CN106292296B - Water island dosing On-Line Control Method and device based on GA SVM - Google Patents
Water island dosing On-Line Control Method and device based on GA SVM Download PDFInfo
- Publication number
- CN106292296B CN106292296B CN201610941724.5A CN201610941724A CN106292296B CN 106292296 B CN106292296 B CN 106292296B CN 201610941724 A CN201610941724 A CN 201610941724A CN 106292296 B CN106292296 B CN 106292296B
- Authority
- CN
- China
- Prior art keywords
- svm
- parameter
- water
- svm models
- max
- 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.)
- Expired - Fee Related
Links
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
- G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
- G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
- G05B13/042—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators in which a parameter or coefficient is automatically adjusted to optimise the performance
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Separation Of Suspended Particles By Flocculating Agents (AREA)
Abstract
The invention discloses a kind of water island dosing On-Line Control Method and device based on GA SVM.This method includes：Obtain that effluent index is satisfactory, water monitoring index M group historical parameter values, and the adding medicine corresponding to every group of historical parameter value history chemical feeding quantity；On-line Control SVM models are set up based on SVMs, wherein, historical parameter value is the input vector of SVM models, and history chemical feeding quantity is the output vector of SVM models；GA algorithms are performed to SVM models, the optimal solution of the parameter in SVM models is solved；Obtained optimal solution input SVM models will be solved, obtain the dosing model on water island；Water monitoring index is monitored in real time to obtain one group of real-time parameter value；And real-time parameter value is inputted to the dosing model on water island, to determine the chemical feeding quantity of current time adding medicine.By the way that method of the invention, it is possible to overcome influence of the water water quality to running, real-time update chemical feeding quantity reduces medicament and wastes, reduces cost.
Description
Technical field
The invention belongs to thermal power plant's water-treatment technology field, and in particular to a kind of water island dosing based on GA-SVM is in line traffic control
Method and apparatus processed, available for automatically controlling chemical feeding quantity.
Background technology
Thermal power plant abbreviation thermal power plant, be by the use of coal, oil, natural gas as fuel production electric energy factory, it
The basic process of production is：Combustion heating water produces steam to fuel in the boiler, and the chemical energy of fuel is transformed into heat energy, Ran Houyou
Steam pressure pushing turbine rotates, and converts thermal energy into mechanical energy, and last steam turbine drives generator rotation, mechanical energy is turned
Become electric energy.
In modern more than 300MW thermal power plant, water treatment system realizes DCS automation control systems substantially.
Water treatment system is handled including condensate ammonification, feedwater ammonification plus hydrazine processing, chlorination processing, at the hydrochlorate that phosphorates of bubble stove
Reason etc., because water, the water-quality constituents of water differ greatly, and can occur a variety of reactions, especially between heterogeneity in water
Strong acid and strong base N-process has complicated nonlinear characteristic.
In order to reduce influence of the water water quality to the normal operation of thermal power plant, it can be adjusted by adding medicament.Current water
Processing system judge generally according to the effluent quality handled additive amount of medicament number, and then repeatedly debugging to meet will
Ask, the substantial amounts of time is wasted and artificial, with obvious hysteresis quality；Secondly adding medicine addition once it is determined that after, substantially
Belong to long term constant state, not only cause the invisible waste of medicine, and due to the time-varying characteristics of water water quality, effluent index without
Method meets the requirement of real time of user.
In summary, for the above-mentioned problems in the prior art, effective solution is not yet proposed at present.
The content of the invention
To solve the above problems, it is an object of the invention to provide a kind of water island dosing On-line Control side based on GA-SVM
Method and device, with solve chemical feeding quantity in current thermal power plant's water treatment procedure determine method inefficiency, medicine waste and can not
Meet the problem of standard water discharge is required.
According to one aspect of the present invention, there is provided a kind of water island dosing On-Line Control Method based on GA-SVM, the party
Method includes：Acquisition effluent index is satisfactory, water monitoring index M group historical parameter values, and corresponding to described in every group
The history chemical feeding quantity of the adding medicine of historical parameter value, wherein, the water monitoring index includes multiple indexs, history described in every group
Parameter value includes the parameter value of the multiple index；On-line Control SVM models are set up based on SVMs, wherein, it is described to go through
History parameter value is the input vector of the SVM models, and the history chemical feeding quantity is the output vector of the SVM models；To described
SVM models perform GA algorithms, solve the optimal solution of the parameter in the SVM models；Obtained optimal solution input will be solved
The SVM models, obtain the dosing model on water island；The water monitoring index is monitored in real time to obtain one group of real-time parameter value；
And the real-time parameter value is inputted to the dosing model on the water island, to determine the chemical feeding quantity of adding medicine described in current time.
The step of SVMs sets up On-line Control SVM models is further based on to specifically include：The SVM moulds are set
The input vector of type is the time series x of the data signal of the water monitoring indexi, set the output of the SVM models to
Measure as corresponding to the time series xiAdding medicine history chemical feeding quantity time series yi, wherein, i=1,2 ..., N, N ＜
M；Will be by the time series xiWith the time series yiTime series collection (the x of compositioni, yi) use Nonlinear MappingHold
Row and constructs the function of the SVM models and is to the mapping of higher dimensional spaceWherein, i=1,2 ...,
N, xi∈Rq, yi∈ R, q are the water monitoring index middle finger target number, and w is weight vector, and b is constant；Introduce slack variable ξ
And penalty factor, the object function for setting up the SVM models is Constraints isWherein, ξ >=0, C ＞ 0, i=1,2 ..., N；Determine the kernel function of the SVM models for radially
Basic functionWherein, σ ＞ 0；Setting up the SVM models is
Wherein, αi=C ξi, it is Lagrange coefficient.
