CN105759199B - It moves the generation of verification electric current low-voltage circuit breaker wink and approaches control method - Google Patents
It moves the generation of verification electric current low-voltage circuit breaker wink and approaches control method Download PDFInfo
- Publication number
- CN105759199B CN105759199B CN201410798569.7A CN201410798569A CN105759199B CN 105759199 B CN105759199 B CN 105759199B CN 201410798569 A CN201410798569 A CN 201410798569A CN 105759199 B CN105759199 B CN 105759199B
- Authority
- CN
- China
- Prior art keywords
- voltage
- current
- electric current
- wink
- error
- 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
Landscapes
- Testing Electric Properties And Detecting Electric Faults (AREA)
- Circuit Arrangement For Electric Light Sources In General (AREA)
Abstract
The present invention relates to the generation of low-voltage circuit breaker wink dynamic verification electric current and control method is approached, is comprised the concrete steps that:The electric voltage regulator that main circuit occurs for wink dynamic verification electric current, and the four coil series connection of strong current transformer primary side, counter circuit power-factor angle are adjusted, and 4I is calculated according to initial currentNCorresponding experiment floating voltage U '40;Electric voltage regulator, and the four coil series connection of strong current transformer primary side are adjusted, strong current transformer secondary voltage is adjusted to U '40, closed circuit, electric current is adjusted to 4I by closed loopN, disconnecting circuit detection floating voltage U40;Nerve network controller calculates 8 times or 12 times of voltage compensation coefficient k8Or k12;If wink, dynamic verification test current was unqualified, pass through ∫ K1ΔI、∫KψΔψRespectively to floating voltage U80Or U120It is modified with closing phase angle;Using feedback control, electronic pressure regulation is coordinated to ensure U80Or U120Precision;Carry out wink dynamic verification experiment;After wink dynamic test test, the error analysis of verification electric current is carried out;On-line training is carried out by online network training mechanism.
Description
Technical field
Technical scheme of the present invention is related to the dynamic verification experiment of low-voltage circuit breaker wink, moves school in specifically low-voltage circuit breaker wink
The generation of galvanoscope and approach control method.
Background technology
Generally strong current transformer primary side is used to take out the generation main circuit of low-voltage circuit breaker wink dynamic verification electric current at present
Head ensures that former secondary no-load voltage ratio is 1,8,12.When the contactor being connect with tap is respectively closed, strong current transformer secondary voltage
Approximation keeps 1,8,12 times of relationship, to carry out specified, 8 times of specified, 12 times of rated current experiments respectively, avoids and big electricity
The connected electric voltage regulator of convertor transformer changes on a large scale, but since strong current transformer primary side is using one group of coil,
Primary side current of transformer is very big when 8 times of progress is tested with 12 times of rated current, needs the line footpath of primary coil very big, and is progress
Rated current is tested, and primary side total number of turns value is very big, to increase the volume of strong current transformer so that its production cost increases
Add.
Wink, dynamic verification electric current had the characteristics that amplitude is big, precision is high, aperiodic component is small.Due to wanting for overload protection characteristics
It asks, even if breaker flows through 8INElectric current, INFor rated current, also to thread off when electric current continues 200ms or more.Therefore, accurate
It really determines, 200ms should be strict controlled in by verifying the duration of electric current.In addition, the dropout of buckle releaser is one by electricity
It is converted into heat, has relationship so as to cause the process and size of current and duration of deformation.This will verify electric current defined
The amplitude of requirement should be rapidly achieved in time, transitional processes are short.And conventional feedback control and constant current technology is required for one
Adjustment time, therefore herein and be not suitable for, so dynamic verification current control of general wink is opened loop control, improving test current
Precision on Shortcomings.
It is test product internal resistance on-line checking that it, which tests process, and when experiment generates specified electricity using feedback control adjustment first
Stream carries voltage and floating voltage by detecting band, calculates the internal resistance of breaker and the internal resistance of experiment power supply.Pass between them
System is:
U10=rsIN+U1 (1)
Wherein:U10For floating voltage, rsIt is equivalent internal resistance, the U of power supply1It is that band carries voltage, i.e. test product internal resistance and electric current
Product.
The switching of strong current transformer primary side is combined with electronic pressure regulation, generates the corresponding voltage of verification electric current.It can derive
Go out to generate 8INCorresponding floating voltage U80:
U80=8rsIN+8U1=8U10 (2)
By the switching of contactor, a rough 8I can be obtainedNCorresponding floating voltage, then using feedback control
System coordinates electronic pressure regulation to ensure U80Precision.8I is obtained by phase selectionN。
However, rsIncluding primary side reduced resistance, it is not a constant numerical value, influence it because being known as:High current becomes
No-load voltage ratio the number of turns of depressor, electronic no-load voltage ratio the number of turns from even pressure regulator, are a very complicated functions, can not Precise Representation.It is practical
Processing method when use is to enable U80=8k8U10.Define k8For 8 times of current compensation factors, it is approximately equal to 1, according to practical feelings
Condition is manually set.It is same to define k12For 12 times of current compensation factors.Ensure the precision of wink dynamic debugging electric current, penalty coefficient k8With
k12It is crucial, due to factors such as power supply, electric voltage regulator and strong current transformer capacity, what the wink streaming current of each grade required
Penalty coefficient k8And k12It is different, therefore the wink streaming current error that system generates is larger.
Invention content
The technical problem to be solved by the present invention is to:The dynamic generation for verifying electric current of low-voltage circuit breaker wink is provided and approaches control
Method, strong current transformer primary side are four coil L1、L2、L3、L4, the turn ratio of four coils is 2:2:1:1, it was testing
Electronic pressure regulation and coil string are taken in journey and the mode combined reduces strong current transformer primary side line footpath, it is cost-effective;Wink moves school
Testing experiment can repeat to re-start, and the current error of this experiment can form one as the amendment of experiment next time
Broad sense closed-loop control.Control uses the complex neural network Step wise approximation control method based on on-line training, to aperiodic component
Successive elimination is realized with amplitude error, improves the control accuracy of wink dynamic verification electric current.
Technical solution is used by the present invention solves the technical problem:Low-voltage circuit breaker wink dynamic verification electric current generation with
Control method is approached, is comprised the concrete steps that:
The first step adjusts the electric voltage regulator that main circuit occurs for wink dynamic verification electric current, and strong current transformer primary side four
The series connection of a coil generates an initial voltage on strong current transformer pair side, corresponding to generate initial current, counter circuit power because
Number angle namely required closing phase angle, and 4I is calculated according to initial currentNCorresponding experiment floating voltage U '40;
Second step adjusts electric voltage regulator, and the four coil series connection of strong current transformer primary side, by strong current transformer
Secondary voltage is adjusted to U '40, closed circuit, electric current is adjusted to 4I by closed loopN, disconnecting circuit detection floating voltage U40;
Third walks, and nerve network controller is according to the desired wink dynamic expectation electric current I debugged*, circuit power factor angle, 4
Times corresponding floating voltage U of rated current40, calculate 8 times or 12 times of voltage compensation coefficient k8Or k12, and according to formula
(3) or (4) calculate separately 8 times or 12 times of rated no-load voltage U80Or U120;
U80=2k8U40 (3)
U120=3k12U40 (4)
4th step, if wink dynamic verification test current it is unqualified, pass through ∫ K1ΔI、∫KψΔ ψ is respectively to floating voltage U80Or
Person U120It is modified with closing phase angle, ∫ K1Δ I is the accumulation item of current amplitude error;∫KψΔψFor the tired of closing phase angle error
Product item;
5th step, strong current transformer primary coil L1And L3Series connection, L2And L4After series connection, then by the line after two groups of series connection
Circle is in parallel, obtains 8INCorresponding experiment floating voltage U80, feedback control is then used, electronic pressure regulation is coordinated to ensure U80Precision;
Or two coil workings of strong current transformer primary side, coil L1And L2Parallel connection obtains 12INCorresponding experiment is unloaded
Then voltage uses feedback control, electronic pressure regulation is coordinated to ensure floating voltage U120Precision;
6th step, silicon-controlled connection circuit is controlled when voltage-phase is circuit power factor angle, and phase selection carries out wink
Dynamic verification experiment;
7th step after wink dynamic test test, carries out the error analysis of verification electric current, the error of high-ranking officers' galvanoscope be decomposed by
In the aperiodic component Δ ψ that phase selection is not allowed to bring, as transient error and test voltage inaccurately brings current amplitude error delta
I can simultaneously be respectively provided by being integrated to current-responsive waveform segments with decomposition computation current amplitude error and transient error
The tolerance of current amplitude error and transient error is made whether the electric current of generation qualified evaluation;
8th step carries out on-line training by online network training mechanism, and training and sample adjustment rule are set to (∫ KIΔ
I)2+(∫KψΔψ)2More than being carried out after a certain numerical value, successive elimination is realized to aperiodic component and current amplitude error.
The beneficial effects of the invention are as follows:Wink, the primary side of main circuit strong current transformer occurred for dynamic verification electric current for four lines
Circle carries out 4 times of rated current and tests four coil series connection;It carries out 8 times of rated current and tests L1And L3Series connection, L2And L4After series connection,
Again by the coils from parallel connection of coils after two groups of series connection;Carry out 12 times of rated current experiments, coil L1With L2Parallel connection increases by above-mentioned variation
The utilization rate of strong current transformer primary coil, while the line footpath of primary coil is reduced, it is substantially reduced strong current transformer
Volume, saved production cost;According to the repeatable feature of wink dynamic verification experiment, it is proposed that verification electric current broad sense closed-loop control
Thought, i.e., the control accuracy of wink dynamic verification electric current can be improved in the correction regulated quantity tested the error of last time as this.
Compared with prior art, it moves the generation of verification electric current low-voltage circuit breaker wink of the present invention and approaches the notable of control method
Progress be:
(1) error of high-ranking officers' galvanoscope is decomposed into since phase selection is not allowed the aperiodic component brought and test voltage inaccurate
Current amplitude error is brought, by being integrated to current-responsive waveform segments, current amplitude error and transient error can be decomposed, carried
Go out the synthesis tolerance of the two errors, whether evaluation verification electric current is qualified.
(2) design realizes the verification electric current neural network control method of broad sense closed loop, using the controlling party of Step wise approximation
Method solves the problems, such as that verification electric current requires amplitude it is difficult to ensure that being rapidly achieved.
(3) nerve network controller can be with on-line training, when model output error is more than tolerance, according to new data sample
This online re -training.
Description of the drawings
Present invention will be further explained below with reference to the attached drawings and examples.
Fig. 1 is low-voltage circuit breaker wink dynamic verification electric current generation main circuit diagram.
Fig. 2 is the dynamic generation for verifying electric current of low-voltage circuit breaker wink and approaches control method work flow diagram.
Specific implementation mode
It moves the generation of verification electric current low-voltage circuit breaker wink of the present invention and approaches control method, comprise the concrete steps that:
The first step adjusts the electric voltage regulator that main circuit occurs for wink dynamic verification electric current, and strong current transformer primary side four
The series connection of a coil generates an initial voltage on strong current transformer pair side, corresponding to generate initial current, counter circuit power because
Number angle namely required closing phase angle, and 4I is calculated according to initial currentNCorresponding experiment floating voltage U '40;
Second step adjusts electric voltage regulator, and the four coil series connection of strong current transformer primary side, by strong current transformer
Secondary voltage is adjusted to U '40, closed circuit, electric current is adjusted to 4I by closed loopN, disconnecting circuit detection floating voltage U40;
Third walks, and nerve network controller is according to the desired wink dynamic expectation electric current I debugged*, circuit power factor angle, 4
Times corresponding floating voltage U of rated current40, calculate 8 times or 12 times of voltage compensation coefficient k8Or k12, and according to formula
(3) or (4) calculate separately 8 times or 12 times of rated no-load voltage U80Or U120;
U80=2k8U40 (3)
U120=3k12U40 (4)
4th step, if wink dynamic verification test current it is unqualified, pass through ∫ K1ΔI、∫KψΔ ψ is respectively to floating voltage U80Or
Person U120It is modified with closing phase angle, ∫ K1Δ I is the accumulation item of current amplitude error;∫KψΔ ψ is the tired of closing phase angle error
Product item;
5th step, strong current transformer primary coil L1And L3Series connection, L2And L4After series connection, then by the line after two groups of series connection
Circle is in parallel, obtains 8INCorresponding experiment floating voltage U80, feedback control is then used, electronic pressure regulation is coordinated to ensure U80Precision;
Or two coil workings of strong current transformer primary side, coil L1And L2Parallel connection obtains 12INCorresponding experiment is unloaded
Then voltage uses feedback control, electronic pressure regulation is coordinated to ensure floating voltage U120Precision;
6th step, silicon-controlled connection circuit is controlled when voltage-phase is circuit power factor angle, and phase selection carries out wink
Dynamic verification experiment;
7th step after wink dynamic test test, carries out the error analysis of verification electric current, the error of high-ranking officers' galvanoscope be decomposed by
In the aperiodic component Δ ψ that phase selection is not allowed to bring, as transient error and test voltage inaccurately brings current amplitude error delta
I can simultaneously be respectively provided by being integrated to current-responsive waveform segments with decomposition computation current amplitude error and transient error
The tolerance of current amplitude error and transient error is made whether the electric current of generation qualified evaluation;
8th step carries out on-line training by online network training mechanism, and training and sample adjustment rule are set to (∫ KIΔ
I)2+(∫KψΔψ)2More than being carried out after a certain numerical value, successive elimination is realized to aperiodic component and current amplitude error.
Ensure the precision of wink dynamic debugging electric current, penalty coefficient k8Or k12It is a crucial factor.Wink dynamic debugging
Expectation electric current, the power-factor angle in circuit, test product impedance and impedance loop are the key parameters for influencing penalty coefficient.These ginsengs
Number connects each other, influences each other and mutually restricts, and is a multi-data fusion system, and neural network is most suitable for solving this kind of ask
Topic.BP networks are current most widely used neural network models, it has powerful non-linear mapping capability and fault tolerance,
Its learning rules are to keep network error quadratic sum minimum to adjust network weight and threshold value by backpropagation.BP neural network
Controller is according to the desired wink dynamic expectation electric current I debugged*, the corresponding unloaded electricity of the power-factor angle in circuit, 4 times of rated current
Press U40, calculate 8 times or 12 times of voltage compensation coefficient k8Or k12, and calculate 8 times or 12 times of rated no-load voltage U80Or
Person U120。
Wink, dynamic verification experiment can repeat to re-start, and the current error of this experiment can be as experiment next time
It corrects, forms a broad sense closed-loop control.Control uses the complex neural network Step wise approximation control method based on on-line training,
Successive elimination is realized to aperiodic component and amplitude error.Wherein ∫ K1Δ I is the accumulation item of current amplitude error;∫KψΔ ψ is to close
The accumulation item of lock phase angle error.Pass through ∫ K1ΔI、∫KψAmendments of the Δ ψ to test voltage and closing phase angle, it is ensured that controlled volume
Step wise approximation.
The error for carrying out verification electric current is decomposed and evaluation
1) amplitude error of electric current:
Electric current is after alternating current circuit is closed a floodgate:
Wherein:ψ-voltage closing phase angle,- hookup power-factor angle, ImThe amplitude of-current cycle component.
The verification electric current of generation has steady-state component i1With transient state component i2Two parts form.Wherein transient state component i2It is to present
Exponential damping.It is generally acknowledged that working asI afterwards2≈0.Therefore, one section rearward is taken in the verification electric current array detected
Can think to include only steady-state component i1.Root-mean-square value, as the virtual value I of electric current are asked with this segment data.It moves with wink and adjusts
The expectation electric current I of examination*Between error delta I=I*- I, as amplitude error.
2) aperiodic component (transient error) of electric current:
The method for seeking may be used integral of the aperiodic component of electric current.Theory deduction is as follows.By (5) both sides from 0~
tNIntegral
It allowsK is integer, is ensured simultaneouslyThen the first item on formula (6) right side is 0.
IfSubstitution formula (6) can obtain after arranging
Formula (7) left side is to verifying on current time from 0~tNIntegral, it can use verification the current sampling data time on
0~tNThe sum of replace.It can be obtained after arrangement
T is the sampling period for verifying electric current in above formula.
According to the above error analysis decomposition result, obtained respectively steady-state error, transient error analytic method, use them
The voltage and phase selection phase angle of experiment next time can be adjusted, to correct error respectively.
3) evaluation of electric current is verified:
Verify the error of electric current phase selection error and amplitude, the different links from testing equipment, closing phase angle error
It is voltage compensation and electronic relief frame section, amplitude error is that power factor detection executes link with phase selection.Verification electric current is commented
Valence is also required to carry out respectively.There are one link error it is exceeded can think verify electric current precision do not reach requirement.
Relative error E may be used in the evaluation of verification current amplitude errorI.It is assumed that relative error is no more than 3% for qualification.
I.e.
It can be according to the amplitude I of aperiodic component and current cycle component that formula (7) obtainmRelative error Eψ, to evaluate
Phase selection error.It is assumed that maximum relative error is no more than 5% for qualification.I.e.
I.e.
From formula (10) as it can be seen that amplitude I to aperiodic component and current cycle componentmRelative error EψRestriction, simultaneously
It is to phase selection angular accuracy requirement.
With ageing equipment the problems such as, some parameters will change, in neural metwork training mechanism always use one at
Constant training sample and object vector, long-play will produce large error.Constantly changed using on-line tuning method defeated
Enter sample and object vector, makes the actual value of their continuous proximity devices.The core of on-line training mechanism is according to actual current
With the error transfer factor training data sample of expectation electric current.It replaces training sample and need to follow and constantly expand training sample, make in sample
Contain different grades of expectation electric current;For the data sample for avoiding input vector from becoming single, ensure that expectation electric current is different etc.
Grade, new input vector in former input vector with its immediate one group to being replaced;Too many event will be differed with training sample
Barrier data are rejected, that is, reject the principle of unusual sample.A training sample is often changed, network re -training is primary, calculates
It obtains voltage compensation coefficient, then wink dynamic debugging electric current is generated by electric current generation mechanism.Training rules and sample adjustment rule are set to
(∫KIΔI)2+(∫KψΔψ)2More than being carried out after a certain numerical value.
Fig. 2 shows that the generation of low-voltage circuit breaker wink of the present invention dynamic verification electric current is with the workflow for approaching control method:
(1) start → whether start to test, be continued waiting for if not starting experiment;It is entered step if starting experiment
(2);
(2) small voltage, low current are exported, the power-factor angle of counter circuit simultaneously calculates 4INCorresponding 4 times of rated no-load electricity
Press U '40;
(3) it adjusts electric voltage regulator and exports 4 times of rated no-load voltage U with coil combination40;
(4) nerve network controller calculates 8 times or 12 times of rated no-load voltage compensation coefficient k of output8Or k12, and calculate
8 times or 12 times of rated no-load voltage U80Or U120;
(5) whether verification electric current mark of conformity is 0, if verification electric current mark of conformity is 0, to test voltage and combined floodgate phase
Angle is modified, i.e. closing phase angle and 8 times or 12 times of rated no-load voltage U80Or U120Error correction;If it is qualified to verify electric current
Mark is not 0, then adjusts electric voltage regulator and export 8 times or 12 times of rated no-load voltages with coil combination;
(6) phase selection carries out wink dynamic verification experiment;
(7) decomposition and evaluation for verifying electric current judge whether verification electric current is qualified, and electric current mark of conformity is verified if qualified
1 is set, then judges whether to test next time;Electric current mark of conformity is verified if unqualified to set to 0, then judges neural network sample
Whether this updates, if update, is updated sample on-line training, judges whether to test later again next time;If not more
Newly, then it directly judges whether to test next time;
(8) if judge to need to be tested, return to step (2) next time;If judgement need not be tested next time,
Then directly exit.
Embodiment 1
It moves the generation of verification electric current the low-voltage circuit breaker wink of the present embodiment and approaches control method, choosing rated current is
8 times of rated current winks of low-voltage circuit breaker dynamic verification experiment of 250A carries out.
The first step adjusts the electric voltage regulator that main circuit occurs for wink dynamic verification electric current, and strong current transformer primary side four
A coil series connection generates a small voltage on strong current transformer pair side, corresponding to generate electric current, the power-factor angle of counter circuit
For 0.4rad namely required closing phase angle, and it is 10.20V according to the corresponding floating voltages of this Current calculation 1000A;
Second step adjusts electric voltage regulator, and the 4 coil series connection of strong current transformer primary side, by strong current transformer
Secondary voltage is adjusted to 10.20V, closed circuit, and electric current is adjusted to 1000A by closed loop, and disconnecting circuit detection floating voltage is
10.54V;
Third walks, and nerve network controller is according to the expectation electric current 2000A of desired wink dynamic debugging, the power factor in circuit
Angle 0.4rad, the corresponding floating voltage 10.54V of 4 times of rated current, calculate 8 times of voltage compensation coefficient K8It is 1.01, and calculates
8 times of rated no-load voltage U80For 21.30V.
4th step is experiment for the first time without being modified to test voltage and closing phase angle;
5th step, strong current transformer primary coil L1、L3Series connection, L2、L4After series connection, then by above two groups of coils from parallel connection of coils,
Then feedback control is used, electronic pressure regulation is coordinated so that the floating voltage of test loop is 21.19V;
6th step controls silicon-controlled connection circuit when voltage-phase is circuit power factor angle 0.4rad, phase selection,
Carry out wink dynamic verification experiment;
7th step is practical after testing to verify electric current 2080.5A, Δ I=80.5A;EI> 3%, Eψ< 5%, so verification electricity
It is unqualified to flow precision.
8th step calculates (∫ KIΔI)2+(∫KψΔψ)2No more than tolerance, neural network sample does not have to online without update
Training;
It is unqualified to verify electric current, is tested next time
Repeat above first, second and third step.
4th step passes through ∫ KIΔ I is to test voltage U80It is modified;
5th step, strong current transformer primary coil L1、L3Series connection, L2、L4After series connection, then by above two groups of coils from parallel connection of coils,
Then feedback control is used, electronic pressure regulation is coordinated so that the floating voltage of test loop is 20.39V;
6th step controls silicon-controlled connection circuit when voltage-phase is circuit power factor angle 0.4rad, phase selection,
Carry out wink dynamic verification experiment;
7th step is practical after testing to verify electric current 2001.5A, Δ I=1.5A;EI=0.075%, Eψ< 5%, so school
Galvanoscope precision is qualified.
Embodiment 2
It moves the generation of verification electric current the low-voltage circuit breaker wink of the present embodiment and approaches control method, choosing rated current is
12 times of rated current winks of low-voltage circuit breaker dynamic verification experiment of 200A carries out.
The first step adjusts the electric voltage regulator that main circuit occurs for wink dynamic verification electric current, and strong current transformer primary side four
A coil series connection generates a small voltage on strong current transformer pair side, corresponding to generate electric current, the power-factor angle of counter circuit
For 1.51rad namely required closing phase angle, and it is 8.16V according to the corresponding floating voltages of this Current calculation 800A;
Second step adjusts electric voltage regulator, and the 4 coil series connection of strong current transformer primary side, by strong current transformer
Secondary voltage is adjusted to 8.16V, closed circuit, and electric current is adjusted to 800A by closed loop, and disconnecting circuit detects floating voltage 8.17V;
Third walks, and nerve network controller is according to the expectation electric current 2400A of desired wink dynamic debugging, the power factor in circuit
Angle 1.51rad, the corresponding floating voltage 8.17V of 4 times of rated current, calculate 12 times of voltage compensation coefficient K12It is 1.004, and counts
Calculate 12 times of rated no-load voltage U120For 24.62V.
4th step is experiment for the first time without being modified to test voltage and closing phase angle;
5th step, strong current transformer primary coil L1、L3Series connection, L2、L4After series connection, then by above two groups of coils from parallel connection of coils,
Then feedback control is used, electronic pressure regulation is coordinated so that the floating voltage of test loop is 24.49V;
6th step, silicon-controlled connection circuit is controlled when voltage-phase is circuit power factor angle 1.51rad, and phase selection is closed
Lock carries out wink dynamic verification experiment;
7th step is practical after testing to verify electric current 2405A, Δ I=5A;EI=0.208%, Eψ> 5%, so verification electricity
It is unqualified to flow precision.
8th step calculates (∫ KIΔI)2+(∫KψΔψ)2No more than tolerance, neural network sample does not have to online without update
Training;
It is unqualified to verify electric current, is tested next time
Repeat above first, second and third step.
4th step passes through ∫ KψΔ ψ is modified closing phase angle.
5th step, strong current transformer primary coil L1、L3Series connection, L2、L4After series connection, then by above two groups of coils from parallel connection of coils,
Then feedback control is used, electronic pressure regulation is coordinated so that the floating voltage of test loop is 24.49V;
6th step, silicon-controlled connection circuit is controlled when voltage-phase is circuit power factor angle 1.525rad, and phase selection is closed
Lock carries out wink dynamic verification experiment;
7th step is practical after testing to verify electric current 2404.4A, Δ I=4.4A;EI=0.183%, Eψ< 5%, so school
Galvanoscope precision is qualified.
Embodiment 3
It moves the generation of verification electric current the low-voltage circuit breaker wink of the present embodiment and approaches control method, choosing rated current is
8 times of rated current winks of low-voltage circuit breaker dynamic verification experiment of 100A carries out.
Expectation electric current is 800A in the present embodiment, carries out test of many times, is divided into two classes, and one kind is neural metwork training sample
This is not updated and neural network is without on-line training, and another kind is neural network according to low-voltage circuit breaker wink dynamic verification electric current
The 8th step of control method is generated and approached when meeting condition, carries out Sample Refreshment and neural metwork training.When by one section
Between under the not newer mode of sample reality output wink dynamic verification electric current be 787.2A, and Sample Refreshment carries out the mode of on-line training
Dynamic verification electric current will be 800.9A in lower reality output wink, it is seen that has higher control essence after Neural Network Online training institution is added
Degree.
Strong current transformer primary coil is divided into four in above-described embodiment, difference is carried out by the way of coil combination
Experiment.
In above-described embodiment, 8 times are calculated with 12 times of rated voltage penalty coefficients by nerve network controller.
In above-described embodiment, after wink dynamic test test, error analysis is carried out, the error of high-ranking officers' galvanoscope is decomposed into due to choosing
Mutually it is not allowed the aperiodic component Δ ψ brought and test voltage inaccurately to bring amplitude error Δ I, to be respectively processed;Together
When respectively provide their tolerance, qualified evaluation is made whether to the electric current of generation.
The current error that this in above-described embodiment is tested can form a broad sense closed loop as the amendment of experiment next time
Control, control use the complex neural network Step wise approximation control method based on on-line training, are missed to aperiodic component and amplitude
Difference realizes successive elimination.
Claims (1)
1. moving the generation of verification electric current low-voltage circuit breaker wink and approaching control method, comprise the concrete steps that:
The first step adjusts the electric voltage regulator that main circuit occurs for wink dynamic verification electric current, and four lines of strong current transformer primary side
Circle series connection generates an initial voltage on strong current transformer pair side, corresponding to generate initial current, counter circuit power factor
Angle namely required closing phase angle, and 4I is calculated according to initial currentNCorresponding experiment floating voltage U '40;
Second step adjusts electric voltage regulator, and the four coil series connection of strong current transformer primary side, by strong current transformer pair side
Voltage is adjusted to U '40, closed circuit, electric current is adjusted to 4I by closed loopN, disconnecting circuit detection floating voltage U40;
Third walks, and nerve network controller is according to the desired wink dynamic expectation electric current I debugged*, circuit power factor angle, 4 times it is specified
The corresponding floating voltage U of electric current40, calculate 8 times or 12 times of voltage compensation coefficient k8Or k12, and according to formula (3) or (4)
Calculate separately 8 times or 12 times of rated no-load voltage U80Or U120;
U80=2k8U40 (3)
U120=3k12U40 (4)
4th step, if wink dynamic verification test current it is unqualified, pass through ∫ K1ΔI、∫KψΔ ψ is respectively to floating voltage U80Or U120
It is modified with closing phase angle, ∫ K1Δ I is the accumulation item of current amplitude error;∫KψΔ ψ is the accumulation item of closing phase angle error;
5th step, strong current transformer primary coil L1And L3Series connection, L2And L4After series connection, then simultaneously by the coil after two groups of series connection
Connection, obtains 8INCorresponding experiment floating voltage U80, feedback control is then used, electronic pressure regulation is coordinated to ensure U80Precision;
Or two coil workings of strong current transformer primary side, coil L1And L2Parallel connection obtains 12INThe unloaded electricity of corresponding experiment
Then pressure uses feedback control, electronic pressure regulation is coordinated to ensure floating voltage U120Precision;
6th step, silicon-controlled connection circuit is controlled when voltage-phase is circuit power factor angle, and phase selection carries out moving school in wink
Test experiment;
7th step after wink dynamic test test, carries out the error analysis of verification electric current, and the error of high-ranking officers' galvanoscope is decomposed into due to choosing
Mutually it is not allowed the aperiodic component Δ ψ brought, as transient error and test voltage inaccurately to bring current amplitude error delta I, leads to
It crosses and current-responsive waveform segments is integrated, can be with decomposition computation current amplitude error and transient error, while respectively providing electric current
The tolerance of amplitude error and transient error is made whether the electric current of generation qualified evaluation;
8th step carries out on-line training by online network training mechanism, and training and sample adjustment rule are set to (∫ KIΔI)2+(∫
KψΔψ)2More than being carried out after a certain numerical value, successive elimination is realized to aperiodic component and current amplitude error.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410798569.7A CN105759199B (en) | 2014-12-19 | 2014-12-19 | It moves the generation of verification electric current low-voltage circuit breaker wink and approaches control method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410798569.7A CN105759199B (en) | 2014-12-19 | 2014-12-19 | It moves the generation of verification electric current low-voltage circuit breaker wink and approaches control method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN105759199A CN105759199A (en) | 2016-07-13 |
CN105759199B true CN105759199B (en) | 2018-08-17 |
Family
ID=56339373
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410798569.7A Expired - Fee Related CN105759199B (en) | 2014-12-19 | 2014-12-19 | It moves the generation of verification electric current low-voltage circuit breaker wink and approaches control method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105759199B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111030041A (en) * | 2019-12-27 | 2020-04-17 | 中国人民解放军海军潜艇学院 | Current control method and control device for checking circuit breaker |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN2821593Y (en) * | 2005-09-29 | 2006-09-27 | 河北工业大学 | Instantaneous property detector for plastic shell breaker |
CN102087338A (en) * | 2011-01-14 | 2011-06-08 | 刘湘如 | Overload characteristic test bed of intelligent circuit breaker and control method of overload characteristic test bed |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8032316B2 (en) * | 2008-04-16 | 2011-10-04 | Phoenix Broadband Technologies, Llc | Measuring and monitoring a power source |
US8914249B2 (en) * | 2008-06-30 | 2014-12-16 | Hioki Denki Kabushiki Kaisha | Resistance measuring apparatus |
-
2014
- 2014-12-19 CN CN201410798569.7A patent/CN105759199B/en not_active Expired - Fee Related
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN2821593Y (en) * | 2005-09-29 | 2006-09-27 | 河北工业大学 | Instantaneous property detector for plastic shell breaker |
CN102087338A (en) * | 2011-01-14 | 2011-06-08 | 刘湘如 | Overload characteristic test bed of intelligent circuit breaker and control method of overload characteristic test bed |
Non-Patent Citations (2)
Title |
---|
基于神经网络的瞬动校验电流控制技术的研究;杜太行等;《电力***保护与控制》;20090516;第37卷(第10期);第44-47、58页 * |
塑壳断路器瞬动保护可靠性检测技术;王金伟等;《低压电器》;20111231(第12期);第48-52页 * |
Also Published As
Publication number | Publication date |
---|---|
CN105759199A (en) | 2016-07-13 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106655144A (en) | Dual-closed-loop control-based active voltage arc extinction method and apparatus for power distribution network in fault | |
US10061276B2 (en) | Machine learning system and magnetizer for motor | |
CN102545210A (en) | Method and system for optimizing and regulating distribution network power supply reliability indexes | |
CN109713688A (en) | Bulk power grid static voltage stability online preventive control method and system | |
CN110045317B (en) | Mutual inductor metering error online detection method and system | |
CN108107716A (en) | A kind of Parameter Measuring method based on improved BP neural network | |
CN106019093B (en) | A kind of online soft sensor method of three-phawse arc furnace arc length | |
CN103024959B (en) | Intelligent electrode regulator and intelligent electrode control method | |
CN105759199B (en) | It moves the generation of verification electric current low-voltage circuit breaker wink and approaches control method | |
CN103595058B (en) | Method for determining capacity of reactive power compensation device of refining furnace | |
CN104037766B (en) | Method for self-adaptive neural inversion control of three-phase parallel connection type active filter | |
CN106056218A (en) | Equipment monthly maintenance scheduling optimization method considering overload and transient stability constraint | |
Gabash et al. | On the control of main substations between transmission and distribution systems | |
Oubbati et al. | Transient stability constrained optimal power flow using teaching learning based optimization | |
WO2016023579A1 (en) | Method for estimating status of ac networks and subsequent adaptive control | |
CN101082661A (en) | Current mutual inductor for intelligent electrical apparatus release and control method thereof | |
CN111985091A (en) | Rapid safety correction method for power system containing UPFC | |
Reddy et al. | PSO based optimal reactive power dispatch for voltage profile improvement | |
CN109936134A (en) | A kind of thermostabilization passway for transmitting electricity power limit section recognition methods and equipment | |
CN109829178A (en) | A kind of substation's real-time emulation method based on model split plot design | |
CN109886316A (en) | Transformer state parameter combination forecasting method based on cloud system similarity weight distribution | |
CN109038558A (en) | Simulate the power distribution network planning scheme security assessment method of optimum control process | |
CN207051355U (en) | High-current generation device | |
CN106709664A (en) | User-oriented power quality evaluation method | |
CN102708261B (en) | The modeling method of automatism voltage control model in a kind of electric system simulation |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20180817 Termination date: 20211219 |