CN110108992B - Cable partial discharge fault identification method and system based on improved random forest algorithm - Google Patents
Cable partial discharge fault identification method and system based on improved random forest algorithm Download PDFInfo
- Publication number
- CN110108992B CN110108992B CN201910440682.0A CN201910440682A CN110108992B CN 110108992 B CN110108992 B CN 110108992B CN 201910440682 A CN201910440682 A CN 201910440682A CN 110108992 B CN110108992 B CN 110108992B
- Authority
- CN
- China
- Prior art keywords
- partial discharge
- cable
- random forest
- data
- forest algorithm
- 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.)
- Active
Links
- 238000004422 calculation algorithm Methods 0.000 title claims abstract description 73
- 238000007637 random forest analysis Methods 0.000 title claims abstract description 61
- 238000000034 method Methods 0.000 title claims abstract description 54
- 238000012937 correction Methods 0.000 claims abstract description 13
- 238000013507 mapping Methods 0.000 claims abstract description 13
- 238000010276 construction Methods 0.000 claims abstract description 10
- 238000012549 training Methods 0.000 claims description 56
- 238000003066 decision tree Methods 0.000 claims description 36
- 238000012360 testing method Methods 0.000 claims description 30
- 238000009826 distribution Methods 0.000 claims description 29
- 238000004590 computer program Methods 0.000 claims description 16
- 230000007547 defect Effects 0.000 claims description 16
- 238000003860 storage Methods 0.000 claims description 13
- 238000012545 processing Methods 0.000 claims description 8
- 238000012952 Resampling Methods 0.000 claims description 5
- 238000013138 pruning Methods 0.000 claims description 5
- 238000009966 trimming Methods 0.000 claims description 5
- 238000005457 optimization Methods 0.000 claims description 4
- 230000000007 visual effect Effects 0.000 claims description 4
- 238000005065 mining Methods 0.000 abstract description 3
- 238000010586 diagram Methods 0.000 description 10
- 230000006870 function Effects 0.000 description 5
- 238000009413 insulation Methods 0.000 description 3
- 238000003909 pattern recognition Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 238000003745 diagnosis Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 239000000126 substance Substances 0.000 description 2
- 230000032683 aging Effects 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000009960 carding Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/12—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
- G01R31/1263—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 solid or fluid materials, e.g. insulation films, bulk material; of semiconductors or LV electronic components or parts; of cable, line or wire insulation
- G01R31/1272—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 solid or fluid materials, e.g. insulation films, bulk material; of semiconductors or LV electronic components or parts; of cable, line or wire insulation of cable, line or wire insulation, e.g. using partial discharge measurements
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Testing Relating To Insulation (AREA)
Abstract
The invention discloses a cable partial discharge fault identification method, a system and a medium based on an improved random forest algorithm, wherein the method comprises the steps of collecting a partial discharge signal of a cable; extracting characteristic data of the partial discharge signal; inputting the extracted feature data into a pre-constructed and trained classifier to obtain a cable partial discharge fault type corresponding to a cable partial discharge signal, wherein the classifier is based on an improved random forest algorithm, the improved random forest algorithm is used for mining high-dimensional attribute data by using a feature construction method and guiding the selection of an attribute subset by using a data weight correction idea based on an Adaboost algorithm, and the classifier is pre-trained to establish a mapping relation between the features of the partial discharge signal and the cable partial discharge fault type. The invention can ensure the identification accuracy, improve the identification efficiency and realize the combination of the identification accuracy and the identification efficiency.
Description
Technical Field
The invention relates to the field of cable aging detection, in particular to a cable partial discharge fault identification method, a system and a medium based on an improved random forest algorithm.
Background
Electric energy plays an indispensable role in real life, and a cable is used as an important part for transmitting the electric energy and has been widely applied, and particularly occupies a main position in an urban power distribution network. The partial discharge amount of the cable subjected to partial discharge is closely related to the current insulation condition of the cable, so that the most effective and intuitive way for evaluating the insulation condition of the cable is to measure the partial discharge amount of the cable. Therefore, the diagnosis method for the partial discharge of the cable in operation is perfected, and the diagnosis method has important significance for improving the reliability and safety of the power system.
When the cable generates partial discharge, physical phenomena such as light, heat, electromagnetic waves, electric pulses and the like are accompanied, and the physical phenomena are the basis for detecting the partial discharge of the cable. Although the insulation state of the cable can be known from the partial discharge generated in the cable, pattern recognition of the partial discharge signal is required to determine the defect type of the partial discharge fault of the cable. At present, the main method of partial discharge pattern recognition is to set up corresponding types of tests according to several common partial discharge types of cables, perform a large number of tests, extract feature data from each group of types of tests, establish a map library for the obtained feature data, train several types of obtained map data by adopting an intelligent algorithm, and finally achieve the effect of classification recognition. The most commonly adopted intelligent algorithms in the classification and identification process comprise a distance-based pattern random forest classifier, a support vector machine, a neural network, a fuzzy random forest classifier and the like. On the premise of a large amount of training data and training time, the intelligent algorithms can achieve good effects, but the partial discharge pattern recognition cannot be achieved in terms of accuracy and time.
