CN109407508A - Sulfur hexafluoride gas-insulating combined electrical apparatus operating status diagnostic method and system - Google Patents
Sulfur hexafluoride gas-insulating combined electrical apparatus operating status diagnostic method and system Download PDFInfo
- Publication number
- CN109407508A CN109407508A CN201811197527.2A CN201811197527A CN109407508A CN 109407508 A CN109407508 A CN 109407508A CN 201811197527 A CN201811197527 A CN 201811197527A CN 109407508 A CN109407508 A CN 109407508A
- Authority
- CN
- China
- Prior art keywords
- sample
- cluster
- original state
- operating status
- model
- 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
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
- G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
- G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
- G05B13/042—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators in which a parameter or coefficient is automatically adjusted to optimise the performance
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Investigating Or Analyzing Materials By The Use Of Electric Means (AREA)
Abstract
This application discloses a kind of sulfur hexafluoride gas-insulating combined electrical apparatus operating status diagnostic method, system, device and computer readable storage mediums, comprising: is analyzed using original state disaggregated model gas sample, obtains original state result;Using operating status diagnostic model to the further optimizing of original state result, diagnostic result is obtained;Wherein, original state disaggregated model advances with ISODATA algorithm and history gas sample is trained to obtain, and operating status diagnostic model advances with Chaos Ant Colony Optimization and history original state result is trained to obtain;The application analyzes gas sample to be detected using original state disaggregated model, obtain original state result, recycle operating status diagnostic model to the further optimizing of original state result, obtain final more accurately diagnostic result, the accuracy of analysis to gas sample is improved, the running state analysis result of more accurately sulfur hexafluoride gas-insulating combined electrical apparatus can be obtained.
Description
Technical field
The present invention relates to detection technique field, in particular to a kind of sulfur hexafluoride gas-insulating combined electrical apparatus operating status is examined
Disconnected method, system, device and computer readable storage medium.
Background technique
Sulfur hexafluoride gas-insulating combined electrical apparatus is due to the features such as its occupied area is small, compact-sized and excellent insulating property
It is widely used in each voltage class substation, longtime running is experience have shown that GIS (GAS in all kinds of interruption of services of substation
Insulated SWITCHGEAR, Cubicle Gas-Insulated Switchgear) failure accounts for larger specific gravity, and pertinent literature shows
GIS fault type can be summarized as typical fault type, including free metal grain defect, metal tip, solid insulation crackle
Or air blister defect etc..Under different insulative defect, there are certain differences for sound caused by GIS combination electric appliance equipment, light, thermal signal
Not, fault type and its development degree can be identified by collected sound, light, thermal signal.
More intelligent optimization method is proposed for fault type recognition at present, including neural network, support vector machine etc.
More novel algorithm, but its precision is limited.
Therefore, the operating status of six-component force balance equipment how is effectively detected, consequently facilitating the latent event of subsequent early stage
Barrier analysis is those skilled in the art's problem to be solved.
Summary of the invention
In view of this, the purpose of the present invention is to provide a kind of diagnosis of sulfur hexafluoride gas-insulating combined electrical apparatus operating status
Method, system, device and computer readable storage medium can obtain more accurately sulfur hexafluoride gas-insulating combined electrical apparatus
Running state analysis result.Its concrete scheme is as follows:
A kind of sulfur hexafluoride gas-insulating combined electrical apparatus operating status diagnostic method, comprising:
Gas sample is analyzed using original state disaggregated model, obtains original state result;
Using operating status diagnostic model to the further optimizing of original state result, diagnostic result is obtained;
Wherein, the original state disaggregated model is to advance with ISODATA algorithm and history gas sample is trained
It obtains, the operating status diagnostic model is to advance with Chaos Ant Colony Optimization and history original state result is trained
It arrives.
Optionally, the training process of the original state disaggregated model, comprising:
N is chosen from history gas sampleCA cluster centre, obtains NCA cluster, NCLess than or equal to history gas sample
Quantity;
N number of history gas sample is assigned to N by closest principleCIn a cluster;
Judge whether the historical sample quantity in each cluster is less than preset sample threshold;
If it is, deleting history sample size is less than the unqualified cluster of the sample threshold, and institute will be belonged to again
The historical sample stated in unqualified cluster is assigned in remaining cluster by closest principle;
It repeats to divide or combined iteration cluster using cluster centre number, until reaching preset first iteration
Threshold value.
Optionally, the training process of the operating status diagnostic model, comprising:
Using the pheromones and heuristic information numerical value of the historical sample in history original state result to cluster centre, obtain
Select path probability;
Using selection path probability and preset probability threshold value, historical sample is sorted out into corresponding cluster;
Until total deviator error of all clusters meets preset statistical error.
Optionally, it is described gas sample is analyzed using original state disaggregated model before, further includes:
Using Optimization about control parameter model to multiple groups include cluster in sample range distribution standard deviation threshold method, two classes cluster in
The parameter configuration sample of minimum range, customized parameter, pheromones threshold value and pheromones evaporation rate carries out optimizing in the heart, obtains most
Excellent parameter configuration;
It is configured using the optimized parameter, obtains sample range distribution standard in the cluster of the original state disaggregated model
The control pheromones and inspiration number of minimum range and the operating status diagnostic model can between poor threshold value and two class cluster centres
Adjustment parameter, pheromones threshold value and pheromones evaporation rate;
Wherein, Optimization about control parameter model is to be constructed based on simulated annealing.
Optionally, described that sample range distribution standard deviation threshold in cluster is included to multiple groups using Optimization about control parameter model
Value, minimum range between two class cluster centres, customized parameter, pheromones threshold value and pheromones evaporation rate parameter configuration sample into
Row optimizing obtains the process of optimized parameter configuration, comprising:
S51: K mean cluster is carried out to multiple groups parameter configuration sample, obtains initial division result;
S52: using the initialization point as a result, obtaining initial target functional value;
S53: utilizing the initial target functional value and annealing temperature calculating formula, adjusts annealing speed, obtains Current Temperatures;
S54: under Current Temperatures, the cluster classification of any parameter configuration sample is changed, current goal functional value is obtained;
S55: judge whether the error sum of squares of current goal functional value is minimum;
S56: if it is, exporting the optimized parameter configuration, end loop;
S57: if it is not, then judging whether the difference of initial target functional value and current goal functional value is less than or equal to 0;
S58: if it is, using current goal functional value as initial target functional value, return step S52;
S59: if it is not, then using Metropolis criterion select current goal functional value or initial target functional value as
New initial target functional value, return step S52;
S510: until reach preset secondary iteration threshold value, end loop, the corresponding parameter of output initial target functional value
Configuration is configured as optimized parameter.
Optionally, it is described gas sample is analyzed using original state disaggregated model before, further includes:
Original gas sample data is normalized, gas sample is obtained.
The invention also discloses a kind of sulfur hexafluoride gas-insulating combined electrical apparatus operating status diagnostic systems, comprising:
Initial diagnosis module obtains original state for analyzing using original state disaggregated model gas sample
As a result;
Condition diagnosing module, for, to the further optimizing of original state result, being obtained using operating status diagnostic model
To diagnostic result;
Wherein, the original state disaggregated model is to advance with ISODATA algorithm and history gas sample is trained
It obtains, the operating status diagnostic model is to advance with Chaos Ant Colony Optimization and history original state result is trained
It arrives.
Optionally, further includes:
Parameter optimization module, for including sample range distribution standard in cluster to multiple groups using Optimization about control parameter model
The parameter configuration sample of minimum range, customized parameter, pheromones threshold value and pheromones evaporation rate between poor threshold value, two class cluster centres
This progress optimizing obtains optimized parameter configuration;
Parameter extraction module obtains the cluster of the original state disaggregated model for configuring using the optimized parameter
The control of minimum range and the operating status diagnostic model between middle sample range distribution standard deviation threshold method and two class cluster centres
Pheromones and customized parameter, pheromones threshold value and the pheromones evaporation rate for inspiring number;
Wherein, Optimization about control parameter model is to be constructed based on simulated annealing.
The invention also discloses a kind of sulfur hexafluoride gas-insulating combined electrical apparatus operating status diagnostic devices, comprising:
Memory, for storing computer program;
Processor realizes sulfur hexafluoride gas-insulating combined electrical apparatus fortune as the aforementioned for executing the computer program
Row method for diagnosing status.
The invention also discloses a kind of computer readable storage medium, meter is stored on the computer readable storage medium
Calculation machine program realizes that aforementioned sulfur hexafluoride gas-insulating combined electrical apparatus such as runs shape when the computer program is executed by processor
The step of state diagnostic method.
In the present invention, sulfur hexafluoride gas-insulating combined electrical apparatus operating status diagnostic method, comprising: utilize original state point
Class model analyzes gas sample, obtains original state result;Using operating status diagnostic model to original state result
Further optimizing, obtains diagnostic result;Wherein, original state disaggregated model is to advance with ISODATA algorithm and history gas
What sample was trained, operating status diagnostic model be advance with Chaos Ant Colony Optimization and history original state result into
Row training obtains.
The present invention carries out gas sample to be detected using the original state disaggregated model established based on ISODATA algorithm
Analysis obtains original state as a result, recycling the operating status diagnostic model established based on Chaos Ant Colony Optimization to original state
As a result further optimizing obtains final more accurately diagnostic result, improves the accuracy of analysis to gas sample, can obtain
To the more accurately running state analysis result of sulfur hexafluoride gas-insulating combined electrical apparatus.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
The embodiment of invention for those of ordinary skill in the art without creative efforts, can also basis
The attached drawing of offer obtains other attached drawings.
Fig. 1 is a kind of sulfur hexafluoride gas-insulating combined electrical apparatus operating status diagnostic method stream disclosed by the embodiments of the present invention
Journey schematic diagram;
Fig. 2 is a kind of original state disaggregated model history gas sample schematic diagram disclosed by the embodiments of the present invention;
Fig. 3 is a kind of original state disaggregated model Clustering Effect figure disclosed by the embodiments of the present invention;
Fig. 4 is that a kind of original state disaggregated model disclosed by the embodiments of the present invention runs convergence curve;
Fig. 5 is a kind of sulfur hexafluoride gas-insulating combined electrical apparatus operating status diagnostic system knot disclosed by the embodiments of the present invention
Structure schematic diagram.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
The embodiment of the invention discloses a kind of sulfur hexafluoride gas-insulating combined electrical apparatus operating status diagnostic methods, referring to figure
Shown in 1, this method comprises:
S11: analyzing gas sample using original state disaggregated model, obtains original state result.
Specifically, original state disaggregated model is to advance with ISODATA algorithm (ISODATA, Iterative
Selforganizing Data Analysis Techniques Algorithm, iteration self-organizing data analysis algorithm) and go through
History gas sample is trained, and in the training process, a large amount of history gas sample is made for original state disaggregated model
It is input in the model established based on ISODATA algorithm for training data, which is trained, wherein history gas
Body sample is the gas sample of known operating status and each gas concentration, enables the model using ISODATA algorithm constantly to not
Same history gas sample is clustered and is classified, and constantly enables the history gas sample of identical operating status be classified as same poly-
Class, so that corresponding operation shape can be judged according to the gas sample of gas with various concentration by training original state disaggregated model
State, and then complete training and obtain original state disaggregated model.
Wherein, gas sample is the gas generated when sulfur hexafluoride gas-insulating combined electrical apparatus is run, main according to gas
Five kinds of vikane (SO2F2), thionyl fluoride (SOF2), hydrofluoric acid (HF), hydrogen sulfide (H2S) and carbon disulfide (CS2) in sample
The concentration of gas, the corresponding five kinds of operating statuses of the different concentration of five kinds of gases, including bubble-discharge, suspended discharge, metallic particles
Electric discharge, point discharge and normal operating condition.
S12: using operating status diagnostic model to the further optimizing of original state result, diagnostic result is obtained.
Specifically, for the further accuracy for reinforcing diagnostic result, in utilization original state disaggregated model tentatively to gas
After body sample obtains original state result, then by operating status diagnostic model original state result is diagnosed again, thus
Obtain final more accurate diagnostic result.
Wherein, operating status diagnostic model is to advance with Chaos Ant Colony Optimization and history original state result is trained
Obtain, operating status diagnostic model in the training process, by the history original state knot of a large amount of original state disaggregated model
Fruit is input in the model established based on Chaos Ant Colony Optimization as training data, is trained to the model, by history
Gas sample in original state result, will be poly- in history original state result as the data object in Chaos Ant Colony Optimization
Class center is still used as the cluster centre in Chaos Ant Colony Optimization, and the corresponding ant of each data object utilizes pheromones and road
Diameter probability etc. the continuous iteration of parameters, constantly each data object is sorted out into different cluster centres, until all poly-
Total biased error of class meets preset deviation threshold value, and each cluster centre represents a kind of failure cause, and then can be using just
Beginning state outcome further refines, and obtains more accurate diagnostic result.
As it can be seen that the embodiment of the present invention is using the original state disaggregated model established based on ISODATA algorithm to be detected
Gas sample is analyzed, and obtains original state as a result, the operating status established based on Chaos Ant Colony Optimization is recycled to diagnose mould
Type obtains final more accurately diagnostic result, improves the analysis to gas sample to the further optimizing of original state result
Accuracy can obtain the running state analysis result of more accurately sulfur hexafluoride gas-insulating combined electrical apparatus.
The embodiment of the invention discloses a kind of specific sulfur hexafluoride gas-insulating combined electrical apparatus operating status diagnostic method,
Relative to a upper embodiment, the present embodiment has made further instruction and optimization to technical solution.It is specific:
Specifically, the training process of original state disaggregated model, shown referring to figs. 2 and 3, including S21 to S25;Wherein,
S21: N is chosen from history gas sampleCA cluster centre, obtains NCA cluster, NCLess than or equal to history gas sample
This quantity;
Specifically, preselecting N from history gas sampleCA cluster centre { Z1,Z2,...,ZNC}。
S22: N number of history gas sample is assigned to N by closest principleCIn a cluster.
Specifically, each history gas sample is calculated at a distance from polymerization site, N number of history gas sample by closest
Principle is assigned in NC cluster, i.e., | | X-Zj| |=min | | X-Zi| |, i=1,2 ..., NC, then X ∈ Zj, in formula, X table
Show history gas sample, ZjIndicate the polymerization site nearest apart from history gas sample.
S23: judge whether the historical sample quantity in each cluster is less than preset sample threshold.
S24: if it is, deleting history sample size is less than the unqualified cluster of sample threshold, and will belong to again not
Historical sample in qualification cluster is assigned in remaining cluster by closest principle.
Specifically, judging ZiIn number of samples, if Nj< θN, then the cluster, and N are deletedC1 is subtracted, step is executed
The historical sample belonged in unqualified cluster is assigned in remaining cluster, θ by S22 by closest principle againNIndicate sample threshold
Value.
S25: dividing cluster using the repetition of cluster centre number or combined iteration, until reaching preset first
Iteration threshold.
If specifically, NC≤ K/2, K indicate the hope number of cluster centre, i.e., cluster centre number, which is not more than, wishes number
Purpose half then enters step toward division;If NC>=2K, i.e. cluster centre number are not less than to wish twice of number, or repeatedly
Generation number is even number, then enters and merge step, otherwise enters step toward division;If the number of iterations is to the first iteration threshold, i.e.,
Last time iteration, then end loop, completes training, obtains original state disaggregated model.
Wherein, step toward division is as follows:
Calculate the standard difference vector σ of all kinds of inter- object distancesj=[σj1,σj2,...,σjn]T, j=1,2 ..., NC;
Each component are as follows:
In formula, σjIndicate the standard difference vector of j inter- object distance, i=1,2 ..., n are dimension, xjiIt is ZjThe history gas of class
I-th of component of sample X, zjiIt is ZjThe cluster centre Z of classjI-th of component, σijIndicate the component of history gas sample.
Solve the largest component σ of each standard deviationjmax, in set { σjmaxIn, if there is σjmax≥θSIllustrate ZjClass sample exists
Standard deviation on corresponding direction is greater than sample range distribution standard deviation threshold method θ in clusterSIf meet simultaneously following two condition it
One:
①And Nj2 (θ of >N+1);
②
Then ZjIt splits intoWithNCAdd 1;ZjInCorresponding σjmaxK σ on componentjmax;ZjInCorresponding σjmaxComponent subtracts
Remove k σjmax, 0 < k < 1, k are bundle factor,WithRespectively indicate ZjTwo cluster centres after division.
If completing division, the number of iterations adds one, returns to S22, if the number of iterations is maximum value, end loop.
Wherein, steps are as follows for merging:
Calculate the distance between all cluster centres Dij=| | Zi-Zj| |, i=1,2 ..., NC-1;J=i+1, i+
2,...,NCCompare all DijWith θCValue, θ will be less thanCDijIt is arranged by ascending order, forms set
It will setIn corresponding two class of each element merge, new cluster is obtained, in cluster
The heartForEvery merging is a pair of, NCSubtract 1.
If completing to merge, the number of iterations adds one, returns to S22, if the number of iterations is maximum value, end loop.
Wherein, cluster centre wishes number K, sample threshold θN, sample range distribution standard deviation threshold method θ in clusterSWith two classes
Minimum range θ between cluster centreCIt is preset.
Wherein, center of a sample Z is respectively clusteredj:
Average distance in class
Population mean distance
Specifically, initial input parameter of the original state result as ant group algorithm, wherein original state result can lead to
Cross arrayIt characterizes, in formula,For decomposition product feature vector, DISOFor original state classification results.
Wherein, the operation convergence curve of original state disaggregated model may refer to shown in Fig. 4.
Specifically, the training process of operating status diagnostic model, including S31 to S33;Wherein,
S31: using the pheromones and heuristic information numerical value of the historical sample in history original state result to cluster centre,
Obtain selection path probability;
S32: using selection path probability and preset probability threshold value, historical sample is sorted out into corresponding cluster;
S33: until total deviator error of all clusters meets preset statistical error.
Specifically, having N number of, each data object as data object using the historical sample in history original state result
There are m attribute, data object is defined as: X={ Xi|Xi=(xi1,xi2,...,xim), i=1,2 ..., N }, classification number is to go through
Cluster centre number in history original state result is indicated with K.
An ant is placed at any data object i, i.e. data object is equivalent to an ant, data object i distribution
To j-th of cluster centre Cj(j=1,2 ..., K), ant is just in data object i to cluster centre CjPath (i, j) on leave
Pheromones τij(t), d (Xi,Cj) indicate XiTo cluster centre CjBetween Euclidean distance;PijIt (t) is ant selection path (i, j)
Probability;Wherein,
Select path probability calculation formula are as follows:
Pheromones expression formula are as follows:
Euclidean distance calculation formula are as follows:
R is preset cluster radius;S=s | d (Xs,Cj)≤R, s=1,2 ..., N and s ≠ j } it indicates to be distributed in cluster
Center CjThe set of data object in neighborhood;ηij(t)=1/d (Xi,Cj) indicate that t moment data object i distributes to j-th of cluster
Center CjHeuristic information numerical value;α and β is for controlling pheromones and inspiring the customized parameter of number;If Pij(t) it is greater than general
Rate threshold value P0, just by XiIt is integrated into CjField.
Model termination condition is that all total biased error ξ of cluster are less than given statistical error ε0;Wherein,
Total biased error ξ calculation formula of all clusters are as follows:
In formula, ξjIndicate the biased error of j-th of cluster;C'jFor new cluster centre, XiIt is all to be integrated into C'jIn class
Data object, it may be assumed that Xi∈{Xh|d(Xs,C'j)≤R;H=1,2 ..., N }, J is of all data objects in the cluster
Number.
Specifically, Chaos Ant Colony Optimization strengthens the effect of positive feedback, the search efficiency of ant is improved, reduction falls into part
A possibility that optimization, more utilizes optimal solution information during algorithm operation, i.e., only allows a suboptimum after each iteration
Ant increase pheromones, the optimal ant can be the present age it is optimal, be also possible to global optimum.
Stopping phenomenon to avoid shrinking as far as possible, Chaos Ant Colony Optimization limits pheromones, before starting search,
The pheromones level on all sides can be set as pheromones maximum value, initialization in this way is conducive to algorithm and searches in the initial period
More solutions, in this way at search initial stage, Ant Search range is larger, so that reducing search stagnates the hair in the local optimum the case where
It is raw.
Wherein, pheromones threshold value can be 0.9, q=0.9, and pheromones evaporation rate can be 0.1, ρ=0.1.
On the basis of previous embodiment, the embodiment of the invention also discloses a kind of sulfur hexafluoride gas-insulating combined electrical apparatuses
Operating status diagnostic method, specific:
Specifically, sample range distribution standard deviation threshold in being clustered in original state disaggregated model and operating status diagnostic model
Minimum range, customized parameter, pheromones threshold value and pheromones evaporation rate are to pre-enter between value, two class cluster centres
Value, the accuracy that final operating status diagnostic result is respectively directly affected between value will cluster for that can obtain optimal input value
Minimum range, customized parameter, pheromones threshold value and pheromones between middle sample range distribution standard deviation threshold method, two class cluster centres
Evaporation rate carries out optimizing as one group of parameter configuration, for this purpose, above-mentioned S11: using original state disaggregated model to gas sample into
Before row analysis, can also include:
S41: it is birdsed of the same feather flock together to multiple groups including sample range distribution standard deviation threshold method, two in cluster using Optimization about control parameter model
The parameter configuration sample progress optimizing of minimum range, customized parameter, pheromones threshold value and pheromones evaporation rate, obtains between class center
It is configured to optimized parameter.
S42: it is configured using optimized parameter, obtains sample range distribution standard deviation threshold in the cluster of original state disaggregated model
The customized parameter of the control pheromones and inspiration number of minimum range and operating status diagnostic model between being worth two class cluster centres,
Pheromones threshold value and pheromones evaporation rate.
Specifically, obtaining optimized parameter with postponing, sample range distribution standard deviation threshold method, two in cluster therein are birdsed of the same feather flock together
Minimum range, customized parameter, pheromones threshold value and pheromones evaporation rate are respectively as original state disaggregated model between class center
With the input of operating status diagnostic model.
Wherein, Optimization about control parameter model is to be constructed based on simulated annealing.
Further, above-mentioned S41 includes sample range distribution standard in cluster to multiple groups using Optimization about control parameter model
The parameter configuration sample of minimum range, customized parameter, pheromones threshold value and pheromones evaporation rate between poor threshold value, two class cluster centres
This progress optimizing obtains the process of optimized parameter configuration, can specifically include S51 to S510;Wherein,
S51: K mean cluster is carried out to multiple groups parameter configuration sample, obtains initial division result.
Specifically, carrying out K mean cluster to parameter configuration sample, initial division result is obtained.
S52: using initialization point as a result, obtaining initial target functional value.
Specifically, obtaining initial target functional value J using initial division result as initial solution ωw。
S53: utilizing initial target functional value and annealing temperature calculating formula, adjusts annealing speed, obtains Current Temperatures;
Specifically, initialization temperature T0, enable T0=Jw, initialize annealing speed α and maximum annealing times, annealing temperature meter
Formula are as follows:Temperature T will be initialized0Annealing temperature calculating formula is substituted into initialization annealing speed α, is calculated
Current Temperatures.
In formula, α is adjustable parameter, can improve the form of annealing curve, and above formula shows to decline comparatively fast in high-temperature area temperature,
It is slower in the decline of low-temperature region temperature, i.e., optimizing mainly is carried out in low-temperature region.
S54: under Current Temperatures, the cluster classification of any parameter configuration sample is changed, current goal functional value is obtained.
Specifically, input random perturbation generates new clustering ω ', i.e., it is random to change working as a parameter configuration sample
Preceding generic obtains current goal functional value J'w。
S55: judge whether the error sum of squares of current goal functional value is minimum;
S56: if it is, output optimized parameter configuration, end loop.
Specifically, judging new current goal functional value J'wIt whether is optimal objective function value, i.e., whether error sum of squares
Minimum is to save clustering ω ' as optimum cluster division, saves J'wFor optimal objective function value, obtains optimized parameter and match
It sets, and then optimized parameter configuration can be exported with end loop.
Wherein, the variance of the corresponding defect type of parameter configuration vector and its practical defect type reaches minimum value, that is, misses
Poor quadratic sum;
Error sum of squares calculation formula are as follows:
In formula, M is the number of cluster centre, less than the total number of parameter configuration, ωiIndicate that the i-th class clusters,It indicates
The cluster centre vector of i-th class.
S57: if it is not, then judging whether the difference of initial target functional value and current goal functional value is less than or equal to 0.
S58: if it is, using current goal functional value as initial target functional value, return step S52;
S59: if it is not, then using Metropolis criterion select current goal functional value or initial target functional value as
New initial target functional value, return step S52;
S510: until reach preset secondary iteration threshold value, end loop, the corresponding parameter of output initial target functional value
Configuration is configured as optimized parameter.
Specifically, passing through discriminant function value difference Δ J=J'w-Jw.Whether Δ J is judged less than 0, if Δ J≤0, is received new
Solution, i.e., using current goal functional value as initial target functional value, return goods step S52;If Δ J > 0, according to Metropolis
Criterion receives new explanation, i.e., is to enable current goal functional value as new initial target function according to the judgement of Metropolis criterion
Value, return step S52 still continue to retain original initial target function value as new initial target functional value, return step
S52;If continuing do not occur optimal objective function value, return step S52 reduces temperature using new initial target functional value,
Continue iteration, until reaching secondary iteration threshold value, that is, reaches maximum number of iterations, end loop exports and finally judges to make
It is configured for the corresponding parameter configuration of initial target functional value of new explanation as optimized parameter.
It, can be with before S11 analyzes gas sample using original state disaggregated model in the embodiment of the present invention
Including original gas sample data is normalized, gas sample is obtained.
For example, five kinds of characteristic gas of SO2F2, SOF2, HF, H2S, CS2 are under different operating statuses in original gas sample
Content, carry out coding and data normalization processing, enable bubble-discharge (0,0,0,0,1);Suspended discharge (0,0,0,1,0);Gold
Metal particles discharge (0,0,1,0,0);Point discharge (0,1,0,0,0);Normal operating condition (1,0,0,0,0) is in favor of quantitative table
Different typical operations is levied, for example, shown in the sample data of 1 normalized of table;For SO2F2, SOF2, HF, H2S,
The various gas content difference of the concentration of five kinds of characteristic gas of CS2 are larger, small to prevent the big number of the data for being input to model appearance from gulping down
Several situation and classification iterative process is allowed preferably to restrain, initial data is normalized, the specific following institute of formula
Show:
μ (SO in above formula2F2), μ (SOF2), μ (H2S), μ (CS2), μ (HF) respectively represents containing for all kinds of feature decomposition gases
Amount;μ represents gas total content;μSRepresent sulfide gas total content.
Table 1
Correspondingly, the embodiment of the invention also discloses a kind of sulfur hexafluoride gas-insulating combined electrical apparatus operating status diagnosis systems
System, shown in Figure 5, which includes:
Initial diagnosis module 11 obtains initial shape for analyzing using original state disaggregated model gas sample
State result;
Condition diagnosing module 12, for, to the further optimizing of original state result, being obtained using operating status diagnostic model
Diagnostic result;
Wherein, original state disaggregated model is to advance with ISODATA algorithm and history gas sample is trained to obtain
, operating status diagnostic model advances with Chaos Ant Colony Optimization and history original state result is trained to obtain.
Wherein, the training process of original state disaggregated model, comprising:
N is chosen from history gas sampleCA cluster centre, obtains NCA cluster, NCLess than or equal to history gas sample
Quantity;
N number of history gas sample is assigned to N by closest principleCIn a cluster;
Judge whether the historical sample quantity in each cluster is less than preset sample threshold;
If it is, deleting history sample size is less than the unqualified cluster of sample threshold, and will belong to again unqualified
Historical sample in cluster is assigned in remaining cluster by closest principle;
It repeats to divide or combined iteration cluster using cluster centre number, until reaching preset first iteration
Threshold value.
Wherein, the training process of operating status diagnostic model, comprising:
Using the pheromones and heuristic information numerical value of the historical sample in history original state result to cluster centre, obtain
Select path probability;
Using selection path probability and preset probability threshold value, historical sample is sorted out into corresponding cluster;
Until total deviator error of all clusters meets preset statistical error.
Specifically, can also include:
Parameter optimization module, for including sample range distribution standard in cluster to multiple groups using Optimization about control parameter model
The parameter configuration sample of minimum range, customized parameter, pheromones threshold value and pheromones evaporation rate between poor threshold value, two class cluster centres
This progress optimizing obtains optimized parameter configuration;
Parameter extraction module, for being configured using optimized parameter, obtain in the cluster of original state disaggregated model sample away from
Control pheromones and inspiration from minimum range between distribution standard deviation threshold value and two class cluster centres and operating status diagnostic model
Several customized parameter, pheromones threshold value and pheromones evaporation raties;
Wherein, Optimization about control parameter model is to be constructed based on simulated annealing.
Specifically, parameter optimization module, comprising:
Division unit obtains initial division result for carrying out K mean cluster to multiple groups parameter configuration sample;
Initial function value computing unit, for being divided using initialization as a result, obtaining initial target functional value;
Temperature calculation unit adjusts annealing speed, obtains for utilizing initial target functional value and annealing temperature calculating formula
Current Temperatures;
Unit is disturbed, for the cluster classification of any parameter configuration sample being changed, obtaining current goal under Current Temperatures
Functional value;
Whether quadratic sum judging unit, the error sum of squares for judging current goal functional value are minimum;
First output unit, for when quadratic sum judging unit decision errors quadratic sum minimum, then exporting optimized parameter and matching
It sets, end loop;
Function value difference judging unit, for not being minimum when quadratic sum judging unit decision errors quadratic sum, then judgement is first
Whether the difference of beginning target function value and current goal functional value is less than or equal to 0;
First assignment unit, for determining that difference is less than or equal to 0 when function value difference judging unit, then by current goal functional value
As initial target functional value, initial function value computing unit is re-called;
Second assignment unit is then selected using Metropolis criterion for determining that difference is greater than 0 when function value difference judging unit
Current goal functional value or initial target functional value are selected as new initial target functional value, re-calls the calculating of initial function value
Unit;
Second output unit, for until reaching preset secondary iteration threshold value, end loop to export initial target function
It is worth corresponding parameter configuration to configure as optimized parameter.
Specifically, can also include, normalization module be obtained for original gas sample data to be normalized
To gas sample.
In addition, the embodiment of the invention also discloses a kind of diagnosis of sulfur hexafluoride gas-insulating combined electrical apparatus operating status to fill
It sets, comprising:
Memory, for storing computer program;
Processor realizes sulfur hexafluoride gas-insulating combined electrical apparatus operation shape as the aforementioned for executing computer program
State diagnostic method.
In addition, the embodiment of the invention also discloses a kind of computer readable storage medium, on computer readable storage medium
It is stored with computer program, realizes that aforementioned sulfur hexafluoride gas-insulating combined electrical apparatus such as is transported when computer program is executed by processor
The step of row method for diagnosing status.
Finally, it is to be noted that, herein, relational terms such as first and second and the like be used merely to by
One entity or operation are distinguished with another entity or operation, without necessarily requiring or implying these entities or operation
Between there are any actual relationship or orders.Moreover, the terms "include", "comprise" or its any other variant meaning
Covering non-exclusive inclusion, so that the process, method, article or equipment for including a series of elements not only includes that
A little elements, but also including other elements that are not explicitly listed, or further include for this process, method, article or
The intrinsic element of equipment.In the absence of more restrictions, the element limited by sentence "including a ...", is not arranged
Except there is also other identical elements in the process, method, article or apparatus that includes the element.
Professional further appreciates that, unit described in conjunction with the examples disclosed in the embodiments of the present disclosure
And algorithm steps, can be realized with electronic hardware, computer software, or a combination of the two, in order to clearly demonstrate hardware and
The interchangeability of software generally describes each exemplary composition and step according to function in the above description.These
Function is implemented in hardware or software actually, the specific application and design constraint depending on technical solution.Profession
Technical staff can use different methods to achieve the described function each specific application, but this realization is not answered
Think beyond the scope of this invention.
Above to a kind of sulfur hexafluoride gas-insulating combined electrical apparatus operating status diagnostic method provided by the present invention, be
System, device and computer readable storage medium are described in detail, and specific case used herein is to the principle of the present invention
And embodiment is expounded, the above embodiments are only used to help understand, and method and its core of the invention is thought
Think;At the same time, for those skilled in the art, according to the thought of the present invention, in specific embodiments and applications
There will be changes, in conclusion the contents of this specification are not to be construed as limiting the invention.
Claims (10)
1. a kind of sulfur hexafluoride gas-insulating combined electrical apparatus operating status diagnostic method characterized by comprising
Gas sample is analyzed using original state disaggregated model, obtains original state result;
Using operating status diagnostic model to the further optimizing of original state result, diagnostic result is obtained;
Wherein, the original state disaggregated model is to advance with ISODATA algorithm and history gas sample is trained to obtain
, the operating status diagnostic model is to advance with Chaos Ant Colony Optimization and history original state result is trained to obtain
's.
2. sulfur hexafluoride gas-insulating combined electrical apparatus operating status diagnostic method according to claim 1, which is characterized in that
The training process of the original state disaggregated model, comprising:
N is chosen from history gas sampleCA cluster centre, obtains NCA cluster, NCLess than or equal to the number of history gas sample
Amount;
N number of history gas sample is assigned to N by closest principleCIn a cluster;
Judge whether the historical sample quantity in each cluster is less than preset sample threshold;
If it is, deleting history sample size is less than the unqualified cluster of the sample threshold, and will belong to again it is described not
Historical sample in qualification cluster is assigned in remaining cluster by closest principle;
It repeats to divide or combined iteration cluster using cluster centre number, until reaching preset first iteration threshold
Value.
3. sulfur hexafluoride gas-insulating combined electrical apparatus operating status diagnostic method according to claim 2, which is characterized in that
The training process of the operating status diagnostic model, comprising:
Using the pheromones and heuristic information numerical value of the historical sample in history original state result to cluster centre, selected
Path probability;
Using selection path probability and preset probability threshold value, historical sample is sorted out into corresponding cluster;
Until total deviator error of all clusters meets preset statistical error.
4. sulfur hexafluoride gas-insulating combined electrical apparatus operating status diagnostic method according to any one of claims 1 to 3,
Be characterized in that, it is described gas sample is analyzed using original state disaggregated model before, further includes:
Using Optimization about control parameter model between multiple groups include cluster in sample range distribution standard deviation threshold method, two class cluster centres
The parameter configuration sample progress optimizing of minimum range, customized parameter, pheromones threshold value and pheromones evaporation rate, obtains optimal ginseng
Number configuration;
It is configured using the optimized parameter, obtains sample range distribution standard deviation threshold in the cluster of the original state disaggregated model
The control pheromones and inspiration number of minimum range and the operating status diagnostic model is adjustable between value and two class cluster centres
Parameter, pheromones threshold value and pheromones evaporation rate;
Wherein, Optimization about control parameter model is to be constructed based on simulated annealing.
5. sulfur hexafluoride gas-insulating combined electrical apparatus operating status diagnostic method according to claim 4, which is characterized in that
It is described using Optimization about control parameter model between multiple groups include cluster in sample range distribution standard deviation threshold method, two class cluster centres
The parameter configuration sample progress optimizing of minimum range, customized parameter, pheromones threshold value and pheromones evaporation rate, obtains optimal ginseng
The process of number configuration, comprising:
S51: K mean cluster is carried out to multiple groups parameter configuration sample, obtains initial division result;
S52: using the initialization point as a result, obtaining initial target functional value;
S53: utilizing the initial target functional value and annealing temperature calculating formula, adjusts annealing speed, obtains Current Temperatures;
S54: under Current Temperatures, the cluster classification of any parameter configuration sample is changed, current goal functional value is obtained;
S55: judge whether the error sum of squares of current goal functional value is minimum;
S56: if it is, exporting the optimized parameter configuration, end loop;
S57: if it is not, then judging whether the difference of initial target functional value and current goal functional value is less than or equal to 0;
S58: if it is, using current goal functional value as initial target functional value, return step S52;
S59: if it is not, then selecting current goal functional value or initial target functional value as new using Metropolis criterion
Initial target functional value, return step S52;
S510: until reach preset secondary iteration threshold value, end loop, the corresponding parameter configuration of output initial target functional value
It is configured as optimized parameter.
6. sulfur hexafluoride gas-insulating combined electrical apparatus operating status diagnostic method according to any one of claims 1 to 3,
Be characterized in that, it is described gas sample is analyzed using original state disaggregated model before, further includes:
Original gas sample data is normalized, gas sample is obtained.
7. a kind of sulfur hexafluoride gas-insulating combined electrical apparatus operating status diagnostic system characterized by comprising
Initial diagnosis module obtains original state result for analyzing using original state disaggregated model gas sample;
Condition diagnosing module, for, to the further optimizing of original state result, being examined using operating status diagnostic model
Disconnected result;
Wherein, the original state disaggregated model is to advance with ISODATA algorithm and history gas sample is trained to obtain
, the operating status diagnostic model is to advance with Chaos Ant Colony Optimization and history original state result is trained to obtain
's.
8. sulfur hexafluoride gas-insulating combined electrical apparatus operating status diagnostic system according to claim 7, which is characterized in that
Further include:
Parameter optimization module, for including sample range distribution standard deviation threshold in cluster to multiple groups using Optimization about control parameter model
Value, minimum range between two class cluster centres, customized parameter, pheromones threshold value and pheromones evaporation rate parameter configuration sample into
Row optimizing obtains optimized parameter configuration;
Parameter extraction module obtains sample in the cluster of the original state disaggregated model for configuring using the optimized parameter
The control information of minimum range and the operating status diagnostic model between this range distribution standard deviation threshold method and two class cluster centres
Element and customized parameter, pheromones threshold value and the pheromones evaporation rate for inspiring number;
Wherein, Optimization about control parameter model is to be constructed based on simulated annealing.
9. a kind of sulfur hexafluoride gas-insulating combined electrical apparatus operating status diagnostic device characterized by comprising
Memory, for storing computer program;
Processor, for executing the computer program to realize such as sulfur hexafluoride gas as claimed in any one of claims 1 to 6
Insulation in combined electric appliance operating status diagnostic method.
10. a kind of computer readable storage medium, which is characterized in that be stored with computer on the computer readable storage medium
Program realizes the sulfur hexafluoride gas-insulating as described in any one of claim 1 to 6 when the computer program is executed by processor
The step of combined electrical apparatus operating status diagnostic method.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811197527.2A CN109407508A (en) | 2018-10-15 | 2018-10-15 | Sulfur hexafluoride gas-insulating combined electrical apparatus operating status diagnostic method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811197527.2A CN109407508A (en) | 2018-10-15 | 2018-10-15 | Sulfur hexafluoride gas-insulating combined electrical apparatus operating status diagnostic method and system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109407508A true CN109407508A (en) | 2019-03-01 |
Family
ID=65467979
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811197527.2A Pending CN109407508A (en) | 2018-10-15 | 2018-10-15 | Sulfur hexafluoride gas-insulating combined electrical apparatus operating status diagnostic method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109407508A (en) |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101299861A (en) * | 2008-04-23 | 2008-11-05 | 南京大学 | Base station system polling path automatization determination method based on shortest cycle |
CN106055918A (en) * | 2016-07-26 | 2016-10-26 | 天津大学 | Power system load data identification and recovery method |
-
2018
- 2018-10-15 CN CN201811197527.2A patent/CN109407508A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101299861A (en) * | 2008-04-23 | 2008-11-05 | 南京大学 | Base station system polling path automatization determination method based on shortest cycle |
CN106055918A (en) * | 2016-07-26 | 2016-10-26 | 天津大学 | Power system load data identification and recovery method |
Non-Patent Citations (4)
Title |
---|
JIANXIA CHEN等: "Real - time Gas Insulation State Evaluation Model", 《IEEE》 * |
YING WANG等: "Research on an Ant Colony ISODATA Algorithm for Clustering Analysis in Real Time Computer Simulation", 《SECOND WORKSHOP ON DIGITAL MEDIA AND ITS APPLICATION IN MUSEUM & HERITAGE》 * |
朱晔: "配电变压器的故障诊断和实例分析", 《发输变电》 * |
陶红 等: "基于蚁群和模拟退火算法的聚类新方法", 《微电子学与计算机》 * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105589806B (en) | A kind of software defect tendency Forecasting Methodology based on SMOTE+Boosting algorithms | |
CN109145949A (en) | Non-intrusive electrical load monitoring and decomposition method and system based on integrated study | |
CN106896219B (en) | The identification of transformer sub-health state and average remaining lifetime estimation method based on Gases Dissolved in Transformer Oil data | |
CN110400231B (en) | Failure rate estimation method for electric energy metering equipment based on weighted nonlinear Bayes | |
CN103197983A (en) | Service component reliability online time sequence predicting method based on probability graph model | |
CN107037306A (en) | Transformer fault dynamic early-warning method based on HMM | |
CN110363354A (en) | Wind field wind power prediction method, electronic device and storage medium | |
CN110321940A (en) | The feature extraction of aircraft telemetry and classification method and device | |
CN105471647A (en) | Power communication network fault positioning method | |
CN115204698A (en) | Real-time analysis method for power supply stability of low-voltage transformer area | |
CN109697332A (en) | A kind of exception monitoring scheme of the stream calculation system based on unsupervised learning method | |
CN116011158A (en) | Topology identification method, system and device for low-voltage transformer area | |
CN110059938B (en) | Power distribution network planning method based on association rule driving | |
CN114186733A (en) | Short-term load prediction method and device | |
CN109858667A (en) | It is a kind of based on thunder and lightning weather to the short term clustering method of loading effects | |
CN109255189A (en) | The parallel real-time mode recognizing method of voltage dip based on streaming computing | |
CN109407508A (en) | Sulfur hexafluoride gas-insulating combined electrical apparatus operating status diagnostic method and system | |
Yang et al. | Data domain adaptation for voltage stability evaluation considering topology changes | |
CN117034149A (en) | Fault processing strategy determining method and device, electronic equipment and storage medium | |
CN115118015A (en) | Platform district power supply stability monitoring system based on fuse terminal | |
CN111239484A (en) | Non-invasive load electricity consumption information acquisition method for non-resident users | |
CN116361628A (en) | Fault category intelligent analysis method and device based on VFD room | |
CN115203873A (en) | Topological relation construction method, device, equipment and medium applied to power distribution network | |
CN115577290A (en) | Distribution network fault classification and source positioning method based on deep learning | |
CN115221955A (en) | Multi-depth neural network parameter fusion system and method based on sample difference analysis |
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: 20190301 |
|
RJ01 | Rejection of invention patent application after publication |