CN108879582A - A kind of control system and control method of intelligent type permanent magnet vacuum circuit breaker - Google Patents
A kind of control system and control method of intelligent type permanent magnet vacuum circuit breaker Download PDFInfo
- Publication number
- CN108879582A CN108879582A CN201810627889.4A CN201810627889A CN108879582A CN 108879582 A CN108879582 A CN 108879582A CN 201810627889 A CN201810627889 A CN 201810627889A CN 108879582 A CN108879582 A CN 108879582A
- Authority
- CN
- China
- Prior art keywords
- signal
- weight
- module
- neural network
- fault
- 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
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02H—EMERGENCY PROTECTIVE CIRCUIT ARRANGEMENTS
- H02H1/00—Details of emergency protective circuit arrangements
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01H—ELECTRIC SWITCHES; RELAYS; SELECTORS; EMERGENCY PROTECTIVE DEVICES
- H01H33/00—High-tension or heavy-current switches with arc-extinguishing or arc-preventing means
- H01H33/02—Details
- H01H33/59—Circuit arrangements not adapted to a particular application of the switch and not otherwise provided for, e.g. for ensuring operation of the switch at a predetermined point in the ac cycle
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01H—ELECTRIC SWITCHES; RELAYS; SELECTORS; EMERGENCY PROTECTIVE DEVICES
- H01H33/00—High-tension or heavy-current switches with arc-extinguishing or arc-preventing means
- H01H33/60—Switches wherein the means for extinguishing or preventing the arc do not include separate means for obtaining or increasing flow of arc-extinguishing fluid
- H01H33/66—Vacuum switches
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01H—ELECTRIC SWITCHES; RELAYS; SELECTORS; EMERGENCY PROTECTIVE DEVICES
- H01H33/00—High-tension or heavy-current switches with arc-extinguishing or arc-preventing means
- H01H33/60—Switches wherein the means for extinguishing or preventing the arc do not include separate means for obtaining or increasing flow of arc-extinguishing fluid
- H01H33/66—Vacuum switches
- H01H33/666—Operating arrangements
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02H—EMERGENCY PROTECTIVE CIRCUIT ARRANGEMENTS
- H02H1/00—Details of emergency protective circuit arrangements
- H02H1/0007—Details of emergency protective circuit arrangements concerning the detecting means
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02H—EMERGENCY PROTECTIVE CIRCUIT ARRANGEMENTS
- H02H1/00—Details of emergency protective circuit arrangements
- H02H1/0092—Details of emergency protective circuit arrangements concerning the data processing means, e.g. expert systems, neural networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computational Linguistics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Evolutionary Biology (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Power Engineering (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention belongs to breaker technical fields, disclose the control system and control method of a kind of intelligent type permanent magnet vacuum circuit breaker, and control system includes:Signal acquisition module, switch-on controlling module, central control module, photoelectric isolation module, fault identification module, alarm, display screen.The present invention can compensate the closing time of breaker by switch-on controlling module, so that the branch wire time of breaker can keep stable under different operating conditions, thus to realize that synchroswitching provides reliable basis;Support vector machines is combined by fault identification module with incremental learning simultaneously, identifies permanent magnet mechanism high-pressure vacuum breaker most common failure.It makes full use of Small Sample Database to carry out training pattern, improves the generalization ability of fault identification, and enriched constantly sample database by learning newly-increased data, classification and recognition can be increased to 90% or more.
Description
Technical field
The invention belongs to the control system of breaker technical field more particularly to a kind of intelligent type permanent magnet vacuum circuit breaker and
Control method.
Background technique
Currently, the prior art commonly used in the trade is such:
Breaker is the electric current referred to close, carry and cut-off under normal loop condition and can close, when defined
Interior carrying and the switching device for cut-offfing the electric current under abnormal loop condition.Breaker is divided into high-voltage circuitbreaker by its use scope
With low-voltage circuit breaker, high-low pressure identification of boundaries is relatively fuzzyyer, and 3kV or more is generally known as High-Voltage Electrical Appliances.Breaker can be used to
Electric energy is distributed, asynchronous motor is infrequently started, the implementations such as power circuit and motor are protected, when their generations are serious
When overload or short circuit and the failures such as under-voltage can automatic shutdown circuit, function is equivalent to Fuse switch and crosses underheat relay
The combination of device etc..And change components are not needed generally after disjunction fault current.Currently, having had been widely used.
However, existing breaker does not account for the influence because of temperature change to breaker mechanical structure, so that closing time is caused to change,
It is unable to synchroswitching, brings inconvenience to actual use;Simultaneously in use, defy capture equipment fault when status information, therefore
Hinder data sample to lack, data are not enriched;Failure analysis methods specific aim is not strong, and is difficult to set up corresponding fault identification mathematics
Model.
In conclusion problem of the existing technology is:
Existing breaker does not account for the influence because of temperature change to breaker mechanical structure, so that closing time be caused to change
Become, is unable to synchroswitching, brings inconvenience to actual use;
Simultaneously in use, defy capture equipment fault when status information, fault data sample lacks, and data do not enrich;
Failure analysis methods specific aim is not strong, and is difficult to set up corresponding fault identification mathematical model.
It is existing to improve the method for classifier learning performance in the presence of the thought that apply Active Learning using unmarked example
And the specific way of particular problem is not solved, need to accomplish " particular problem, concrete analysis ";Algorithm computation complexity is high,
Algorithm is caused to be difficult to apply in the quick demand for control of breaker.
But one kind up to the present, there is no to carry out analysis mining to influence factor under big data driving, and will dynamic spy
Property incorporate prediction model adjustable strategies prediction technique.
Summary of the invention
In view of the problems of the existing technology, the present invention provides a kind of control systems of intelligent type permanent magnet vacuum circuit breaker
And control method.
The invention is realized in this way a kind of control method of intelligent type permanent magnet vacuum circuit breaker, including:
By signal acquisition module Acquisition Circuit signal, signal generates corresponding driving signal based on the received;
Coarse adjustment is carried out to the breaker closing time by switch-on controlling module and fine tuning controls;
By photoelectric isolation module using light as media transmission electric signal;
By fault identification module Acquisition Circuit fault-signal data, establishes fault signature data sample and carry out identification failure
Feature;If identifying fault-signal, triggers alarm and alarm;Signal message and event by display screen display acquisition
Hinder status information.
Central control module is to switching signal, optical transport electric signal, fault identification signal after driving signal, coarse adjustment and fine tuning
Carry out control processing, wherein believe first switching signal, optical transport electric signal, fault identification after driving signal, coarse adjustment and fine tuning
It number carries out carrying out sample pre-selection using k means clustering algorithm, according to data pre-selection as a result, to important in unmarked example set
Each example in sample set L* is inquired, and the category that inquiry is obtained returns;Obtain the part " weight in unmarked example set
Want " example and label is obtained, to replace all unmarked examples;Using sample pre-selection as a result, significant samples collection L* is as support
The training set of vector machine Active Learning constructs neural network model BP, generates the initial weight of neural network model BP;By weight
The random number being initialized as between [a, b], wherein a, b are the integer for meeting following equation:
Wherein H is network node in hidden layer;Weight and threshold value to neural network model BP carry out
It is dynamically refined, obtains dynamic neural network model DBP, generate the weight and threshold value of dynamic neural network model DBP,
With adaptive immune genetic AIGA algorithm optimization dynamic neural network model DBP, prediction model AIGA- is obtained
DBP knows according to switching signal, optical transport electric signal, failure after prediction model AIGA-DBP calculating driving signal, coarse adjustment and fine tuning
The predicted value of level signal;Specifically have:Adjust the weight w between neural network model BP hidden layer and output layerkj;Adjust wkj's
Purpose is desirable to the new output o of output node j* pjO is exported than currentlypjCloser to target value tpj, whereinα represents the degree of approach, remains unchanged in each cycle of training, and with node in hidden layer
The adjustment of H and become smaller, do not consider threshold value, have:Wherein wkjWithPoint
The weight of front and back, y Wei not updatedpkFor hidden layer output, △ wkjFor wkjKnots modification;Obtain △ wkjSolution equation:Wherein,It is asked according to least square and error principle
It solves equationObtain △ wkjApproximate solution:
The hidden layer node k of output node j is connected to each, the weight calculated between k and j changes △ wkj, update
Weight simultaneously calculates error of sum square E, then in one optimal k of k ∈ [1, H] interval selection, so that E is minimum;
Adjust the weight v between neural network model BP input layer and hidden layerik;Adjust vikPurpose be once nerve
Network algorithm falls into local minimum point, and modification weight can jump out the minimal point, judge that neural network algorithm falls into local minimum
The condition of point is the change rate △ E=0 of error E, and E>0;Do not consider threshold value, the change of the weight of hidden layer node k by with
Lower equation obtains:Wherein δpj=f-1(ypk+Δypk)-f-1(ypk), M is natural number, and hidden layer exports ypkIt solves
Formula is:Wherein △ ypkFor ypkKnots modification, have:
According to least square and error principle solution formulaThe matrix equation of building, obtains:Aggregative formulaWith
Calculate the consecutive mean knots modification of weight between hidden layer and output layer
Calculate the consecutive mean knots modification of weight between input layer and hidden layerM
The natural number between 10~20 is taken, the consecutive mean weight of neural network model BP is obtained, according to the dynamic of neural network model BP
State average weight obtains dynamic neural network model DBP;Judge whether the error of predicted value and desired value meets the condition of setting,
Export control instruction.
Further, K mean cluster algorithm includes:
Given sample set D={ x1,x2,;;;,xm, " k mean value " algorithm divides C={ c for cluster gained cluster1,c2,;;;,
ckMinimize square error:
Wherein, x is cluster ciMean vector;The cluster number K=2 of generation;
Further, with adaptive immune genetic AIGA algorithm optimization dynamic neural network model DBP power plant and threshold value,
It obtains prediction model AIGA-DBP and specifically includes following steps:
Antigen recognizing is carried out first, i.e. the identification of mean square error generates initial antibody (dynamic neural network model DBP
Weight and threshold value) after, carry out antibody coding, then calculating antibody fitness and antibody concentration are carried out according to the result of calculating
Adaptive antibody selection operation based on concentration obtains suitable weight and threshold value if this operation meets condition, and by result
Output;If this operation is unsatisfactory for condition, need to carry out adaptive crossover operation and TSP question operation, and carry out weight and threshold
The update of value, then antibody fitness and antibody concentration are recalculated, weight and threshold value until obtaining the condition that meets, last basis
Meet the weight and threshold value of condition, obtains prediction model AIGA-DBP.
Further, shutting-brake control method includes:
Firstly, determining the initiating of movable contact of vacuum circuit breaker according to the closing coil electric current of known permanent magnetic vacuum breaker
The current value I1 of moment t1, rigid conjunction ta and movable contact of vacuum circuit breaker initiating moment coil, coil is in open position
The timeconstantτ 2 of timeconstantτ 1 and closing position, and calculate parameter S described in the text;
Then, according to aforementioned parameters, the parameter of VVVE fitting current equation is determined, to show that VVVE reference current is bent
Line;
Finally, recalculating fitting current equation parameter by constantly regulate coefficient k and η, obtaining modified reference current
Curve realizes the coarse adjustment and fine tuning to the breaker closing time using coil current tracing, with optimized breaker moving contact fortune
Dynamic characteristic.
Further, fault recognition method includes::
1) fault signature data are normalized, and data are randomly divided into A, B, C3 group data, establish fault signature data sample
This;
2) disaggregated model M1 is obtained using support vector cassification algorithm training data A, utilizes support vector machines increment
It practises algorithm training data B and obtains model M 2;
3) accuracy of data C verifying model M 1 and M2 is utilized.
Another object of the present invention is to provide a kind of control methods for realizing the intelligent type permanent magnet vacuum circuit breaker
Computer program.
Another object of the present invention is to provide a kind of control methods for realizing the intelligent type permanent magnet vacuum circuit breaker
Computer.
Another object of the present invention is to provide a kind of computer readable storage mediums, including instruction, when it is in computer
When upper operation, so that computer executes the control method of the intelligent type permanent magnet vacuum circuit breaker.
A kind of control system another object of the present invention is to provide intelligent type permanent magnet vacuum circuit breaker includes:
Signal acquisition module is connect with central control module, is used for Acquisition Circuit signal, and signal generates phase based on the received
The driving signal answered;
Switch-on controlling module is connect with central control module, for carrying out coarse adjustment and fine tuning control to the breaker closing time
System;
Central control module, with signal acquisition module, switch-on controlling module, photoelectric isolation module, fault identification module, report
Alert device, display screen connection, work normally for controlling modules;
Photoelectric isolation module is connect with central control module, for using light as media transmission electric signal;
Fault identification module, connect with central control module, is used for Acquisition Circuit fault-signal data, establishes fault signature
Data sample carries out identification fault signature;
Alarm is connect with central control module, for carrying out triggering alarm according to the fault-signal of identification;
Display screen is connect with central control module, for showing the signal message and fault status information of acquisition.
Another object of the present invention is to provide a kind of control systems for carrying the intelligent type permanent magnet vacuum circuit breaker
Permanent magnetic vacuum breaker controls equipment
Advantages of the present invention and good effect are:
The present invention can compensate the closing time of breaker by switch-on controlling module, so that the divide-shut brake of breaker
Time can keep stable under different operating conditions, thus to realize that synchroswitching provides reliable basis;Known simultaneously by failure
Other module combines support vector machines with incremental learning, identifies permanent magnet mechanism high-pressure vacuum breaker most common failure.Sufficiently benefit
With Small Sample Database come training pattern, the generalization ability of fault identification is improved, and enriched constantly sample by learning newly-increased data
Classification and recognition can be increased to 90% or more by this library.
Traditional Active Learning Method does not provide perfect solution to how to select the data needed.And this
Invention provides a set of solution model assumed based on cluster using the method for integrating member variable, for Active Learning Method
Formula;So that the sample category inquiry in Active Learning is evidence-based and applied widely.
The SVM model generation technique based on sample pre-selection that the present invention announces, ensure that SVM category of model accuracy
On the basis of, greatly reduce demand of the classifier to category.The experiment proves that feasibility of the invention, classification accuracy are higher than
Similar SVM method, not available for high precision, stability and existing SVM model generating method.
The present invention constructs neural network model BP, is dynamically refined to the weight and threshold value of neural network model BP, from
And dynamic neural network model DBP is obtained, then use adaptive immune genetic AIGA algorithm optimization dynamic neural network model DBP
Prediction model AIGA-DBP is obtained, the predicted value for needing adjustment signal is finally calculated with prediction model AIGA-DBP, it can
The output order accuracy of optimal control signal.
Big data analysis technology has been used in the present invention so that the excavation of influence factor is highly efficient and accurate, influence because
Element consideration is more comprehensive, effectively improves the accuracy of prediction.
Detailed description of the invention
Fig. 1 is the Control system architecture block diagram of intelligent type permanent magnet vacuum circuit breaker provided in an embodiment of the present invention.
In figure:1, signal acquisition module;2, switch-on controlling module;3, central control module;4, photoelectric isolation module;5, event
Hinder identification module;6, alarm;7, display screen.
Fig. 2 is fault recognition method flow chart provided in an embodiment of the present invention.
Specific embodiment
In order to further understand the content, features and effects of the present invention, the following examples are hereby given, and cooperate attached drawing
Detailed description are as follows.
As shown in Figure 1, the control system of intelligent type permanent magnet vacuum circuit breaker provided in an embodiment of the present invention, including:Signal
Acquisition module 1, central control module 3, photoelectric isolation module 4, fault identification module 5, alarm 6, is shown switch-on controlling module 2
Display screen 7.
Signal acquisition module 1 is connect with central control module 3, is used for Acquisition Circuit signal, and signal generates based on the received
Corresponding driving signal;
Switch-on controlling module 2 is connect with central control module 3, for carrying out coarse adjustment and fine tuning to the breaker closing time
Control;
Central control module 3, with signal acquisition module 1, switch-on controlling module 2, photoelectric isolation module 4, fault identification mould
Block 5, alarm 6, display screen 7 connect, and work normally for controlling modules;
Photoelectric isolation module 4 is connect with central control module 3, for using light as media transmission electric signal;
Fault identification module 5 is connect with central control module 3, is used for Acquisition Circuit fault-signal data, establishes failure spy
Sign data sample carries out identification fault signature;
Alarm 6 is connect with central control module 3, for carrying out triggering alarm according to the fault-signal of identification;
Display screen 7 is connect with central control module 3, for showing the signal message and fault status information of acquisition.
2 control method of switch-on controlling module provided by the invention is as follows:
Firstly, determining the initiating of movable contact of vacuum circuit breaker according to the closing coil electric current of known permanent magnetic vacuum breaker
The current value I1 of moment t1, rigid conjunction ta and movable contact of vacuum circuit breaker initiating moment coil, coil is in open position
The timeconstantτ 2 of timeconstantτ 1 and closing position, and calculate parameter S described in the text;
Then, according to aforementioned parameters, the parameter of VVVE fitting current equation is determined, to show that VVVE reference current is bent
Line;
Finally, recalculating fitting current equation parameter by constantly regulate coefficient k and η, obtaining modified reference current
Curve realizes the coarse adjustment and fine tuning to the breaker closing time using coil current tracing, with optimized breaker moving contact fortune
Dynamic characteristic.
5 recognition methods of fault identification module provided by the invention is as follows:
1) fault signature data are normalized, and data are randomly divided into A, B, C3 group data, establish fault signature data sample
This;
2) disaggregated model M1 is obtained using support vector cassification algorithm training data A, utilizes support vector machines increment
It practises algorithm training data B and obtains model M 2;
3) accuracy of data C verifying model M 1 and M2 is utilized.
When the invention works, by 1 Acquisition Circuit signal of signal acquisition module, signal generates corresponding drive based on the received
Dynamic signal;Coarse adjustment is carried out to the breaker closing time by switch-on controlling module 2 and fine tuning controls;Central control module 3 is dispatched
Photoelectric isolation module 4 is using light as media transmission electric signal;By 5 Acquisition Circuit fault-signal data of fault identification module, establish
Fault signature data sample carries out identification fault signature;If identifying fault-signal, triggers alarm 6 and alarm;It is logical
Cross the signal message and fault status information of the display acquisition of display screen 7.
A kind of control method of intelligent type permanent magnet vacuum circuit breaker, including:
By signal acquisition module Acquisition Circuit signal, signal generates corresponding driving signal based on the received;
Coarse adjustment is carried out to the breaker closing time by switch-on controlling module and fine tuning controls;
By photoelectric isolation module using light as media transmission electric signal;
By fault identification module Acquisition Circuit fault-signal data, establishes fault signature data sample and carry out identification failure
Feature;If identifying fault-signal, triggers alarm and alarm;Signal message and event by display screen display acquisition
Hinder status information.
Central control module is to switching signal, optical transport electric signal, fault identification signal after driving signal, coarse adjustment and fine tuning
Carry out control processing, wherein believe first switching signal, optical transport electric signal, fault identification after driving signal, coarse adjustment and fine tuning
It number carries out carrying out sample pre-selection using k means clustering algorithm, according to data pre-selection as a result, to important in unmarked example set
Each example in sample set L* is inquired, and the category that inquiry is obtained returns;Obtain the part " weight in unmarked example set
Want " example and label is obtained, to replace all unmarked examples;Using sample pre-selection as a result, significant samples collection L* is as support
The training set of vector machine Active Learning constructs neural network model BP, generates the initial weight of neural network model BP;By weight
The random number being initialized as between [a, b], wherein a, b are the integer for meeting following equation:
Wherein H is network node in hidden layer;Weight and threshold value to neural network model BP carry out
It is dynamically refined, obtains dynamic neural network model DBP, generate the weight and threshold value of dynamic neural network model DBP,
With adaptive immune genetic AIGA algorithm optimization dynamic neural network model DBP, prediction model AIGA- is obtained
DBP knows according to switching signal, optical transport electric signal, failure after prediction model AIGA-DBP calculating driving signal, coarse adjustment and fine tuning
The predicted value of level signal;Specifically have:Adjust the weight w between neural network model BP hidden layer and output layerkj;Adjust wkj's
Purpose is desirable to the new output o of output node j* pjO is exported than currentlypjCloser to target value tpj, whereinα represents the degree of approach, remains unchanged in each cycle of training, and with node in hidden layer
The adjustment of H and become smaller, do not consider threshold value, have:Wherein wkjWithPoint
The weight of front and back, y Wei not updatedpkFor hidden layer output, △ wkjFor wkjKnots modification;Obtain △ wkjSolution equation:Wherein,It is asked according to least square and error principle
It solves equationObtain △ wkjApproximate solution:
The hidden layer node k of output node j is connected to each, the weight calculated between k and j changes △ wkj, update
Weight simultaneously calculates error of sum square E, then in one optimal k of k ∈ [1, H] interval selection, so that E is minimum;
Adjust the weight v between neural network model BP input layer and hidden layerik;Adjust vikPurpose be once nerve
Network algorithm falls into local minimum point, and modification weight can jump out the minimal point, judge that neural network algorithm falls into local minimum
The condition of point is the change rate △ E=0 of error E, and E>0;Do not consider threshold value, the change of the weight of hidden layer node k by with
Lower equation obtains:Wherein δpj=f-1(ypk+Δypk)-f-1(ypk), M is natural number, and hidden layer exports ypkIt solves
Formula is:Wherein △ ypkFor ypkKnots modification, have:
According to least square and error principle solution formulaThe matrix equation of building, obtains:Aggregative formulaWith
Calculate the consecutive mean knots modification of weight between hidden layer and output layer?Meter
Calculate the consecutive mean knots modification of weight between input layer and hidden layer,M takes 10~
Natural number between 20 obtains the consecutive mean weight of neural network model BP, according to the consecutive mean of neural network model BP
Weight obtains dynamic neural network model DBP;Judge whether the error of predicted value and desired value meets the condition of setting, output control
System instruction.
K mean cluster algorithm includes:
Given sample set D={ x1,x2,;;;,xm, " k mean value " algorithm divides C={ c for cluster gained cluster1,c2,;;;,
ckMinimize square error:
Wherein, x is cluster ciMean vector;The cluster number K=2 of generation;
It is planted with the power of adaptive immune genetic AIGA algorithm optimization dynamic neural network model DBP and threshold value, acquisition is pre-
It surveys model AIGA-DBP and specifically includes following steps:
Antigen recognizing is carried out first, i.e. the identification of mean square error generates initial antibody (dynamic neural network model DBP
Weight and threshold value) after, carry out antibody coding, then calculating antibody fitness and antibody concentration are carried out according to the result of calculating
Adaptive antibody selection operation based on concentration obtains suitable weight and threshold value if this operation meets condition, and by result
Output;If this operation is unsatisfactory for condition, need to carry out adaptive crossover operation and TSP question operation, and carry out weight and threshold
The update of value, then antibody fitness and antibody concentration are recalculated, weight and threshold value until obtaining the condition that meets, last basis
Meet the weight and threshold value of condition, obtains prediction model AIGA-DBP.
Shutting-brake control method includes:
Firstly, determining the initiating of movable contact of vacuum circuit breaker according to the closing coil electric current of known permanent magnetic vacuum breaker
The current value I1 of moment t1, rigid conjunction ta and movable contact of vacuum circuit breaker initiating moment coil, coil is in open position
The timeconstantτ 2 of timeconstantτ 1 and closing position, and calculate parameter S described in the text;
Then, according to aforementioned parameters, the parameter of VVVE fitting current equation is determined, to show that VVVE reference current is bent
Line;
Finally, recalculating fitting current equation parameter by constantly regulate coefficient k and η, obtaining modified reference current
Curve realizes the coarse adjustment and fine tuning to the breaker closing time using coil current tracing, with optimized breaker moving contact fortune
Dynamic characteristic.
Such as Fig. 2, fault recognition method provided in an embodiment of the present invention includes::
S101:Fault signature data are normalized, and data are randomly divided into A, B, C3 group data, establish fault signature data
Sample;
S102:Disaggregated model M1 is obtained using support vector cassification algorithm training data A, utilizes support vector machines increment
Learning algorithm training data B obtains model M 2;
S103:Utilize the accuracy of data C verifying model M 1 and M2.
Verification method includes:
Algorithm 1;Active SVM;
Input:Sample set U, GroundTrue
Output:Optimal classification hyperplane (ω, b)
Obtain the L* under current p
Utilize L* Training Support Vector Machines model M *
If M* is better than the model M generated based on U, Groundtrue, then
In the above-described embodiments, can come wholly or partly by software, hardware, firmware or any combination thereof real
It is existing.When using entirely or partly realizing in the form of a computer program product, the computer program product include one or
Multiple computer instructions.When loading on computers or executing the computer program instructions, entirely or partly generate according to
Process described in the embodiment of the present invention or function.The computer can be general purpose computer, special purpose computer, computer network
Network or other programmable devices.The computer instruction may be stored in a computer readable storage medium, or from one
Computer readable storage medium is transmitted to another computer readable storage medium, for example, the computer instruction can be from one
A web-site, computer, server or data center pass through wired (such as coaxial cable, optical fiber, Digital Subscriber Line (DSL)
Or wireless (such as infrared, wireless, microwave etc.) mode is carried out to another web-site, computer, server or data center
Transmission).The computer-readable storage medium can be any usable medium or include one that computer can access
The data storage devices such as a or multiple usable mediums integrated server, data center.The usable medium can be magnetic Jie
Matter, (for example, floppy disk, hard disk, tape), optical medium (for example, DVD) or semiconductor medium (such as solid state hard disk Solid
State Disk (SSD)) etc..
The above is only the preferred embodiments of the present invention, and is not intended to limit the present invention in any form,
Any simple modification made to the above embodiment according to the technical essence of the invention, equivalent variations and modification, belong to
In the range of technical solution of the present invention.
Claims (10)
1. a kind of control method of intelligent type permanent magnet vacuum circuit breaker, which is characterized in that the intelligent type permanent magnet vacuum circuit breaker
Control method include:
By signal acquisition module Acquisition Circuit signal, signal generates corresponding driving signal based on the received;
Coarse adjustment is carried out to the breaker closing time by switch-on controlling module and fine tuning controls;
By photoelectric isolation module using light as media transmission electric signal;
By fault identification module Acquisition Circuit fault-signal data, establishes fault signature data sample and carry out identification failure spy
Sign;If identifying fault-signal, triggers alarm and alarm;Pass through the signal message and failure of display screen display acquisition
Status information.
Central control module carries out switching signal, optical transport electric signal, fault identification signal after driving signal, coarse adjustment and fine tuning
Control processing, wherein first to switching signal after driving signal, coarse adjustment and fine tuning, optical transport electric signal, fault identification signal into
Row carries out sample pre-selection using k means clustering algorithm, according to data pre-selection as a result, to significant samples in unmarked example set
Each example in collection L* is inquired, and the category that inquiry is obtained returns;Obtain the part " important " in unmarked example set
Example simultaneously obtains label, to replace all unmarked examples;Using sample pre-selection as a result, significant samples collection L* is as supporting vector
The training set of machine Active Learning constructs neural network model BP, generates the initial weight of neural network model BP;Weight is initial
The random number between [a, b] is turned to, wherein a, b is the integer for meeting following equation:Wherein H is that network is hidden
Number containing node layer;The weight and threshold value of neural network model BP are dynamically refined, dynamic neural network model DBP is obtained,
The weight and threshold value of dynamic neural network model DBP are generated,
With adaptive immune genetic AIGA algorithm optimization dynamic neural network model DBP, prediction model AIGA-DBP, root are obtained
It is predicted that model AIGA-DBP calculates switching signal after driving signal, coarse adjustment and fine tuning, optical transport electric signal, fault identification signal
Predicted value;Specifically have:Adjust the weight w between neural network model BP hidden layer and output layerkj;Adjust wkjPurpose be desirable to
The new output of output node jO is exported than currentlypjCloser to target value tpj, wherein
α represents the degree of approach, remains unchanged in each cycle of training, and becomes smaller with the adjustment of node in hidden layer H, does not consider threshold value,
Have:Wherein wkjWithRespectively update the weight of front and back, ypkFor
Hidden layer output, △ wkjFor wkjKnots modification;Obtain △ wkjSolution equation:Wherein,Equation is solved according to least square and error principle
Obtain △ wkjApproximate solution:
The hidden layer node k of output node j is connected to each, the weight calculated between k and j changes △ wkj, update weight
And error of sum square E is calculated, then in one optimal k of k ∈ [1, H] interval selection, so that E is minimum;
Adjust the weight v between neural network model BP input layer and hidden layerik;Adjust vikPurpose be once neural network calculate
Method falls into local minimum point, and modification weight can jump out the minimal point, judge that neural network algorithm falls into the item of local minimum point
Part is the change rate △ E=0 of error E, and E>0;Do not consider threshold value, the change of the weight of hidden layer node k passes through following equation
?:Wherein δpj=f-1(ypk+Δypk)-f-1(ypk), M is natural number, and hidden layer exports ypkSolution formula is:Wherein △ ypkFor ypkKnots modification, have:According to least square and mistake
Poor principle solution formulaThe matrix equation of building, obtains:
Aggregative formulaWithCalculate hidden layer and output layer
Between weight consecutive mean knots modification?It calculates between input layer and hidden layer
The consecutive mean knots modification of weight,M takes the natural number between 10~20, obtains
The consecutive mean weight of neural network model BP obtains dynamic neural network according to the consecutive mean weight of neural network model BP
Model DBP;Judge whether the error of predicted value and desired value meets the condition of setting, exports control instruction.
2. the control method of intelligent type permanent magnet vacuum circuit breaker as described in claim 1, which is characterized in that K mean cluster algorithm
Including:
Given sample set D={ x1,x2,;;;,xm, " k mean value " algorithm divides C={ c for cluster gained cluster1,c2,;;;,ck}
Minimize square error:
Wherein, x is cluster ciMean vector;The cluster number K=2 of generation.
3. the control method of intelligent type permanent magnet vacuum circuit breaker as described in claim 1, which is characterized in that
It is planted with the power of adaptive immune genetic AIGA algorithm optimization dynamic neural network model DBP and mould is predicted in threshold value, acquisition
Type AIGA-DBP specifically includes following steps:
Antigen recognizing is carried out first, i.e. the identification of mean square error generates the initial antibody (power of dynamic neural network model DBP
Value and threshold value) after, antibody coding is carried out, then calculating antibody fitness and antibody concentration, are based on according to the result of calculating
The adaptive antibody selection operation of concentration obtains suitable weight and threshold value, and result is defeated if this operation meets condition
Out;If this operation is unsatisfactory for condition, need to carry out adaptive crossover operation and TSP question operation, and carry out weight and threshold value
Update, then recalculate antibody fitness and antibody concentration, weight and threshold value until obtaining the condition that meets are last according to full
The weight and threshold value of sufficient condition obtain prediction model AIGA-DBP.
4. the control method of intelligent type permanent magnet vacuum circuit breaker as described in claim 1, which is characterized in that shutting-brake control method packet
It includes:
Firstly, determining the initiating moment of movable contact of vacuum circuit breaker according to the closing coil electric current of known permanent magnetic vacuum breaker
The current value I1 of t1, rigid conjunction ta and movable contact of vacuum circuit breaker initiating moment coil, time of the coil in open position
The timeconstantτ 2 of constant, τ 1 and closing position, and calculate parameter S described in the text;
Then, according to aforementioned parameters, the parameter of VVVE fitting current equation is determined, to obtain VVVE reference current curve;
Finally, recalculate fitting current equation parameter by constantly regulate coefficient k and η, it is bent to obtain modified reference current
Line realizes the coarse adjustment and fine tuning to the breaker closing time using coil current tracing, with the movement of optimized breaker moving contact
Characteristic.
5. the control method of intelligent type permanent magnet vacuum circuit breaker as described in claim 1, which is characterized in that fault recognition method packet
It includes::
1) fault signature data are normalized, and data are randomly divided into A, B, C3 group data, establish fault signature data sample;
2) disaggregated model M1 is obtained using support vector cassification algorithm training data A, is calculated using support vector machines incremental learning
Method training data B obtains model M 2;
3) accuracy of data C verifying model M 1 and M2 is utilized.
6. a kind of computer for realizing the control method of intelligent type permanent magnet vacuum circuit breaker described in Claims 1 to 5 any one
Program.
7. a kind of computer for realizing the control method of intelligent type permanent magnet vacuum circuit breaker described in Claims 1 to 5 any one.
8. a kind of computer readable storage medium, including instruction, when run on a computer, so that computer is executed as weighed
Benefit requires the control method of intelligent type permanent magnet vacuum circuit breaker described in 1-5 any one.
9. a kind of intelligent type permanent magnet vacuum circuit breaker for realizing the control method of intelligent type permanent magnet vacuum circuit breaker described in claim 1
The control system of device, which is characterized in that the control system of the intelligent type permanent magnet vacuum circuit breaker includes:
Signal acquisition module is connect with central control module, is used for Acquisition Circuit signal, and signal generates corresponding based on the received
Driving signal;
Switch-on controlling module is connect with central control module, for carrying out coarse adjustment and fine tuning control to the breaker closing time;
Central control module, with signal acquisition module, switch-on controlling module, photoelectric isolation module, fault identification module, alarm
Device, display screen connection, work normally for controlling modules;
Photoelectric isolation module is connect with central control module, for using light as media transmission electric signal;
Fault identification module, connect with central control module, is used for Acquisition Circuit fault-signal data, establishes fault signature data
Sample carries out identification fault signature;
Alarm is connect with central control module, for carrying out triggering alarm according to the fault-signal of identification;
Display screen is connect with central control module, for showing the signal message and fault status information of acquisition.
10. a kind of permanent magnetic vacuum breaker control for the control system for carrying intelligent type permanent magnet vacuum circuit breaker described in claim 9
Control equipment.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810627889.4A CN108879582A (en) | 2018-06-19 | 2018-06-19 | A kind of control system and control method of intelligent type permanent magnet vacuum circuit breaker |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810627889.4A CN108879582A (en) | 2018-06-19 | 2018-06-19 | A kind of control system and control method of intelligent type permanent magnet vacuum circuit breaker |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108879582A true CN108879582A (en) | 2018-11-23 |
Family
ID=64339563
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810627889.4A Pending CN108879582A (en) | 2018-06-19 | 2018-06-19 | A kind of control system and control method of intelligent type permanent magnet vacuum circuit breaker |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108879582A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110415670A (en) * | 2019-06-05 | 2019-11-05 | 新沂市承翔电子有限公司 | A kind of buzzer that multi-angle sluices |
CN113315099A (en) * | 2021-05-20 | 2021-08-27 | 国网宁夏电力有限公司检修公司 | Fault early warning method, medium and system for converter station alternating current filter circuit breaker |
CN115081625A (en) * | 2022-07-21 | 2022-09-20 | 常安集团有限公司 | Intelligent control method and system for miniature circuit breaker |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102931023A (en) * | 2012-11-13 | 2013-02-13 | 北京交通大学 | Realization of control method and device for vacuum circuit breaker based on coil current |
CN103311041A (en) * | 2013-03-19 | 2013-09-18 | 北京交通大学 | Switching control method of VVVE permanent-magnet vacuum circuit breaker |
CN104636479A (en) * | 2015-02-15 | 2015-05-20 | 西安电子科技大学 | Industrial big data driven total completion time prediction method |
CN105938776A (en) * | 2016-05-30 | 2016-09-14 | 安徽征圣智能科技有限公司 | Controller for permanent-magnet vacuum circuit breaker |
CN107818340A (en) * | 2017-10-25 | 2018-03-20 | 福州大学 | Two-stage Air-conditioning Load Prediction method based on K value wavelet neural networks |
-
2018
- 2018-06-19 CN CN201810627889.4A patent/CN108879582A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102931023A (en) * | 2012-11-13 | 2013-02-13 | 北京交通大学 | Realization of control method and device for vacuum circuit breaker based on coil current |
CN103311041A (en) * | 2013-03-19 | 2013-09-18 | 北京交通大学 | Switching control method of VVVE permanent-magnet vacuum circuit breaker |
CN104636479A (en) * | 2015-02-15 | 2015-05-20 | 西安电子科技大学 | Industrial big data driven total completion time prediction method |
CN105938776A (en) * | 2016-05-30 | 2016-09-14 | 安徽征圣智能科技有限公司 | Controller for permanent-magnet vacuum circuit breaker |
CN107818340A (en) * | 2017-10-25 | 2018-03-20 | 福州大学 | Two-stage Air-conditioning Load Prediction method based on K value wavelet neural networks |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110415670A (en) * | 2019-06-05 | 2019-11-05 | 新沂市承翔电子有限公司 | A kind of buzzer that multi-angle sluices |
CN113315099A (en) * | 2021-05-20 | 2021-08-27 | 国网宁夏电力有限公司检修公司 | Fault early warning method, medium and system for converter station alternating current filter circuit breaker |
CN115081625A (en) * | 2022-07-21 | 2022-09-20 | 常安集团有限公司 | Intelligent control method and system for miniature circuit breaker |
CN115081625B (en) * | 2022-07-21 | 2022-11-11 | 常安集团有限公司 | Intelligent control method and system for miniature circuit breaker |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108879582A (en) | A kind of control system and control method of intelligent type permanent magnet vacuum circuit breaker | |
KR101916926B1 (en) | Smart meter system and detection method applying abnormal information detection technology | |
CN101099281B (en) | Intelligent configuration system for power distribution feeder reclosers and switches | |
US9753477B2 (en) | Load forecasting for residential sector demand response | |
CN107566163A (en) | A kind of alarm method and device of user behavior analysis association | |
US9250674B2 (en) | Methods and systems for restoring power based on forecasted loads | |
RU2601957C2 (en) | Method and apparatus for controlling energy services based on market data | |
Anwar et al. | Ensuring data integrity of OPF module and energy database by detecting changes in power flow patterns in smart grids | |
CN109492857A (en) | A kind of distribution network failure risk class prediction technique and device | |
CN102682406B (en) | Power distribution network improved method and system for plug-in electric vehicle | |
JP2020501491A (en) | System and method for dynamic energy storage system control | |
US11383612B2 (en) | Detection of electric vehicle charging at premise | |
WO2013023178A1 (en) | Adaptive stochastic controller for dynamic treatment of cyber-physical systems | |
CN106451418B (en) | Photovoltaic plant divides group's equivalent modeling method online | |
CN109523190A (en) | A kind of management method of electric energy meter, managing device and terminal | |
CN111967738B (en) | Risk early warning method, system and medium of power grid information energy fusion system | |
US20170364839A1 (en) | Information processing device, model construction method, and program recording medium | |
Lundstrom et al. | Fast primary frequency response using coordinated DER and flexible loads: Framework and residential-scale demonstration | |
CN109687426A (en) | Failure rate parameter modeling method, device, equipment and storage medium | |
CN111131237A (en) | Microgrid attack identification method based on BP neural network and grid-connected interface device | |
Li et al. | Islanding fault detection based on data-driven approach with active developed reactive power variation | |
CN108512222A (en) | A kind of intelligent substation complex automatic system | |
Jindal et al. | A data analytical approach using support vector machine for demand response management in smart grid | |
CN110474327A (en) | Power distribution network CPS information-physical combination forecast failure generation method and system | |
Liu et al. | A novel probabilistic framework with interpretability for generator coherency identification |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181123 |
|
RJ01 | Rejection of invention patent application after publication |