep neural
- dissolved oxygen
- data
- parameter
- 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
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D11/00—Control of flow ratio
- G05D11/02—Controlling ratio of two or more flows of fluid or fluent material
- G05D11/13—Controlling ratio of two or more flows of fluid or fluent material characterised by the use of electric means
Abstract
The invention discloses a kind of Dissolved Oxygen concentration Control method and system based on deep neural network, methods described includes:Detect and gather multigroup initial data in waste water;Deep neural network model is established according to the initial data;Control parameter is calculated using the deep neural network model;According to the control parameter, aeration quantity is adjusted.Dissolved oxygen model of the invention by establishing deep neural network, realize and dissolved oxygen is precisely controlled.
Description
Technical field
The present invention relates to control technology field, more particularly to a kind of Dissolved Oxygen concentration Control method and system.
Background technology
Dissolved oxygen be sewage aerobic biological treatment system operation key factor, oxygen supply total amount directly determine sewage disposal into
This, dissolved oxygen concentration level is too low, reduces activity of activated sludge, suppresses degraded of the biology to organic matter, produces sludge bulking;
Dissolved oxygen is too high to be accelerated to consume the organic matter in sewage, microorganism is caused the aging of activated sludge due to a lack of nutrition, long
Phase too high dissolved oxygen can reduce the flocculating property and adsorption capacity of activated sludge, increase energy consumption, cause suspended solid settleability
It is deteriorated.Therefore, the control of dissolved oxygen is extremely important.
The concentration control of sewage active sludge aerobic processing system dissolved oxygen is with complicated non-linear, hysteresis quality and certain
Uncertainty, therefore, take into full account processing load in the case of, seek one kind by controlling aeration quantity and then to dissolved oxygen
Concentration realizes the method being precisely controlled, and has very for improving the gentle effect of Automated water of sewage disposal control, saving the energy
Big meaning.
At present, the control to sewage treatment plant's aeration quantity relies primarily on design standard and the experience of technician is come artificially
Implement, still, flow of inlet water and water quality parameter change at any time, and this change has very big randomness, and this causes to sewage
The operation control difficulty of factory is very big.In addition, the sewage disposal system that sewage treatment plant uses at present is a kind of multivariable nonlinearity
System, concentration control, sewage quality and the service condition relation complexity of dissolved oxygen, it is difficult to described with linear relationship, therefore nothing
Method establishes corresponding model with the method for traditional Analysis on Mechanism and mathematical derivation.
The content of the invention
For in the prior art the defects of, the present invention provides a kind of Dissolved Oxygen concentration Control side based on deep neural network
Method and system, by applying the dissolved oxygen model based on deep neural network, by realizing control aeration quantity to dissolving
Oxygen is precisely controlled.
In a first aspect, the invention provides a kind of Dissolved Oxygen concentration Control method based on deep neural network, the side
Method includes:
Detect and gather multigroup initial data in waste water;
Deep neural network model is established according to the initial data;
Control parameter is calculated using the deep neural network model;
According to the control parameter, aeration quantity is adjusted.
Further, it is described that deep neural network model is established according to the initial data, specifically include:
A part of data in the initial data are taken out as training sample;
Call the parameter training function of the deep neural network model;
Using the parameter training function, the training sample is learnt and trained;
Deep neural network model is established according to the result of study and training.
Further, it is described that deep neural network model is established according to the initial data, in addition to:Utilize the depth
Neural network model carries out the emulation of wastewater treatment.
Further, the emulation that wastewater treatment is carried out using the deep neural network model, is specifically included:
Remaining data are taken out in the initial data as test sample;
Call the simulated function of the deep neural network model;
Sewage disposal is emulated using the simulated function and the test sample, obtains prediction data;
The Stability and veracity of the model is detected using the prediction data.
Further, the Stability and veracity of the proof depth neural network model, is specifically included:
The variation tendency of the prediction data and measured data is contrasted, examines error between the two.