Further, to SVM models execution GA algorithms, the step of the optimal solution of the parameter in the solution SVM models
Suddenly include：Step A：The parameter of GA algorithms is set, determines that initial population scale G, global iterative number of times, mutation probability, intersection are general
Rate, determines the decision variable that slack variable ξ, penalty factor and nuclear parameter σ are GA algorithms；Step B：It is 1 to make iterations, just
The parameter of the beginningization SVM models, and binary coding is carried out to the parameter of the SVM models after initialization, based on chaos
Theory produces G group initial populations；Step C：Using following circulation steps, slack variable ξ, penalty factor and nuclear parameter σ are determined
Optimal solution：Step C1：Each individual fitness function value, step C2 in population are calculated according to fitness function：Judge current change
Whether generation number reaches global iterative number of times, if reaching, exports optimum individual, end step C, if being not up to, adaptation
Degree functional value highest optimum individual is preserved, and records the minimum worst individual sequence number of fitness function value, makes iteration time
Number Jia 1, is selected, intersected, mutation genetic is operated, and replacing worst individual sequence number described in serial number with the optimum individual preserved
New individual, produce new population, return to step C1.
Further, the fitness function is For with the time series collection
(xi, yi) in i-th group of data as the input vector of the SVM models calculate obtained computer mould analog values,For the time
Sequence sets (xi, yi) in yiAverage value.
Further, the step of optimal solution for solving the parameter in the SVM models, also includes：Step D：Solution is obtained
The optimal solution as the SVM models parameter, by the time series x of the data signal of the water monitoring indexpIt is defeated
Enter the SVM models, obtain corresponding to the time series xpAdding medicine simulation chemical feeding quantity time seriesWherein, p
=N+1, N+2 ..., M；Step E：According to corresponding to the time series xpAdding medicine history chemical feeding quantity time series yp
With the time seriesCalculate the deterministic coefficient of the SVM models；Step F：Judge whether the deterministic coefficient is more than
Predetermined coefficient value, if being more than, the step of terminating to solve the optimal solution of the parameter in the SVM models, if being not more than, returns
Step A.
Further, step B is specifically included：Step B1, randomly selects 3N [0,1] value ranges and constitutes initiation sequence ε0=
(ε1,0, ε2,0..., ε3N, 0), ε is mapped by LogisticN, j+1=μ εN, j(1-εN, j), obtain the different chaos sequence in G track
Arrange εj, wherein, εN, 0For initial parameter, εN, j∈ (0,1), μ ∈ (0,4) are control parameter, n=1,2 ..., 3N-1, j=1,
2 ..., G；Step B2, defines decision variable scope Xmin, Xmax：
Xmin=[ξ1, min, ξ2, min..., ξN, min, C1, min, C2, min..., CN, min, σ1, min, σ2, min..., σN, min],
Xmax=[ξ1, max, ξ2, max..., ξN, max, C1, max, C2, max..., CN, max, σ1, max, σ2, max..., σN, max]；
Step B3, X is pressed by the chaos sequenceN, j=XN, min+(XN, max-XN, min)εN, jAmplify the value of each decision variable
Scope, obtains G and represents SVM model parameter Sequence composition initial populations：
(X1,1, X1,2..., X1,3N), (X2,1, X2,2..., X2,3N) ..., (XG, 1, XG, 2..., XG, 3N),
Wherein, XN, jFor n-th of decision variable value of j-th of body, XN, maxFor the maximum of n-th of body, XN, minFor n-th
The minimum value of body, j=1,2 ..., G, j=1,2 ..., 3N.
Further, in step C2, selection operation is selected using wheel disc, and the crossover operator of crossover operation uses single-point,
The mutation probability selection adaptive probability of mutation operation, wherein,
Crossover probability is calculated using below equation：
Mutation probability is calculated using below equation：
Wherein, f is the non-dominant grade of current individual,fmaxFor individual non-dominant grade in current population most
Big value, Pc1For constant, Pm1To limit the minimum value of mutation probability, k1∈ (0,1), k2∈ (0,1).
Further, ε in step B1N, 0Do not include 0,0.25,0.5,0.75 and 1.
Further, the water monitoring index includes water-carrying capacity, pH value, turbidity value, chloride ion content and ammonia-nitrogen content,
The adding medicine is sulfuric acid, sodium hydroxide, bactericide, flocculation aid, coagulant or reducing agent.
, should according to another aspect of the present invention there is provided a kind of water island dosing On-line Control device based on GA-SVM
Device includes：Historical data acquisition module, for obtaining, effluent index is satisfactory, M groups history of water monitoring index ginseng
Numerical value, and corresponding to the history chemical feeding quantity of the adding medicine of historical parameter value described in every group, wherein, the water monitoring index bag
Multiple indexs are included, historical parameter value includes the parameter value of the multiple index described in every group；Model building module, for based on branch
Hold vector machine and set up On-line Control SVM models, wherein, the historical parameter value is the input vector of the SVM models, described to go through
History chemical feeding quantity is the output vector of the SVM models；Parameter optimization module, for performing GA algorithms to the SVM models, is solved
The optimal solution of parameter in the SVM models, and the obtained optimal solution input SVM models will be solved, obtain water island
Dosing model；Real-time monitoring modular, one group of real-time parameter value for monitoring the water monitoring index in real time；And dosing
Control module is measured, the dosing model for the real-time parameter value to be inputted to the water island, to determine dosing described in current time
The chemical feeding quantity of agent.
By the solution of the present invention, the dosing mould on water island is set up using the data of history water monitoring index and chemical feeding quantity
Type, wherein setting up model using SVMs, and determines the parameter of model by GA algorithms, make use of the overall situation of GA algorithms to search
Suo Nengli, obtains more preferable SVMs parameter, enhances solution to model and releases and predictive ability, and dosing is carried out using the model
The determination of amount, overcomes the adverse effect that water water quality time-varying characteristics are caused to running, and can be according to current water quality
Situation dynamically updates optimal chemical feeding quantity in real time, realizes the On-line Control of dosing process, reduces the waste of medicament, reduces people
Work and operating cost, meet the water quality requirement of water outlet in real time.
Brief description of the drawings
Fig. 1 is the flow chart for the water island dosing On-Line Control Method based on GA-SVM that the embodiment of the present invention one is provided；
Fig. 2 is the stream of step 3 in the water island dosing On-Line Control Method based on GA-SVM that the embodiment of the present invention two is provided
Cheng Tu；
Fig. 3 to Fig. 7 is pH value in the water monitoring index that the embodiment of the present invention two is provided respectively, to come water-carrying capacity, water turbid
The change curve of degree, chloride ion content and ammonia-nitrogen content；
Fig. 8 is reality-matched curve figure of the chemical feeding quantity for the sulfuric acid that the embodiment of the present invention two is provided；
Fig. 9 is reality-prediction curve figure of the chemical feeding quantity for the sulfuric acid that the embodiment of the present invention two is provided；And
Figure 10 is the block diagram for the water island dosing On-line Control device based on GA-SVM that the embodiment of the present invention three is provided.
Embodiment
To make the object, technical solutions and advantages of the present invention clearer.Technical scheme will be carried out below
Clear complete description, it is clear that described embodiment is a part of embodiment of the present invention, rather than whole embodiments.It is based on
Embodiments of the invention, one of ordinary skill in the art obtained on the premise of creative work is not made it is all its
His embodiment, belongs to the scope of protection of the invention.
Before the embodiment of the present application is introduced, just method used herein is described as follows first：
SVMs (Support Vector Machine, SVM) is that Corinna Cortes and Vapnik etc. are proposed
A kind of solution small sample, the learning method of non-linear and high problem of dimension, can preferably gram with solid theoretical foundation
Take the intrinsic local minimum of neural net method, cross the selection too defect such as dependence experience of study and structure and type.
Genetic algorithm (Genetic Algorithm, GA) passes through to selecting, intersecting in biological heredity and evolutionary process, become
The imitation of different mechanism, to complete the adaptable search process to Optimum Solution.The automation that SVM parameters are chosen can be achieved, has
Effect improves precision of prediction.
Embodiment one
Reference picture 1, shows a kind of embodiment of the water island dosing On-Line Control Method based on GA-SVM, the embodiment pin
There is obvious hysteresis quality and constancy to current thermal power plant's water treatment dosing system, it is pre- with reference to SVM Nonlinear Time Series
Survey and the preferred characteristic of GA parameters, select the water monitoring index of history effluent index satisfactory correspondence period and add
Dose is sample, and sample training is carried out using SVM models, and application GA algorithms carry out the preferred of SVM model parameters, finally give
The dosing model on water island, by the real-time dosing of model realization Intelligent Dynamic, the method for the embodiment specifically includes following step
Rapid S102 to step S112.
Step S102：Acquisition effluent index is satisfactory, water monitoring index M group historical parameter values, and correspondingly
In the history chemical feeding quantity of the adding medicine of every group of historical parameter value.
Water monitoring index refers to the monitoring index during water inlet of water island, preferably includes to come water-carrying capacity, pH value, turbidity value, chlorine
Multiple indexs such as ion concentration and ammonia-nitrogen content；The species of adding medicine includes sulfuric acid, sodium hydroxide, bactericide, flocculation aid, coagulation
Agent or reducing agent etc..
In the historical data of thermal power plant's water treatment dosing system, filter out effluent index and meet the requirements the data of period,
Include the parameter value of each index in the water monitoring index at each time point in the period, and each adding medicine at time point
Chemical feeding quantity.
Wherein, one group of historical parameter value of water monitoring index includes the parameter value of each index, and M group history is obtained altogether
Parameter value.
Step S104：On-line Control SVM models are set up based on SVMs, wherein, historical parameter value is SVM models
Input vector, history chemical feeding quantity is the output vector of SVM models.
In this step, a kind of On-line Control SVM models are set up by SVMs theory, above-mentioned M groups history is joined
In numerical value partly or entirely as the input vector of the model, correspondingly, using corresponding to the history parameters as input vector
The history chemical feeding quantity of value obtains a kind of On-line Control SVM models as the output vector of model.
Step S106：GA algorithms are performed to SVM models, the optimal solution of the parameter in SVM models is solved.
The model obtained to above-mentioned steps S104 performs GA algorithms, realizes the global optimizing of the parameter in SVM models, finally
It can obtain the optimal solution of the parameter in SVM models.
Step S108：Obtained optimal solution input SVM models will be solved, obtain the dosing model on water island.
Step S110：The real-time parameter value of water monitoring index is monitored in real time.
Step S112：Real-time parameter value is inputted to the dosing model on water island, to determine the chemical feeding quantity of current time adding medicine.
After the dosing model on water island is obtained by above-mentioned steps, for the ginseng of the water monitoring index real-time monitored
Numerical value, only need to be input to dosing model, you can obtain the chemical feeding quantity of current time adding medicine, realize dosing process
Line traffic control, can overcome water water quality time-varying characteristics to running according to the current optimal chemical feeding quantity of water quality situation real-time update
Cheng Zaocheng adverse effect, reduces the waste of medicament, reduces artificial and operating cost, and the water quality that water outlet is met in real time will
Ask.Meanwhile, during the foundation of the model, model is set up using SVMs, and determine by GA algorithms the ginseng of model
Number, make use of the ability of searching optimum of GA algorithms, obtains more preferable SVMs parameter, enhance solution to model and release and predict
Ability.
Embodiment two
This example show a kind of preferred embodiment on the basis of above-described embodiment one, provided in the embodiment
Water island dosing On-Line Control Method based on GA-SVM, for thermal power plant's water process production process is relied solely at present experience and
Traditional dosing model be difficult to ensure that chemical feeding quantity on-line optimization analysis, and GA algorithms ripe problem, local convergence, initial population
The problem of distributivity is poor, generates initial population using chaology, realizes that GA is global using adaptive crossover and mutation probability operation
The ability of optimizing, and GA is referred into training sample among SVM determine optimized parameter, realize intelligent dynamic water treatment medicine.
It is described as follows：
Step 1, data signal and its correspondence dosing of the satisfactory M groups history water monitoring index of effluent index are obtained
The chemical feeding quantity of agent, it is determined that the input vector and output vector of the GA-SVM dosing models on digitlization water island, wherein, sample number is M,
The sample number of training set is N, and the sample number of test set is M-N.Wherein, water monitoring index includes coming water-carrying capacity, pH value, turbidity
Value, chloride ion content, ammonia-nitrogen content, adding medicine species include sulfuric acid, sodium hydroxide, bactericide, flocculation aid, coagulant, reduction
Agent etc..Five index parameters in water monitoring index are as the input vector of model, and output vector is one of which dosing
The chemical feeding quantity of agent, the chemical feeding quantity of every kind of adding medicine can be controlled using the method for the embodiment.
