A distribution network fault sample expansion method, device, terminal and storage medium

Abstract

The present application discloses a distribution network fault sample expansion method, device, terminal and storage medium. Through the method provided by the present application, based on the maximum value, minimum value and arithmetic mean of the first characteristic interval and the second characteristic interval, combined with the numerical random generation formula, a waveform sample is randomly generated by calculation to obtain the first expanded feature data and the second expanded feature data as the expanded sample. No experimental site support is required. Compared with the existing laboratory simulation expansion method, the processing difficulty is smaller and the sample expansion generation efficiency is higher. This solves the technical problems that the existing technology needs experimental site support, the waveform generation rate is slow, the fault conditions are limited, and the fault sample expansion generation efficiency is low.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a method, device, terminal and storage medium for expanding distribution network fault samples. Background Art

[0002] Due to natural causes such as strong winds, lightning strikes, and external force damage, high-resistance grounding faults may occur in distribution networks. The grounding medium is diverse, and the grounding resistance is generally from a few hundred ohms to a few thousand ohms. The characteristics of such faults are not obvious, and traditional fault protection measures and devices cannot effectively identify them. In recent years, machine learning methods have been gradually used to identify high-resistance grounding faults. Traditional learning tools such as artificial neural networks, extreme learning machines, and Bayesian networks are widely used to learn implicit knowledge of high-resistance grounding faults, mine fault characteristics, and generate discrimination rules.

[0003] The effect of machine learning methods depends on the size of the sample. At present, due to the small number of high-resistance grounding faults, there is less data on high-resistance grounding faults in the distribution network, and the original data is limited, sparse, and insufficient. At present, in view of the small number of distribution network grounding fault samples, a real laboratory method is generally used to generate recording waveforms by artificially simulating high-resistance grounding faults, or PSCAD, ETAP and other simulation software are used to establish a high-resistance grounding model to generate simulation waveforms to enrich high-resistance grounding fault samples. However, this method requires experimental site support, the waveform generation rate is slow, and the fault conditions are limited. There are technical problems such as low efficiency in fault sample expansion. Summary of the invention

[0004] The present application provides a distribution network fault sample expansion method, device, terminal and storage medium, which are used to solve the technical problems that the prior art requires experimental site support, has a slow waveform generation rate, has limited fault conditions, and has low efficiency in fault sample expansion generation.

[0005] The first aspect of the present application provides a method for expanding distribution network fault samples, comprising:

[0006] Acquire sample waveform data of the distribution network, wherein the sample waveform data includes: target fault sample waveform data and non-target fault sample waveform data;

[0007] Based on the target fault sample waveform data and the non-target fault sample waveform data, combined with a preset waveform feature type, a first feature quantity and a second feature quantity are obtained, wherein the first feature quantity is a feature quantity extracted from the target fault sample waveform data according to the waveform feature type, and the second feature quantity is a feature quantity extracted from the non-target fault sample waveform data according to the waveform feature type;

[0008] Based on a first feature quantity interval formed by the first feature quantity and a second feature quantity interval formed by the second feature quantity, when the first feature quantity interval and the second feature quantity interval do not overlap, based on the maximum value, minimum value and arithmetic mean of the first feature interval and the second feature interval, combined with a numerical random generation formula, first extended feature data and second extended feature data are obtained by calculation, wherein the first extended feature data is calculated based on the first feature interval and the numerical random generation formula, and the second extended feature data is calculated based on the second feature interval and the numerical random generation formula.

[0009] Preferably, the random value generation formula is specifically:

[0010]

[0011]

[0012] Where D s is the first extended feature data, F s is the second extended feature data, D is the first feature quantity, is the maximum value of the first feature value interval, D is the minimum value of the first feature value interval, is the arithmetic mean of the first feature value interval, is the maximum value of the second feature value interval, F is the minimum value of the second feature value interval, is the arithmetic mean of the second feature value interval, and u and v are random numbers between 0 and 1.

[0013] Preferably, it also includes:

[0014] When the first feature quantity interval overlaps with the second feature quantity interval, the maximum value and minimum value of the first feature quantity interval and the second feature quantity interval are updated based on the maximum value, minimum value and arithmetic mean of the first feature quantity interval and the second feature quantity interval, combined with a preset interval boundary adjustment equation, so that the first feature quantity interval and the second feature quantity interval do not overlap.

[0015] Preferably, the waveform feature type specifically includes: the mean and variance of the sample waveform data in time domain analysis, or the energy and coefficient after wavelet transformation.

[0016] A second aspect of the present application provides a distribution network fault sample expansion device, comprising:

[0017] A sample waveform data acquisition unit, used to acquire sample waveform data of the distribution network, wherein the sample waveform data includes: target fault sample waveform data and non-target fault sample waveform data;

[0018] A feature extraction unit, configured to obtain a first feature quantity and a second feature quantity based on the target fault sample waveform data and the non-target fault sample waveform data in combination with a preset waveform feature type, wherein the first feature quantity is a feature quantity extracted from the target fault sample waveform data according to the waveform feature type, and the second feature quantity is a feature quantity extracted from the non-target fault sample waveform data according to the waveform feature type;

[0019] A sample expansion processing unit is used to obtain first expanded feature data and second expanded feature data by calculation based on a first feature interval formed by the first feature and a second feature interval formed by the second feature, when the first feature interval does not overlap with the second feature interval, based on the maximum value, minimum value and arithmetic mean of the first feature interval and the second feature interval, combined with a numerical random generation formula, wherein the first expanded feature data is calculated based on the first feature interval and the numerical random generation formula, and the second expanded feature data is calculated based on the second feature interval and the numerical random generation formula.

[0020] Preferably, the random value generation formula is specifically:

[0021]

[0022]

[0023] Where D s is the first extended feature data, F s is the second extended feature data, D is the first feature quantity, is the maximum value of the first feature value interval, D is the minimum value of the first feature value interval, is the arithmetic mean of the first feature value interval, F is the second feature value, is the maximum value of the second feature value interval, F is the minimum value of the second feature value interval, is the arithmetic mean of the second feature value interval, and u and v are random numbers between 0 and 1.

[0024] Preferably, it also includes:

[0025] A feature quantity interval boundary processing unit is used to update the maximum value and minimum value of the first feature quantity interval and the second feature quantity interval based on the maximum value, minimum value and arithmetic mean of the first feature quantity interval and the second feature quantity interval in combination with a preset interval boundary adjustment equation when the first feature quantity interval overlaps with the second feature quantity interval, so that the first feature quantity interval and the second feature quantity interval do not overlap.

[0026] Preferably, the waveform feature type specifically includes: the mean and variance of the sample waveform data in time domain analysis, or the energy and coefficient after wavelet transformation.

[0027] A third aspect of the present application provides a distribution network fault sample expansion terminal, including: a memory and a processor;

[0028] The memory is used to store program codes, wherein the program codes correspond to the method for expanding distribution network fault samples provided in the first aspect of the present application;

[0029] The processor is configured to execute the program code.

[0030] A fourth aspect of the present application provides a computer-readable storage medium, in which a program code corresponding to the distribution network fault sample expansion method provided in the first aspect of the present application is stored.

[0031] It can be seen from the above technical solutions that this application has the following advantages:

[0032] Through the method provided in the present application, based on the maximum value, minimum value and arithmetic mean of the first characteristic interval and the second characteristic interval, combined with the numerical random generation formula, the waveform sample is randomly generated by calculation to obtain the first extended feature data and the second extended feature data as the extended sample. No experimental site support is required, and the processing difficulty is smaller than the existing laboratory simulation expansion method, and the sample expansion generation efficiency is higher. This solves the technical problems that the existing technology requires experimental site support, the waveform generation rate is slow, the fault conditions are limited, and the fault sample expansion generation efficiency is low. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0034] Figure 1A schematic flow chart of a first embodiment of a method for expanding fault samples in a distribution network provided in the present application.

[0035] Figure 2 A schematic flow chart of a second embodiment of a method for expanding fault samples in a distribution network provided in the present application.

[0036] Figure 3 A schematic structural diagram of an embodiment of a distribution network fault sample expansion device provided in the present application. DETAILED DESCRIPTION

[0037] The embodiments of the present application provide a distribution network fault sample expansion method, device, terminal and storage medium, which are used to solve the technical problems that the prior art requires experimental site support, has a slow waveform generation rate, has limited fault conditions, and has low efficiency in fault sample expansion generation.

[0038] In order to make the purpose, features, and advantages of the invention of this application more obvious and easy to understand, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the embodiments described below are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0039] See also Figure 1 The first embodiment of the present application provides a method for expanding distribution network fault samples, comprising:

[0040] Step 101 : Acquire sample waveform data of a distribution network, where the sample waveform data includes: target fault sample waveform data and non-target fault sample waveform data.

[0041] It should be noted that, first, sample waveform data of the distribution network is obtained. This sample waveform data can be obtained from the distribution network fault waveform library, including: target fault sample waveform data and non-target fault sample waveform data. The non-target fault sample waveform data refers to waveform sample data of other fault types and normal waveform sample data.

[0042] Step 102 : Based on the target fault sample waveform data and the non-target fault sample waveform data and in combination with a preset waveform feature type, a first feature quantity and a second feature quantity are obtained.

[0043] According to the target fault sample waveform data and the non-target fault sample waveform data obtained in step 101, feature extraction is performed on the two types of sample waveform data respectively, so as to obtain a first feature value and a second feature value.

[0044] The first feature quantity is a feature quantity extracted from target fault sample waveform data according to the waveform feature type, and the second feature quantity is a feature quantity extracted from non-target fault sample waveform data according to the waveform feature type.

[0045] Step 103: Based on the first feature quantity interval formed by the first feature quantity and the second feature quantity interval formed by the second feature quantity, when the first feature quantity interval and the second feature quantity interval do not overlap, based on the maximum value, minimum value and arithmetic mean of the first feature interval and the second feature interval, combined with a numerical random generation formula, the first extended feature data and the second extended feature data are obtained by calculation.

[0046] The first extended feature data is calculated based on the first feature interval and a random numerical generation formula, and the second extended feature data is calculated based on the second feature interval and a random numerical generation formula.

[0047] It should be noted that a first feature quantity interval is formed based on the maximum and minimum values ​​of the first feature quantity, and a second feature quantity interval is formed based on the maximum and minimum values ​​of the second feature quantity. When the first feature quantity interval does not overlap with the second feature quantity interval, the first extended feature data and the second extended feature data are obtained by calculation based on the maximum and minimum values ​​and the arithmetic mean of the first and second feature quantities in combination with a numerical random generation formula. The calculated extended feature data are used as extended samples and can be used for neural network training. Compared with the existing laboratory simulation expansion method, the processing difficulty is smaller and the sample expansion generation efficiency is higher. This solves the technical problems that the existing technology requires experimental site support, the waveform generation rate is slow, the fault conditions are limited, and the fault sample expansion generation efficiency is low.

[0048] The above is a detailed description of a first embodiment of a distribution network fault sample expansion method provided by the present application. The following is a detailed description of a second embodiment of a distribution network fault sample expansion method provided by the present application.

[0049] See also Figure 2 The second embodiment of the present application is a method for expanding distribution network fault samples based on the first embodiment, including:

[0050] More specifically, the random value generation formula mentioned in step 103 is specifically expressed as follows:

[0051]

[0052]

[0053] Where D s is the first extended feature data, F sis the second extended feature data, D is the first feature quantity, is the maximum value of the first feature value interval, D is the minimum value of the first feature value interval, is the arithmetic mean of the first feature value interval, is the maximum value of the second feature value interval, F is the minimum value of the second feature value interval, is the arithmetic mean of the second feature value interval, and u and v are random numbers between 0 and 1.

[0054] Assuming that the expected number of samples is n, and the number of existing samples is S, at this time, only nS steps are needed to randomly generate new expanded samples, and the generated samples are mixed with the original samples to form an expanded new sample library.

[0055] In some possible embodiments, it further includes:

[0056] Step 100: When the first feature quantity interval overlaps with the second feature quantity interval, based on the maximum value, minimum value and arithmetic mean of the first feature quantity interval and the second feature quantity interval, combined with a preset interval boundary adjustment equation, update the maximum value and minimum value of the first feature quantity interval and the second feature quantity interval so that the first feature quantity interval and the second feature quantity interval do not overlap.

[0057] It should be noted that when the first feature value interval overlaps with the second feature value interval, the feature boundary distances in the fault type D and the interference type F can be increased according to different overlapping situations to expand the feature distinction between the two types of fault types. The specific methods include:

[0058] 1) Condition 1: Next, if If established, D =DDC q otherwise

[0059] 2) Condition 2: Next, if If established, D =DDC q otherwise

[0060] 3) Condition 3: Next, if If established, F =FDC q otherwise

[0061] 4) Condition 4: Next, if If established, F=FDC q otherwise

[0062] 5) Condition 5: or Next, judge Is it true? If so, increase the maximum value of the sample otherwise

[0063] 6) Condition 6: or Next, judge D ≤ F Is it true? If so, reduce the maximum value of the sample otherwise

[0064] Among them, DDC q The first element of the first characteristic quantity in descending order, DAC q The first element of the first characteristic quantity in ascending order, FDC q The first element of the second feature value in descending order, FAC q The first element of the second feature value in ascending order.

[0065] More specifically, the waveform feature types specifically include: mean and variance of sample waveform data in time domain analysis, or energy and coefficients after wavelet transformation.

[0066] The above is a detailed description of a second embodiment of a distribution network fault sample expansion method provided by the present application. The following is a detailed description of an embodiment of a distribution network fault sample expansion device provided by the present application.

[0067] See also Figure 3 The third embodiment of the present application provides a distribution network fault sample expansion device, comprising:

[0068] The sample waveform data acquisition unit 201 is used to acquire sample waveform data of the distribution network, the sample waveform data including: target fault sample waveform data and non-target fault sample waveform data;

[0069] A feature extraction unit 202 is used to obtain a first feature quantity and a second feature quantity based on the target fault sample waveform data and the non-target fault sample waveform data in combination with a preset waveform feature type, wherein the first feature quantity is a feature quantity extracted from the target fault sample waveform data according to the waveform feature type, and the second feature quantity is a feature quantity extracted from the non-target fault sample waveform data according to the waveform feature type;

[0070] The sample expansion processing unit 203 is used to obtain the first expanded feature data and the second expanded feature data by calculation based on the maximum value, minimum value and arithmetic mean of the first feature interval and the second feature interval, combined with the numerical random generation formula, based on the first feature interval formed by the first feature quantity and the second feature interval formed by the second feature quantity when the first feature quantity interval and the second feature quantity interval do not overlap, wherein the first expanded feature data is calculated based on the first feature interval and the numerical random generation formula, and the second expanded feature data is calculated based on the second feature interval and the numerical random generation formula.

[0071] Furthermore, the numerical random generation formula is specifically as follows:

[0072]

[0073]

[0074] Where D s is the first extended feature data, F s is the second extended feature data, D is the first feature quantity, is the maximum value of the first feature value interval, D is the minimum value of the first feature value interval, is the arithmetic mean of the first feature value interval, F is the second feature value, is the maximum value of the second feature value interval, F is the minimum value of the second feature value interval, is the arithmetic mean of the second feature value interval, and u and v are random numbers between 0 and 1.

[0075] Furthermore, it also includes:

[0076] The feature quantity interval boundary processing unit 200 is used to update the maximum value and minimum value of the first feature quantity interval and the second feature quantity interval based on the maximum value, minimum value and arithmetic mean of the first feature quantity interval and the second feature quantity interval in combination with a preset interval boundary adjustment equation when there is overlap between the first feature quantity interval and the second feature quantity interval, so that the first feature quantity interval and the second feature quantity interval do not overlap.

[0077] Furthermore, the waveform feature types specifically include: mean and variance of sample waveform data in time domain analysis, or energy and coefficients after wavelet transformation.

[0078] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the terminals, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0079] In the several embodiments provided in the present application, it should be understood that the disclosed terminals, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0080] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can, for example, be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0081] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0082] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0083] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0084] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for expanding distribution network fault samples, characterized in that: include: Acquire sample waveform data of the distribution network, wherein the sample waveform data includes: target fault sample waveform data and non-target fault sample waveform data; Based on the target fault sample waveform data and the non-target fault sample waveform data, combined with a preset waveform feature type, a first feature quantity and a second feature quantity are obtained, wherein the first feature quantity is a feature quantity extracted from the target fault sample waveform data according to the waveform feature type, and the second feature quantity is a feature quantity extracted from the non-target fault sample waveform data according to the waveform feature type; Based on a first feature quantity interval formed by the first feature quantity and a second feature quantity interval formed by the second feature quantity, when the first feature quantity interval and the second feature quantity interval do not overlap, based on the maximum value, the minimum value and the arithmetic mean of the first feature quantity interval and the second feature quantity interval, combined with a numerical random generation formula, first extended feature data and second extended feature data are obtained by calculation, wherein the first extended feature data is calculated based on the first feature quantity interval and the numerical random generation formula, and the second extended feature data is calculated based on the second feature quantity interval and the numerical random generation formula; The random value generation formula is specifically: ； In the formula, is the first extended feature data, is the second extended feature data, D is the first feature quantity, is the maximum value of the first feature value interval, is the minimum value of the first feature value interval, is the arithmetic mean of the first feature value interval, is the maximum value of the second feature value interval, is the minimum value of the second feature value interval, is the arithmetic mean of the second feature value interval, and A random number in [0,1].

2. A distribution network fault sample expansion method according to claim 1, characterized in that: Also includes: When the first feature quantity interval overlaps with the second feature quantity interval, the maximum value and minimum value of the first feature quantity interval and the second feature quantity interval are updated based on the maximum value, minimum value and arithmetic mean of the first feature quantity interval and the second feature quantity interval, combined with a preset interval boundary adjustment equation, so that the first feature quantity interval and the second feature quantity interval do not overlap.

3. A distribution network fault sample expansion method according to claim 1, characterized in that: The waveform feature types specifically include: the mean and variance of the sample waveform data in the time domain analysis, or the energy and coefficient after wavelet transformation.

4. A distribution network fault sample expansion device, characterized in that: include: A sample waveform data acquisition unit, used to acquire sample waveform data of the distribution network, wherein the sample waveform data includes: target fault sample waveform data and non-target fault sample waveform data; A feature extraction unit, configured to obtain a first feature quantity and a second feature quantity based on the target fault sample waveform data and the non-target fault sample waveform data in combination with a preset waveform feature type, wherein the first feature quantity is a feature quantity extracted from the target fault sample waveform data according to the waveform feature type, and the second feature quantity is a feature quantity extracted from the non-target fault sample waveform data according to the waveform feature type; A sample expansion processing unit, configured to obtain first expanded feature data and second expanded feature data by calculation based on a first feature interval formed by the first feature and a second feature interval formed by the second feature, when the first feature interval does not overlap with the second feature interval, based on a maximum value, a minimum value and an arithmetic mean value of the first feature interval and the second feature interval, in combination with a numerical random generation formula, wherein the first expanded feature data is calculated based on the first feature interval and the numerical random generation formula, and the second expanded feature data is calculated based on the second feature interval and the numerical random generation formula; The random value generation formula is specifically: ； In the formula, is the first extended feature data, is the second extended feature data, D is the first feature quantity, is the maximum value of the first feature value interval, is the minimum value of the first feature value interval, is the arithmetic mean of the first feature value interval, is the maximum value of the second feature value interval, is the minimum value of the second feature value interval, is the arithmetic mean of the second feature value interval, and A random number in [0,1].

5. A distribution network fault sample expansion device according to claim 4, characterized in that: Also includes: A feature quantity interval boundary processing unit is used to update the maximum value and minimum value of the first feature quantity interval and the second feature quantity interval based on the maximum value, minimum value and arithmetic mean of the first feature quantity interval and the second feature quantity interval in combination with a preset interval boundary adjustment equation when the first feature quantity interval overlaps with the second feature quantity interval, so that the first feature quantity interval and the second feature quantity interval do not overlap.

6. A distribution network fault sample expansion device according to claim 4, characterized in that: The waveform feature types specifically include: the mean and variance of the sample waveform data in the time domain analysis, or the energy and coefficient after wavelet transformation.

7. A distribution network fault sample expansion terminal, characterized in that: include: Memory and processor; The memory is used to store program codes, wherein the program codes are used to implement the method for expanding distribution network fault samples according to any one of claims 1 to 3; The processor is configured to execute the program code.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores program code for implementing the method for expanding distribution network fault samples according to any one of claims 1 to 3.

Images