Further, it is described to calculate control parameter using the deep neural network model, specifically include:
The input parameter of the model is set, the output parameter corresponding with input parameter is calculated using the model, is made
For control parameter.
Further, the input parameter includes:Wastewater flow, hydraulic detention time, pollutant concentration of intaking, water outlet are dirty
Contaminate thing concentration, activated sludge concentration, dissolved oxygen concentration;The output parameter is aeration quantity.
Further, it is described to detect and after gathering multigroup initial data in waste water, establish deep neural network model
Before, in addition to:
Efficiency analysis is carried out to the initial data, therefrom extracted valid data, described in valid data renewal
Initial data.
Further, the initial data includes:Wastewater flow, hydraulic detention time, pollutant concentration of intaking, water outlet are dirty
Contaminate thing concentration, activated sludge concentration, aeration quantity parameter and corresponding dissolved oxygen concentration.
Second aspect, it is described present invention also offers a kind of Dissolved Oxygen concentration Control system based on deep neural network
System includes:Data acquisition module, model building module, parameter calculating module, control module;
The data acquisition module, for detecting and gathering multigroup initial data in waste water;
The model building module, for establishing deep neural network model according to the initial data;
The parameter calculating module, for calculating control parameter using the deep neural network model;
The control module, for according to the control parameter, being adjusted to aeration quantity.
As shown from the above technical solution, the present invention provides a kind of Dissolved Oxygen concentration Control method based on deep neural network
And system, deep neural network is applied to dissolved oxygen monitoring with Controlling model, establishing the dissolving based on deep neural network
Oxygen model, by being precisely controlled to control aeration quantity with realizing to dissolved oxygen.
Brief description of the drawings
Fig. 1 shows the schematic flow sheet of Dissolved Oxygen concentration Control method provided by the invention.
Fig. 2 shows the schematic network structure of restricted Boltzmann machine.
Fig. 3 shows the schematic flow sheet that deep neural network model is established in the present invention.
Fig. 4 shows the schematic flow sheet emulated in the present invention using deep neural network model.
Fig. 5 shows the structural representation of Dissolved Oxygen concentration Control system provided by the invention.
Embodiment
The embodiment of technical solution of the present invention is described in detail below in conjunction with accompanying drawing.Following examples are only used for
Clearly illustrate technical scheme, therefore be intended only as example, and the protection of the present invention can not be limited with this
Scope.
Embodiment one
Fig. 1 shows the stream for the Dissolved Oxygen concentration Control method based on deep neural network that the embodiment of the present invention one provides
Journey schematic diagram.As shown in figure 1, methods described includes:
Step S100, detect and gather multigroup initial data in waste water;
Step S200, deep neural network model is established according to the initial data;
Step S300, control parameter is calculated using the deep neural network model;
Step S400, according to the control parameter, aeration quantity is adjusted.
The concrete technical scheme of the embodiment of the present invention one is:
Step S100, detect and gather multigroup initial data in waste water.
The initial data includes but is not limited to:Wastewater flow, hydraulic detention time, pollutant concentration of intaking, water outlet are dirty
Contaminate thing concentration, activated sludge concentration, aeration quantity parameter and corresponding dissolved oxygen concentration.
After collecting initial data, need to carry out efficiency analysis to initial data if necessary, therefrom extracted valid data,
The initial data is updated with the valid data.
The concrete mode of efficiency analysis is:Initial data is contrasted with the correction data monitored in advance, examines two
The error size of person, if a certain initial data, compared with corresponding correction data, error is excessive, or more than some setting
Threshold value, then it is assumed that the initial data is invalid data, on the contrary then be valid data, after completing all contrasts, is had with what is filtered out
Data are imitated to replace initial data, turn into new initial data, in case subsequent step uses.
Step S200, deep neural network model is established according to the initial data.
The principle of the embodiment of the present invention is:Deep neural network is applied in dissolved oxygen monitoring and Controlling model, specifically
For, utilize dissolved oxygen model of the repetitive exercise Algorithm for Training in deep neural network based on deep neural network.
In general, deep neural network model and shallow-layer neural network model all use hierarchy, both include
Input layer, hidden layer and output layer, there is connection between adjacent layer, do not connected between same layer and the node of cross-layer.Compared to
Shallow-layer neural network model, deep neural network model are adjusted using the method rather than back-propagation algorithm successively trained in feedback
When whole, gradient is more and more sparse, and, error correction signal is less and less from top layer more down, and restrains easy Local Minimum,
This is mainly closely related without label with the randomness of initial value and data.
Deep neural network (Deep Neural Network, DNN) refers to having the expression of multilayer hidden layer non-linear
The deep structure of relation, the function of complexity can be approached in theory using the model structure.Each neural unit of deep neural network
Between represent is a kind of non-linear relation, can be used for approaching complicated function and being fitted Observable data, in learning data
Powerful modeling ability and Extracting Ability are shown on substantive characteristics.
Meanwhile deep neural network has multiple hidden layers, the network of more hidden layers has excellent feature learning ability, this
Learn obtained feature for data have it is more essential portray, beneficial to classification.The essence of deep learning is how hidden by building
Layer is hidden, learns more useful feature using substantial amounts of training data, so as to improve the accuracy of prediction, therefore, depth model
It is means, feature learning is purpose.Complexity can be achieved by learning the nonlinear network structure of deep layer in deep neural network model
Function approximation represents the distributed expression of input data.
Because the influence factor of dissolved oxygen is various and complicated, the modeling ability of the model structure such as gauss hybrid models of shallow-layer
It is limited, higher order dependencies that can not be between accurate expression dissolved oxygen and its influence factor.The powerful modeling of deep neural network
Ability is then more suitable for the dissolved oxygen data modeling to complexity.
It is respectively depth conviction pessimistic concurrency control that deep neural network common model, which has three kinds, convolutional neural networks model, automatic
Encoder model, the dissolved oxygen concentration in waste water is predicted using depth conviction net by the present invention and control model.
Depth conviction net (DBN) model is laminated by multiple restricted Boltzmann machines (RBM), and RBM is Deep
Learning-individual critically important network structure.Depth belief network is a kind of generation model, it by train neuron it
Between weight carry out training data.
Limited Boltzmann machine is made up of two layers of neuron, and one layer is visual layers (V), and another layer is hidden layer (H).Wherein
Every layer of node is no link, what the node of layer and interlayer linked entirely.As shown in Figure 2.Wherein, n represents the number of neuron,
V, h represent the state vector of visible layer and hidden layer, and a, b represent the bias vector of visible layer and hidden layer, W be visible layer with it is hidden
Hide the weight matrix between layer.
RBM is a model based on energy, and for RBM, the energy function between hidden layer and visible layer is:
According to energy function, the joint probability between visible layer and hidden layer can be provided:
Wherein Z is normalization factor:
For a practical problem, it is necessary to know the probability distribution of observation data, the edge distribution of its corresponding p (v, h), tool
Body is:
Similar, it can obtain:
Consider afterwards when all neuron states on given hidden layer (or visible layer), it is seen that layer (or hide
Layer) probability that some neural unit is activated, in order to simplify derivation, it is denoted as:
Above formula represents the last year h in hkThe vector obtained afterwards, is subsequently introduced:
It can be obtained using two formulas above:
E (v, h)=- β (v, h-k)-hkak(v)
Then by deriving, obtain:
DBN is made up of multiple RBM layers, and DBN training process is as follows:
(1) first RBM is trained up first;
(2) weight for the RBM that first is trained and offset are fixed, and second is used as using the state of recessive neuron
Individual RBM input vector;
(3) second RBM is trained up, is then stacked on first RBM top;
(4) it is multiple to repeat above step;
(5) if the packet in training set contains label, then the RBM of top will have representative when training
The neuron of tag along sort is trained together.