Step 2, the On-line Control SVM models of chemical feeding quantity and water monitoring index are set up based on SVMsWherein, K (xi, yi) it is kernel function, b is constant, αiFor Lagrange multiplier, i=1,
2 ..., N, wherein, N ＜ M.The step is specifically included：
Step 21, the time series x of the data signal of water monitoring in future indexiAs mode input vector, and by numeral
The time series y of the corresponding history chemical feeding quantity of signaliAs the output vector of model, wherein, i=1,2 ..., N.
Step 22, by time series collection (xi, yi) use Nonlinear MappingPerform the mapping to higher dimensional space, and structure
The function for making On-line Control SVM models isWherein, i=1,2 ..., N, xi∈Rq, yi∈ R, q are next
Water monitoring index middle finger target number, in this embodiment, q=5, w is weight vector, and b is constant.
Step 23, slack variable ξ and penalty factor are introduced, the object function of SVM models is set upIt is with constraintsWherein, ξ >=0, C ＞ 0, i=1,
2 ..., N, wherein it is desired to explanation, according to the common knowledge of those skilled in the art, the meaning of st namely constraints
Justice expression.
Step 24, radial direction base RBF kernel functions are introducedWherein, σ ＞ 0.
Wherein, in addition to radial direction base RBF kernel functions, also selectable kernel function includes linear kernel function, polynomial kernel letter
Number and Sigmoid kernel functions.Wherein, linear kernel function is a special case of Radial basis kernel function；Polynomial kernel function is general
Apply in the feature space of higher dimensionality, it may appear that amount of calculation surge phenomenon, or even the Xie Buzheng obtained in some cases
Really；Sigmoid kernel functions only have could turn into effective kernel function under given conditions, and its accuracy will be less than radial direction base
Kernel function.Therefore, kernel function is used as using RBF in the embodiment.
Step 25, setting up On-line Control SVM models isWherein, αi=C ξi, it is glug
Bright day coefficient.
Step 3, to On-line Control SVM models perform GA algorithms, to the slack variable ξ in SVM models, penalty factor and
Nuclear parameter σ carries out global optimizing, obtains the optimal solution of each parameter, and by the SVM models in the input step 2 of optimized parameter, obtain
To the GA-SVM dosing models on digitlization water island.As shown in Fig. 2 specifically including：
Step 31, the parameter of GA algorithms is set, initial population scale G, global iterative number of times MAXGEN is determined, namely it is maximum
Genetic algebra, mutation probability Pm1, crossover probability Pc1, select the parameter area of SVM models：Slack variable [ξmin, ξmax], punishment
The factor [Cmin, Cmax] and [σmin, σmax] nuclear parameter, wherein, ξminFor slack variable minimum value, ξmaxFor slack variable maximum,
CminFor penalty factor minimum value, CmaxFor penalty factor maximum, σminFor nuclear parameter minimum value, σmaxFor nuclear parameter maximum,
It is decision variable to determine slack variable ξ, penalty factor and nuclear parameter σ.
Step 32, GEN=1 is made, the decision parameters of the SVM models are initialized, and binary system volume is carried out to these parameters
Code, produces G group initial populations based on chaology, specifically includes：
Step S1, randomly selects 3N [0,1] value ranges and constitutes initiation sequence ε0=(ε1,0, ε2,0..., ε3N, 0), pass through
Logistic maps εN, j+1=μ εN, j(1-εN, j), obtain the different chaos sequence ε in G trackj, wherein εN, 0For initial parameter value
(initial value of selection does not include 0,0.25,0.5,0.75 and 1 this 5 values), εN, j∈ (0,1), μ ∈ (0,4) are control parameter, n
=1,2 ..., 3N-1, j=1,2 ..., G.
Step S2, defines decision variable scope Xmin, Xmax：
Xmin=[ξ1, min, ξ2, min,..., ξN, min, C1, min, C2, min..., CN, min, σ1, min, σ2, min..., σN, min],
Xmax=[ξ1, max, ξ2, max..., ξN, max, C1, max, C2, max..., CN, max, σ1, max, σ2, max..., σN, max]。
Step S3, X is pressed by chaos sequenceI, j=XI, min+(XI, max-XI, min)εI, jAmplify the span of each decision parameters,
Obtain G and represent SVM model parameter Sequence composition initial populations：
(X1,1, X1,2..., X1,3N), (X2,1, X2,2..., X2,3N) ..., (XG, 1, XG, 2..., XG, 3N),
Wherein, XI, jFor i-th of body, j-th of decision variable value, XI, maxFor the maximum of i-th of body, XI, minFor i-th of body
Minimum value, i=1,2 ..., G, j=1,2 ..., 3N.
Step 33, population feeding SVM models are trained, and each individual in population fit is calculated according to fitness function
Response functional value, wherein, deterministic coefficient DC of the SVM models on training sample set1Inverse be fitness functionyiFor i-th of training sample value,For i-th of computer mould analog values, that is, with xiAs
Obtained computer mould analog values are calculated during the input vector of the SVM models,For training sample average value, i=1,2 ...,
N。
Step 34, judge whether current iteration number of times GEN reaches global iterative number of times MAXGEN, required if meeting and terminating,
Optimum individual is then exported, optimized parameter ξ, C and σ is obtained, and carries out step 37, if it is not satisfied, then carrying out step 35.
Step 35, fitness function value highest optimum individual is preserved, prevents outstanding gene genetic operator from operating
It is lost, and records the minimum worst individual sequence number Index of fitness function value.
Step 36, GEN=GEN+1 is made, is selected, intersected, mutation genetic is operated, and produces new population, and use step
35 optimum individuals preserved replace serial number Index new individual, produce new population, return to step 33.
Wherein, selection operation is selected using wheel disc, and the crossover operator of crossover operation is general using the variation of single-point, mutation operation
Rate selects adaptive probability.Wherein, crossover probability is calculated：
Wherein, in formula：F is the non-dominant grade of current individual；For the non-dominant grade of selection, typically it can use
fmaxRepresent the maximum of individual non-dominant grade in current population；Pc1For a constant, the minimum value for limiting crossover probability,
Be conducive to the stability and validity evolved；k1∈ (0,1), the amplitude of variation for controlling crossover probability.
Mutation probability is calculated：
In formula：Pm1Minimum value for limiting mutation probability；k2∈ (0,1), the amplitude of variation for controlling crossover probability；
The meaning of remaining variables is the same.
Step 37, optimized parameter is inputted into SVM models, Verification is carried out to data using test sample collection, that is, will
The time series x of the data signal of the water monitoring indexpThe SVM models are inputted, obtain corresponding to the time series xp
Adding medicine simulation chemical feeding quantity time seriesWherein, p=N+1, N+2 ..., M, and calculate the determination of test sample collection
Property coefficient DC2：
If DC2＞ 0.75, meets and requires, then export optimized parameter, be otherwise transferred to and enter 32.
Step 4, obtained water monitoring index will be monitored in real time as input vector input digitlization water island GA-SVM to add
Medicine model, determines the optimal chemical feeding quantity at current time.
In a kind of concrete application, using the method for above-described embodiment, thermal power plant's water treatment medicine amount is predicted, should
The water monitoring index of factory includes coming water-carrying capacity, pH value, turbidity value, chloride ion content and ammonia-nitrogen content, chooses in October, 2014
The time series of 5 monitoring indexes of water treatment system water is used as GA-SVM dosing moulds between 11 days~24 days July in 2015
The input variable of type, in the period time series of the chemical feeding quantity of sulfuric acid as GA-SVM dosing models output variable.Wherein,
PH value, change curve such as Fig. 3, Fig. 4, Fig. 5, Fig. 6 and Fig. 7 institute for carrying out water-carrying capacity, coming water turbidity, chloride ion content and ammonia-nitrogen content
Show.
Calculated using the method for above-described embodiment, the optimized parameter for obtaining model by GA algorithm optimizing first is：C
=62.206, ξ=3.164 and σ=1.16.These optimized parameters are inputted into data of the GA-SVM dosings model to the above-mentioned period
Learning training is carried out, and output vector is obtained to calculating being fitted, wherein, the deterministic coefficient DC of training set models fitting1
=0.889, the deterministic coefficient DC of test set models fitting2=0.873, obtain the reality of sulfuric acid chemical feeding quantity --- fitting is bent
Line, as shown in Figure 8.
Using the water island 25 days~in Augusts, 2015 water of 25 days July in 2015 monitoring index as model input
Variable, inputs the GA-SVM dosing models of above-mentioned foundation, solves the optimal chemical feeding quantity of sulfuric acid, obtains prediction deterministic coefficient DC=
0.773, obtain the reality of the chemical feeding quantity of sulfuric acid in the period --- prediction curve, as shown in Figure 9.As shown in Figure 8, Figure 9 and
The deterministic coefficient of fitting prediction can be seen that the fitting precision of GA-SVM dosing models is higher, and with preferable extensive energy
Power.
The calculating of other medicaments such as optimal addition dosage of sodium hydroxide, bactericide, flocculation aid, coagulant, reducing agent etc. is equal
It can be obtained according to the method for the present embodiment to calculate.
Embodiment three
As shown in Figure 10, this embodiment offers a kind of water island dosing On-line Control device based on GA-SVM, the control
Device is applied in the water treatment system of thermal power plant, specifically includes historical data acquisition module 10, model building module 20, parameter
Optimization module 30, real-time monitoring modular 40 and chemical feeding quantity control module 50.
Wherein, historical data acquisition module 10 is used to obtaining that effluent index to be satisfactory, water monitoring index M groups are gone through
History parameter value, and the adding medicine corresponding to every group of historical parameter value history chemical feeding quantity；Model building module 20 is used to be based on
SVMs sets up On-line Control SVM models, wherein, historical parameter value is the input vector of SVM models, and history chemical feeding quantity is
The output vector of SVM models；Parameter optimization module 30 is used to perform SVM models GA algorithms, the parameter in solution SVM models
Optimal solution, and obtained optimal solution input SVM models will be solved, obtain the dosing model on water island；Real-time monitoring modular 40 is used for
The real-time parameter value of water monitoring index is monitored in real time；And chemical feeding quantity control module 50 is used to real-time parameter value inputting water island
Dosing model, to determine the chemical feeding quantity of current time adding medicine.
Preferably, model building module 20 is when setting up On-line Control SVM models, and performing specific steps includes：SVM is set
The input vector of model is the time series x of the data signal of water monitoring indexi, set SVM models output vector for pair
Should be in time series xiAdding medicine history chemical feeding quantity time series yi, wherein, i=1,2 ..., N, N ＜ M；Will be by the time
Sequence xiWith time series yiTime series collection (the x of compositioni, yi) use Nonlinear MappingPerform reflecting to higher dimensional space
Penetrate, and construct the function of SVM models and beWherein, i=1,2 ..., N, xi∈Rq, yi∈ R, q are water
Monitoring index middle finger target number, w is weight vector, and b is constant；Slack variable ξ and penalty factor are introduced, SVM models are set up
Object function isConstraints isWherein, ξ >=0, C ＞
0, i=1,2 ..., N；The kernel function for determining SVM models is RBFWherein, σ ＞ 0；
Setting up SVM models isWherein, αi=C ξi, it is Lagrange coefficient.
Preferably, parameter optimization module 30 is specifically wrapped the step of execution in the optimal solution of the parameter in solving SVM models
Include：Step A：The parameter of GA algorithms is set, initial population scale G, global iterative number of times, mutation probability, crossover probability is determined, really
Determine the decision variable that slack variable ξ, penalty factor and nuclear parameter σ are GA algorithms；Step B：It is 1 to make iterations, initialization
The parameter of SVM models, and binary coding is carried out to the parameter of the SVM models after initialization, at the beginning of producing G groups based on chaology
Beginning population；Step C：Using following circulation steps, slack variable ξ, penalty factor and nuclear parameter σ optimal solution are determined：Step
C1：Each individual fitness function value, step C2 in population are calculated according to fitness function：Judge whether current iteration number of times reaches
To global iterative number of times, if reaching, optimum individual, end step C are exported, if being not up to, fitness function value highest
Optimum individual preserve, and record the minimum worst individual sequence number of fitness function value, make iterations plus 1, selected
Select, intersect, mutation genetic is operated, and the new individual of the worst individual sequence number of serial number is replaced with the optimum individual preserved, producing new
Population, return to step C1.
Preferably, in above-mentioned steps C1, fitness function is For with time series
Collect (xi, yi) in i-th group of data as the input vector of SVM models calculate obtained computer mould analog values,For time series collection
(xi, yi) in yiAverage value.
Preferably, parameter optimization module 30 is in the optimal solution of the parameter in solving SVM models, the step of execution specifically also
Including：Step D：The optimal solution that solution is obtained is used as the parameter of SVM models, the time of the data signal of water monitoring in future index
Sequence xpSVM models are inputted, obtain corresponding to time series xpAdding medicine simulation chemical feeding quantity time seriesWherein, p
=N+1, N+2 ..., M；Step E：According to corresponding to time series xpAdding medicine history chemical feeding quantity time series ypAnd when
Between sequenceCalculate the deterministic coefficient of SVM models；Step F：Judge whether deterministic coefficient is more than predetermined coefficient value, if greatly
In the step of then terminating to solve the optimal solution of the parameter in SVM models, if being not more than, return to step A.
Preferably, parameter optimization module 30 is initializing the parameter of SVM models, and produces G groups initially based on chaology
During population, the step of specifically performing includes：Step B1, randomly selects 3N [0,1] value ranges and constitutes initiation sequence ε0=(ε1,0,
ε2,0..., ε3N, 0), wherein, εN, 0Do not include 0,0.25,0.5,0.75 and 1.ε is mapped by LogisticN, j+1=μ εN, j(1-
εN, j), obtain the different chaos sequence ε in G trackj, wherein, εN, 0For initial parameter, εN, j∈ (0,1), μ ∈ (0,4) are control
Parameter, n=1,2 ..., 3N-1, j=1,2 ..., G；Step B2, defines decision variable scope Xmin, Xmax：
Xmin=[ξ1, min, ξ2, min..., ξN, min, C1, min, C2, min..., CN, min, σ1, min, σ2, min..., σN, min],
Xmax=[ξ1, max, ξ2, max..., ξN, max, C1, max, C2, max..., CN, max, σ1, max, σ2, max..., σN, max]；
Step B3, X is pressed by chaos sequenceN, j=XN, min+(XN, max-XN, min)εN, jAmplify the span of each decision variable,
Obtain G and represent SVM model parameter Sequence composition initial populations：
(X1,1, X1,2..., X1,3N), (X2,1, X2,2..., X2,3N) ..., (XG, 1, XG, 2..., XG, 3N), wherein,
XN, jFor n-th of decision variable value of j-th of body, XN, maxFor the maximum of n-th of body, XN, minFor the minimum value of n-th body, j
=1,2 ..., G, j=1,2 ..., 3N.
Preferably, selection operation is selected using wheel disc in parameter optimization module 30, and the crossover operator of crossover operation is using single
Point, the mutation probability of mutation operation, which is washed, to be selected vernacular and answers probability, wherein,
Crossover probability is calculated using below equation：
Mutation probability is calculated using below equation：
Wherein, f is the non-dominant grade of current individual,fmaxFor individual non-dominant grade in current population most
Big value, Pc1For constant, Pm1To limit the minimum value of mutation probability, k1∈ (0,1), k2∈ (0,1).
Preferably, water monitoring index includes water-carrying capacity, pH value, turbidity value, chloride ion content and ammonia-nitrogen content, adding medicine
For sulfuric acid, sodium hydroxide, bactericide, flocculation aid, coagulant or reducing agent.
For foregoing each method embodiment, in order to be briefly described, therefore it is all expressed as to a series of combination of actions, but
It is that those skilled in the art should know, the present invention is not limited by described sequence of movement, because according to the present invention,
Some steps can be using other along going or while perform；Secondly, those skilled in the art should also know, the above method is implemented
Example belongs to preferred embodiment, and involved action and the module not necessarily present invention are necessary.
For foregoing each device embodiment, in order to be briefly described, therefore it is all expressed as to a series of block combiner, but
It is that those skilled in the art should know, the present invention is not limited by described block combiner, because according to the present invention,
Certain module can be performed using other modules；Secondly, those skilled in the art should also know, said apparatus embodiment belongs to
Necessary to preferred embodiment, the involved module not necessarily present invention.Each embodiment in this specification is adopted
Described with progressive mode, what each embodiment was stressed is the difference with other embodiment, each embodiment it
Between identical similar part mutually referring to.For device embodiment, because it is substantially similar to embodiment of the method, institute
With the fairly simple of description, the relevent part can refer to the partial explaination of embodiments of method.
A kind of water island dosing On-Line Control Method and device based on GA-SVM provided by the present invention are carried out above
It is discussed in detail, specific case used herein is set forth to the principle and embodiment of the present invention, above example
Illustrate the method and its core concept for being only intended to help to understand the present invention；Simultaneously for those of ordinary skill in the art, according to
According to the thought of the present invention, it will change in specific embodiments and applications, to sum up, this specification content should not
It is interpreted as limitation of the present invention.
Claims (7)
1. a kind of water island dosing On-Line Control Method based on GA-SVM, it is characterised in that including：
Acquisition effluent index is satisfactory, water monitoring index M group historical parameter values, and corresponding to history described in every group
The history chemical feeding quantity of the adding medicine of parameter value, wherein, the water monitoring index includes multiple indexs, history parameters described in every group
Value includes the parameter value of the multiple index；
On-line Control SVM models are set up based on SVMs, wherein, the historical parameter value is the input of the SVM models
Vector, the history chemical feeding quantity is the output vector of the SVM models；
GA algorithms are performed to the SVM models, the optimal solution of the parameter in the SVM models is solved, the step includes：
Step A：The parameter of GA algorithms is set, initial population scale, global iterative number of times, mutation probability, crossover probability is determined, really
Determine the decision variable that slack variable, penalty factor and nuclear parameter are GA algorithms,
Step B：It is 1 to make iterations, initializes the parameter of the SVM models, and to the ginseng of the SVM models after initialization
Number carries out binary coding, and G group initial populations are produced based on chaology,
Step C：Using following circulation steps, the optimal solution of slack variable, penalty factor and nuclear parameter is determined：Step C1：According to
Fitness function calculates each individual fitness function value, step C2 in population：Judge whether current iteration number of times reaches the overall situation
Iterations, if reaching, exports optimum individual, end step C, if being not up to, fitness function value highest is optimal
Individual is preserved, and records the minimum worst individual sequence number of fitness function value, is made iterations plus 1, is selected, handed over
Pitch, mutation genetic is operated, and the new individual of worst individual sequence number described in serial number is replaced with the optimum individual preserved, produce newly
Population, return to step C1；
The obtained optimal solution input SVM models will be solved, obtain the dosing model on water island；
The water monitoring index is monitored in real time to obtain one group of real-time parameter value；And
The real-time parameter value is inputted to the dosing model on the water island, to determine the chemical feeding quantity of adding medicine described in current time,
Wherein, in step C2, selection operation is selected using wheel disc, and the crossover operator of crossover operation uses single-point, mutation operation
Mutation probability selection adaptive probability, wherein,
Crossover probability is calculated using below equation：
Mutation probability is calculated using below equation：
Wherein, f is the non-dominant grade of current individual,fmaxFor the maximum of individual non-dominant grade in current population,
Pc1For constant, Pm1To limit the minimum value of mutation probability, k1∈ (0,1), k2∈ (0,1).
