CN116595459A - Pollution flashover early warning method and system based on electric field signals - Google Patents
Pollution flashover early warning method and system based on electric field signals Download PDFInfo
- Publication number
- CN116595459A CN116595459A CN202310393819.8A CN202310393819A CN116595459A CN 116595459 A CN116595459 A CN 116595459A CN 202310393819 A CN202310393819 A CN 202310393819A CN 116595459 A CN116595459 A CN 116595459A
- Authority
- CN
- China
- Prior art keywords
- electric field
- early warning
- pollution flashover
- pollution
- field signal
- 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
- 230000005684 electric field Effects 0.000 title claims abstract description 77
- 238000000034 method Methods 0.000 title claims abstract description 57
- 238000011161 development Methods 0.000 claims abstract description 42
- 238000012549 training Methods 0.000 claims abstract description 23
- 239000012212 insulator Substances 0.000 claims abstract description 19
- 238000012544 monitoring process Methods 0.000 claims abstract description 17
- 230000008859 change Effects 0.000 claims abstract description 12
- 238000003066 decision tree Methods 0.000 claims description 19
- 238000007637 random forest analysis Methods 0.000 claims description 16
- 238000013528 artificial neural network Methods 0.000 claims description 12
- 238000005070 sampling Methods 0.000 claims description 8
- 238000000605 extraction Methods 0.000 claims description 7
- 238000011478 gradient descent method Methods 0.000 claims description 7
- 238000012360 testing method Methods 0.000 claims description 5
- 238000004140 cleaning Methods 0.000 claims description 4
- 230000002776 aggregation Effects 0.000 claims description 3
- 238000004220 aggregation Methods 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 3
- 230000008569 process Effects 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 3
- 210000003205 muscle Anatomy 0.000 claims 3
- 238000010276 construction Methods 0.000 abstract description 8
- 238000005516 engineering process Methods 0.000 abstract description 6
- 238000001514 detection method Methods 0.000 abstract description 5
- 238000009422 external insulation Methods 0.000 abstract description 3
- 230000018109 developmental process Effects 0.000 description 26
- 238000004422 calculation algorithm Methods 0.000 description 8
- 238000010586 diagram Methods 0.000 description 5
- 230000005540 biological transmission Effects 0.000 description 4
- 238000005192 partition Methods 0.000 description 4
- 238000009826 distribution Methods 0.000 description 3
- 238000009413 insulation Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000012423 maintenance Methods 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 230000007797 corrosion Effects 0.000 description 1
- 238000005260 corrosion Methods 0.000 description 1
- 239000000428 dust Substances 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 150000003839 salts Chemical class 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000009736 wetting Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
-
- 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
- G01R31/12—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing
- G01R31/1227—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing of components, parts or materials
- G01R31/1245—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing of components, parts or materials of line insulators or spacers, e.g. ceramic overhead line cap insulators; of insulators in HV bushings
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0499—Feedforward networks
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J13/00—Circuit arrangements for providing remote indication of network conditions, e.g. an instantaneous record of the open or closed condition of each circuitbreaker in the network; Circuit arrangements for providing remote control of switching means in a power distribution network, e.g. switching in and out of current consumers by using a pulse code signal carried by the network
- H02J13/00002—Circuit arrangements for providing remote indication of network conditions, e.g. an instantaneous record of the open or closed condition of each circuitbreaker in the network; Circuit arrangements for providing remote control of switching means in a power distribution network, e.g. switching in and out of current consumers by using a pulse code signal carried by the network characterised by monitoring
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J13/00—Circuit arrangements for providing remote indication of network conditions, e.g. an instantaneous record of the open or closed condition of each circuitbreaker in the network; Circuit arrangements for providing remote control of switching means in a power distribution network, e.g. switching in and out of current consumers by using a pulse code signal carried by the network
- H02J13/00006—Circuit arrangements for providing remote indication of network conditions, e.g. an instantaneous record of the open or closed condition of each circuitbreaker in the network; Circuit arrangements for providing remote control of switching means in a power distribution network, e.g. switching in and out of current consumers by using a pulse code signal carried by the network characterised by information or instructions transport means between the monitoring, controlling or managing units and monitored, controlled or operated power network element or electrical equipment
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Power Engineering (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Chemical & Material Sciences (AREA)
- Ceramic Engineering (AREA)
- Probability & Statistics with Applications (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
A pollution flashover early warning method based on an electric field signal comprises the following steps: step 1: extracting characteristic parameters of a space electric field signal of one cycle; step 2: the running insulator wets from the surface to the flashover, the pollution flashover development section can be divided into three stages of a safety zone, a forecasting zone and a dangerous zone according to the change of the amplitude of the space electric field, specific characteristic parameters are input into a pre-constructed and training prediction model, and the pollution flashover development section is identified; step 3: judging according to the identification result of the pollution flashover development section, and if the pollution flashover development section is identified as a forecast area and a dangerous area, sending pollution flashover early warning information; if the pollution flashover development section is identified as a safety area, a pollution degree prediction model is established, and pollution degree prediction is carried out. The method provided by the application can enrich the online monitoring means of the external insulation of the overhead line, thereby expanding the intelligent operation and detection technology of the power system and meeting the requirements of the current ubiquitous power Internet of things construction and the strong intelligent power grid construction.
Description
Technical Field
The application belongs to the field of high-voltage and insulation technology and power equipment state monitoring, and particularly relates to a pollution flashover early warning method and system based on an electric field signal.
Background
An overhead line insulator is one of electrical equipment with larger application amount of a power system, the running state of the overhead line insulator influences the safe running of a power grid, and insulator pollution flashover is an important factor causing power failure of the power grid. Under the influence of factors such as electricity, light, heat, stress, pollution and the like, the insulation performance of the electric power system can be reduced, so that a discharge phenomenon is caused, and the safe and stable operation of the electric power system is seriously threatened. In recent years, pollution flashover tripping accidents occur in coastal areas, high-temperature and high-humidity areas, chemical industry parks, heavy industrial dust areas and other areas, and the external insulation of power transmission and distribution in China still faces serious tests.
Pollution flashover early warning can be effectively realized by means of carrying out on-line monitoring on the insulation weak part. However, the current pollution flashover early warning means is quite single, and the only leakage current on-line monitoring technology used in the actual engineering is obtained. However, the method still has certain defects, such as selection of characteristic parameters of leakage current signals, pollution degree evaluation algorithm and corresponding early warning threshold setting thereof have no unified theorem. In addition, although the leakage current on-line monitoring technology is applied to overhead lines, the technology needs to collect current by installing a collector ring, long-term operation can cause corrosion of an insulating surface, and the cost is increased for operation and maintenance of the power transmission line.
As one of important parameters of the electrical equipment, the electric field strength can well reflect the operation state of the electrical equipment. Compared with a leakage current monitoring method, the method has the advantages that a collector ring is not required to be installed for monitoring the space electric field signal, the insulator surface is not required to be contacted, and the method has obvious advantages in installation and maintenance. Compared with indirect detection methods such as optics and acoustics, the electric measurement method based on the electric field is low in cost, little in environmental influence and suitable for on-line monitoring of pollution degree. At present, no method is available for realizing pollution flashover early warning by utilizing a space electric field signal.
The prior published patent name is 'cable partial discharge fault identification method, system and medium based on improved random forest algorithm' (publication number is CN 110108992A), which discloses that 'acquisition of partial discharge signals of cables' is included; extracting characteristic data of the partial discharge signals; inputting the extracted characteristic data into a classifier which is built in advance and is trained to obtain the cable partial discharge fault type corresponding to the cable partial discharge signal, wherein the classifier is a classifier based on an improved random forest algorithm, the improved random forest algorithm utilizes a characteristic building method to mine high-dimensional attribute data and a data weight correction idea based on an Adaboost algorithm to guide the selection of an attribute subset, and the classifier is trained in advance to establish the mapping relation between the characteristic of the partial discharge signal and the cable partial discharge fault type.
The disclosure document aims at identifying local faults of the cable, monitoring the partial discharge signals of the cable, and does not solve the problem of early warning of pollution flashover accidents of insulators of overhead lines.
Disclosure of Invention
In order to solve the defects in the prior art, the application provides the pollution flashover early warning method based on the electric field signal, which can enrich the on-line monitoring means of the external insulation of the overhead line, thereby expanding the intelligent operation and detection technology of the power system and meeting the requirements of the current ubiquitous power Internet of things construction and the construction of a strong intelligent power grid.
The application adopts the following technical scheme.
In one aspect, the present patent provides a pollution flashover early warning method based on electric field signals, comprising the steps of:
step 1: extracting characteristic parameters of a space electric field signal of one cycle;
step 2: the running insulator wets from the surface to the flashover, the pollution flashover development section can be divided into three stages of a safety zone, a forecasting zone and a dangerous zone according to the change of the amplitude of the space electric field, specific characteristic parameters are input into a pre-constructed and training prediction model, and the pollution flashover development section is identified;
step 3: judging according to the identification result of the pollution flashover development section, and if the pollution flashover development section is identified as a forecast area and a dangerous area, sending pollution flashover early warning information; if the pollution flashover development section is identified as a safety area, a pollution degree prediction model is established, and pollution degree prediction is carried out.
Further, the space electric field signal at the cross arm of the tower is monitored, and the distance between the monitoring position and the insulator string is 0.5-1m.
Further, in the step 1, the characteristic parameter includes a spatial electric field signal peak value E max Effective value E rms Total harmonic waveWave distortion rate alpha THD And third harmonic factor K 3 。
Further, the peak value E max Is the amplitude of 50Hz cycle of a spatial electric field signal waveform, E rms The effective value is calculated by the following steps:
wherein e (i) is electric field waveform sampling data, and N is the number of data points of one cycle of sampling;
the sampled data points of one cycle are subjected to fast Fourier transformation, and then the total harmonic distortion rate alpha can be obtained THD ：
G in k Representing the magnitude of the kth harmonic;
based further on the fast fourier transform result, third harmonic factors can be obtained:
further, the characteristic parameters are subjected to further dimensionless treatment:
k max ＝E max /E' max
g＝E max /E rms
in E' max For E in the previous early warning period max Minimum value as E of current early warning period max Comparing the reference, the early warning period can take one day; k in max Namely, the peak value coefficient can represent the change characteristic of the waveform amplitude of the spatial electric field in the current early warning period; g is a peak factor, i.e., the ratio of peak to effective value, that characterizes the degree of distortion of the waveform.
Further, in the step 2, a prediction mode is inputThe specific characteristic parameter is alpha THD 、K 3 、g、k max ；
And training a prediction model by adopting a cluster recognition method of a random forest and a decision tree.
Further, the training of the predictive model includes the steps of:
s1: selecting space electric field signal characteristic parameter data of three areas to perform random forest training;
randomly sampling the training sample set by a Bootstrap aggregation method to form n sub-sample sets, wherein the data quantity in each sub-sample set is the same as that of the training sample set, namely M;
s2: the total number of the attributes of the training set is N, and the total number of the attributes is the total number of judgment bases corresponding to the specific characteristic parameters respectively;
dividing the sample into two nodes according to whether the sample meets classification attributes or not, and calculating the coefficient of the foundation classified according to different attributes;
selecting an attribute from the root node as a splitting attribute of the node, wherein the basis of selecting the splitting attribute is to follow the principle of minimum coefficient of the foundation so that the sum of the coefficient of the foundation of each node after splitting is minimum; then selecting the splitting attribute of the next splitting point, and so on until a decision tree meeting the requirement is obtained;
s3: repeating the step S2 to form n decision trees, thereby forming a random forest; each decision tree will have a vote result, the most voted category being the final model prediction result.
Further, in S2, the calculation formula of the coefficient of ken is as follows:
where P (i) is the ratio of the number of samples divided into class i to the total samples of the node.
Further, in the step 3, when the identification result is a non-safe area, pollution flashover early warning information is sent out;
when the identification result is that the safety zone is transited to the forecast zone for the first time, the early warning information is attractive;
when the identification result is that the safety zone and the forecast zone are transited back and forth, the early warning information is focused;
when the identification result is in the forecast area and the duration time is longer, the early warning information is that cleaning measures are taken immediately to the site;
when the identification result is a dangerous area, the early warning information is immediately sent to the site to take cleaning measures.
Further, in the step 3, training the BP neural network by a gradient descent method to establish a pollution degree prediction model;
at modeling, 0.01mg/cm 2 、0.05mg/cm 2 、0.10mg/cm 2 、0.20mg/cm 2 、0.35mg/cm 2 And establishing a sample database for the space electric field characteristic parameters under five pollution degrees so as to train and test, and ensuring the accuracy of the model.
Further, after modeling is completed, inputting the specific characteristic parameters obtained in real time into a neural network prediction model to obtain a pollution degree prediction result.
On the other hand, the patent provides a pollution flashover early warning system based on an electric field signal, which comprises a characteristic parameter extraction module, a pollution flashover development section identification module and a pollution flashover early warning judgment module;
the characteristic parameter extraction module is used for extracting the characteristic parameter of the electric field signal of one cycle to obtain a specific characteristic parameter alpha THD 、K 3 、g、k max ；
The pollution flashover development section identification module inputs specific characteristic parameters into a pre-constructed and training prediction model to finish the identification of the pollution flashover development section, wherein the pollution flashover development section is divided into three stages of a safety zone, a forecasting zone and a danger zone according to the change of the spatial electric field amplitude;
the pollution flashover early warning judging module is used for sending out pollution flashover early warning information when the pollution flashover development section is identified as a forecast area and a dangerous area; when the pollution flashover development section is identified as a safety zone, a pollution degree prediction model is established, pollution degree prediction is carried out, and a BP neural network is trained by a gradient descent method in a modeling process.
Compared with the prior art, the method has the beneficial effects that the key characteristic parameters are extracted by utilizing the electric field signal of the operation insulator space, and the identification of the pollution flashover development section and the prediction of the pollution degree of the surface of the insulator are carried out, so that the pollution flashover early warning is realized. The method is expected to realize insulated non-contact pollution flashover early warning outside power transmission and distribution, so that intelligent operation and detection means and measures of the power system are expanded, and the requirements of current ubiquitous power Internet of things construction and strong intelligent power grid construction are met.
Drawings
FIG. 1 is a schematic diagram of an electric field signal monitoring position in a pollution flashover early warning method based on an electric field signal;
fig. 2 is a schematic diagram of an electric field waveform of an overhead line insulator in the pollution flashover early warning method based on an electric field signal;
FIG. 3 is a flow chart of pollution flashover early warning in a pollution flashover early warning method based on an electric field signal;
FIG. 4 is a characteristic diagram of a pollution flashover development section in a pollution flashover early warning method based on an electric field signal;
FIG. 5 is a flow chart of a random forest algorithm in a pollution flashover early warning method based on an electric field signal;
FIG. 6 is a schematic diagram of decision tree classification in a pollution flashover early warning method based on electric field signals provided by the application;
fig. 7 is a diagram of a neural network prediction structure in a pollution flashover early warning method based on an electric field signal.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the technical solutions of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. The described embodiments of the application are only some, but not all, embodiments of the application. All other embodiments, which can be made by those skilled in the art without inventive faculty, are within the scope of the application, based on the spirit of the application.
As shown in fig. 1, the electric field signal monitoring position may be a certain point of three points A, B, C in the drawing, and the distance between the electric field signal monitoring position and the insulator string may be between 0.5 and 1m.
A typical overhead line insulator space electric field waveform is shown in fig. 2. According to the waveform change characteristics, the application provides that the characteristic parameter extraction is carried out on the space electric field signal of one cycle, and the space electric field signal comprises four characteristic parameters, namely a space electric field signal peak E max Effective value E rms Total harmonic distortion rate alpha THD And third harmonic factor K 3 . Wherein the peak value Emax is the amplitude of a 50Hz cycle of the waveform of the spatial electric field signal, E rms The effective value is calculated by the following steps:
wherein e (i) is electric field waveform sampling data, and N is the number of data points of one cycle sampling.
The sampled data points of one cycle are subjected to fast Fourier transformation, and then the total harmonic distortion rate alpha can be obtained THD ：
G in k The magnitude of the kth harmonic is indicated. Based further on the fast fourier transform result, third harmonic factors can be obtained:
the characteristic parameters are further processed in a dimensionless mode to eliminate the influence caused by the difference of electric field intensity basic values under different voltage levels and monitoring point distances, and therefore the general change rule of the space electric field signals can be embodied:
k max ＝E max /E' max (4)
g＝E max /E rms (5)
in E' max For E in the previous early warning period max And compared with a standard, the early warning period can take one day. Then get k according to equation (4) max Namely, the peak value coefficient, can represent the change characteristic of the waveform amplitude of the spatial electric field in the current early warning period. g is a peak factor, i.e., the ratio of peak to effective value, that characterizes the degree of distortion of the waveform.
The application provides a pollution flashover early warning method based on an electric field signal, which comprises the following steps:
step 1: extracting characteristic parameters of a space electric field signal of a cycle to obtain k max 、g、K 3 、α THD 。
Step 2: as shown in fig. 4, the running insulator from surface wetting to flashover can divide the pollution flashover development section into three stages of a safety zone, a forecasting zone and a dangerous zone according to the change of the amplitude of the space electric field, and specific characteristic parameters are input into a pre-constructed and trained prediction model to identify the pollution flashover development section.
Data samples in the predictive model are obtained in the laboratory and in the field. The application adopts a cluster recognition method of random forests and decision trees to recognize the pollution flashover development sections. Random Forest (RF) pattern recognition involves a combined classifier set consisting of multiple decision trees. It can learn the characteristics of the sample comprehensively and perform recognition well. The method selects 100 groups of spatial electric field signal characteristic parameter data of three areas for RF training, wherein the data comprises 35 groups of safety zone data, 43 groups of forecast zone data and 22 groups of danger zone data. A flow chart of the RF algorithm is shown in fig. 5.
According to the application, the training sample set is randomly sampled by a Bootstrap aggregation method to form n sub-sample sets, and the data quantity in each sub-sample set is the same as that in the training sample set, namely 80. A decision tree is then constructed based on each sub-sample set. The classification is performed herein using a decision tree approach. An example of a decision tree is shown in fig. 6. Parameter k in frame max ≤0.3、K 3 And the node classification attribute is equal to or less than 0.8 or g is equal to or less than 3.5. "Samples" is the number of Samples of the current node, and "Value" is the classification of the current node, i.e., the number of Samples of different regions. In the order of the safe area, the dangerous area and the critical area, the Class is the classification result of the current node, namely the area where most samples of the node are located.
The sample is divided into two nodes according to whether the sample satisfies the classification attribute. And calculating the coefficient of the Kernel classified according to different parameters, and selecting the parameter with the smallest coefficient of the Kernel to divide the decision tree. Gini is the Gini coefficient of the current node, and the calculation formula is as follows:
where P (i) is the ratio of the number of samples divided into class i to the total samples of the node. The smaller the coefficient of kunity, the lower the purity and the better the properties.
One attribute is selected from the root node (first node) as a split attribute of the node. As shown in FIG. 6, the root node has a partition attribute of k max Less than or equal to 0.3. The basis for selecting the segmentation attribute is to follow the principle of minimum kunit so that the sum of the kunit of each node after segmentation is minimal. Then select the split attribute of the second split node, in the example K 3 Less than or equal to 0.8. Then a third split is performed and so on until a decision tree is obtained that meets the requirements. Since the number of samples in each partition is different in different subsamples, the selection order of node partition attributes is also different, that is, the root node partition attribute of other decision trees may be K 3 0.7 or g.ltoreq.3, determined according to the principle of minimum coefficient of kunity.
Repeating the steps to form n decision trees, thereby forming a random forest. For the test set, there will be one voting result per decision tree, with the most voted class being the final model prediction result.
Step 3: judging according to the identification result of the pollution flashover development section, and if the pollution flashover development section is identified as a forecast area and a dangerous area, sending pollution flashover early warning information; if the pollution flashover development section is identified as a safety area, a pollution degree prediction model is established, and pollution degree prediction is carried out.
The early warning mode comprises the following steps:
TABLE 1
The safety section has longer duration and belongs to a state in which the insulator is relatively stable, so the application provides that the pollution degree prediction is carried out in the safety section. The application trains BP neural network to build pollution degree prediction model by gradient descent method. The neural network has high classification precision and strong learning capability, and can fully approximate complex nonlinear relations. The gradient descent method is an optimization algorithm that finds the minimum of the loss function between the output value and the desired value, and the lower the minimum of the loss function, the closer the output result is to the desired result. In modeling, the equivalent salt density is 0.01mg/cm 2 、0.05mg/cm 2 、0.10mg/cm 2 、0.20mg/cm 2 、0.35mg/cm 2 And establishing a sample database for the space electric field characteristic parameters under five pollution degrees so as to train and test, and ensuring the accuracy of the model. After modeling is completed, the characteristic parameters obtained in real time are input into a neural network prediction model, and then a pollution degree prediction result can be obtained. The neural network prediction structure is shown in fig. 7.
A pollution flashover early warning system based on an electric field signal comprises a characteristic parameter extraction module, a pollution flashover development section identification module and a pollution flashover early warning judgment module;
the characteristic parameter extraction module is used for extracting the characteristic parameter of the electric field signal of one cycle to obtain a specific characteristic parameter alpha THD 、K 3 、g、k max ；
The pollution flashover development section identification module inputs specific characteristic parameters into a pre-constructed and training prediction model to finish the identification of the pollution flashover development section, wherein the pollution flashover development section is divided into three stages of a safety zone, a forecasting zone and a danger zone according to the change of the spatial electric field amplitude;
the pollution flashover early warning judging module is used for sending out pollution flashover early warning information when the pollution flashover development section is identified as a forecast area and a dangerous area; when the pollution flashover development section is identified as a safety zone, a pollution degree prediction model is established, pollution degree prediction is carried out, and a BP neural network is trained by a gradient descent method in a modeling process.
Compared with the prior art, the method has the beneficial effects that the key characteristic parameters are extracted by utilizing the electric field signal of the operation insulator space, and the identification of the pollution flashover development section and the prediction of the pollution degree of the surface of the insulator are carried out, so that the pollution flashover early warning is realized. The method is expected to realize insulated non-contact pollution flashover early warning outside power transmission and distribution, so that intelligent operation and detection means and measures of the power system are expanded, and the requirements of current ubiquitous power Internet of things construction and strong intelligent power grid construction are met.
Finally, it should be noted that the above embodiments are only for illustrating the technical solution of the present application and not for limiting the same, and although the present application has been described in detail with reference to the above embodiments, it should be understood by those skilled in the art that: modifications and equivalents may be made to the specific embodiments of the application without departing from the spirit and scope of the application, which is intended to be covered by the claims.
Claims (12)
1. The pollution flashover early warning method based on the electric field signal is characterized by comprising the following steps of:
step 1: extracting characteristic parameters of a space electric field signal of one cycle;
step 2: the running insulator wets from the surface to the flashover, the pollution flashover development section can be divided into three stages of a safety zone, a forecasting zone and a dangerous zone according to the change of the amplitude of the space electric field, specific characteristic parameters are input into a pre-constructed and training prediction model, and the pollution flashover development section is identified;
step 3: judging according to the identification result of the pollution flashover development section, and if the pollution flashover development section is identified as a forecast area and a dangerous area, sending pollution flashover early warning information; if the pollution flashover development section is identified as a safety area, a pollution degree prediction model is established, and pollution degree prediction is carried out.
2. The pollution flashover early warning method based on the electric field signal according to claim 1, wherein the method comprises the following steps of:
and monitoring a space electric field signal at the cross arm of the tower, wherein the distance between the monitoring position and the insulator string is 0.5-1m.
3. The pollution flashover early warning method based on the electric field signal according to claim 1, wherein the method comprises the following steps of:
in the step 1, the characteristic parameter includes a spatial electric field signal peak value E max Effective value E rms Total harmonic distortion rate alpha THD And third harmonic factor K 3 。
4. The pollution flashover pre-warning method based on the electric field signal according to claim 3, wherein the method comprises the following steps of:
the peak value E max Is the amplitude of 50Hz cycle of a spatial electric field signal waveform, E rms The effective value is calculated by the following steps:
wherein e (i) is electric field waveform sampling data, and N is the number of data points of one cycle of sampling;
the sampled data points of one cycle are subjected to fast Fourier transformation, and then the total harmonic distortion rate alpha can be obtained THD ：
G in k Representing the magnitude of the kth harmonic;
based further on the fast fourier transform result, third harmonic factors can be obtained:
5. the pollution flashover early warning method based on the electric field signal according to claim 4, wherein the method comprises the following steps:
and carrying out further dimensionless treatment on the characteristic parameters:
k max ＝E max /E' max
g＝E max /E rms
in E' max For E in the previous early warning period max Minimum value as E of current early warning period max Comparing the reference, the early warning period can take one day; k in max Namely, the peak value coefficient can represent the change characteristic of the waveform amplitude of the spatial electric field in the current early warning period; g is a peak factor, i.e., the ratio of peak to effective value, that characterizes the degree of distortion of the waveform.
6. The pollution flashover early warning method based on the electric field signal according to claim 5, wherein the method comprises the following steps:
in the step 2, the specific characteristic parameter input into the prediction model is alpha THD 、K 3 、g、k max ；
And training a prediction model by adopting a cluster recognition method of a random forest and a decision tree.
7. The pollution flashover early warning method based on the electric field signal according to claim 6, wherein the method comprises the following steps:
the training of the prediction model comprises the following steps:
s1: selecting space electric field signal characteristic parameter data of three areas to perform random forest training;
randomly sampling the training sample set by a Bootstrap aggregation method to form n sub-sample sets, wherein the data quantity in each sub-sample set is the same as that of the training sample set, namely M;
s2: the total number of the attributes of the training set is N, and the total number of the attributes is the total number of judgment bases corresponding to the specific characteristic parameters respectively;
dividing the sample into two nodes according to whether the sample meets classification attributes or not, and calculating the coefficient of the foundation classified according to different attributes;
selecting an attribute from the root node as a splitting attribute of the node, wherein the basis of selecting the splitting attribute is to follow the principle of minimum coefficient of the foundation so that the sum of the coefficient of the foundation of each node after splitting is minimum; then selecting the splitting attribute of the next splitting point, and so on until a decision tree meeting the requirement is obtained;
s3: repeating the step S2 to form n decision trees, thereby forming a random forest; each decision tree will have a vote result, the most voted category being the final model prediction result.
8. The pollution flashover early warning method based on the electric field signal according to claim 7, wherein the method comprises the following steps:
in S2, the calculation formula of the coefficient of ken is as follows:
where P (i) is the ratio of the number of samples divided into class i to the total samples of the node.
9. The pollution flashover early warning method based on the electric field signal according to claim 1, wherein the method comprises the following steps of:
in the step 3, when the identification result is a non-safety zone, pollution flashover early warning information is sent out;
when the identification result is that the safety zone is transited to the forecast zone for the first time, the early warning information is attractive;
when the identification result is that the safety zone and the forecast zone are transited back and forth, the early warning information is focused;
when the identification result is in the forecast area and the duration time is longer, the early warning information is that cleaning measures are taken immediately to the site;
when the identification result is a dangerous area, the early warning information is immediately sent to the site to take cleaning measures.
10. The pollution flashover early warning method based on the electric field signal according to claim 1, wherein the method comprises the following steps of:
in the step 3, training a BP neural network by a gradient descent method to establish a pollution degree prediction model;
at modeling, 0.01mg/cm 2 、0.05mg/cm 2 、0.10mg/cm 2 、0.20mg/cm 2 、0.35mg/cm 2 And establishing a sample database for the space electric field characteristic parameters under five pollution degrees so as to train and test, and ensuring the accuracy of the model.
11. The pollution flashover pre-warning method based on the electric field signal according to claim 10, wherein the method comprises the following steps of:
after modeling is completed, inputting the specific characteristic parameters obtained in real time into a neural network prediction model to obtain a pollution degree prediction result.
12. The utility model provides a dirty sudden strain of a muscle early warning system based on electric field signal, includes feature parameter extraction module, dirty sudden strain of a muscle development district identification module and dirty sudden strain of a muscle early warning judgement module, its characterized in that:
the characteristic parameter extraction module is used for extracting the characteristic parameter of the electric field signal of one cycle to obtain a specific characteristic parameter alpha THD 、K 3 、g、k max ；
The pollution flashover development section identification module inputs specific characteristic parameters into a pre-constructed and training prediction model to finish the identification of the pollution flashover development section, wherein the pollution flashover development section is divided into three stages of a safety zone, a forecasting zone and a danger zone according to the change of the spatial electric field amplitude;
the pollution flashover early warning judging module is used for sending out pollution flashover early warning information when the pollution flashover development section is identified as a forecast area and a dangerous area; when the pollution flashover development section is identified as a safety zone, a pollution degree prediction model is established, pollution degree prediction is carried out, and a BP neural network is trained by a gradient descent method in a modeling process.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310393819.8A CN116595459A (en) | 2023-04-13 | 2023-04-13 | Pollution flashover early warning method and system based on electric field signals |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310393819.8A CN116595459A (en) | 2023-04-13 | 2023-04-13 | Pollution flashover early warning method and system based on electric field signals |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116595459A true CN116595459A (en) | 2023-08-15 |
Family
ID=87603410
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310393819.8A Pending CN116595459A (en) | 2023-04-13 | 2023-04-13 | Pollution flashover early warning method and system based on electric field signals |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116595459A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117330883A (en) * | 2023-12-01 | 2024-01-02 | 国网山西省电力公司电力科学研究院 | Overhead line insulator running state monitoring system and method |
| CN117368797A (en) * | 2023-11-17 | 2024-01-09 | 国网青海省电力公司海南供电公司 | Composite insulator flashover early warning method based on leakage current and EFS |
-
2023
- 2023-04-13 CN CN202310393819.8A patent/CN116595459A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117368797A (en) * | 2023-11-17 | 2024-01-09 | 国网青海省电力公司海南供电公司 | Composite insulator flashover early warning method based on leakage current and EFS |
| CN117330883A (en) * | 2023-12-01 | 2024-01-02 | 国网山西省电力公司电力科学研究院 | Overhead line insulator running state monitoring system and method |
| CN117330883B (en) * | 2023-12-01 | 2024-02-27 | 国网山西省电力公司电力科学研究院 | Overhead line insulator running state monitoring system and method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116595459A (en) | Pollution flashover early warning method and system based on electric field signals | |
| CN106908674B (en) | Transformer state evaluation method based on multi-state quantity prediction | |
| CN103983757B (en) | Based on the transformer insulated heat ageing state reliability estimation method of mixture Weibull distribution | |
| CN111768082A (en) | Power equipment state evaluation method based on big data analysis | |
| CN108832619A (en) | Transient stability evaluation in power system method based on convolutional neural networks | |
| CN105740546A (en) | Transmission tower and ground wire security evaluation and service life prediction system based on corrosion state | |
| CN103440497B (en) | A kind of GIS insulation defect shelf depreciation collection of illustrative plates mode identification method | |
| CN112200694A (en) | Dominant instability mode identification model construction and application method based on graph neural network | |
| CN104578053A (en) | Power system transient stability prediction method based on disturbance voltage trajectory cluster features | |
| CN116819250A (en) | Method for detecting fault arc in photovoltaic system | |
| Zhang et al. | Research on the application of Decision Tree and Random Forest Algorithm in the main transformer fault evaluation | |
| CN113065522A (en) | Transformer partial discharge type identification method based on deep neural network | |
| Li et al. | A line-fault cause analysis method for distribution network based on decision-making tree and machine learning | |
| CN109711687B (en) | Insulator state fuzzy evaluation method based on improved entropy method | |
| CN105162413A (en) | Method for evaluating performances of photovoltaic system in real time based on working condition identification | |
| CN116244600A (en) | Method, system and equipment for constructing GIS intermittent discharge mode identification model | |
| CN113689053B (en) | Strong convection weather overhead line power failure prediction method based on random forest | |
| CN108536911B (en) | Converter transformer state evaluation method based on center distance and sample characteristics | |
| CN112348262A (en) | Power cable operation safety early warning method | |
| Yu et al. | BP Neural Network Based Fault Prediction of Distribution Network during Typhoon Disaster | |
| Zhang et al. | Research on main transformer defect detection methods based on Conditional Inference Tree and AdaBoost Algorithm | |
| CN112881827B (en) | Oil-immersed transformer fault diagnosis method based on improved grey correlation analysis | |
| Qi et al. | Research on Transformer Fault Detection Based on CEEMDAN-DBN | |
| Zhang et al. | Icing Image based power system voltage stability evaluation of transmission lines | |
| Pei et al. | Construction of power equipment fault feature model based on unified semantic expression |
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 |