, it is necessary to according to actual conditions to the flow of inlet water Q of waste water when in the scene described in the application embodiment of the present inventionw、
Pollutant concentration Cin, hydraulic detention time T, activated sludge concentration MLSS, processing water outlet pollutant concentration Ceff, aeration quantity QgWith
(it is single DO values that integral type continuously stirs formula reaction member to dissolved oxygen concentration DO, and gallery type reaction member is divided into head end by along journey
DO1, middle-end DO2With end DO3) carry out one group of sample measures (Qw、Cin、T、MLSS、Ceff, DO is respectively as input vector x1、x2、
x3、x4、x5And x6, aeration quantity QgAs output vector), that is, one group of input, output sample are obtained, then the DNN is trained,
So that control errors are within the specific limits, that is, obtaining one stablizes available DNN networks, and this network can be used in waste water
Dissolved oxygen concentration is monitored and controlled it by changing aeration quantity, for according to Qw、Cin、T、MLSS、CeffWith DO pairs
Aeration quantity is controlled, the computation model by for realize wastewater biochemical handle in dissolved oxygen be precisely controlled provide it is great operable
The scientific guidance opinion of property.
It is the principle that DNN models are established by step S200 above, it is preferable that as shown in figure 3, step S200 is specifically wrapped
Include:
Step S210, a part of data in the initial data are taken out as training sample;
Step S220, call the parameter training function of the deep neural network (DNN) model;
Step S230, using the parameter training function, the training sample is learnt and trained;
Step S240, DNN models are established according to the result of study and training;
Step S250, the emulation of wastewater treatment is carried out using the DNN models.
Wherein, in order to more effectively train the model, the training sample taken out from initial data should be not less than 500 groups,
After taking out training sample, remaining initial data can be used as test sample.
For each group of training sample of taking-up, input data and output data are classified as, wherein, input data includes
Wastewater flow, hydraulic detention time, intake pollutant concentration, water outlet pollutant concentration, activated sludge concentration and dissolved oxygen concentration,
Output data is aeration quantity.
Learnt and trained using repetitive exercise algorithm according to input data and output data, by control errors certain
In the range of, stablize available DNN models so as to obtain one, the model reflects input data and inputs out the inherence of sample
Contact, can be used for being monitored the dissolved oxygen concentration in sewage, and control it by changing aeration quantity.
In order to eliminate influence of the data of different measurement units to network model, the study and training that make network effectively may be used
Lean on, input data and output data can be normalized, such as normalized to [0,1].
It is further preferred that as shown in figure 4, step S250 is specifically included:
Step S251, remaining data are taken out in the initial data as test sample;
Step S252, call the simulated function of the deep neural network model;
Step S253, sewage disposal is emulated using the simulated function and the test sample, obtain predicting number
According to;
Step S254, the Stability and veracity of the model is detected using the prediction data.
Wherein, the step S254 is specially:The variation tendency of obtained prediction data and measured data is contrasted,
Compare error size between the two, for detecting and evaluating the Stability and veracity of the deep neural network model,
Fitness of generalization, data for detection model etc..If error is larger, the accuracy for the DNN models established and steady
It is qualitative poor, it is necessary to relearn, train and establish model by step S200;If error is smaller, the DNN models are proved
Stability and veracity is preferable, available for subsequent applications.
Wherein, the prediction data and measured data are aeration quantity.
Step S300, control parameter is calculated using the deep neural network model.
Specifically, in this step, the input parameter for setting model according to being actually needed, is calculated accordingly by model
Output parameter;Wherein, input parameter includes wastewater flow, hydraulic detention time, pollutant concentration of intaking, it is dense to go out water pollutant
Degree, activated sludge concentration and dissolved oxygen concentration;Output parameter is aeration quantity.
Step S400, according to the control parameter, aeration quantity is adjusted.
Specifically, the output of aeration quantity is adjusted according to output parameter, reaches the numerical value indicated by output parameter.Pass through
Aeration quantity is adjusted, the ratio of air in waste water can be changed, so as to reach the purpose of adjustment dissolved oxygen.
Based on above content, the technique effect that the embodiment of the present invention one can be realized is:Using deep neural network to dirt
Dissolved oxygen depth is modeled in water treatment procedure, can it is very flexible according to Practical Project build one meet concrete engineering will
The stable model asked, the aeration quantity in aerator more accurately can be predicted and control by the foundation of stable model, from
And being precisely controlled to the dissolved oxygen concentration in waste water is realized, improve the automatization level of sewage disposal, the effectively save energy.
Embodiment two
To the embodiment of the present invention one accordingly, Fig. 5 shows that one kind provided in an embodiment of the present invention is based on depth nerve net
The structural representation of the Dissolved Oxygen concentration Control system of network.As shown in figure 5, the system includes:Data acquisition module 101, mould
Type establishes module 102, parameter calculating module 103, control module 104;
The data acquisition module 101, preferably intelligent gateway device, it is multigroup original in waste water for detecting and gathering
Data.The data acquisition module 101 is additionally operable to carry out efficiency analysis to initial data if necessary, therefrom extracted valid data
To replace initial data.
Wherein, the initial data includes but is not limited to:Wastewater flow, hydraulic detention time, pollutant concentration of intaking, go out
Water pollutant concentration, activated sludge concentration, aeration quantity parameter and corresponding dissolved oxygen concentration.
The model building module 102, for establishing deep neural network model according to the initial data;And specifically use
In:A part of data in the initial data are taken out as training sample;Call the deep neural network (DNN) model
Parameter training function;Using the parameter training function, the training sample is learnt and trained;According to study and training
Result establish DNN models;The emulation of wastewater treatment is carried out using the DNN models.
The parameter calculating module 103, for calculating control parameter using the deep neural network model.Specifically,
The parameter calculating module 103 sets the input parameter of model, and Boot Model calculates corresponding output parameter.Wherein, input
Parameter includes wastewater flow, hydraulic detention time, pollutant concentration of intaking, water outlet pollutant concentration, activated sludge concentration and molten
Solve oxygen concentration;Output parameter is aeration quantity.
The control module 104, for according to the control parameter, being adjusted to aeration quantity.
Wherein, the aeration quantity is mainly controlled by air blower and pipeline valve, and the control module 104 receives control
After parameter, the air output of air blower, or the open degree of control pipeline valve, or both control simultaneously are adjusted according to the control parameter
Collective effect, make to reach numerical value indicated by control parameter by the gas flow of air blower and pipeline valve, so that gas can essence
Useless oxygen in water is adjusted accurately to required numerical value.
Based on above content, what the embodiment of the present invention two can reach has the technical effect that:It is right by model building module 102
Dissolved oxygen depth is modeled in sewage disposal process, flexibly can be built one according to Practical Project very and be met concrete engineering
It is required that stable model, more accurately can predict and control the aeration quantity in aerator by the foundation of stable model,
So as to realize being precisely controlled to the dissolved oxygen concentration in waste water, the automatization level of sewage disposal, the effectively save energy are improved.
Technical solution of the present invention is applicable not only to the control to dissolved oxygen in sewage disposal, is also applied for other similar works
Journey is put into practice.
Finally it should be noted that:Various embodiments above is merely illustrative of the technical solution of the present invention, rather than its limitations;To the greatest extent
The present invention is described in detail with reference to foregoing embodiments for pipe, it will be understood by those within the art that:Its according to
The technical scheme described in foregoing embodiments can so be modified, either which part or all technical characteristic are entered
Row equivalent substitution;And these modifications or replacement, the essence of appropriate technical solution is departed from various embodiments of the present invention technology
The scope of scheme, it all should cover among the claim of the present invention and the scope of specification.
Claims (10)
1. the Dissolved Oxygen concentration Control method based on deep neural network, it is characterised in that methods described includes:
Detect and gather multigroup initial data in waste water;
Deep neural network model is established according to the initial data;
Control parameter is calculated using the deep neural network model;
According to the control parameter, aeration quantity is adjusted.
2. Dissolved Oxygen concentration Control method according to claim 1, it is characterised in that described to be built according to the initial data
Vertical deep neural network model, is specifically included:
A part of data in the initial data are taken out as training sample;
Call the parameter training function of the deep neural network model;
Using the parameter training function, the training sample is learnt and trained;
Deep neural network model is established according to the result of study and training.
3. Dissolved Oxygen concentration Control method according to claim 2, it is characterised in that described to be built according to the initial data
Vertical deep neural network model, in addition to:The emulation of wastewater treatment is carried out using the deep neural network model.
4. Dissolved Oxygen concentration Control method according to claim 3, it is characterised in that described to utilize the depth nerve net
Network model carries out the emulation of wastewater treatment, specifically includes:
Remaining data are taken out in the initial data as test sample;
Call the simulated function of the deep neural network model;
Sewage disposal is emulated using the simulated function and the test sample, obtains prediction data;
The Stability and veracity of the model is detected using the prediction data.
5. Dissolved Oxygen concentration Control method according to claim 4, it is characterised in that the proof depth neutral net mould
The Stability and veracity of type, is specifically included:
The variation tendency of the prediction data and measured data is contrasted, examines error between the two.
6. Dissolved Oxygen concentration Control method according to claim 1, it is characterised in that described to utilize the depth nerve net
Network model calculates control parameter, specifically includes:
The input parameter of the model is set, the output parameter corresponding with input parameter is calculated using the model, as control
Parameter processed.
7. Dissolved Oxygen concentration Control method according to claim 6, it is characterised in that the input parameter includes:Waste water
Flow, hydraulic detention time, pollutant concentration of intaking, water outlet pollutant concentration, activated sludge concentration, dissolved oxygen concentration;It is described
Output parameter is aeration quantity.
8. Dissolved Oxygen concentration Control method according to claim 1, it is characterised in that described to detect and gather in waste water
After multigroup initial data, establish before deep neural network model, in addition to:
Efficiency analysis is carried out to the initial data, therefrom extracted valid data, updated with the valid data described original
Data.
9. Dissolved Oxygen concentration Control method according to any one of claim 1 to 3, it is characterised in that the original number
According to including:Wastewater flow, hydraulic detention time, pollutant concentration of intaking, water outlet pollutant concentration, activated sludge concentration, aeration
Measure parameter and corresponding dissolved oxygen concentration.
10. the Dissolved Oxygen concentration Control system based on deep neural network, it is characterised in that the system includes:Data acquisition
Module, model building module, parameter calculating module, control module;
The data acquisition module, for detecting and gathering multigroup initial data in waste water;
The model building module, for establishing deep neural network model according to the initial data;
The parameter calculating module, for calculating control parameter using the deep neural network model;
The control module, for according to the control parameter, being adjusted to aeration quantity.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710756466.8A CN107402586A (en) | 2017-08-29 | 2017-08-29 | Dissolved Oxygen concentration Control method and system based on deep neural network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710756466.8A CN107402586A (en) | 2017-08-29 | 2017-08-29 | Dissolved Oxygen concentration Control method and system based on deep neural network |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107402586A true CN107402586A (en) | 2017-11-28 |
Family
ID=60397578
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710756466.8A Pending CN107402586A (en) | 2017-08-29 | 2017-08-29 | Dissolved Oxygen concentration Control method and system based on deep neural network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107402586A (en) |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108133289A (en) * | 2017-12-21 | 2018-06-08 | 中国铁建电气化局集团有限公司 | Tunnel ventilation control method and system based on environmental forecasting |
CN109019892A (en) * | 2018-08-13 | 2018-12-18 | 重庆工商大学 | A kind of regulation method based on data assimilation on-line optimization aeration quantity |
CN109214513A (en) * | 2018-10-23 | 2019-01-15 | 江南大学 | A kind of solid liquid waste intelligence coupled processing method based on adaptive depth confidence network |
CN110197704A (en) * | 2019-04-18 | 2019-09-03 | 北京工业大学 | A kind of self-organizing BP neural network water outlet total phosphorus prediction technique based on limited Boltzmann machine |
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 |
CN111680449A (en) * | 2020-05-12 | 2020-09-18 | 埃睿迪信息技术(北京)有限公司 | Method for realizing accurate aeration based on deep learning algorithm model |
CN112782975A (en) * | 2020-12-02 | 2021-05-11 | 海天水务集团股份公司 | Sewage treatment aeration intelligent control method and system based on deep learning |
CN113830896A (en) * | 2021-09-27 | 2021-12-24 | 湖南工商大学 | Aerobic segment oxygen early warning method, system and equipment based on graph neural network |
CN114573096A (en) * | 2022-03-25 | 2022-06-03 | 南京大学 | Sewage treatment aeration system control method based on microbial community |
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 |
CN116116181A (en) * | 2023-04-18 | 2023-05-16 | 科扬环境科技有限责任公司 | Model optimization-based waste gas and wastewater treatment method and device |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO1998003434A1 (en) * | 1996-07-19 | 1998-01-29 | Mitsubishi Chemical Corporation | Device for controlling dissolved oxygen concentration of aeration tank, device for controlling temperature of aeration tank, device for controlling flow rate of raw water for homogeneous-flow liquid surface, and wastewater treatment equipment used in activated sludge process |
CN101576734A (en) * | 2009-06-12 | 2009-11-11 | 北京工业大学 | Dissolved oxygen control method based on dynamic radial basis function neural network |
CN102183621A (en) * | 2011-02-28 | 2011-09-14 | 中国农业大学 | Aquaculture dissolved oxygen concentration online forecasting method and system |
CN103064290A (en) * | 2013-01-01 | 2013-04-24 | 北京工业大学 | Dissolved oxygen model prediction control method based on self-organization radial basis function neural network |
CN103499982A (en) * | 2013-09-30 | 2014-01-08 | 北京工业大学 | Self-organization control method of sewage treatment process |
CN107071716A (en) * | 2017-06-14 | 2017-08-18 | 扬州大学 | Monitoring water environment system and method based on TD-LTE private networks |
-
2017
- 2017-08-29 CN CN201710756466.8A patent/CN107402586A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO1998003434A1 (en) * | 1996-07-19 | 1998-01-29 | Mitsubishi Chemical Corporation | Device for controlling dissolved oxygen concentration of aeration tank, device for controlling temperature of aeration tank, device for controlling flow rate of raw water for homogeneous-flow liquid surface, and wastewater treatment equipment used in activated sludge process |
CN101576734A (en) * | 2009-06-12 | 2009-11-11 | 北京工业大学 | Dissolved oxygen control method based on dynamic radial basis function neural network |
CN102183621A (en) * | 2011-02-28 | 2011-09-14 | 中国农业大学 | Aquaculture dissolved oxygen concentration online forecasting method and system |
CN103064290A (en) * | 2013-01-01 | 2013-04-24 | 北京工业大学 | Dissolved oxygen model prediction control method based on self-organization radial basis function neural network |
CN103499982A (en) * | 2013-09-30 | 2014-01-08 | 北京工业大学 | Self-organization control method of sewage treatment process |
CN107071716A (en) * | 2017-06-14 | 2017-08-18 | 扬州大学 | Monitoring water environment system and method based on TD-LTE private networks |
Cited By (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108133289A (en) * | 2017-12-21 | 2018-06-08 | 中国铁建电气化局集团有限公司 | Tunnel ventilation control method and system based on environmental forecasting |
CN109019892A (en) * | 2018-08-13 | 2018-12-18 | 重庆工商大学 | A kind of regulation method based on data assimilation on-line optimization aeration quantity |
CN109214513A (en) * | 2018-10-23 | 2019-01-15 | 江南大学 | A kind of solid liquid waste intelligence coupled processing method based on adaptive depth confidence network |
CN109214513B (en) * | 2018-10-23 | 2021-08-13 | 江南大学 | Solid-liquid waste intelligent coupling treatment method based on adaptive deep belief network |
CN110197704B (en) * | 2019-04-18 | 2021-06-04 | 北京工业大学 | Self-organizing BP neural network water output total phosphorus prediction method based on limited Boltzmann machine |
CN110197704A (en) * | 2019-04-18 | 2019-09-03 | 北京工业大学 | A kind of self-organizing BP neural network water outlet total phosphorus prediction technique based on limited Boltzmann machine |
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 |
CN111680449A (en) * | 2020-05-12 | 2020-09-18 | 埃睿迪信息技术(北京)有限公司 | Method for realizing accurate aeration based on deep learning algorithm model |
CN112782975A (en) * | 2020-12-02 | 2021-05-11 | 海天水务集团股份公司 | Sewage treatment aeration intelligent control method and system based on deep learning |
CN113830896A (en) * | 2021-09-27 | 2021-12-24 | 湖南工商大学 | Aerobic segment oxygen early warning method, system and equipment based on graph neural network |
CN113830896B (en) * | 2021-09-27 | 2022-12-27 | 湖南工商大学 | Aerobic segment oxygen early warning method, system and equipment based on graph neural network |
CN114573096A (en) * | 2022-03-25 | 2022-06-03 | 南京大学 | Sewage treatment aeration system control method based on microbial community |
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 |
CN116116181A (en) * | 2023-04-18 | 2023-05-16 | 科扬环境科技有限责任公司 | Model optimization-based waste gas and wastewater treatment method and device |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107402586A (en) | Dissolved Oxygen concentration Control method and system based on deep neural network | |
CN105510546B (en) | A kind of biochemical oxygen demand (BOD) BOD intelligent detecting methods based on self-organizing Recurrent RBF Neural Networks | |
CN104376380B (en) | A kind of ammonia nitrogen concentration Forecasting Methodology based on recurrence self organizing neural network | |
Han et al. | Data-driven intelligent monitoring system for key variables in wastewater treatment process | |
CN107358021B (en) | DO prediction model establishment method based on BP neural network optimization | |
CN103744293B (en) | Wastewater treatment monitoring method and system based on fuzzy neural network | |
CN104007243B (en) | Application back propagation artificial neural network model evaluates the method for swimming pool water | |
CN110824915B (en) | GA-DBN network-based intelligent monitoring method and system for wastewater treatment | |
CN104182794B (en) | Method for soft measurement of effluent total phosphorus in sewage disposal process based on neural network | |
CN108469507A (en) | A kind of water outlet BOD flexible measurement methods based on Self organizing RBF Neural Network | |
CN102854296A (en) | Sewage-disposal soft measurement method on basis of integrated neural network | |
CN103728431A (en) | Industrial sewage COD (chemical oxygen demand) online soft measurement method based on ELM (extreme learning machine) | |
CN107664682A (en) | A kind of water quality hard measurement Forecasting Methodology of ammonia nitrogen | |
KR102311657B1 (en) | Smart management system for wastewater treatment | |
CN114037163A (en) | Sewage treatment effluent quality early warning method based on dynamic weight PSO (particle swarm optimization) optimization BP (Back propagation) neural network | |
CN109657790A (en) | A kind of Recurrent RBF Neural Networks water outlet BOD prediction technique based on PSO | |
CN104063609B (en) | Method of assisting in judging pollution source monitoring data validity by utilizing neural network | |
Mubarak et al. | Metro-environmental data approach for the prediction of chemical oxygen demand in new Nicosia wastewater treatment plant | |
CN115713448A (en) | Catchment unit-based method for quickly tracing conventional factors of basin water pollution | |
CN112417765A (en) | Sewage treatment process fault detection method based on improved teacher-student network model | |
CN107665288A (en) | A kind of water quality hard measurement Forecasting Methodology of COD | |
CN106769748B (en) | Intelligent detection system for water permeability of membrane bioreactor-MBR (Membrane bioreactor) | |
CN116029612A (en) | Precise control method for additional carbon source of town sewage plant based on deep learning | |
Zhang et al. | Effluent Quality Prediction of Wastewater Treatment System Based on Small-world ANN. | |
CN107664683A (en) | A kind of water quality hard measurement Forecasting Methodology of total nitrogen |
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 |