2. the water island dosing On-Line Control Method according to claim 1 based on GA-SVM, it is characterised in that based on support
The step of vector machine sets up On-line Control SVM models specifically includes：
It is the time series x of the data signal of the water monitoring index to set the input vector of the SVM modelsi, set described
The output vector of SVM models is corresponding to the time series xiAdding medicine history chemical feeding quantity time series yi, wherein, i
=1,2 ..., N, N ＜ M；
Will be by the time series xiWith the time series yiTime series collection (the x of compositioni, yi) use Nonlinear Mapping
The mapping to higher dimensional space is performed, and constructs the function of the SVM models and isWherein, i=1,
2 ..., N, xi∈Rq, yi∈ R, q are the water monitoring index middle finger target number, and w is weight vector, and b is constant；
Slack variable ξ and penalty factor are introduced, the object function for setting up the SVM models is
Constraints isWherein, ξ >=0, C ＞ 0, i=1,2 ..., N；
The kernel function for determining the SVM models is RBFWherein, σ ＞ 0；
Setting up the SVM models isWherein, αi=C ξi, it is Lagrange coefficient.
3. the water island dosing On-Line Control Method according to claim 1 based on GA-SVM, it is characterised in that the adaptation
Spending function is For with the time series collection (xi, yi) in i-th group of data as described
The input vector of SVM models calculates obtained computer mould analog values,For the time series collection (xi, yi) in yiAverage value.
4. the water island dosing On-Line Control Method according to claim 3 based on GA-SVM, it is characterised in that solve described
The step of optimal solution of parameter in SVM models, also includes：
Step D：The parameter of the obtained optimal solution as the SVM models will be solved, by the number of the water monitoring index
The time series x of word signalpThe SVM models are inputted, obtain corresponding to the time series xpAdding medicine simulation chemical feeding quantity
Time seriesWherein, p=N+1, N+2 ..., M；
Step E：According to corresponding to the time series xpAdding medicine history chemical feeding quantity time series ypWith the time sequence
RowCalculate the deterministic coefficient of the SVM models；
Step F：Judge whether the deterministic coefficient is more than predetermined coefficient value, if being more than, terminate to solve in the SVM models
Parameter optimal solution the step of, if be not more than, return to step A.
5. the water island dosing On-Line Control Method according to claim 4 based on GA-SVM, it is characterised in that step B has
Body includes：
Step B1, randomly selects 3N [0,1] value ranges and constitutes initiation sequence ε0=(ε1,0, ε2,0..., ε3N, 0), pass through
Logistic maps εN, j+1=μ εN, j(1-εN, j), obtain the different chaos sequence ε in G trackj, wherein, εN, 0For initial parameter,
εN, j∈ (0,1), μ ∈ (0,4) are control parameter, n=1,2 ..., 3N-1, j=1,2 ..., G；
Step B2, defines decision variable scope Xmin, Xmax：
Xmin=[ξ1, min, ξ2, min..., ξN, min, C1, min, C2, min..., CN, min, σ1, min, σ2, min..., σN, min],
Xmax=[ξ1, max, ξ2, max..., ξN, max, C1, max, C2, max..., CN, max, σ1, max, σ2, max..., σN, max]；
Step B3, X is pressed by the chaos sequenceN, j=XN, min+(XN, max-XN, min)εN, jAmplify the span of each decision variable,
Obtain G and represent SVM model parameter Sequence composition initial populations：
(X1,1, X1,2..., X1,3N), (X2,1, X2,2..., X2,3N) ..., (XG, 1, XG, 2..., XG, 3N),
Wherein, XN, jFor n-th of decision variable value of j-th of body, XN, maxFor the maximum of n-th of body, XN, minFor n-th body
Minimum value, j=1,2 ..., G, j=1,2 ..., 3N.
6. the water island dosing On-Line Control Method according to claim 5 based on GA-SVM, it is characterised in that in step B1
εN, 0Do not include 0,0.25,0.5,0.75 and 1.
7. the water island dosing On-Line Control Method according to claim 1 based on GA-SVM, it is characterised in that the water
Monitoring index includes water-carrying capacity, pH value, turbidity value, chloride ion content and ammonia-nitrogen content, and the adding medicine is sulfuric acid, hydroxide
Sodium, bactericide, flocculation aid, coagulant or reducing agent.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610941724.5A CN106292296B (en) | 2016-10-25 | 2016-10-25 | Water island dosing On-Line Control Method and device based on GA SVM |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610941724.5A CN106292296B (en) | 2016-10-25 | 2016-10-25 | Water island dosing On-Line Control Method and device based on GA SVM |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106292296A CN106292296A (en) | 2017-01-04 |
| CN106292296B true CN106292296B (en) | 2017-11-03 |
Family
ID=57720369
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610941724.5A Expired - Fee Related CN106292296B (en) | 2016-10-25 | 2016-10-25 | Water island dosing On-Line Control Method and device based on GA SVM |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106292296B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106502093B (en) * | 2016-10-25 | 2017-09-05 | 大唐（北京）水务工程技术有限公司 | Water island dosing On-Line Control Method based on GA SVR |
| CN108830035A (en) * | 2018-05-28 | 2018-11-16 | 华东交通大学 | A kind of novel water process coagulant dosage control method, computer, computer program |
| CN113261064B (en) * | 2019-01-23 | 2024-03-08 | 国立研究开发法人科学技术振兴机构 | Dosage management support system |
| CN112163674A (en) * | 2019-12-13 | 2021-01-01 | 哈尔滨理工大学 | Bidirectional optimization BP neural network sliding mode variable structure control method |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102005135A (en) * | 2010-12-09 | 2011-04-06 | 上海海事大学 | Genetic algorithm-based support vector regression shipping traffic flow prediction method |
| CN102249411A (en) * | 2011-05-17 | 2011-11-23 | 中国科学技术大学 | Method for optimizing sewage treatment process |
| CN102841979A (en) * | 2012-06-26 | 2012-12-26 | 广州市新之地环保产业有限公司 | Sludge reflux prediction method based on PCA-GA-SVR (principal component analysis-genetic algorithm-support vector regression) |
| CN104933310A (en) * | 2015-06-19 | 2015-09-23 | 河海大学 | Method for predicting membrane pollution tendency in membrane distilled water processing system on the basis of GA-LSSVM (Genetic Algorithm- Least Squares Support Vector Machine) model |
| CN105676814A (en) * | 2016-01-11 | 2016-06-15 | 大唐环境产业集团股份有限公司 | SFLA-SVM-based digital water island online agent adding control method |
-
2016
- 2016-10-25 CN CN201610941724.5A patent/CN106292296B/en not_active Expired - Fee Related
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102005135A (en) * | 2010-12-09 | 2011-04-06 | 上海海事大学 | Genetic algorithm-based support vector regression shipping traffic flow prediction method |
| CN102249411A (en) * | 2011-05-17 | 2011-11-23 | 中国科学技术大学 | Method for optimizing sewage treatment process |
| CN102841979A (en) * | 2012-06-26 | 2012-12-26 | 广州市新之地环保产业有限公司 | Sludge reflux prediction method based on PCA-GA-SVR (principal component analysis-genetic algorithm-support vector regression) |
| CN104933310A (en) * | 2015-06-19 | 2015-09-23 | 河海大学 | Method for predicting membrane pollution tendency in membrane distilled water processing system on the basis of GA-LSSVM (Genetic Algorithm- Least Squares Support Vector Machine) model |
| CN105676814A (en) * | 2016-01-11 | 2016-06-15 | 大唐环境产业集团股份有限公司 | SFLA-SVM-based digital water island online agent adding control method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106292296A (en) | 2017-01-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106292296B (en) | Water island dosing On-Line Control Method and device based on GA SVM | |
| CN105676814B (en) | The online Adding medicine control method in digitlization water island based on SFLA SVM | |
| CN111553468B (en) | Method for accurately predicting water quality of effluent of sewage treatment plant | |
| CN106371316B (en) | Water island dosing On-Line Control Method based on PSO LSSVM | |
| CN109084415A (en) | Central air-conditioning operating parameter optimization method based on artificial neural network and genetic algorithms | |
| CN105069220A (en) | Back-propagation (BP) neural network immune genetic algorithm based microbial fermentation optimization method | |
| CN106502093B (en) | Water island dosing On-Line Control Method based on GA SVR | |
| CN107993723A (en) | A kind of warfarin dose prediction modeling method based on integrated evolutionary learning | |
| CN106682316A (en) | Real-time effluent total-phosphorus monitoring system based on peak radial basis function neural network | |
| Wang et al. | A full-view management method based on artificial neural networks for energy and material-savings in wastewater treatment plants | |
| CN115986728A (en) | Power grid situation prediction method considering uncertainty factors and terminal | |
| CN108830035A (en) | A kind of novel water process coagulant dosage control method, computer, computer program | |
| CN109408896B (en) | Multi-element intelligent real-time monitoring method for anaerobic sewage treatment gas production | |
| CN110929809B (en) | Sewage key water quality index soft measurement method of characteristic self-enhanced cyclic neural network | |
| CN114971090A (en) | Electric heating load prediction method, system, equipment and medium | |
| CN113051828B (en) | Online prediction method for natural gas water dew point driven by technological parameters | |
| CN114707692A (en) | Wetland effluent ammonia nitrogen concentration prediction method and system based on hybrid neural network | |
| CN114119277A (en) | Artificial intelligent neural network-based flocculation dosing decision analysis method | |
| CN113762591A (en) | Short-term electric quantity prediction method and system based on GRU and multi-core SVM counterstudy | |
| CN109086887A (en) | Method for early warning of the depth RBF neural in conjunction with the AHP based on entropy weight | |
| CN113591078A (en) | Industrial control intrusion detection system and method based on convolutional neural network architecture optimization | |
| CN117291069A (en) | LSTM sewage water quality prediction method based on improved DE and attention mechanism | |
| Sun et al. | Research on prediction of housing prices based on GA-PSO-BP neural network model: evidence from Chongqing, China | |
| CN110674460B (en) | E-Seq2Seq technology-based data driving type unit combination intelligent decision method | |
| CN115206444A (en) | Optimal drug dosage prediction method based on FCM-ANFIS model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB03 | Change of inventor or designer information |
Inventor after: Qiao Yongcheng Inventor after: Liu Linhu Inventor after: Qin Dong Inventor after: Chang Xudong Inventor after: Wang Yanyang Inventor after: Jiang Yingbiao Inventor after: Di Ruobing Inventor after: Song Jinfeng Inventor after: Wang Shuangxian Inventor before: Qin Dong Inventor before: Yang Chunlong Inventor before: Wei Lai |
|
| CB03 | Change of inventor or designer information | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20170927 Address after: 030012 Shanxi city of Taiyuan province Pingyang West Lane Road, No. 19 Applicant after: YANGCHENG INTERNATIONAL POWER GENERATION CO., LTD. Applicant after: Datang (Beijing) Water Engineering Technology Co., Ltd. Address before: 100097 Beijing Haidian District Zizhuyuan Road No. 120 3 301 Applicant before: Datang (Beijing) Water Engineering Technology Co., Ltd. |
|
| TA01 | Transfer of patent application right | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20171103 Termination date: 20201025 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |