CN109031114A - A kind of modeling of spring actuator mechanism circuit-breaker and method for diagnosing faults - Google Patents
A kind of modeling of spring actuator mechanism circuit-breaker and method for diagnosing faults Download PDFInfo
- Publication number
- CN109031114A CN109031114A CN201811147478.1A CN201811147478A CN109031114A CN 109031114 A CN109031114 A CN 109031114A CN 201811147478 A CN201811147478 A CN 201811147478A CN 109031114 A CN109031114 A CN 109031114A
- Authority
- CN
- China
- Prior art keywords
- data
- current
- breaker
- energy storage
- spring
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/327—Testing of circuit interrupters, switches or circuit-breakers
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Testing Electric Properties And Detecting Electric Faults (AREA)
Abstract
The invention discloses a kind of modeling of spring actuator mechanism circuit-breaker and method for diagnosing faults, comprising steps of 1) establish signal transmitting system and data collection system using current sensor, data collecting card, and the environment of the storing data based on LabView is built in computer end；2) circuit-breaker switching on-off coil current and energy storage motor current model are built in Simulink；3) Model Parameter Optimization is carried out using the Stochastic Optimization Algorithms based on genetic algorithm, generates the model emulation signal that may replace actual signal；4) it is clustered using the data that K- means clustering algorithm generates emulation, forms java standard library, fault diagnosis is carried out using the method for fast Template Matching；5) fault diagnosis is carried out using the circuit breaker failure diagnostic method that depth confidence network DBN and softmax classifier combines.The invention enables emulation signals to replace physical fault data, can carry out effective fault diagnosis, and fast Template Matching method operand is small, easy to operate.
Description
Technical field
The present invention relates to the technical fields of breaker modeling and fault diagnosis, refer in particular to a kind of spring operating mechanism open circuit
Device modeling and method for diagnosing faults.
Background technique
Modern power systems scale is more and more huger, and structure becomes increasingly complex, and can people safe and reliable to power equipment
Operation is also increasingly paid close attention to, and breaker attracts attention always as a kind of important power equipment.Investigation display, contact of breaker
Abrasion, non-Switching Synchronization, shelf depreciation etc. can all seriously affect the safe and stable operation of power grid, and cause huge economic damage
It loses.When carrying out Analysis on Fault Diagnosis to breaker, only by collection site data or in Physical Experiment platform emulation, obtained number
Be according to amount it is far from being enough, to obtain all significant conditions of various kinds of equipment and unrealistic, and the difficulty of field conduct and at
This is also unacceptable.For this purpose, this paper presents build circuit-breaker switching on-off coil current, energy storage motor in Simulink
The model of electric current and electric arc generates simulation model signal, and passes through genetic algorithm and the Stochastic Optimization Algorithms based on genetic algorithm
Breaker model parameter is optimized, so that the signal that Simulink is emulated can replace actual signal progress failure and examine
Disconnected analysis.Hereafter, only it can be obtained the number of various the High Voltage Circuit Breaker Conditions by modifying the relevant parameter of above-mentioned breaker model
According to the exploitation extracted for subsequent characteristics with fault diagnosis algorithm provides reasonable, sufficient data.
The present invention proposes that, using the circuit breaker failure diagnostic method based on fast Template Matching, this method is equal first with K-
Value clustering algorithm clusters breaker related data, is formed after java standard library, is carried out using the method for fast Template Matching
Fault diagnosis.Since the corresponding waveform of monitoring data of every class failure is distinguishing, so, this method can carry out effective event
Barrier diagnosis, and fast Template Matching method operand is small, it is easy to operate.The present invention is also proposed using depth confidence network (DBN)
The circuit breaker failure diagnostic method combined with softmax classifier can not only extract signal high level using depth confidence network
Secondary characteristic information, and data dimension can be reduced, effectively prevent the influence caused by classification results of excessive dimension.It is based on
The circuit breaker failure diagnostic method of fast Template Matching is few suitable for circuit-breaker status type and big field is distinguished between class and class
Scape, the method for diagnosing faults based on depth confidence network are suitable for the big scene of every kind of quantity of state of breaker.Pass through the above method
The state of comprehensive analysis breaker provides good basis for the safe operation of breaker.
Summary of the invention
It is an object of the invention to overcome the deficiencies in the prior art, propose a kind of spring actuator mechanism circuit-breaker modeling with
Method for diagnosing faults breaks through conventional on-site acquisition data or Physical Experiment platform emulation, and obtained data volume is much insufficient, and
All significant conditions of various kinds of equipment and unrealistic are obtained, the difficulty and cost of field conduct are also unacceptable disadvantage,
Circuit-breaker switching on-off coil current, energy storage motor current model are built in proposition in Simulink, generate simulation model signal simultaneously
Optimization makes it to replace real data, it is further proposed that the fast Template Matching method based on K- means clustering algorithm and use are deep
It spends the circuit breaker failure diagnostic method that confidence network (DBN) and softmax classifier combine and carries out fault diagnosis, be breaker
Fault data simulation proposes new method, improves the deficiency of existing method for diagnosing faults.
To achieve the above object, technical solution provided by the present invention are as follows: a kind of modeling of spring actuator mechanism circuit-breaker with
Method for diagnosing faults, comprising the following steps:
1) signal transmitting system and data collection system are established using current sensor, data collecting card, and in computer end
Build the environment of the storing data based on LabView；
2) circuit-breaker switching on-off coil current and energy storage motor current model are built in Simulink；
3) Model Parameter Optimization is carried out using the Stochastic Optimization Algorithms based on genetic algorithm, generation may replace actual signal
Model emulation signal；
4) it is clustered using the data that K- means clustering algorithm generates emulation, java standard library is formed, using fast Template
Matched method carries out fault diagnosis；
5) event is carried out using the circuit breaker failure diagnostic method that depth confidence network (DBN) and softmax classifier combine
Barrier diagnosis.
In step 1), signal transmitting system and data collection system are established using current sensor, data collecting card, and
The environment of the storing data based on LabView is built in computer end, the type produced using Xi'an Xi electricity high-voltage switch gear Co., Ltd
Number be LW-40.5 (being hung under mechanism)/T4000-50 high-voltage circuitbreaker, it is good using Hall current sensor linear dynamics,
It is small in size and light-weight, transmit accurate feature, be installed in divide-shut brake magnetic bobbin core inlet-outlet line and energy storage motor into
On line；The current signal that Hall sensor obtains is acquired by data collecting card, and converts digital signal for analog signal；?
It is programmed in LabView software, generates friendly graphical interfaces, coding downloads in controller FPGA to control input and output
Acquisition modes and to acquisite approachs of the module to data.
In step 2), circuit-breaker switching on-off coil current and energy storage motor current model are built in Simulink, it is right
In divide-shut brake coil, it can be equivalent to the series loop of an inductance and resistance, therefore the differential equation of circuit may be expressed as:
In formula, U, R, i are respectively the voltage, resistance and electric current of equivalent circuit, and ψ is magnetic linkage, the air gap of ψ=Li, L and iron core
X is related, and L=L (x) is obtained:
Wherein, v indicates iron core movement velocity, equation Section 3Indicate counter electromotive force, it can score closing coil
Action process be broadly divided into following four stage:
①t∈t0~t1: divide-shut brake coil is in t0Moment is powered, and due to the presence of inductance, the value of coil current is not moment
Reach stable state, but is gradually increased from 0.Meanwhile the attraction of iron core also gradually increases, but its attraction is also insufficient in this stage
To allow iron core to act, so, v=0 solves the differential equation, can get the divide-shut brake coil current expression formula in the stageIron core can be obtained before starting movement, the electric current of divide-shut brake coil is increased with faster speed, in t1
Moment, the size of current value drive iron core to move enough；
②t∈t1~t2: iron core setting in motion generates counter electromotive force, promotes coil current to be gradually reduced, t2It is the stage
Finish time, indicate that iron core has touched operating mechanism and noticeable deceleration or stop motion；
③t∈t2~t3: iron core stop motion, size of current index rise, and expression formula is as follows:
Wherein, Lm> L0, therefore, the rate of climb in the stage is lower than the first stage；
4. t=t3~t4: this stage auxiliary switch K is disconnected, and produces electric arc between the contact of auxiliary switch, arc voltage is fast
Speed increases, and is reduced rapidly coil current to 0；
Energy storage motor electric current is not only influenced by inherent parameters, but also is influenced by the fluctuation of load, typical energy storage motor electricity
Stream fluctuation is broadly divided into following several stages:
1. energy storage motor powers on, start to do non-loaded starting after powering on, armature electric current follows following rule:Wherein, ia、Ua、RaRespectively energy storage motor armature supply, armature voltage, armature resistance, TMIt is normal for the electromechanical time
Number, TM=JRa/C2, J is the rotary inertia of rotor and bindiny mechanism, and C is electromechanical constant, and the π of C=pN φ/2 a, p are electricity
Machine number of pole-pairs, N are armature winding number of effective conductors, and φ is the magnetic flux of every pole, and a is branch logarithm；
2. motor starts turning in current of electric stationary process, but not yet extension spring is done work, so motor is in without negative
Rotary state is carried, electric current tends towards stability:Wherein, RωIt is rotational resistance coefficient；
3. energy storage motor does work, switching-in spring is driven to be allowed to energy storage, the size of the stage current and loads related, the bullet of variation
Spring thermal energy storage process can do following simplification: set the undeformed initial length of spring as l, the radius of spring moved end to axle center is r, is closed
One end of lock spring is circled around energy storage axle, when using the line in the axle center of energy storage axle and spring attachment point as benchmark axis
When, if the angle that turns over of transmission shaft is α, the angle that spring turns over is β, energy storage axle by spring pulling force circumference tangential side
To component beWherein f is spring tension, and Δ l is spring type variable, f⊥After transmission ratio
The load change situation of energy storage motor in the process is obtained, the current variation value of this process is further obtained.
In step 3), Model Parameter Optimization is carried out using the Stochastic Optimization Algorithms based on genetic algorithm, generation can take
For the model emulation signal of actual signal, optimization process is divided into following six step:
3.1) one group of individual is generated at random as initial population, and calculates the fitness size of each individual, and definition adapts to
Spend function are as follows:Wherein yiFor measured value,For the calculated value after model optimization；
3.2) judge whether newborn individual meets the condition of convergence, if meeting 3.3) output is arrived as a result, executing if being unsatisfactory for
3.6), the condition of convergence includes two, and one is that the number of iterations reaches preset the number of iterations, the other is former and later two individuals are suitable
Changing value should be worth；
3.3) it is ranked up by fitness value, and carries out duplication operation；
3.4) according to crossover probability PcCarry out crossover operation；
3.5) according to mutation probability PmCarry out mutation operation；
3.6) it returns and 3.2) is rejudged, until meeting the condition of convergence, circulation terminates, and exports optimum results.
It in step 4), is clustered using the data that K- means clustering algorithm generates emulation, forms java standard library, used
The method of fast Template Matching carries out fault diagnosis, including following five steps:
4.1) initialize: a shared k kind data type selectes k cluster centre (m at random1、m2、…、mk), such as mkIt indicates
K-th of cluster centre；
4.2) x is distributedi: to each sample of the optimization x generated in step 3)i, find the cluster centre nearest from it
Afterwards, it assigns it in the cluster；
4.3) cluster centre is corrected, the center of each cluster is recalculated:Wherein NiIt is i-th
The data point number that a signal includes, xijFor j-th of data point of i-th of signal；
4.4) existing classification deviation is calculated:Wherein miFor the cluster centre of i-th of signal；
4.5) convergence judgement: if J restrains, (m is returned to1、m2、…、mk), algorithm terminates；Otherwise, turn 4.2).
In step 5), using the circuit breaker failure diagnosis side of depth confidence network (DBN) and the combination of softmax classifier
Method carries out fault diagnosis, including following four step:
5.1) characteristic signal (divide-shut brake coil current, energy storage motor electric current) is obtained, set evidence and survey is respectively set
Examination group data；
5.2) training data is inputted into depth confidence network first tier, successively trained from first layer to the second layer；
5.3) according to training label and the classifying rules of softmax classifier, then the basis of the 5.2) step training result
On, from top to low layer trim network parameter, complete the training process of entire depth confidence network；
5.4) test data is input to the model that training finishes in 5.3), output category result, and multi-group data is surveyed
After examination, the accuracy rate of fault distinguishing is obtained.
Compared with prior art, the present invention have the following advantages that with the utility model has the advantages that
1, the present invention, which realizes to model in Simulink for the first time, generates circuit-breaker switching on-off coil current, energy storage motor electric current
Emulation signal, and utilize Stochastic Optimization Algorithms Optimized model parameter so that obtain fault data may replace real data into
Row Analysis on Fault Diagnosis.
2, the present invention realizes for the first time only can be obtained different types of breaker event by changing breaker model parameter
Hinder data, greatly reduces experimental cost while expanding fault data amount.
3, for the present invention by establishing java standard library using K- means clustering algorithm, operand is small, easy to operate, on this basis
It is proposed that fast Template Matching method carries out fault diagnosis.
4, the circuit breaker failure diagnostic method that the present invention is combined using depth confidence network (DBN) and softmax classifier,
The high-level characteristic information of signal can not only be extracted, but also substantially reduce data dimension, effectively prevent excessive dimension to point
Influence caused by class result carries out accurate fault diagnosis.
5, the method for the present invention has extensive use space, behaviour in the generation of circuit breaker failure data and method for diagnosing faults
Make simple, adaptable, there are bright prospects in efficient diagnosis circuit breaker failure type.
Detailed description of the invention
Fig. 1 is logical flow diagram of the present invention.
Fig. 2 is the divide-shut brake coil current illustraton of model that the present invention is built using Simulink.
Fig. 3 is the energy storage motor current model figure that the present invention is built using Simulink.
Fig. 4 is that the present invention builds the resulting divide-shut brake coil current of modeling using Simulink and energy storage motor emulates
The comparison diagram of data and truthful data.
Fig. 5 is that the divide-shut brake coil current that the present invention is optimized using Stochastic Optimization Algorithms and energy storage motor emulate data
With the comparison diagram of truthful data.
Fig. 6 is the cluster and fault diagnosis result of the Fast template matching algorithm proposed by the present invention based on K- mean algorithm
Figure.
Fig. 7 is the circuit breaker failure proposed by the present invention combined using depth confidence network (DBN) and softmax classifier
Diagnostic result figure.
Specific embodiment
The present invention is further explained in the light of specific embodiments.
As shown in Figures 1 to 6, the modeling of spring actuator mechanism circuit-breaker provided by the present embodiment and method for diagnosing faults,
Model LW-40.5 (hanging under the mechanism)/T4000-50 height for having used Xi'an Xi electricity high-voltage switch gear Co., Ltd to produce is broken
Road device, the equipment such as Hall sensor sample true fault data, with Simulink simulated fault data, and propose that two kinds of failures are examined
Disconnected method comprising following steps:
1) signal transmitting system and data collection system are established using current sensor, data collecting card, and in computer end
The environment for building the storing data based on LabView, the model LW- produced using Xi'an Xi electricity high-voltage switch gear Co., Ltd
The high-voltage circuitbreaker of 40.5 (being hung under mechanism)/T4000-50, it is good, small in size using Hall current sensor linear dynamics and
It is light-weight, accurate feature is transmitted, is installed on the inlet-outlet line of divide-shut brake magnetic bobbin core and the inlet wire of energy storage motor；Pass through
Data collecting card acquires the current signal that Hall sensor obtains, and converts digital signal for analog signal；It is soft in LabView
It is programmed in part, generates friendly graphical interfaces, coding downloads in controller FPGA to control input/output module logarithm
According to acquisition modes and to acquisite approachs, the physical fault data of acquisition generate emulation data and require emulator for modeling
Height can replace truthful data.
2) circuit-breaker switching on-off coil current and energy storage motor current model, Fig. 2, Fig. 3 difference are built in Simulink
The division generated for the present invention with the Simulink divide-shut brake coil former built and energy storage motor model, Fig. 4 for simulation model
The comparison diagram of brake cable loop current and energy storage motor electric current and truthful data a, wherein electricity can be equivalent to for divide-shut brake coil
The series loop of sense and resistance, therefore the differential equation of circuit may be expressed as:
In formula, U, R, i are respectively the voltage, resistance and electric current of equivalent circuit, and ψ is magnetic linkage, the air gap of ψ=Li, L and iron core
X is related, and L=L (x) is obtained:
Wherein, v indicates iron core movement velocity, equation Section 3Indicate counter electromotive force, score closing coil
Action process is broadly divided into following four stage:
①t∈t0~t1: divide-shut brake coil is in t0Moment is powered, and due to the presence of inductance, the value of coil current is not moment
Reach stable state, but is gradually increased from 0, meanwhile, the attraction of iron core also gradually increases, but its attraction is also insufficient in this stage
To allow iron core to act, so, v=0 solves the differential equation, obtains the divide-shut brake coil current expression formula in the stageIron core is obtained before starting movement, the electric current of divide-shut brake coil increases rapidly, in t1Moment, current value
Size drive enough iron core move；
②t∈t1~t2: iron core setting in motion generates counter electromotive force, promotes coil current to be gradually reduced, t2It is the stage
Finish time, indicate that iron core has touched operating mechanism and noticeable deceleration or stop motion；
③t∈t2~t3: iron core stop motion, size of current index rise, and expression formula is as follows:
Wherein, Lm> L0, therefore, the rate of climb in the stage is lower than the first stage；
④t∈t3~t4: this stage auxiliary switch K is disconnected, and produces electric arc between the contact of auxiliary switch, arc voltage is fast
Speed increases, and is reduced rapidly coil current to 0；
Energy storage motor electric current is not only influenced by inherent parameters, but also is influenced by the fluctuation of load, typical energy storage motor electricity
Stream fluctuation is broadly divided into following several stages:
Energy storage motor electric current modeling process is as follows: energy storage motor electric current is not only influenced by inherent parameters, but also is loaded
The influence of fluctuation, typical energy storage motor current fluctuation are broadly divided into following several stages:
1. energy storage motor powers on, start to do non-loaded starting after powering on, armature electric current follows following rule:Wherein, ia、Ua、RaRespectively energy storage motor armature supply, armature voltage, armature resistance, TMIt is normal for the electromechanical time
Number, TM=JRa/C2, J is the rotary inertia of rotor and bindiny mechanism, and C is electromechanical constant, and the π of C=pN φ/2 a, p are electricity
Machine number of pole-pairs, N are armature winding number of effective conductors, and φ is the magnetic flux of every pole, and a is branch logarithm；
2. motor starts turning in current of electric stationary process, but not yet extension spring is done work, so motor is in without negative
Rotary state is carried, electric current tends towards stability:Wherein, RωIt is rotational resistance coefficient；
3. energy storage motor does work, switching-in spring is driven to be allowed to energy storage, the size of the stage current and loads related, the bullet of variation
Spring thermal energy storage process can do following simplification: the undeformed initial length of spring is set as l, the radius of spring moved end to axle center is r,
One end of switching-in spring is circled around energy storage axle, when using the line in the axle center of energy storage axle and spring attachment point as benchmark axis
When, if the angle that turns over of transmission shaft is α, the angle that spring turns over is β, energy storage axle by spring pulling force circumference tangential side
To component beWherein, f is spring tension, and Δ l is spring type variable, f⊥By transmission ratio it
The load change situation of energy storage motor in the process is obtained afterwards, further obtains the current variation value of this process.
3) Model Parameter Optimization is carried out using the Stochastic Optimization Algorithms based on genetic algorithm, generation may replace actual signal
Model emulation signal, Fig. 5 is the comparison diagram of the emulation data and real data that are generated later using Stochastic Optimization Algorithms optimization, excellent
Change process is divided into following six step:
3.1) one group of individual is generated at random as initial population, and calculates the fitness size of each individual, and definition adapts to
Spend function are as follows:Wherein yiFor measured value,For the calculated value after model optimization；
3.2) judge whether newborn individual meets the condition of convergence, if meeting 3.3) output is arrived as a result, executing if being unsatisfactory for
3.6), the condition of convergence includes two, and one is that the number of iterations reaches preset the number of iterations, the other is former and later two individuals are suitable
Changing value should be worth；
3.3) it is ranked up by fitness value, and carries out duplication operation；
3.4) according to crossover probability PcCarry out crossover operation；
3.5) according to mutation probability PmCarry out mutation operation；
3.6) it returns and 3.2) is rejudged, until meeting the condition of convergence, circulation terminates, and exports optimum results.
4) it is clustered using the data that K- means clustering algorithm generates emulation, java standard library is formed, using fast Template
Matched method carries out fault diagnosis, and Fig. 6 is cluster result and fault diagnosis result figure, and process includes following five steps:
4.1) initialize: a shared k kind data type selectes k cluster centre (m at random1、m2、…、mk), such as mkIt indicates
K-th of cluster centre；
4.2) x is distributedi: to each sample of the optimization x generated in step 3)i, find the cluster centre nearest from it
Afterwards, it assigns it in the cluster；
4.3) cluster centre is corrected, the center of each cluster is recalculated:Wherein NiIt is i-th
The data point number that a signal includes, xijFor j-th of data point of i-th of signal；
4.4) existing classification deviation is calculated:Wherein miFor the cluster centre of i-th of signal；
4.5) convergence judgement: if J restrains, (m is returned to1、m2、…、mk), algorithm terminates；Otherwise, turn 4.2).
5) event is carried out using the circuit breaker failure diagnostic method that depth confidence network (DBN) and softmax classifier combine
Barrier diagnosis, Fig. 7 are diagnostic result figure, and it is 100% that the diagnostic method, which can obtain fault diagnosis precision, as seen from the figure, algorithmic procedure packet
Include following four step:
5.1) characteristic signal, including divide-shut brake coil current and energy storage motor electric current are obtained, set evidence is respectively set
With test group data；
5.2) training data is inputted into depth confidence network first tier, successively trained from first layer to the second layer；
5.3) according to training label and the classifying rules of softmax classifier, then the basis of the 5.2) step training result
On, from top to low layer trim network parameter, complete the training process of entire depth confidence network；
5.4) test data is input to the model that training finishes in 5.3), output category result, and multi-group data is surveyed
After examination, the accuracy rate of fault distinguishing is obtained.
In conclusion the present invention provides new method to obtain circuit breaker failure data after using method, by
Simulink simulated fault data, and allow fault data that real data is replaced to do diagnosis point by Stochastic Optimization Algorithms
Analysis, it is also proposed that two kinds of method for diagnosing faults, fast Template Matching method operand is small, easy to operate, using depth confidence net
The circuit breaker failure diagnostic method that network (DBN) and softmax classifier combine can not only extract the high-level feature letter of signal
Breath, and data dimension can be reduced, the influence caused by classification results of excessive dimension is effectively prevented, is worthy to be popularized.
Embodiment described above is only the preferred embodiments of the invention, and but not intended to limit the scope of the present invention, therefore
All shapes according to the present invention change made by principle, should all be included within the scope of protection of the present invention.
Claims (6)
1. a kind of spring actuator mechanism circuit-breaker modeling and method for diagnosing faults, it is characterised in that: the breaker modeling is directed to
Circuit-breaker switching on-off coil current, energy storage motor current model utilize after building experiment porch and obtaining physical fault data
Simulink modeling generates emulation signal, by genetic algorithm and based on the Stochastic Optimization Algorithms of genetic algorithm to breaker model
Parameter optimizes, and enables model that must emulate signal and hereafter only passes through instead of actual signal progress Analysis on Fault Diagnosis
The relevant parameter for modifying above-mentioned breaker model can be obtained the operating status of various breakers, propose to be based on fast Template Matching
Circuit breaker failure diagnostic method, breaker related data is clustered using K- means clustering algorithm, formed java standard library it
Afterwards, fault diagnosis is carried out using the method for fast Template Matching, furthermore, it is also proposed that use depth confidence network DBN and softmax
The circuit breaker failure diagnostic method that classifier combines；Itself the following steps are included:
1) signal transmitting system and data collection system are established using current sensor, data collecting card, and is built in computer end
The environment of storing data based on LabView；
2) circuit-breaker switching on-off coil current and energy storage motor current model are built in Simulink；
3) Model Parameter Optimization is carried out using the Stochastic Optimization Algorithms based on genetic algorithm, generates the mould that can replace actual signal
Type emulates signal；
4) it is clustered using the data that K- means clustering algorithm generates emulation, java standard library is formed, using fast Template Matching
Method carry out fault diagnosis；
5) fault diagnosis is carried out using the circuit breaker failure diagnostic method that depth confidence network DBN and softmax classifier combines.
2. a kind of spring actuator mechanism circuit-breaker modeling according to claim 1 and method for diagnosing faults, it is characterised in that:
In step 1), signal transmitting system and data collection system are established using current sensor, data collecting card, and in computer end
The environment for building the storing data based on LabView, the model LW- produced using Xi'an Xi electricity high-voltage switch gear Co., Ltd
The high-voltage circuitbreaker of 40.5 (being hung under mechanism)/T4000-50, by Hall current sensor be mounted on divide-shut brake magnetic bobbin core into
On the inlet wire of outlet and energy storage motor；The current signal that Hall current sensor obtains is acquired by data collecting card, and by mould
Quasi- signal is converted into digital signal；It is programmed in LabView software, generates graphical interfaces, coding downloads to controller
To control input/output module to the acquisition modes and to acquisite approachs of data in FPGA.
3. a kind of spring actuator mechanism circuit-breaker modeling according to claim 1 and method for diagnosing faults, it is characterised in that:
In step 2), circuit-breaker switching on-off coil current and energy storage motor current model are built in Simulink, wherein divide-shut brake
Coil current modeling is as follows: divide-shut brake coil is equivalent to the series loop of an inductance and resistance, therefore the differential equation table of circuit
It is shown as:
In formula, U, R, i are respectively the voltage, resistance and electric current of equivalent circuit, and ψ is magnetic linkage, and the air gap x of ψ=Li, L and iron core has
It closes, L=L (x) is obtained:
Wherein, v indicates iron core movement velocity, equation Section 3Indicate counter electromotive force, the movement of score closing coil
Process is broadly divided into following four stage:
①t∈t0~t1: divide-shut brake coil is in t0Moment is powered, and due to the presence of inductance, the value of coil current is not to reach moment
Stable state, but gradually increased from 0, meanwhile, the attraction of iron core also gradually increases, but its attraction is also not enough to allow in this stage
Iron core movement, so, v=0 solves the differential equation, obtains the divide-shut brake coil current expression formula in the stageIron core is obtained before starting movement, the electric current of divide-shut brake coil increases rapidly, in t1Moment, current value
Size drive enough iron core move；
②t∈t1~t2: iron core setting in motion generates counter electromotive force, promotes coil current to be gradually reduced, t2It is the knot in the stage
The beam moment indicates that iron core has touched operating mechanism and noticeable deceleration or stop motion；
③t∈t2~t3: iron core stop motion, size of current index rise, and expression formula is as follows:
Wherein, Lm> L0, therefore, the rate of climb in the stage is lower than the first stage；
④t∈t3~t4: this stage auxiliary switch K is disconnected, and produces electric arc between the contact of auxiliary switch, arc voltage increases rapidly
Greatly, it is reduced rapidly coil current to 0；
Energy storage motor electric current modeling process is as follows: energy storage motor electric current is not only influenced by inherent parameters, but also by the fluctuation of load
Influence, typical energy storage motor current fluctuation is broadly divided into following several stages:
1. energy storage motor powers on, start to do non-loaded starting after powering on, armature electric current follows following rule:Wherein, ia、Ua、RaRespectively energy storage motor armature supply, armature voltage, armature resistance, TMIt is normal for the electromechanical time
Number, TM=JRa/C2, J is the rotary inertia of rotor and bindiny mechanism, and C is electromechanical constant, and the π of C=pN φ/2 a, p are electricity
Machine number of pole-pairs, N are armature winding number of effective conductors, and φ is the magnetic flux of every pole, and a is branch logarithm；
2. motor starts turning in current of electric stationary process, but not yet extension spring is done work, so motor is in non-loaded turn
Dynamic state, electric current tend towards stability:Wherein, RωIt is rotational resistance coefficient；
3. energy storage motor does work, switching-in spring is driven to be allowed to energy storage, the size of the stage current is related to load variation, spring storage
Energy process can do following simplification: set the undeformed initial length of spring as l, the radius of spring moved end to axle center is r, is closed a floodgate
One end of spring is circled around energy storage axle, when using the line in the axle center of energy storage axle and spring attachment point as benchmark axis,
If the angle that transmission shaft turns over be α, the angle that spring turns over be β, energy storage axle by spring pulling force circumference tangential direction
Component beWherein, f is spring tension, and Δ l is spring type variable, f⊥After transmission ratio
The load change situation of energy storage motor in the process is obtained, the current variation value of this process is further obtained.
4. a kind of spring actuator mechanism circuit-breaker modeling according to claim 1 and method for diagnosing faults, it is characterised in that:
In step 3), Model Parameter Optimization is carried out using the Stochastic Optimization Algorithms based on genetic algorithm, generation can replace true letter
Number model emulation signal, optimization process is divided into following six step:
3.1) one group of individual is generated at random as initial population, and calculates the fitness size of each individual, defines fitness letter
Number are as follows:Wherein yiFor measured value,For the calculated value after model optimization；
3.2) judge whether newborn individual meets the condition of convergence, exported if meeting as a result, executed if being unsatisfactory for 3.3) to 3.6),
The condition of convergence includes two, and one is that the number of iterations reaches preset the number of iterations, the other is former and later two individual fitnesses
Changing value；
3.3) it is ranked up by fitness value, and carries out duplication operation；
3.4) according to crossover probability PcCarry out crossover operation；
3.5) according to mutation probability PmCarry out mutation operation；
3.6) it returns and 3.2) is rejudged, until meeting the condition of convergence, circulation terminates, and exports optimum results.
5. a kind of spring actuator mechanism circuit-breaker modeling according to claim 1 and method for diagnosing faults, it is characterised in that:
It in step 4), is clustered using the data that K- means clustering algorithm generates emulation, java standard library is formed, using fast Template
Matched method carries out fault diagnosis, including following five steps:
4.1) initialize: a shared k kind data type selectes k cluster centre (m at random1、m2、…、mk), mkIt indicates k-th to gather
Class center；
4.2) x is distributedi: to each sample of the optimization x generated in step 3)i, will after finding the cluster centre nearest from it
It is assigned in the cluster；
4.3) cluster centre is corrected, the center of each cluster is recalculated:Wherein NiBelieve for i-th
Number data point number for including, xijFor j-th of data point of i-th of signal；
4.4) existing classification deviation is calculated:Wherein miFor the cluster centre of i-th of signal；
4.5) convergence judgement: if J restrains, (m is returned to1、m2、…、mk), algorithm terminates；Otherwise, turn 4.2).
6. a kind of spring actuator mechanism circuit-breaker modeling according to claim 1 and method for diagnosing faults, it is characterised in that:
In step 5), failure is carried out using the circuit breaker failure diagnostic method that depth confidence network DBN and softmax classifier combines
Diagnosis, including following four step:
5.1) characteristic signal, including divide-shut brake coil current and energy storage motor electric current are obtained, set evidence and survey is respectively set
Examination group data；
5.2) training data is inputted into depth confidence network first tier, successively trained from first layer to the second layer；
5.3) according to training label and the classifying rules of softmax classifier, then the 5.2) on the basis of step training result, from
It is top to low layer trim network parameter, complete the training process of entire depth confidence network；
5.4) test data is input in 5.3) model that training finishes, output category result, and to multi-group data test after,
Obtain the accuracy rate of fault distinguishing.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811147478.1A CN109031114A (en) | 2018-09-29 | 2018-09-29 | A kind of modeling of spring actuator mechanism circuit-breaker and method for diagnosing faults |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811147478.1A CN109031114A (en) | 2018-09-29 | 2018-09-29 | A kind of modeling of spring actuator mechanism circuit-breaker and method for diagnosing faults |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109031114A true CN109031114A (en) | 2018-12-18 |
Family
ID=64615003
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811147478.1A Pending CN109031114A (en) | 2018-09-29 | 2018-09-29 | A kind of modeling of spring actuator mechanism circuit-breaker and method for diagnosing faults |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109031114A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110263837A (en) * | 2019-06-13 | 2019-09-20 | 河海大学 | A kind of circuit breaker failure diagnostic method based on multilayer DBN model |
| CN111913103A (en) * | 2020-08-06 | 2020-11-10 | 国网福建省电力有限公司 | Fault detection method for spring energy storage operating structure circuit breaker |
| CN115291092A (en) * | 2022-03-26 | 2022-11-04 | 福州大学 | Mechanism fault judgment method for switching-on spring energy storage process of spring operating mechanism |
| CN116910594A (en) * | 2023-09-14 | 2023-10-20 | 青岛明思为科技有限公司 | Rolling bearing fault diagnosis method based on impulse neural network |
| CN117250496A (en) * | 2023-11-17 | 2023-12-19 | 国网浙江省电力有限公司双创中心 | Spring energy storage detection method and system of GIS circuit breaker |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107219457A (en) * | 2017-06-15 | 2017-09-29 | 河北工业大学 | Frame-type circuit breaker fault diagnosis and degree assessment method based on operation annex electric current |
| CN107490760A (en) * | 2017-08-22 | 2017-12-19 | 西安工程大学 | The circuit breaker failure diagnostic method of fuzzy neural network is improved based on genetic algorithm |
-
2018
- 2018-09-29 CN CN201811147478.1A patent/CN109031114A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107219457A (en) * | 2017-06-15 | 2017-09-29 | 河北工业大学 | Frame-type circuit breaker fault diagnosis and degree assessment method based on operation annex electric current |
| CN107490760A (en) * | 2017-08-22 | 2017-12-19 | 西安工程大学 | The circuit breaker failure diagnostic method of fuzzy neural network is improved based on genetic algorithm |
Non-Patent Citations (1)
| Title |
|---|
| 石梦洁: "弹簧操动机构断路器建模与故障诊断方法研究", 《中国优秀硕士论文全文数据库 信息科技辑》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110263837A (en) * | 2019-06-13 | 2019-09-20 | 河海大学 | A kind of circuit breaker failure diagnostic method based on multilayer DBN model |
| CN111913103A (en) * | 2020-08-06 | 2020-11-10 | 国网福建省电力有限公司 | Fault detection method for spring energy storage operating structure circuit breaker |
| CN111913103B (en) * | 2020-08-06 | 2022-11-08 | 国网福建省电力有限公司 | Fault detection method for spring energy storage operating structure circuit breaker |
| CN115291092A (en) * | 2022-03-26 | 2022-11-04 | 福州大学 | Mechanism fault judgment method for switching-on spring energy storage process of spring operating mechanism |
| CN115291092B (en) * | 2022-03-26 | 2024-06-04 | 福州大学 | Mechanism fault judging method in spring energy storage process of switching-on spring of spring operating mechanism |
| CN116910594A (en) * | 2023-09-14 | 2023-10-20 | 青岛明思为科技有限公司 | Rolling bearing fault diagnosis method based on impulse neural network |
| CN116910594B (en) * | 2023-09-14 | 2023-12-01 | 青岛明思为科技有限公司 | Rolling bearing fault diagnosis method based on impulse neural network |
| CN117250496A (en) * | 2023-11-17 | 2023-12-19 | 国网浙江省电力有限公司双创中心 | Spring energy storage detection method and system of GIS circuit breaker |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109031114A (en) | A kind of modeling of spring actuator mechanism circuit-breaker and method for diagnosing faults | |
| Amraee et al. | Transient instability prediction using decision tree technique | |
| CN108551167A (en) | A kind of electric power system transient stability method of discrimination based on XGBoost algorithms | |
| CN109061463A (en) | A kind of monitoring of mechanical state of high-voltage circuit breaker and method for diagnosing faults | |
| CN109270442A (en) | High-voltage circuitbreaker fault detection method based on DBN-GA neural network | |
| CN106295153B (en) | A kind of Fault Diagnosis of Aircraft Engine Gas Path method based on twin support vector machines | |
| CN109086817A (en) | A kind of Fault Diagnosis for HV Circuit Breakers method based on deepness belief network | |
| CN109298330A (en) | Fault Diagnosis for HV Circuit Breakers method based on GHPSO-BP | |
| CN110535146A (en) | The Method for Reactive Power Optimization in Power of Policy-Gradient Reinforcement Learning is determined based on depth | |
| CN111060815B (en) | GA-Bi-RNN-based high-voltage circuit breaker fault diagnosis method | |
| CN105242205A (en) | Aviation three-level AC power generator rotary rectifier online fault diagnosis method | |
| CN107392304A (en) | A kind of Wind turbines disorder data recognition method and device | |
| CN109284672A (en) | A kind of Mechanical Failure of HV Circuit Breaker diagnostic method based on PSO-Kmeans algorithm | |
| CN107450016A (en) | Fault Diagnosis for HV Circuit Breakers method based on RST CNN | |
| CN108876163A (en) | The transient rotor angle stability fast evaluation method of comprehensive causality analysis and machine learning | |
| CN110263837A (en) | A kind of circuit breaker failure diagnostic method based on multilayer DBN model | |
| CN110766313A (en) | Cable tunnel comprehensive state evaluation method based on operation and maintenance system | |
| CN114839531A (en) | Motor fault detection method based on group type sparse self-coding and group intelligence | |
| CN113030789A (en) | Series arc fault diagnosis and line selection method based on convolutional neural network | |
| Zhang et al. | Improved GWO-MCSVM algorithm based on nonlinear convergence factor and tent chaotic mapping and its application in transformer condition assessment | |
| Darabian et al. | Combined use of sensitivity analysis and hybrid Wavelet-PSO-ANFIS to improve dynamic performance of DFIG-based wind generation | |
| CN115456106A (en) | High-voltage circuit breaker fault diagnosis model optimization method | |
| Ma et al. | Fault diagnosis of fan gearboxes based on EEMD energy entropy and SOM neural networks | |
| Liu et al. | Fault Diagnosis of Jointless Track Circuit Based on ReliefF-C4. 5 Decision Tree | |
| CN111178617A (en) | Multi-sensor management method based on perception decision guidance |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20181218 |