Disclosure of Invention
The technical problems to be solved by the invention are as follows: aiming at the problems in the prior art, the cable partial discharge fault identification method, the system and the medium based on the improved random forest algorithm are provided, which can ensure the identification accuracy, improve the identification efficiency, and realize the identification accuracy and efficiency.
In order to solve the technical problems, the invention adopts the technical scheme that:
a cable partial discharge fault identification method based on an improved random forest algorithm comprises the following implementation steps:
1) collecting a partial discharge signal of the cable;
2) extracting characteristic data of the partial discharge signal;
3) inputting the extracted feature data into a pre-constructed and trained classifier to obtain a cable partial discharge fault type corresponding to the cable partial discharge signal, wherein the classifier is based on an improved random forest algorithm, the improved random forest algorithm utilizes a feature construction 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 pre-trained to establish a mapping relation between the features of the partial discharge signal and the cable partial discharge fault type.
Preferably, the feature data in step 2) includes at least one of 24 high-dimensional attributes, and the 24 high-dimensional attributes include
And
the asymmetry asymmetries Asy and the cross-correlation coefficient Cc of the three,
and
skewness Sk, steepness Ku, Peak Peak; wherein the content of the first and second substances,
a two-dimensional spectrogram of the average discharge capacity phase distribution is shown,
a two-dimensional spectrogram of the phase distribution of the maximum discharge capacity is shown,
a phase distribution spectrogram representing the discharge times;
a two-dimensional spectrogram representing the phase distribution of the average discharge capacity of the positive half cycle,
a two-dimensional spectrogram of the phase distribution of the average discharge capacity of the negative half cycle is shown,
a two-dimensional spectrogram showing the phase distribution of the maximum discharge capacity of the positive half cycle,
a two-dimensional spectrogram representing the phase distribution of the maximum discharge capacity of the negative half cycle,
a phase distribution spectrogram showing the positive half-cycle discharge times,
indicating number of negative half cyclesA digital phase distribution spectrum.
Preferably, the cable partial discharge fault types include internal discharge, creeping discharge and corona discharge.
Preferably, step 3) is preceded by a step of training a classifier, and the detailed steps include:
s1) respectively collecting cable partial discharge signals aiming at various cable partial discharge fault types and extracting characteristic data, and respectively selecting one part as a training set and the other part as a test set; resampling the training set by adopting a Bootstrap method, and randomly generating the training set, wherein the total number of attributes of the training set is M, the total number of attributes is the total number of judgment bases corresponding to the characteristic data respectively judged, and the number of sub-attributes is N;
s2) adopting a data set weight correction idea of an Adaboost algorithm to guide the selection of the attribute subset, and carrying out optimization processing on the attribute subset, wherein the attribute subset is a set of judgment bases corresponding to characteristic parameters used in a classification process; selecting an attribute subset Q on the basis of the training set, wherein the training set is a characteristic parameter used in the classification process, and the sub-attributes are judgment bases corresponding to the characteristic parameters respectively; directly trimming the selected probability of the attribute subset Q (i +1) through fault data in Ci of the previous sub-decision tree, wherein the sub-decision tree is an visual description of each group of data from the beginning to the end of classification, and selecting the attribute subset according to the trimmed probability value and a conventional random method;
s3) generating a corresponding decision tree by utilizing the constructed training set and the corresponding attribute subset, taking the information gain as a division principle, and splitting the node in the mode that the splitting mode when the information gain is maximum is the optimal splitting mode of the decision tree; and each tree grows completely without pruning;
s4) testing the test set by utilizing the decision tree to obtain corresponding categories, and taking the category with the most output in the decision tree as the category to which the test set belongs, thereby establishing the mapping relation between the characteristics of the local discharge signal and the cable local discharge fault type.
The invention also provides a cable defect identification system based on the improved random forest algorithm, which comprises computer equipment, wherein the computer equipment is programmed or configured to execute the steps of the cable partial discharge fault identification method based on the improved random forest algorithm.
The invention also provides a cable defect identification system based on the improved random forest algorithm, which comprises computer equipment, wherein a storage medium of the computer equipment is stored with a computer program which is programmed or configured to execute the cable partial discharge fault identification method based on the improved random forest algorithm.
The invention also provides a cable defect identification system based on the improved random forest algorithm, which comprises partial discharge signal acquisition equipment and an upper computer which are connected with each other, wherein the upper computer is programmed or configured to execute the steps of the cable partial discharge fault identification method based on the improved random forest algorithm.
The invention also provides a cable defect identification system based on the improved random forest algorithm, which comprises partial discharge signal acquisition equipment and an upper computer which are connected with each other, wherein a storage medium of the upper computer is stored with a computer program which is programmed or configured to execute the cable partial discharge fault identification method based on the improved random forest algorithm.
The present invention also provides a computer readable storage medium having stored thereon a computer program programmed or configured to execute the cable partial discharge fault identification method based on the improved random forest algorithm of the present invention.
The invention also provides a cable defect identification system based on the improved random forest algorithm, which comprises the following steps:
the signal acquisition program unit is used for acquiring partial discharge signals of the cable;
the characteristic extraction program unit is used for extracting characteristic data of the partial discharge signal;
and the classification identification program unit is used for inputting the extracted feature data into a pre-constructed and trained classifier to obtain a cable partial discharge fault type corresponding to the cable partial discharge signal, the classifier is based on an improved random forest algorithm, the improved random forest algorithm utilizes a feature construction 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 pre-trained to establish a mapping relation between the features of the partial discharge signal and the cable partial discharge fault type.
Compared with the prior art, the invention has the following advantages: the invention inputs the extracted characteristic data into a pre-constructed classifier which completes training to obtain the cable partial discharge fault type corresponding to the cable partial discharge signal, the classifier is based on an improved random forest algorithm, the improved random forest algorithm excavates high-dimensional attribute data by using the characteristic construction method and guides the selection of the attribute subset by using the data weight correction thought based on the Adaboost algorithm, the classifier is pre-trained to establish the mapping relation between the characteristics of the partial discharge signal and the cable partial discharge fault type, the classifier identifies the cable partial discharge fault type by using the improved random forest algorithm, excavates the high-dimensional attribute data by using the characteristic construction method, guides the selection of the attribute subset by using the data set weight correction thought based on the Adaboost algorithm, identifies the cable partial discharge fault type, and greatly shortens the time under the condition of ensuring the accuracy, the recognition efficiency is improved.
Drawings
FIG. 1 is a schematic diagram of a basic flow of a method according to an embodiment of the present invention.
FIG. 2 is a graph illustrating cable partial discharge fault type identification profiles for 300 test data sets in accordance with an embodiment of the present invention.
Detailed Description
As shown in fig. 1, the implementation steps of the cable partial discharge fault identification method based on the improved random forest algorithm in this embodiment include:
1) collecting a partial discharge signal of the cable;
2) extracting characteristic data of the partial discharge signal;
3) inputting the extracted feature data into a pre-constructed and trained classifier to obtain a cable partial discharge fault type corresponding to a cable partial discharge signal, wherein the classifier is based on an improved random forest algorithm, the improved random forest algorithm is used for mining high-dimensional attribute data by using a feature construction method and guiding the selection of an attribute subset by using a data weight correction idea based on an Adaboost algorithm, and the classifier is pre-trained to establish a mapping relation between the features of the partial discharge signal and the cable partial discharge fault type.
In this embodiment, the feature data in step 2) includes 24 high-dimensional attributes, see table 1:
as shown in Table 1, the above-mentioned 24 high-dimensional attributes include
And
the asymmetry asymmetries Asy and the cross-correlation coefficient Cc of the three,
and
skewness Sk, steepness Ku, Peak Peak; wherein the content of the first and second substances,
a two-dimensional spectrogram of the average discharge capacity phase distribution is shown,
a two-dimensional spectrogram of the phase distribution of the maximum discharge capacity is shown,
a phase distribution spectrogram representing the discharge times;
a two-dimensional spectrogram representing the phase distribution of the average discharge capacity of the positive half cycle,
a two-dimensional spectrogram of the phase distribution of the average discharge capacity of the negative half cycle is shown,
a two-dimensional spectrogram showing the phase distribution of the maximum discharge capacity of the positive half cycle,
a two-dimensional spectrogram representing the phase distribution of the maximum discharge capacity of the negative half cycle,
a phase distribution spectrogram showing the positive half-cycle discharge times,
and (3) showing a phase distribution spectrogram of the number of negative half-cycle discharges. In addition, one or more of the above 24 attributes can be selected as required to construct the feature data, but from the aspect of accuracy, the more attributes are adopted, the better the completeness of the constructed feature data is, and the more accurate the cable partial discharge fault identification result is. The mining of the high-dimensional attribute data by using the feature construction method in the step 3) specifically refers to constructing corresponding feature data, namely 24 high-dimensional attributes shown in table 1, by taking the basic partial discharge waveform as an object.
In this embodiment, the types of the cable partial discharge fault include internal discharge, surface discharge, and corona discharge.
In this embodiment, step 3) is preceded by a step of training a classifier, and the detailed steps include:
s1) respectively collecting cable partial discharge signals aiming at various cable partial discharge fault types and extracting characteristic data, and respectively selecting one part as a training set and the other part as a test set; resampling the training set by adopting a Bootstrap method, and randomly generating the training set, wherein the total number of attributes of the training set is M, the total number of attributes is the total number of judgment bases corresponding to the characteristic data respectively judged, and the number of sub-attributes is N;
s2) adopting a data set weight correction idea of an Adaboost algorithm to guide the selection of the attribute subset, and carrying out optimization processing on the attribute subset, wherein the attribute subset is a set of judgment bases corresponding to characteristic parameters used in the classification process; selecting an attribute subset Q on the basis of the training set, wherein the training set is a characteristic parameter used in the classification process, and the sub-attributes are judgment bases corresponding to the characteristic parameters respectively; directly trimming the selected probability of the attribute subset Q (i +1) through fault data in Ci of the previous sub-decision tree, wherein the sub-decision tree is an visual description of each group of data from the beginning to the end of classification, and selecting the attribute subset according to the trimmed probability value and a conventional random method;
in this embodiment, through three types of partial discharge tests, namely, internal discharge, creeping discharge and corona discharge, after a large number of tests, 600 groups of feature data with 24 attributes are respectively extracted from each group of tests, and after 1800 groups of feature data are obtained, 1500 groups of feature data are randomly selected as training data, and the remaining 300 groups of data are used as a test set. Creating a random forest classifier: resampling the training data by using a Bootstrap method to obtain 1500 training sets, S1，S2，S3，…，S1500The total number of attributes is M, and the number of sub-attributes is N.
The selection of the attribute subset Q is performed on the basis of the 1500 training sets described above, C of the last sub-decision treeiThe fault data in (1) directly affects the selected probability of a subset of attributes Q (i +1), each attribute in the subset of attributes being Qi,jIs the jth attribute qjWherein j is 1,2,3, …, m; . Hypothesis CiThe data set for diagnosing errors in the training subset in the sub-decision tree is miWhen training the subset Si+1In the set miThe number of data in (1) is niThen the next attribute subset Qi+1The selected probability P of each attribute in (a) changes. In this embodiment, a calculation function expression of the probability P of directly trimming the attribute subset Q (i +1) by the fault data in Ci of the previous sub-decision tree is as follows:
in the above formula, M is the total number of attributes of the training set, N is the number of sub-attributes, and k is the space variable, and the calculation method is as follows:
in the above formula, niFor the next training subset Si+1In the last diagnostic fault set miNumber of data in, miIs CiA data set for diagnosing errors in the training subsets in the sub-decision tree; q. q.si,jIs the jth attribute qjWherein j is 1,2,3, …, m; pjIs the jth attribute qjThe selected probability value of (a); m is the total number of attributes; n is the number of attributes in the sub-attribute set; qiSetting the attribute set of the last sub-decision tree; and alpha is an accuracy threshold value of the decision tree training.
S3) generating a corresponding decision tree (C1, C2, C3, … and C1500) by utilizing the constructed training set and the corresponding attribute subset, taking the information gain as a division principle, and splitting the node in the mode, wherein the splitting mode when the information gain is maximum is the optimal splitting mode of the decision tree; and each tree grows completely without pruning;
in this embodiment, each training set is used to generate a corresponding decision tree C1、C2、C3、…、C1500Before selecting attributes on each non-leaf node, randomly extracting m attributes from the 24 attributes as a split attribute set of the current node, wherein m is<And 24, splitting the node in the best splitting mode in the m attributes, wherein each tree grows completely without pruning.
The method takes information gain as a division principle, and comprises the following specific steps:
dividing m data into n classes, wherein the proportion of each class is PiNumber of i-th class/m;
assuming that D is the sample training set, the entropy inf (D) of the sample training set D is:
in the above formula, PiThe proportion of the ith class obtained by dividing m data into n classes is shown, and n is the number of the data division classes.
Suppose that the samples in the sample training set D are divided according to the attribute A, the attribute A divides the sample training set D into v different classes, and the entropy inf of the sample training set D after the divisionA(D) Comprises the following steps:
in the above formula, DjThe j class number in the class V is divided, V is the class carding of the attribute A dividing the sample training set D, Info (D)j) Is DjD is the sample training set D.
From this, the information gain (a) is:
Gain(A)＝inf(D)-infA(D)
in the above formula, inf (D) is the entropy of the sample training set D, infA(D) The entropy of the training set D is obtained by dividing the samples in the sample training set D according to the attribute A, and the subtraction of the two is the information gain (A).
S4) testing the test set by utilizing the decision tree to obtain corresponding categories, and taking the category with the most output in the decision tree as the category to which the test set belongs, thereby establishing the mapping relation between the characteristics of the local discharge signal and the cable local discharge fault type. In this embodiment, data to be tested is used as a test set sample X, each decision tree is used for testing, 1500 decision trees established by a random forest method are used for testing, corresponding categories C1(X), C2(X), C3(X), … and C1500(X) are obtained, and a voting method is used to take the category with the largest output from the 1500 decision trees as the category to which the test set sample X belongs, so that a mapping relationship between the characteristics of a local discharge signal and the cable local discharge fault type is established.
Fig. 2 is a distribution diagram of cable partial discharge fault type identification of 300 sets of test data according to an embodiment of the present invention, and as can be seen from fig. 2, in the identification result of the 300 sets of test data, among three decision tree categories of internal discharge (air gap discharge), creeping discharge and corona discharge, the number of decision trees of internal discharge (air gap discharge) is the largest, so that the final cable partial discharge fault type is identified as internal discharge (air gap discharge). By adopting the classifier in the embodiment to identify the partial discharge fault mode of the transformer, the time can be greatly shortened while the accuracy is ensured.
In addition, the present embodiment further provides a cable defect identification system based on the improved random forest algorithm, which includes a computer device programmed or configured to execute the steps of the cable partial discharge fault identification method based on the improved random forest algorithm according to the present embodiment. In addition, the present embodiment further provides a cable defect identification system based on an improved random forest algorithm, which includes a computer device, where a storage medium of the computer device stores a computer program programmed or configured to execute the cable partial discharge fault identification method based on the improved random forest algorithm according to the present embodiment. In addition, the embodiment further provides a cable defect identification system based on the improved random forest algorithm, which includes a partial discharge signal acquisition device and an upper computer connected to each other, where the upper computer is programmed or configured to execute the steps of the cable partial discharge fault identification method based on the improved random forest algorithm. In addition, the embodiment further provides a cable defect identification system based on the improved random forest algorithm, which includes a partial discharge signal acquisition device and an upper computer connected to each other, where a storage medium of the upper computer stores a computer program programmed or configured to execute the cable partial discharge fault identification method based on the improved random forest algorithm. Furthermore, the present embodiment also provides a computer readable storage medium, which stores thereon a computer program programmed or configured to execute the cable partial discharge fault identification method based on the improved random forest algorithm according to the present embodiment. In addition, the embodiment further provides a cable defect identification system based on the improved random forest algorithm, which includes:
the signal acquisition program unit is used for acquiring partial discharge signals of the cable;
the characteristic extraction program unit is used for extracting characteristic data of the partial discharge signal;
and the classification identification program unit is used for inputting the extracted feature data into a pre-constructed and trained classifier to obtain a cable partial discharge fault type corresponding to the cable partial discharge signal, the classifier is based on an improved random forest algorithm, the improved random forest algorithm utilizes a feature construction 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 pre-trained to establish a mapping relation between the features of the partial discharge signal and the cable partial discharge fault type.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The above description is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments, and all technical solutions belonging to the idea of the present invention belong to the protection scope of the present invention. It should be noted that modifications and embellishments within the scope of the invention may occur to those skilled in the art without departing from the principle of the invention, and are considered to be within the scope of the invention.
Claims (9)
1. A cable partial discharge fault identification method based on an improved random forest algorithm is characterized by comprising the following implementation steps:
1) collecting a partial discharge signal of the cable;
2) extracting characteristic data of the partial discharge signal;
3) inputting the extracted feature data into a pre-constructed and trained classifier to obtain a cable partial discharge fault type corresponding to the cable partial discharge signal, wherein the classifier is based on an improved random forest algorithm, the improved random forest algorithm utilizes a feature construction 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 pre-trained to establish a mapping relation between the features of the partial discharge signal and the cable partial discharge fault type;
step 3) is also preceded by a step of training a classifier, and the detailed steps comprise:
s1) respectively collecting cable partial discharge signals aiming at various cable partial discharge fault types and extracting characteristic data, and respectively selecting one part as a training set and the other part as a test set; resampling the training set by adopting a Bootstrap method, and randomly generating the training set, wherein the total number of attributes of the training set is M, the total number of attributes is the total number of judgment bases corresponding to the characteristic data respectively judged, and the number of sub-attributes is N;
s2) adopting a data set weight correction idea of an Adaboost algorithm to guide the selection of the attribute subset, and carrying out optimization processing on the attribute subset, wherein the attribute subset is a set of judgment bases corresponding to characteristic parameters used in a classification process; selecting an attribute subset Q on the basis of the training set, wherein the training set is a characteristic parameter used in the classification process, and the sub-attributes are judgment bases corresponding to the characteristic parameters respectively; directly trimming the selected probability of the attribute subset Q (i +1) through fault data in Ci of the previous sub-decision tree, wherein the sub-decision tree is an visual description of each group of data from the beginning to the end of classification, and selecting the attribute subset according to the trimmed probability value and a conventional random method;
s3) generating a corresponding decision tree by utilizing the constructed training set and the corresponding attribute subset, taking the information gain as a division principle, and splitting the node in the mode that the splitting mode when the information gain is maximum is the optimal splitting mode of the decision tree; and each tree grows completely without pruning;
s4) testing the test set by utilizing the decision tree to obtain corresponding categories, and taking the category with the most output in the decision tree as the category to which the test set belongs, thereby establishing the mapping relation between the characteristics of the local discharge signal and the cable local discharge fault type.
2. Cable office based on improved random forest algorithm according to claim 1The failure identification method is characterized in that the characteristic data in the step 2) comprises at least one of 24 high-dimensional attributes, the 24 high-dimensional attributes comprise asymmetry Asy and cross-correlation coefficient Cc of Hqn (j), Hqm (j) and Hn (j),H + qn (j)、H‾ qn (j)、H + qm (j)、H‾ qm (j)、H + n (j) AndH‾ n (j) Skewness Sk, steepness Ku, Peak Peak; wherein Hqn (j) represents an average discharge capacity phase distribution two-dimensional spectrogram, Hqm (j) represents a maximum discharge capacity phase distribution two-dimensional spectrogram, and Hn (j) represents a discharge frequency phase distribution spectrogram;H + qn (j) A two-dimensional spectrogram representing the phase distribution of the average discharge capacity of the positive half cycle,H‾ qn (j) A two-dimensional spectrogram of the phase distribution of the average discharge capacity of the negative half cycle is shown,H + qm (j) A two-dimensional spectrogram showing the phase distribution of the maximum discharge capacity of the positive half cycle,H‾ qm (j) A two-dimensional spectrogram representing the phase distribution of the maximum discharge capacity of the negative half cycle,H + n (j) A phase distribution spectrogram showing the positive half-cycle discharge times,H + n (j) And (3) showing a phase distribution spectrogram of the number of negative half-cycle discharges.
3. The cable partial discharge fault identification method based on the improved random forest algorithm is characterized in that the cable partial discharge fault types comprise internal discharge, surface discharge and corona discharge.
4. A cable defect identification system based on an improved random forest algorithm, comprising computer equipment, wherein the computer equipment is programmed or configured to execute the steps of the cable partial discharge fault identification method based on the improved random forest algorithm according to any one of claims 1 to 3.
5. A cable defect identification system based on an improved random forest algorithm, comprising a computer device, wherein a storage medium of the computer device stores a computer program programmed or configured to execute the cable partial discharge fault identification method based on the improved random forest algorithm according to any one of claims 1 to 3.
6. A cable defect identification system based on an improved random forest algorithm comprises partial discharge signal acquisition equipment and an upper computer which are connected with each other, and is characterized in that the upper computer is programmed or configured to execute the steps of the cable partial discharge fault identification method based on the improved random forest algorithm as claimed in any one of claims 1-3.
7. A cable defect identification system based on an improved random forest algorithm comprises partial discharge signal acquisition equipment and an upper computer which are connected with each other, and is characterized in that a storage medium of the upper computer stores a computer program which is programmed or configured to execute the cable partial discharge fault identification method based on the improved random forest algorithm according to any one of claims 1-3.
8. A computer-readable storage medium, characterized in that the computer-readable storage medium has stored thereon a computer program programmed or configured to execute the method for identifying partial discharge faults in cables based on an improved random forest algorithm according to any one of claims 1 to 3.
9. A cable defect identification system based on an improved random forest algorithm is characterized by comprising:
the signal acquisition program unit is used for acquiring partial discharge signals of the cable;
the characteristic extraction program unit is used for extracting characteristic data of the partial discharge signal;
the classification identification program unit is used for inputting the extracted feature data into a pre-constructed and trained classifier to obtain a cable partial discharge fault type corresponding to the cable partial discharge signal, the classifier is based on an improved random forest algorithm, the improved random forest algorithm utilizes a feature construction 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 pre-trained to establish a mapping relation between the features of the partial discharge signal and the cable partial discharge fault type;
the training step of the classifier comprises the following steps:
s1) respectively collecting cable partial discharge signals aiming at various cable partial discharge fault types and extracting characteristic data, and respectively selecting one part as a training set and the other part as a test set; resampling the training set by adopting a Bootstrap method, and randomly generating the training set, wherein the total number of attributes of the training set is M, the total number of attributes is the total number of judgment bases corresponding to the characteristic data respectively judged, and the number of sub-attributes is N;
s2) adopting a data set weight correction idea of an Adaboost algorithm to guide the selection of the attribute subset, and carrying out optimization processing on the attribute subset, wherein the attribute subset is a set of judgment bases corresponding to characteristic parameters used in a classification process; selecting an attribute subset Q on the basis of the training set, wherein the training set is a characteristic parameter used in the classification process, and the sub-attributes are judgment bases corresponding to the characteristic parameters respectively; directly trimming the selected probability of the attribute subset Q (i +1) through fault data in Ci of the previous sub-decision tree, wherein the sub-decision tree is an visual description of each group of data from the beginning to the end of classification, and selecting the attribute subset according to the trimmed probability value and a conventional random method;
s3) generating a corresponding decision tree by utilizing the constructed training set and the corresponding attribute subset, taking the information gain as a division principle, and splitting the node in the mode that the splitting mode when the information gain is maximum is the optimal splitting mode of the decision tree; and each tree grows completely without pruning;
s4) testing the test set by utilizing the decision tree to obtain corresponding categories, and taking the category with the most output in the decision tree as the category to which the test set belongs, thereby establishing the mapping relation between the characteristics of the local discharge signal and the cable local discharge fault type.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910440682.0A CN110108992B (en) | 2019-05-24 | 2019-05-24 | Cable partial discharge fault identification method and system based on improved random forest algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910440682.0A CN110108992B (en) | 2019-05-24 | 2019-05-24 | Cable partial discharge fault identification method and system based on improved random forest algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110108992A CN110108992A (en) | 2019-08-09 |
| CN110108992B true CN110108992B (en) | 2021-07-23 |
Family
ID=67492128
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910440682.0A Active CN110108992B (en) | 2019-05-24 | 2019-05-24 | Cable partial discharge fault identification method and system based on improved random forest algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110108992B (en) |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110796187A (en) * | 2019-10-22 | 2020-02-14 | 西安奕斯伟硅片技术有限公司 | Method and device for classifying defects |
| CN111366816B (en) * | 2020-04-26 | 2021-04-13 | 华北电力大学 | Power grid fault diagnosis method based on machine learning |
| CN111815027A (en) * | 2020-06-09 | 2020-10-23 | 山东大学 | Photovoltaic station generated power prediction method and system |
| CN111721834B (en) * | 2020-06-22 | 2023-08-08 | 南京南瑞继保电气有限公司 | Cable partial discharge on-line monitoring defect identification method |
| CN112308146A (en) * | 2020-11-02 | 2021-02-02 | 国网福建省电力有限公司 | Distribution transformer fault identification method based on operation characteristics |
| CN113128568B (en) * | 2021-03-26 | 2023-08-11 | 广州大学 | Excavator activity recognition method, system, device and storage medium |
| CN113283371A (en) * | 2021-06-09 | 2021-08-20 | 上海交通大学 | Local discharge feature extraction and classification method based on BRISK features |
| CN114167238A (en) * | 2021-12-03 | 2022-03-11 | 广东电网有限责任公司广州供电局 | Cable partial discharge detection method, device, system, equipment and storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101923650A (en) * | 2010-08-27 | 2010-12-22 | 北京大学 | Random forest classification method and classifiers based on comparison mode |
| CN102928720A (en) * | 2012-11-07 | 2013-02-13 | 广东电网公司 | Defect rate detecting method of oil immersed type main transformer |
| CN103226835A (en) * | 2013-04-15 | 2013-07-31 | 北京大学深圳研究生院 | Target tracking method and system based on on-line initialization gradient enhancement regression tree |
| CN104391970A (en) * | 2014-12-04 | 2015-03-04 | 深圳先进技术研究院 | Attribute subspace weighted random forest data processing method |
| CN105759177A (en) * | 2016-04-26 | 2016-07-13 | 浙江大学城市学院 | Classified-multi-mode-fusion-based distributed grid island detection method |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101964063B (en) * | 2010-09-14 | 2012-06-27 | 南京信息工程大学 | Method for constructing improved AdaBoost classifier |
| KR20140019950A (en) * | 2012-08-07 | 2014-02-18 | 성균관대학교산학협력단 | Method for generating 3d coordinate using finger image from mono camera in terminal and mobile terminal for generating 3d coordinate using finger image from mono camera |
| CN102937694A (en) * | 2012-11-27 | 2013-02-20 | 广州供电局有限公司 | Device for monitoring external insulation strength of dirty insulator |
| CN105844300A (en) * | 2016-03-24 | 2016-08-10 | 河南师范大学 | Optimized classification method and optimized classification device based on random forest algorithm |
| CN106778853A (en) * | 2016-12-07 | 2017-05-31 | 中南大学 | Unbalanced data sorting technique based on weight cluster and sub- sampling |
| CN107516108A (en) * | 2017-08-15 | 2017-12-26 | 国网四川省电力公司电力科学研究院 | Grader creation method and partial discharge of transformer method of fault pattern recognition |
| US10748040B2 (en) * | 2017-11-20 | 2020-08-18 | Kavya Venkata Kota Sai KOPPARAPU | System and method for automatic assessment of cancer |
-
2019
- 2019-05-24 CN CN201910440682.0A patent/CN110108992B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101923650A (en) * | 2010-08-27 | 2010-12-22 | 北京大学 | Random forest classification method and classifiers based on comparison mode |
| CN102928720A (en) * | 2012-11-07 | 2013-02-13 | 广东电网公司 | Defect rate detecting method of oil immersed type main transformer |
| CN103226835A (en) * | 2013-04-15 | 2013-07-31 | 北京大学深圳研究生院 | Target tracking method and system based on on-line initialization gradient enhancement regression tree |
| CN104391970A (en) * | 2014-12-04 | 2015-03-04 | 深圳先进技术研究院 | Attribute subspace weighted random forest data processing method |
| CN105759177A (en) * | 2016-04-26 | 2016-07-13 | 浙江大学城市学院 | Classified-multi-mode-fusion-based distributed grid island detection method |
Non-Patent Citations (2)
| Title |
|---|
| 不同气隙尺寸的油纸绝缘气隙放电特征及发展阶段识别;龙震泽;《中国优秀硕士学位论文全文数据库工程科技二辑》;20160615(第6期);正文第1章，第4章 * |
| 基于机器学习算法的寿命预测与故障诊断技术的发展综述;郭一帆 等;《计算机测量与控制》;20190331;第27卷(第3期);正文第7至12页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110108992A (en) | 2019-08-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110108992B (en) | Cable partial discharge fault identification method and system based on improved random forest algorithm | |
| CN107273920A (en) | A kind of non-intrusion type household electrical appliance recognition methods based on random forest | |
| CN103675610B (en) | Characterization factor extracting method in shelf depreciation on-line checking | |
| CN109100627A (en) | A kind of power equipment partial discharges fault diagnostic method based on end-to-end mode | |
| CN104198898A (en) | Local discharge development process diagnosis method based on pulse-train analysis | |
| CN114325256A (en) | Power equipment partial discharge identification method, system, equipment and storage medium | |
| CN115166563B (en) | Power battery aging state evaluation and retirement screening method and system | |
| CN108693448B (en) | Partial discharge mode recognition system applied to power equipment | |
| CN103558519A (en) | GIS partial discharge ultrasonic signal identification method | |
| CN102279358A (en) | MCSKPCA based neural network fault diagnosis method for analog circuits | |
| CN106874950A (en) | A kind of method for identifying and classifying of transient power quality recorder data | |
| CN116167010B (en) | Rapid identification method for abnormal events of power system with intelligent transfer learning capability | |
| CN110726898A (en) | Power distribution network fault type identification method | |
| CN107516108A (en) | Grader creation method and partial discharge of transformer method of fault pattern recognition | |
| CN110942098A (en) | Power supply service quality analysis method based on Bayesian pruning decision tree | |
| CN112528774B (en) | Intelligent unknown radar signal sorting system and method in complex electromagnetic environment | |
| CN111508517A (en) | Intelligent micro-motor product control method based on noise characteristics | |
| CN110580492A (en) | Track circuit fault precursor discovery method based on small fluctuation detection | |
| CN116401532A (en) | Method and system for recognizing frequency instability of power system after disturbance | |
| CN115128345A (en) | Power grid safety early warning method and system based on harmonic monitoring | |
| CN107133643A (en) | Note signal sorting technique based on multiple features fusion and feature selecting | |
| CN111898446A (en) | Single-phase earth fault studying and judging method based on multi-algorithm normalization analysis | |
| CN110543675A (en) | Power transmission line fault identification method | |
| CN116340812A (en) | Transformer partial discharge fault mode identification method and system | |
| CN115902557A (en) | Switch cabinet fault diagnosis processing method and device and nonvolatile storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |