CN107490760A - The circuit breaker failure diagnostic method of fuzzy neural network is improved based on genetic algorithm - Google Patents
The circuit breaker failure diagnostic method of fuzzy neural network is improved based on genetic algorithm Download PDFInfo
- Publication number
- CN107490760A CN107490760A CN201710725538.2A CN201710725538A CN107490760A CN 107490760 A CN107490760 A CN 107490760A CN 201710725538 A CN201710725538 A CN 201710725538A CN 107490760 A CN107490760 A CN 107490760A
- Authority
- CN
- China
- Prior art keywords
- mrow
- msub
- layer
- genetic algorithm
- neural network
- 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
- G01R31/3271—Testing of circuit interrupters, switches or circuit-breakers of high voltage or medium voltage devices
- G01R31/3272—Apparatus, systems or circuits therefor
-
- 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
- G01R31/3271—Testing of circuit interrupters, switches or circuit-breakers of high voltage or medium voltage devices
- G01R31/3275—Fault detection or status indication
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Feedback Control In General (AREA)
Abstract
The invention discloses the circuit breaker failure diagnostic method that fuzzy neural network is improved based on genetic algorithm, it is specially：Step 1, the Historical Monitoring characteristic quantity for obtaining breaker and corresponding diagnosis, structure is based on the improved fuzzy neural network model of genetic algorithm, and initializes fuzzy neural network and genetic algorithm, sets the algorithm parameter of correlation；Step 2, trained based on the improved fuzzy neural network model of genetic algorithm according to the breaker historical data in step 1；Step 3, after the completion for the treatment of step 2, obtain the real-time monitoring feature amount of breaker；Step 4, according to obtaining diagnosing monitored the High Voltage Circuit Breaker Condition based on the breaker real-time monitoring feature amount obtained in the improved fuzzy neural network model of genetic algorithm and step 3 in step 2.The circuit breaker failure diagnostic method of the present invention, can comprehensive assessment breaker working condition, and can carries out Precise Diagnosis in independent failure classification.
Description
Technical field
The invention belongs to transformer fault diagnosis technology field, and in particular to one kind improves fuzzy god based on genetic algorithm
Circuit breaker failure diagnostic method through network.
Background technology
Current power industry attention rate highest technological innovation field is the intelligent on-line fault diagnosis of power transmission and transforming equipment
Technology.And because existing on-Line Monitor Device hardware design tends to maturing, so Future Development intelligent online assessment experts
Software by be grid equipment on-line fault diagnosis technical intelligence key.
Following grid equipment intellectuality repair based on condition of component will be born under a kind of new thinking in " internet+" is instructed
Electric power apparatus examination cloud service platform based on big data analysis.Breaker plays protection power equipment and adjusting device networks and transported
Capable effect, it is the part of most critical in power network, therefore to the intelligent requirements of its on-line fault diagnosis also more and more higher.
The content of the invention
It is an object of the invention to provide a kind of circuit breaker failure diagnosis side that fuzzy neural network is improved based on genetic algorithm
Method, can comprehensive assessment breaker working condition, and can carries out Precise Diagnosis in independent failure classification.
The technical solution adopted in the present invention is that the circuit breaker failure diagnosis of fuzzy neural network is improved based on genetic algorithm
Method, specifically implement according to following steps：
Step 1, the Historical Monitoring characteristic quantity for obtaining breaker and corresponding diagnosis, structure are improved based on genetic algorithm
Fuzzy neural network model, and initialize fuzzy neural network and genetic algorithm, set the algorithm parameter of correlation；
Step 2, trained based on the improved fuzzy neural network of genetic algorithm according to the breaker historical data in step 1
Model；
Step 3, after the completion for the treatment of step 2, the real-time monitoring feature amount of breaker is obtained, its monitoring variable includes main circuit current
Signal, divide-shut brake coil current signal, energy storage motor starting current signal and divide-shut brake iron core displacement signal；
Step 4, according to being obtained in step 2 based on being obtained in the improved fuzzy neural network model of genetic algorithm and step 3
Breaker real-time monitoring feature amount diagnose monitored the High Voltage Circuit Breaker Condition.
The features of the present invention also resides in：
In step 1 structure be based on the improved fuzzy neural network model of genetic algorithm, and initialize fuzzy neural network and
Genetic algorithm, specifically implement according to following steps：
Step a, fuzzy neural network is established：Its structure is similar to BP neural network, by input layer, three layers of hidden layer and defeated
Go out layer composition, the calculating process of each layer of node is replaced by fuzzy logic, changes the error-duration model of networking connection weight and threshold value
Process is replaced by genetic algorithm, and specific method is as follows：
First layer is input layer, each node of this layer respectively with each input quantity xiConnection, it is therefore an objective to set x will be inputted
=[x1,x2,…,xn]TNext layer is delivered to, the node layer number is N1=n；
For the second layer equivalent to the blurring device in fuzzy logic, one node correspondingly includes a Linguistic Value, the layer
It is the degree of membership for calculating input quantity
In formula (1), i=1,2 ..., n, j=1,2 ..., mi, n is the number of input quantity, miIt is xiFuzzy partition number, should
Node layer number is
With the 4th layer equivalent to the indistinct logic computer in fuzzy logic, knowledge base acts on this two layers third layer, and the 3rd
A node in layer just represents a fuzzy rule, and the layer is for determining corresponding fuzzy rule, and can calculate often rules and regulations
Relevance grade then, i.e.,：
In formula (2)：i1∈{1,2,…,m1, i2∈{1,2,…,m2..., in∈{1,2,…,mn, j=1,2 ...,
M,The interstitial content of this layer is N3=m；
4th layer completion be normalization calculate, according to formula (3), its nodes is equal with last layer, i.e.,：
N4=N3=m；
Layer 5 is output layer, the anti -fuzzy introductional also corresponded in fuzzy logic, is accomplished that defuzzification calculates,
I.e.：
In formula (4)：wijIt is yiJ-th of Linguistic Value membership function central value, its vector mode is as follows：
In formula (5)：
Step b, after step a, genetic algorithm is established：Encoded using binary system, the conventional " roulette wheel of selection operation
Method "；Crossover operation is used in initial operating stage with mutation operation, crossover operation is carried out in 0.65 crossing-over rate with average value, with flat
Average carries out mutation operation in 0.55 aberration rate, runs the later stage in algorithm, intersection behaviour is carried out in 0.05 crossing-over rate with average value
Make, mutation operation is carried out in 0.09 aberration rate with average value, crossing-over rate P is calculated according to formula (7) and formula (8)cWith aberration rate Pm：
In formula (7) and formula (8)：fmaxFor maximum individual adaptation degree, favgFor average individual fitness, f ' intersects behaviour to perform
Make the maximum adaptation degree in individual, f is to perform the maximum adaptation degree in mutation operation individual；
Object function for modulus type output quantity and training sample desired output poor sum minimum value, i.e.,：
In formula (9), YFNN-GAFor the output valve of FNN-GA models, YdataFor the desired output of training sample, N is sample
Number；
Individual adaptation degree is as follows：
In formula (10), CmaxElect the maximum individual adaptation degree of population as；
Step c, after the completion for the treatment of step b, fuzzy neural network and genetic algorithm are initialized, sets the algorithm parameter of correlation：
Have for fuzzy neural network：For the rule provided by expert, third layer is corresponded to the connection weight of respective rule
wij0.8 is arranged to, threshold θij0 is arranged to, to ensure the priority of expertise；Then by remaining connection weight wijIt is arranged to
The random number and threshold θ of (- 1,0.8)ijThe random number of (- 1,1) is arranged to, learning rate is arranged to 0.8；
Have for genetic algorithm：Setting population is M=100, and it is T=200 to evolve and terminate algebraically, and initial crossing-over rate is Pc0
=0.7, initial aberration rate is Pm0=0.001.
Step 2 is specifically implemented according to following steps：
Step 2.1, by the use of breaker historical data as the input value of training sample, be used as instruction by the use of corresponding diagnosis
Practice the output valve of sample；
Step 2.2, after step 2.1, the input value of training sample is input in network, observing the output valve of network is
It is no identical with the output valve of desired training sample；
Step 2.3, by after step 2.2, if having reached desired output, model training is completed, and all models are joined
Number renewal is that assurance model can carry out accurate fault diagnosis to the breaker of every kind of model into corresponding database, therefore often
The breaker of kind model should all have corresponding database；
If failing to reach desired output, existing connection weight and threshold value are encoded with genetic algorithm；
Step 2.4, after step 2.3, carry out genetic manipulation and produce new colony, further decoding produces new connection weight
Original value is replaced with threshold value；
Step 2.5,2.2~step 2.4 of repeat step, untill output reaches expectation or iterations reaches, extremely
This model training finishes.
In step 3, the real-time monitoring feature amount of breaker of acquisition is gathered by the High Voltage Circuit Breaker Condition monitoring system.
The structure of the High Voltage Circuit Breaker Condition monitoring system is：Including microprocessor, microprocessor is connected to power supply mould
Block, information process unit, data storage cell, communication module, and the input of information process unit is connected with sensor；Sensing
Device includes main circuit current sensor, divide-shut brake coil current sensor, energy storage motor starting current sensor and divide-shut brake
Displacement transducer unshakable in one's determination.
The beneficial effects of the invention are as follows：
(1) circuit breaker failure diagnostic method of the invention, can comprehensive assessment breaker working condition, and can is independent
Precise Diagnosis is carried out on fault category, compensate for the missing of existing the High Voltage Circuit Breaker Condition appraisal procedure；
(2) circuit breaker failure diagnostic method of the invention, there is the efficient habit, excellent adaptivity, quick learnt by oneself
Data-handling capacity and good interactive capability, and quick obtaining and to failure sample can be carried out to existing expertise
This progress analytic learning, complete diagnostic knowledge base is obtained, accurate authority is provided so as to be assessed for the running status of breaker.
Brief description of the drawings
Fig. 1 is the flow chart of the circuit breaker failure diagnostic method of the present invention；
Fig. 2 is built in the circuit breaker failure diagnostic method of the present invention based on the improved fuzzy neural network of genetic algorithm
The structural representation of model；
Fig. 3 is that the structure of the High Voltage Circuit Breaker Condition monitoring system used in the circuit breaker failure diagnostic method of the present invention is shown
It is intended to；
Fig. 4 is the Evaluated effect figure of the circuit breaker failure diagnostic method of the present invention.
In figure, 1. microprocessors, 2. power modules, 3. information process units, 4. sensors, 5. communication modules, 6. data
Memory cell.
Embodiment
The present invention is described in detail with reference to the accompanying drawings and detailed description.
The present invention improves the circuit breaker failure diagnostic method of fuzzy neural network based on genetic algorithm, as shown in figure 1, specifically
Implement according to following steps：
Step 1, the Historical Monitoring characteristic quantity for obtaining breaker and corresponding diagnosis, structure are improved based on genetic algorithm
Fuzzy neural network model, and initialize fuzzy neural network and genetic algorithm, set the algorithm parameter of correlation, specifically according to
Following methods are implemented：
The Historical Monitoring characteristic quantity of breaker is provided with corresponding diagnosis by associate power enterprise；
Structure is based on the improved fuzzy neural network model of genetic algorithm, and initializes fuzzy neural network and calculated with heredity
Method, specifically implement according to following steps：
Step a, fuzzy neural network is established：Its structure is similar to BP neural network, by input layer, three layers of hidden layer and defeated
Go out layer composition, as shown in Fig. 2 the calculating process of each layer of node is replaced by fuzzy logic, change networking connection weight and threshold value
Error-duration model process replaced by genetic algorithm, specific method is as follows：
First layer is input layer, each node of this layer respectively with each input quantity xiConnection, it is therefore an objective to set x will be inputted
=[x1,x2,…,xn]TNext layer is delivered to, the node layer number is N1=n；
The second layer is equivalent to the blurring device in fuzzy logic, and one node correspondingly includes a Linguistic Value, such as：
NB, NS etc., this layer are the degrees of membership for calculating input quantity
In formula (1), i=1,2 ..., n, j=1,2 ..., mi, n is the number of input quantity, miIt is xiFuzzy partition number, should
Node layer number is
With the 4th layer equivalent to the indistinct logic computer in fuzzy logic, knowledge base acts on this two layers third layer, and the 3rd
A node in layer just represents a fuzzy rule, and the layer is for determining corresponding fuzzy rule, and can calculate often rules and regulations
Relevance grade then, i.e.,：
In formula (2)：i1∈{1,2,…,m1, i2∈{1,2,…,m2..., in∈{1,2,…,mn, j=1,2 ...,
M,The interstitial content of this layer is N3=m；
4th layer completion be normalization calculate, as shown in Equation 3, its nodes is equal with last layer, i.e.,：
N4=N3=m；
Layer 5 is output layer, the anti -fuzzy introductional also corresponded in fuzzy logic, is accomplished that defuzzification calculates,
I.e.：
In formula (4)：wijIt is yiJ-th of Linguistic Value membership function central value, its vector mode is as follows：
In formula (5)：
Step b, after step a, genetic algorithm is established：Encoded using binary system, the conventional " roulette wheel of selection operation
Method "；Crossover operation is used in initial operating stage with mutation operation, and crossover operation is carried out in 0.65 or so crossing-over rate with average value,
Mutation operation is carried out in 0.55 or so aberration rate with average value, the later stage is run in algorithm, with average value in 0.05 cross
Rate carries out crossover operation, and mutation operation is carried out in 0.09 or so aberration rate with average value, calculates and intersects according to formula (7) and formula (8)
Rate PcWith aberration rate Pm：
In formula (7) and formula (8)：fmaxFor maximum individual adaptation degree, favgFor average individual fitness, f ' intersects behaviour to perform
Make the maximum adaptation degree in individual, f is to perform the maximum adaptation degree in mutation operation individual；
Object function for modulus type output quantity and training sample desired output poor sum minimum value, i.e.,：
In formula (9), YFNN-GAFor the output valve of FNN-GA models, YdataFor the desired output of training sample, N is sample
Number；
Individual adaptation degree is as follows：
In formula (10), CmaxElect the maximum individual adaptation degree of population as；
Step c, after the completion for the treatment of step b, fuzzy neural network and genetic algorithm are initialized, sets the algorithm parameter of correlation：
Have for fuzzy neural network：For the rule provided by expert, third layer is corresponded to the connection weight of respective rule
wij0.8 is arranged to, threshold θij0 is arranged to, to ensure the priority of expertise；Then by remaining connection weight wijIt is arranged to
The random number and threshold θ of (- 1,0.8)ijThe random number of (- 1,1) is arranged to, learning rate is arranged to 0.8；
Have for genetic algorithm：Setting population is M=100, and it is T=200 to evolve and terminate algebraically, and initial crossing-over rate is Pc0
=0.7, initial aberration rate is Pm0=0.001.
Step 2, trained based on the improved fuzznet of genetic algorithm according to the breaker historical data in step 1
Network, specifically implement according to following steps：
Step 2.1, by the use of breaker historical data as the input value of training sample, be used as instruction by the use of corresponding diagnosis
Practice the output valve of sample；
Step 2.2, after step 2.1, the input value of training sample is input in network, observing the output valve of network is
It is no identical with the output valve of desired training sample；
Step 2.3, by after step 2.2, if having reached desired output, model training is completed, and all models are joined
Number renewal is that assurance model can carry out accurate fault diagnosis to the breaker of every kind of model into corresponding database, therefore often
The breaker of kind model should all have corresponding database；
If failing to reach desired output, existing connection weight and threshold value are encoded with genetic algorithm；
Step 2.4, after step 2.3, carry out genetic manipulation and produce new colony, further decoding produces new connection weight
Original value is replaced with threshold value；
Step 2.5, repeat step 2.2 arrive step 2.4, untill output reaches expectation or iterations reaches, extremely
This model training finishes.
Step 3, after the completion for the treatment of step 2, the real-time monitoring feature amount of breaker is obtained, its monitoring variable includes main circuit current
Signal, divide-shut brake coil current signal, energy storage motor starting current signal and divide-shut brake iron core displacement signal；
The real-time monitoring feature amount of breaker of acquisition is gathered by the High Voltage Circuit Breaker Condition monitoring system；
The structure of the High Voltage Circuit Breaker Condition monitoring system is：As shown in figure 3, including microprocessor 1, microprocessor 1 divides
Power module 2, information process unit 3, data storage cell 6, communication module 5, and the input of information process unit 3 are not connected with
End is connected with sensor 4；Sensor 4 includes main circuit current sensor, divide-shut brake coil current sensor, energy storage motor and opened
Streaming current sensor and divide-shut brake iron core displacement transducer.
The operation principle of the High Voltage Circuit Breaker Condition monitoring system is as follows：
The monitoring system is made up of microprocessor 1 and related peripheral hardware with sensor 4, microprocessor 1 and information processing
Unit 3 can be used for the main circuit current signal, divide-shut brake coil current signal, storage received and preliminary treatment sensor 4 monitors
Energy electric motor starting current signal and divide-shut brake iron core displacement signal, then protected data by data storage cell 6 and communication module
Deposit and be uploaded to host computer and shown and analyzed and processed, host computer can be realized to prison by using the monitoring method of the present invention
Signal is surveyed to be analyzed and processed and judge equipment fault and provide diagnostic result.
Step 4, according to being obtained in step 2 based on being obtained in the improved fuzzy neural network model of genetic algorithm and step 3
Breaker real-time monitoring feature amount diagnose monitored the High Voltage Circuit Breaker Condition.
Experimental verification：
As shown in figure 4, be according on test platform to VJY-12P/T630-25-210 (Z) indoor type vacuum type high pressure
Breaker carries out related simulated experiment, then using obtained data training pattern and verifies the diagnosis effect of model, wherein
It can show that only once diagnostic error, this explanation present invention improve fuzzy neural to model based on genetic algorithm in 50 experiments
The circuit breaker failure diagnostic method of network more accurately can implement fault diagnosis to breaker.
The present invention improves the circuit breaker failure diagnostic method of fuzzy neural network based on genetic algorithm, and energy quick obtaining is existing
Expertise knowledge and can be from historical diagnostic data learning to breaker fault signature, compensate for existing breaker fortune
The missing of row state evaluating method.
Claims (5)
1. based on genetic algorithm improve fuzzy neural network circuit breaker failure diagnostic method, it is characterised in that specifically according to
Lower step is implemented：
Step 1, the Historical Monitoring characteristic quantity for obtaining breaker and corresponding diagnosis, structure are based on the improved mould of genetic algorithm
Neural network model is pasted, and initializes fuzzy neural network and genetic algorithm, sets the algorithm parameter of correlation；
Step 2, trained based on the improved fuzzy neural network mould of genetic algorithm according to the breaker historical data in step 1
Type；
Step 3, after the completion for the treatment of step 2, the real-time monitoring feature amount of breaker is obtained, its monitoring variable is believed including main circuit current
Number, divide-shut brake coil current signal, energy storage motor starting current signal and divide-shut brake iron core displacement signal；
Step 4, according to being obtained in step 2 based on obtaining in genetic algorithm improved fuzzy neural network model and step 3
Breaker real-time monitoring feature amount diagnoses monitored the High Voltage Circuit Breaker Condition.
2. the circuit breaker failure diagnostic method according to claim 1 that fuzzy neural network is improved based on genetic algorithm, its
It is characterised by, structure is based on the improved fuzzy neural network model of genetic algorithm in the step 1, and initializes fuzznet
Network and genetic algorithm, specifically implement according to following steps：
Step a, fuzzy neural network is established：Its structure is similar to BP neural network, by input layer, three layers of hidden layer and output layer
Form, the calculating process of each layer of node is replaced by fuzzy logic, changes the error-duration model process of networking connection weight and threshold value
Replaced by genetic algorithm, specific method is as follows：
First layer is input layer, each node of this layer respectively with each input quantity xiConnection, it is therefore an objective to set x=will be inputted
[x1,x2,…,xn]TNext layer is delivered to, the node layer number is N1=n；
For the second layer equivalent to the blurring device in fuzzy logic, one node correspondingly includes a Linguistic Value, and the layer is to use
To calculate the degree of membership of input quantity
<mrow>
<msubsup>
<mi>&mu;</mi>
<mi>i</mi>
<mi>j</mi>
</msubsup>
<mo>=</mo>
<msub>
<mi>&mu;</mi>
<msubsup>
<mi>A</mi>
<mi>i</mi>
<mi>j</mi>
</msubsup>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
<mo>;</mo>
</mrow>
In formula (1), i=1,2 ..., n, j=1,2 ..., mi, n is the number of input quantity, miIt is xiFuzzy partition number, the layer section
Count out for
Third layer is with the 4th layer equivalent to the indistinct logic computer in fuzzy logic, and knowledge base is acted on this two layers, in third layer
A node just represent a fuzzy rule, the layer is for determining corresponding fuzzy rule, and can be calculated per rule
Relevance grade, i.e.,：
<mrow>
<msub>
<mi>a</mi>
<mi>j</mi>
</msub>
<mo>=</mo>
<mi>min</mi>
<mo>{</mo>
<msubsup>
<mi>&mu;</mi>
<mn>1</mn>
<msub>
<mi>i</mi>
<mn>1</mn>
</msub>
</msubsup>
<mo>,</mo>
<msubsup>
<mi>&mu;</mi>
<mn>2</mn>
<msub>
<mi>i</mi>
<mn>2</mn>
</msub>
</msubsup>
<mo>,</mo>
<mn>...</mn>
<msubsup>
<mi>&mu;</mi>
<mi>n</mi>
<msub>
<mi>i</mi>
<mi>n</mi>
</msub>
</msubsup>
<mo>}</mo>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>2</mn>
<mo>)</mo>
</mrow>
<mo>;</mo>
</mrow>
In formula (2)：i1∈{1,2,…,m1, i2∈{1,2,…,m2..., in∈{1,2,…,mn, j=1,2 ..., m,The interstitial content of this layer is N3=m；
4th layer completion be normalization calculate, according to formula (3), its nodes is equal with last layer, i.e.,：
N4=N3=m；
<mrow>
<msub>
<mover>
<mi>a</mi>
<mo>&OverBar;</mo>
</mover>
<mi>j</mi>
</msub>
<mo>=</mo>
<mfrac>
<msub>
<mi>a</mi>
<mi>j</mi>
</msub>
<mrow>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>m</mi>
</munderover>
<msub>
<mi>a</mi>
<mi>i</mi>
</msub>
</mrow>
</mfrac>
<mo>,</mo>
<mrow>
<mo>(</mo>
<mi>j</mi>
<mo>=</mo>
<mn>1</mn>
<mo>,</mo>
<mn>2</mn>
<mo>,</mo>
<mo>...</mo>
<mo>,</mo>
<mi>m</mi>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>3</mn>
<mo>)</mo>
</mrow>
<mo>;</mo>
</mrow>
Layer 5 is output layer, the anti -fuzzy introductional also corresponded in fuzzy logic, is accomplished that defuzzification calculates, i.e.,：
<mrow>
<msub>
<mi>y</mi>
<mi>i</mi>
</msub>
<mo>=</mo>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>m</mi>
</munderover>
<msub>
<mi>w</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
</msub>
<msub>
<mover>
<mi>a</mi>
<mo>&OverBar;</mo>
</mover>
<mi>j</mi>
</msub>
<mo>,</mo>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
<mo>,</mo>
<mn>2</mn>
<mo>,</mo>
<mo>...</mo>
<mo>,</mo>
<mi>r</mi>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>4</mn>
<mo>)</mo>
</mrow>
<mo>;</mo>
</mrow>
In formula (4)：wijIt is yiJ-th of Linguistic Value membership function central value, its vector mode is as follows：
<mrow>
<mi>y</mi>
<mo>=</mo>
<mi>w</mi>
<mover>
<mi>a</mi>
<mo>&OverBar;</mo>
</mover>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>5</mn>
<mo>)</mo>
</mrow>
<mo>;</mo>
</mrow>
In formula (5)：
Step b, after step a, genetic algorithm is established：Encoded using binary system, selection operation conventional " roulette wheel method "；Hand over
Fork operation is used in initial operating stage with mutation operation, is carried out crossover operation in 0.65 crossing-over rate with average value, is existed with average value
0.55 aberration rate carries out mutation operation, and the later stage is run in algorithm, crossover operation is carried out in 0.05 crossing-over rate with average value, with flat
Average carries out mutation operation in 0.09 aberration rate, and crossing-over rate P is calculated according to formula (7) and formula (8)cWith aberration rate Pm：
<mrow>
<msub>
<mi>P</mi>
<mi>c</mi>
</msub>
<mo>=</mo>
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<mn>0.8</mn>
<mo>-</mo>
<mfrac>
<mrow>
<mo>(</mo>
<mn>0.8</mn>
<mo>-</mo>
<mn>0.5</mn>
<mo>)</mo>
<mo>(</mo>
<msup>
<mi>f</mi>
<mo>&prime;</mo>
</msup>
<mo>-</mo>
<msub>
<mi>f</mi>
<mrow>
<mi>a</mi>
<mi>v</mi>
<mi>g</mi>
</mrow>
</msub>
<mo>)</mo>
</mrow>
<mrow>
<msub>
<mi>f</mi>
<mrow>
<mi>m</mi>
<mi>a</mi>
<mi>x</mi>
</mrow>
</msub>
<mo>-</mo>
<msub>
<mi>f</mi>
<mrow>
<mi>a</mi>
<mi>v</mi>
<mi>g</mi>
</mrow>
</msub>
</mrow>
</mfrac>
<mo>,</mo>
</mrow>
</mtd>
<mtd>
<mrow>
<msup>
<mi>f</mi>
<mo>&prime;</mo>
</msup>
<mo>&GreaterEqual;</mo>
<msub>
<mi>f</mi>
<mrow>
<mi>a</mi>
<mi>v</mi>
<mi>g</mi>
</mrow>
</msub>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mn>0.8</mn>
<mo>,</mo>
</mrow>
</mtd>
<mtd>
<mrow>
<msup>
<mi>f</mi>
<mo>&prime;</mo>
</msup>
<mo><</mo>
<msub>
<mi>f</mi>
<mrow>
<mi>a</mi>
<mi>v</mi>
<mi>g</mi>
</mrow>
</msub>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>7</mn>
<mo>)</mo>
</mrow>
<mo>;</mo>
</mrow>
<mrow>
<msub>
<mi>P</mi>
<mi>m</mi>
</msub>
<mo>=</mo>
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<mn>0.1</mn>
<mo>-</mo>
<mfrac>
<mrow>
<mo>(</mo>
<mn>0.1</mn>
<mo>-</mo>
<mn>0.001</mn>
<mo>)</mo>
<mo>(</mo>
<msub>
<mi>f</mi>
<mrow>
<mi>m</mi>
<mi>a</mi>
<mi>x</mi>
</mrow>
</msub>
<mo>-</mo>
<mi>f</mi>
<mo>)</mo>
</mrow>
<mrow>
<msub>
<mi>f</mi>
<mrow>
<mi>m</mi>
<mi>a</mi>
<mi>x</mi>
</mrow>
</msub>
<mo>-</mo>
<msub>
<mi>f</mi>
<mrow>
<mi>a</mi>
<mi>v</mi>
<mi>g</mi>
</mrow>
</msub>
</mrow>
</mfrac>
<mo>,</mo>
</mrow>
</mtd>
<mtd>
<mrow>
<mi>f</mi>
<mo>&GreaterEqual;</mo>
<msub>
<mi>f</mi>
<mrow>
<mi>a</mi>
<mi>v</mi>
<mi>g</mi>
</mrow>
</msub>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mn>0.1</mn>
<mo>,</mo>
</mrow>
</mtd>
<mtd>
<mrow>
<mi>f</mi>
<mo><</mo>
<msub>
<mi>f</mi>
<mrow>
<mi>a</mi>
<mi>v</mi>
<mi>g</mi>
</mrow>
</msub>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>8</mn>
<mo>)</mo>
</mrow>
<mo>;</mo>
</mrow>
In formula (7) and formula (8)：fmaxFor maximum individual adaptation degree, favgIt is individual to perform crossover operation for average individual fitness, f '
Maximum adaptation degree in body, f are to perform the maximum adaptation degree in mutation operation individual；
Object function for modulus type output quantity and training sample desired output poor sum minimum value, i.e.,：
<mrow>
<mi>J</mi>
<mo>=</mo>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>N</mi>
</munderover>
<mo>|</mo>
<msub>
<mi>Y</mi>
<mrow>
<mi>F</mi>
<mi>N</mi>
<mi>N</mi>
<mo>-</mo>
<mi>G</mi>
<mi>A</mi>
</mrow>
</msub>
<mo>-</mo>
<msub>
<mi>Y</mi>
<mrow>
<mi>d</mi>
<mi>a</mi>
<mi>t</mi>
<mi>a</mi>
</mrow>
</msub>
<mo>|</mo>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>9</mn>
<mo>)</mo>
</mrow>
<mo>;</mo>
</mrow>
In formula (9), YFNN-GAFor the output valve of FNN-GA models, YdataFor the desired output of training sample, N is number of samples；
Individual adaptation degree is as follows：
In formula (10), CmaxElect the maximum individual adaptation degree of population as；
Step c, after the completion for the treatment of step b, fuzzy neural network and genetic algorithm are initialized, sets the algorithm parameter of correlation：
Have for fuzzy neural network：For the rule provided by expert, third layer is corresponded to the connection weight w of respective ruleijIf
0.8 is set to, threshold θij0 is arranged to, to ensure the priority of expertise；Then by remaining connection weight wijBe arranged to (- 1,
0.8) random number and threshold θijThe random number of (- 1,1) is arranged to, learning rate is arranged to 0.8；
Have for genetic algorithm：Setting population is M=100, and it is T=200 to evolve and terminate algebraically, and initial crossing-over rate is Pc0=
0.7, initial aberration rate is Pm0=0.001.
3. the circuit breaker failure diagnostic method according to claim 1 that fuzzy neural network is improved based on genetic algorithm, its
It is characterised by, the step 2 is specifically implemented according to following steps：
Step 2.1, by the use of breaker historical data as the input value of training sample, be used as training sample by the use of corresponding diagnosis
This output valve；
Step 2.2, after step 2.1, the input value of training sample is input in network, observe network output valve whether with
The output valve of desired training sample is identical；
Step 2.3, by after step 2.2, if having reached desired output, model training is completed, and by all model parameters more
It is the accurate fault diagnosis of breaker progress that assurance model can be to every kind of model newly into corresponding database, therefore every kind of type
Number breaker should all have corresponding database；
If failing to reach desired output, existing connection weight and threshold value are encoded with genetic algorithm；
Step 2.4, after step 2.3, carry out genetic manipulation and produce new colony, further decoding produces new connection weight and threshold
Value replaces original value；
Step 2.5,2.2~step 2.4 of repeat step, untill output reaches expectation or iterations reaches, so far mould
Type training finishes.
4. the circuit breaker failure diagnostic method according to claim 1 that fuzzy neural network is improved based on genetic algorithm, its
It is characterised by, in the step 3, the real-time monitoring feature amount of breaker of acquisition is adopted by the High Voltage Circuit Breaker Condition monitoring system
Collection.
5. the circuit breaker failure diagnostic method according to claim 1 that fuzzy neural network is improved based on genetic algorithm, its
It is characterised by, the structure of the High Voltage Circuit Breaker Condition monitoring system is：Including microprocessor (1), the microprocessor (1) point
Power module (2), information process unit (3), data storage cell (6), communication module (5), and information processing list are not connected with
The input of first (3) is connected with sensor (4)；
The sensor (4) includes main circuit current sensor, divide-shut brake coil current sensor, energy storage motor starting current
Sensor and divide-shut brake iron core displacement transducer.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710725538.2A CN107490760A (en) | 2017-08-22 | 2017-08-22 | The circuit breaker failure diagnostic method of fuzzy neural network is improved based on genetic algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710725538.2A CN107490760A (en) | 2017-08-22 | 2017-08-22 | The circuit breaker failure diagnostic method of fuzzy neural network is improved based on genetic algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107490760A true CN107490760A (en) | 2017-12-19 |
Family
ID=60645869
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710725538.2A Pending CN107490760A (en) | 2017-08-22 | 2017-08-22 | The circuit breaker failure diagnostic method of fuzzy neural network is improved based on genetic algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107490760A (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109031114A (en) * | 2018-09-29 | 2018-12-18 | 华南理工大学 | A kind of modeling of spring actuator mechanism circuit-breaker and method for diagnosing faults |
| CN109270442A (en) * | 2018-08-21 | 2019-01-25 | 西安工程大学 | High-voltage circuitbreaker fault detection method based on DBN-GA neural network |
| CN110716133A (en) * | 2019-09-06 | 2020-01-21 | 国网浙江省电力有限公司嘉兴供电公司 | High-voltage circuit breaker fault studying and judging method based on Internet of things and big data technology |
| CN111060815A (en) * | 2019-12-17 | 2020-04-24 | 西安工程大学 | GA-Bi-RNN-based high-voltage circuit breaker fault diagnosis method |
| CN111179576A (en) * | 2019-11-15 | 2020-05-19 | 国网江苏省电力有限公司 | Power utilization information acquisition fault diagnosis method and system with inductive learning function |
| CN111982198A (en) * | 2020-08-21 | 2020-11-24 | 中国南方电网有限责任公司超高压输电公司南宁局 | State maintenance method and system for high-voltage circuit breaker |
| CN112257988A (en) * | 2020-09-29 | 2021-01-22 | 中广核工程有限公司 | Complex accident feature identification and risk early warning system and method for nuclear power plant |
| CN112288153A (en) * | 2020-10-22 | 2021-01-29 | 福州大学 | Automatic optimization method for initial value weight of cerebellum model neural network fault diagnoser |
| CN114793304A (en) * | 2022-06-22 | 2022-07-26 | 深圳市文浩科技有限公司 | Artificial intelligence Internet of things data information measuring, transmitting and analyzing method |
| CN116087692A (en) * | 2023-04-12 | 2023-05-09 | 国网四川省电力公司电力科学研究院 | Distribution network tree line discharge fault identification method, system, terminal and medium |
| CN116628425A (en) * | 2023-06-01 | 2023-08-22 | 常州易宝网络服务有限公司 | Big data real-time monitoring system and method |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20020016665A1 (en) * | 1998-10-22 | 2002-02-07 | Ulyanov Sergei V. | System for intelligent control of an engine based on soft computing |
| WO2008025093A1 (en) * | 2006-09-01 | 2008-03-06 | Innovative Dairy Products Pty Ltd | Whole genome based genetic evaluation and selection process |
| CN101917150A (en) * | 2010-06-24 | 2010-12-15 | 江苏大学 | Robust controller of permanent magnet synchronous motor based on fuzzy-neural network generalized inverse and construction method thereof |
| CN102768361A (en) * | 2012-07-09 | 2012-11-07 | 东南大学 | GPS/INS combined positioning method based on genetic particle filtering and fuzzy neural network |
| CN104133372A (en) * | 2014-07-09 | 2014-11-05 | 河海大学常州校区 | Room temperature control algorithm based on fuzzy neural network |
| CN106099850A (en) * | 2016-06-06 | 2016-11-09 | 南京理工大学 | CT saturation identification improved method based on transient current feature |
| CN106291351A (en) * | 2016-09-20 | 2017-01-04 | 西安工程大学 | Primary cut-out fault detection method based on convolutional neural networks algorithm |
-
2017
- 2017-08-22 CN CN201710725538.2A patent/CN107490760A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20020016665A1 (en) * | 1998-10-22 | 2002-02-07 | Ulyanov Sergei V. | System for intelligent control of an engine based on soft computing |
| WO2008025093A1 (en) * | 2006-09-01 | 2008-03-06 | Innovative Dairy Products Pty Ltd | Whole genome based genetic evaluation and selection process |
| CN101917150A (en) * | 2010-06-24 | 2010-12-15 | 江苏大学 | Robust controller of permanent magnet synchronous motor based on fuzzy-neural network generalized inverse and construction method thereof |
| CN102768361A (en) * | 2012-07-09 | 2012-11-07 | 东南大学 | GPS/INS combined positioning method based on genetic particle filtering and fuzzy neural network |
| CN104133372A (en) * | 2014-07-09 | 2014-11-05 | 河海大学常州校区 | Room temperature control algorithm based on fuzzy neural network |
| CN106099850A (en) * | 2016-06-06 | 2016-11-09 | 南京理工大学 | CT saturation identification improved method based on transient current feature |
| CN106291351A (en) * | 2016-09-20 | 2017-01-04 | 西安工程大学 | Primary cut-out fault detection method based on convolutional neural networks algorithm |
Non-Patent Citations (1)
| Title |
|---|
| 张伟等: "基于遗传算法的动态模糊神经网络城市快速路入口匝道控制", 《公路交通科技》 * |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109270442A (en) * | 2018-08-21 | 2019-01-25 | 西安工程大学 | High-voltage circuitbreaker fault detection method based on DBN-GA neural network |
| CN109031114A (en) * | 2018-09-29 | 2018-12-18 | 华南理工大学 | A kind of modeling of spring actuator mechanism circuit-breaker and method for diagnosing faults |
| CN110716133A (en) * | 2019-09-06 | 2020-01-21 | 国网浙江省电力有限公司嘉兴供电公司 | High-voltage circuit breaker fault studying and judging method based on Internet of things and big data technology |
| CN110716133B (en) * | 2019-09-06 | 2022-03-15 | 国网浙江省电力有限公司嘉兴供电公司 | High-voltage circuit breaker fault studying and judging method based on Internet of things and big data technology |
| CN111179576B (en) * | 2019-11-15 | 2021-08-31 | 国网江苏省电力有限公司 | Power utilization information acquisition fault diagnosis method and system with inductive learning function |
| CN111179576A (en) * | 2019-11-15 | 2020-05-19 | 国网江苏省电力有限公司 | Power utilization information acquisition fault diagnosis method and system with inductive learning function |
| CN111060815B (en) * | 2019-12-17 | 2021-09-14 | 西安工程大学 | GA-Bi-RNN-based high-voltage circuit breaker fault diagnosis method |
| CN111060815A (en) * | 2019-12-17 | 2020-04-24 | 西安工程大学 | GA-Bi-RNN-based high-voltage circuit breaker fault diagnosis method |
| CN111982198A (en) * | 2020-08-21 | 2020-11-24 | 中国南方电网有限责任公司超高压输电公司南宁局 | State maintenance method and system for high-voltage circuit breaker |
| CN112257988A (en) * | 2020-09-29 | 2021-01-22 | 中广核工程有限公司 | Complex accident feature identification and risk early warning system and method for nuclear power plant |
| CN112288153A (en) * | 2020-10-22 | 2021-01-29 | 福州大学 | Automatic optimization method for initial value weight of cerebellum model neural network fault diagnoser |
| CN112288153B (en) * | 2020-10-22 | 2022-06-14 | 福州大学 | Automatic optimization method for initial value weight of cerebellum model neural network fault diagnoser |
| CN114793304A (en) * | 2022-06-22 | 2022-07-26 | 深圳市文浩科技有限公司 | Artificial intelligence Internet of things data information measuring, transmitting and analyzing method |
| CN116087692A (en) * | 2023-04-12 | 2023-05-09 | 国网四川省电力公司电力科学研究院 | Distribution network tree line discharge fault identification method, system, terminal and medium |
| CN116628425A (en) * | 2023-06-01 | 2023-08-22 | 常州易宝网络服务有限公司 | Big data real-time monitoring system and method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107490760A (en) | The circuit breaker failure diagnostic method of fuzzy neural network is improved based on genetic algorithm | |
| CN105354587B (en) | A kind of method for diagnosing faults of wind-driven generator group wheel box | |
| CN113779496B (en) | Power equipment state evaluation method and system based on equipment panoramic data | |
| CN113255848B (en) | Water turbine cavitation sound signal identification method based on big data learning | |
| CN101872165A (en) | Method for fault diagnosis of wind turbines on basis of genetic neural network | |
| CN107066759A (en) | A kind of Vibration Fault Diagnosis of Turbine Rotor method and device | |
| WO2023142424A1 (en) | Power financial service risk control method and system based on gru-lstm neural network | |
| CN112418277A (en) | Method, system, medium, and apparatus for predicting remaining life of rotating machine component | |
| CN108921230A (en) | Method for diagnosing faults based on class mean value core pivot element analysis and BP neural network | |
| CN107862763B (en) | Train safety early warning evaluation model training method, module and monitoring evaluation system | |
| CN106874963B (en) | A kind of Fault Diagnosis Method for Distribution Networks and system based on big data technology | |
| CN107037306A (en) | Transformer fault dynamic early-warning method based on HMM | |
| CN112906764B (en) | Communication safety equipment intelligent diagnosis method and system based on improved BP neural network | |
| CN107145675A (en) | Diagnosing fault of power transformer device and method based on BP neural network algorithm | |
| CN105574589B (en) | Transformer oil chromatographic method for diagnosing faults based on niche genetic algorithm | |
| CN108170994A (en) | A kind of oil-immersed electric reactor method for diagnosing faults based on two-way depth network | |
| CN107274067A (en) | A kind of distribution transformer overloads methods of risk assessment | |
| CN106447092A (en) | Marine reverse osmosis desalination system performance prediction method based on MEA-BP neural network | |
| CN111273125A (en) | RST-CNN-based power cable channel fault diagnosis method | |
| CN106896219A (en) | The identification of transformer sub-health state and average remaining lifetime method of estimation based on Gases Dissolved in Transformer Oil data | |
| CN108959498A (en) | A kind of big data processing platform and its design method for health monitoring | |
| CN110516813A (en) | A method of batteries of electric automobile RDR prediction is carried out based on big data machine learning | |
| CN112990546A (en) | Chemical plant power transformer fault prediction method based on particle swarm and neural network | |
| CN114611372A (en) | Industrial equipment health prediction method based on Internet of things edge calculation | |
| CN115902642A (en) | Battery state of charge estimation method and device, electronic equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20171219 |
|
| RJ01 | Rejection of invention patent application after publication |