Switchgear insulation defect type detection method, device, equipment and program product

Abstract

The present application relates to a method, apparatus, device, and program product for detecting insulation defect types in switchgear. The method comprises: obtaining detection data of a target gas within a target switchgear, the detection data including the concentration of the target gas, the generation rate of the target gas, and the temperature of the target switchgear; inputting the detection data into a target support vector machine to obtain a detection result of the target support vector machine; the detection result is the insulation defect type of the target switchgear; wherein the kernel function parameters and penalty parameters of the target support vector machine are derived based on a cross-validation algorithm. This method enables insulation defect type detection of switchgear while being less susceptible to interference from external environmental factors and improving detection accuracy.

Description

Technical Field

[0001] The present application relates to the technical field of electrical equipment detection, and in particular to a method, device, equipment and program product for detecting insulation defect types in switch cabinets. Background Art

[0002] Switchgear, as a crucial power distribution device, plays a vital role in power systems. However, long-term operation in harsh environments such as high temperature, high voltage, and humidity can easily lead to failures due to insulation degradation. A switchgear failure can not only damage other critical equipment on the line but also potentially cause a power outage. Therefore, it's essential to analyze the types of switchgear insulation defects so that maintenance can be tailored to each defect type to prevent major accidents.

[0003] In the prior art, when a switch cabinet fault occurs, ultrasonic signals, current and other data of each module of the switch cabinet are detected to determine the type of insulation defect causing the switch cabinet fault based on the detection results.

[0004] However, in the prior art, detection methods based on ultrasonic signals, current and other data are easily interfered with by external environmental factors and have low detection accuracy. Summary of the Invention

[0005] Based on this, it is necessary to provide a switch cabinet insulation defect type detection method, device, equipment and program product that is not easily interfered by external environmental factors and has improved detection accuracy to address the above technical problems.

[0006] In a first aspect, the present application provides a method for detecting insulation defect types in a switch cabinet, the method comprising:

[0007] Acquire detection data of a target gas in a target switchgear, the detection data including the concentration of the target gas, the generation rate of the target gas, and the temperature of the target switchgear; input the detection data into a target support vector machine to obtain a detection result of the target support vector machine; the detection result is the insulation defect type of the target switchgear; wherein the kernel function parameters and penalty parameters of the target support vector machine are derived based on a cross-validation algorithm.

[0008] In one embodiment, the process of determining the kernel function parameters and penalty parameters of the target support vector machine includes: constructing K initial support vector machines; obtaining multiple parameter pairs, each parameter pair including different candidate kernel function parameters and candidate penalty parameters; for each parameter pair, assigning the parameter pair to the K initial vector machines respectively, and determining the insulation defect classification quality of the K support vector machines after the assignment; obtaining the target parameter pair corresponding to the highest insulation defect classification quality, and using the candidate kernel function parameters and candidate penalty parameters included in the target parameter pair as the kernel function parameters and penalty parameters of the target support vector machine respectively.

[0009] In one embodiment, constructing K initial support vector machines includes: obtaining experimental data, the experimental data including the concentration of the target gas in the test switch cabinet under different insulation defect types, the production rate of the target gas in the test switch cabinet, and the temperature of the test switch cabinet; dividing the experimental data into K experimental data sets; performing K model training processes based on the K experimental data sets to obtain K initial support vector machines; wherein the model training process includes: selecting K-1 experimental data sets from the K experimental data sets as training sets, and using the unselected experimental data sets as test sets; performing model training based on the training set and the test set to obtain the initial support vector machine.

[0010] In one embodiment, determining the insulation defect classification quality of the K assigned support vector machines includes: for each assigned support vector machine, obtaining a target test set corresponding to the assigned support vector machine, and using the target test set as the input of the assigned support vector machine to obtain the output of the assigned support vector machine; obtaining the mean square error value of the output of the K assigned support vector machines; and determining the insulation defect classification quality of the K assigned support vector machines based on the mean square error value.

[0011] In one embodiment, obtaining the target parameter pair corresponding to the highest insulation defect classification quality includes: obtaining the target parameter pair corresponding to the K support vector machines after the assignment with the minimum mean square error.

[0012] In one embodiment, the method further includes: after determining the kernel function parameters and penalty parameters of the target support vector machine based on a cross-validation algorithm, training using the experimental data based on the determined kernel function parameters and penalty parameters to obtain the target support vector machine.

[0013] In a second aspect, the present application also provides a switch cabinet insulation defect type detection device. The device includes:

[0014] an acquisition module, configured to acquire detection data of a target gas in a target switch cabinet, the detection data including a concentration of the target gas, a generation rate of the target gas, and a temperature of the target switch cabinet;

[0015] A detection module is used to input the detection data into a target support vector machine to obtain a detection result of the target support vector machine; the detection result is the insulation defect type of the target switchgear; wherein the kernel function parameters and penalty parameters of the target support vector machine are derived based on a cross-validation algorithm.

[0016] In one embodiment, the apparatus further comprises:

[0017] Construction module, used to construct K initial support vector machines;

[0018] A second acquisition module is used to acquire multiple parameter pairs, each of which includes different candidate kernel function parameters and candidate penalty parameters;

[0019] A determination module is used to assign the parameter pair to the K initial vector machines for each parameter pair, and determine the insulation defect classification quality of the K support vector machines after the assignment;

[0020] The third acquisition module is used to obtain the target parameter pair corresponding to the highest insulation defect classification quality, and use the candidate kernel function parameters and candidate penalty parameters included in the target parameter pair as the kernel function parameters and penalty parameters of the target support vector machine respectively.

[0021] In one embodiment, the construction module is specifically used to: obtain experimental data, which includes the concentration of the target gas in the test switch cabinet under different insulation defect types, the production rate of the target gas in the test switch cabinet, and the temperature of the test switch cabinet; divide the experimental data into K experimental data sets; based on the K experimental data sets, perform K model training processes to obtain K initial support vector machines; wherein, the model training process includes: selecting K-1 experimental data sets from the K experimental data sets as training sets, and using the unselected experimental data sets as test sets; performing model training based on the training set and the test set to obtain the initial support vector machine.

[0022] In one embodiment, the determination module is specifically used to: for each assigned support vector machine, obtain a target test set corresponding to the assigned support vector machine, and use the target test set as the input of the assigned support vector machine to obtain the output of the assigned support vector machine; obtain the mean square error value of the outputs of the K assigned support vector machines; and determine the insulation defect classification quality of the K assigned support vector machines based on the mean square error value.

[0023] In one embodiment, the third acquisition module is specifically used to obtain the target parameter pair corresponding to the K assigned support vector machines with the smallest mean square error.

[0024] In one embodiment, the apparatus further comprises:

[0025] The training module is used to determine the kernel function parameters and penalty parameters of the target support vector machine based on the cross-validation algorithm, and then use the experimental data to perform training based on the determined kernel function parameters and penalty parameters to obtain the target support vector machine.

[0026] In a third aspect, the present application further provides a computer device. The computer device includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0027] Acquire detection data of a target gas in a target switchgear, the detection data including the concentration of the target gas, the generation rate of the target gas, and the temperature of the target switchgear; input the detection data into a target support vector machine to obtain a detection result of the target support vector machine; the detection result is the insulation defect type of the target switchgear; wherein the kernel function parameters and penalty parameters of the target support vector machine are derived based on a cross-validation algorithm.

[0028] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the following steps:

[0029] Acquire detection data of a target gas in a target switchgear, the detection data including the concentration of the target gas, the generation rate of the target gas, and the temperature of the target switchgear; input the detection data into a target support vector machine to obtain a detection result of the target support vector machine; the detection result is the insulation defect type of the target switchgear; wherein the kernel function parameters and penalty parameters of the target support vector machine are derived based on a cross-validation algorithm.

[0030] In a fifth aspect, the present application further provides a computer program product. The computer program product includes a computer program that, when executed by a processor, implements the following steps:

[0031] Acquire detection data of a target gas in a target switchgear, the detection data including the concentration of the target gas, the generation rate of the target gas, and the temperature of the target switchgear; input the detection data into a target support vector machine to obtain a detection result of the target support vector machine; the detection result is the insulation defect type of the target switchgear; wherein the kernel function parameters and penalty parameters of the target support vector machine are derived based on a cross-validation algorithm.

[0032] The above-mentioned switchgear insulation defect type detection method, device, equipment and program product obtain the detection data of the target gas in the target switchgear and input the detection data into the target support vector machine to obtain the detection result, which is the insulation defect type of the target switchgear; that is, the target support vector machine detects the insulation defect type of the target switchgear according to the gas decomposition characteristics, and is therefore not easily affected by external factors; among them, since the kernel function parameters and penalty parameters of the target support vector machine are derived based on the cross-validation algorithm, the detection accuracy of the target support vector machine is greatly improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 This is a diagram of an application environment of a method for detecting insulation defect types in a switch cabinet according to an embodiment;

[0034] Figure 2 1 is a flow chart of a method for detecting insulation defect types in a switch cabinet according to an embodiment;

[0035] Figure 3 A schematic diagram of a process for determining kernel function parameters and penalty parameters in one embodiment;

[0036] Figure 4 A schematic diagram of a process for constructing K initial support vector machines in one embodiment;

[0037] Figure 5 A schematic diagram of a process for determining insulation defect classification quality in one embodiment;

[0038] Figure 6 A schematic diagram of a process for determining a target support vector machine and performing insulation defect type detection in one embodiment;

[0039] Figure 7 This is a structural block diagram of a switch cabinet insulation defect type detection device in one embodiment;

[0040] Figure 8 is a structural block diagram of a second switch cabinet insulation defect type detection device in one embodiment;

[0041] Figure 9 is a structural block diagram of a third switch cabinet insulation defect type detection device in one embodiment;

[0042] Figure 10 FIG. 1 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0044] Switchgear is a critical electrical equipment used for switching, controlling, and protecting power generation, transmission, distribution, and conversion in power systems. Its safe operation directly impacts the reliability of power supply. However, various insulation defects within switchgear are inevitable during the manufacturing, distribution, installation, operation, and maintenance processes. These defects, under continuous operating voltage, can cause partial discharge, leading to the decomposition of air within the switchgear, producing various gaseous decomposition products. Damage to solid insulation materials can also produce gases such as CO and CO₂. Most of these products are harmful to humans and the environment. Acidic gases, for example, react with water vapor to produce acids that corrode metal equipment and insulating components, further damaging the insulation performance within the switchgear. If these defects are left unchecked and unchecked, they will inevitably lead to switchgear insulation failures, potentially damaging other critical equipment along the line and potentially causing power outages, resulting in significant economic losses for the enterprise. As a crucial power distribution device, switchgear plays a vital role in opening and closing power lines, transmitting and switching operating loads within the power system. However, switchgear, operating in harsh environments such as high temperature, high pressure, and humidity for extended periods, can easily experience insulation degradation, leading to failures. A switchgear failure can damage other critical equipment on the line and potentially cause a power outage. Therefore, it's essential to analyze the types of insulation defects in switchgear so that maintenance can be tailored to each defect type to prevent major accidents.

[0045] In the prior art, when a switch cabinet fault occurs, ultrasonic signals, current and other data of each module of the switch cabinet are detected to determine the type of insulation defect that causes the switch cabinet fault.

[0046] However, in the prior art, detection methods based on ultrasonic signals, current and other data are easily interfered with by external environmental factors and have low detection accuracy.

[0047] The switch cabinet insulation defect type detection method provided in the embodiment of the present application can be applied to Figure 1In the application environment shown, the electrochemical sensor 101 and the high-precision temperature sensor 102 are both connected to a computer device 103. The electrochemical sensor 101 transmits the collected concentration data of each gas component in the target switchgear to the computer device 103, and the high-precision temperature sensor 102 transmits the collected temperature data in the target switchgear to the computer device 103. The computer device 103 performs subsequent processing based on the received concentration data of each gas component and temperature data to determine the degree of degradation of the target switchgear. The computer device 103 can be, but is not limited to, various personal computers, laptops, smartphones, tablet computers, servers, or a server cluster consisting of multiple servers.

[0048] In one embodiment, Figure 2 As shown, a method for detecting insulation defect types in a switch cabinet is provided. Figure 1 Taking the computer device 103 in the example as an example, the method includes the following steps:

[0049] Step 201 : Acquire detection data of a target gas in a target switch cabinet, where the detection data includes the concentration of the target gas, the generation rate of the target gas, and the temperature of the target switch cabinet.

[0050] The switch cabinet that needs to be tested for insulation defect type is taken as the target switch cabinet, and the computer device 103 obtains the test data of the target switch cabinet. According to the test data, the insulation defect type of the target switch cabinet is detected. The test data includes the concentration of the target gas, the generation rate of the target gas and the temperature of the target switch cabinet. When partial discharge occurs in the air of the target switch cabinet, ozone, nitric oxide, nitrogen dioxide, nitrogen pentoxide and other gas decomposition products are generated, wherein nitric oxide, nitrogen pentoxide and other gas decomposition products are reaction intermediates, and nitrogen dioxide and other gases are reaction final products. Moreover, when the target switch cabinet overheats, there will be a significant temperature rise, and then gases such as CO will slowly be generated in the air. Therefore, the volume nitrogen dioxide in the target switch cabinet is generated. , nitric oxide and carbon monoxide are used as target gases, that is, as characteristic parameters for judging the type of insulation defects in the target switchgear; correspondingly, the concentration of the target gas is the gas concentration of nitrogen dioxide, nitric oxide and carbon monoxide in the target switchgear; the generation rate of the target gas refers to the rate at which nitrogen dioxide, nitric oxide and carbon monoxide gases are generated in the target switchgear, and the generation rate can be obtained by dividing the net increase in the concentration of each target gas by the detection time, wherein the detection time can be a fixed time period set manually, and the net increase in the concentration of each target gas is the difference in concentration between the start and end of gas collection; since the target switchgear will have a significant temperature rise when it overheats, the temperature of the target switchgear is also used as one of the detection data, and the temperature of the target switchgear can be collected by a high-precision temperature sensor.

[0051] Step 202: input the detection data into a target support vector machine to obtain a detection result of the target support vector machine; the detection result is the insulation defect type of the target switchgear; wherein the kernel function parameters and penalty parameters of the target support vector machine are obtained based on a cross-validation algorithm.

[0052] Among them, the support vector machine (SVM) is a type of generalized linear classifier that performs binary classification on data in a supervised learning manner. The target support vector machine is a trained support vector machine that can output the type of insulation defect of the switch cabinet based on the input detection data. Among them, the two important parameters that affect the classification accuracy of the target support vector machine are the kernel function parameter and the penalty parameter. The basic idea of ​​the cross-validation algorithm is to group the original data, one part as the training set and the other as the validation set. The classifier is first trained with the training set, and then the trained model is tested with the validation set, which is used as the performance indicator for evaluating the classifier. Since the support vector machine is a classifier, the cross-validation algorithm can be used to determine the optimal kernel function parameters and penalty parameters, and apply them to the target support vector to improve the detection accuracy of the target support vector machine and realize the detection of the insulation defect type of the target switch cabinet.

[0053] In the above-mentioned switchgear insulation defect type detection method, the detection data of the target gas in the target switchgear is obtained, and the detection data is input into the target support vector machine to obtain the detection result, which is the insulation defect type of the target switchgear; that is, the target support vector machine detects the insulation defect type of the target switchgear according to the gas decomposition characteristics, and is therefore not easily affected by external factors; wherein, since the kernel function parameters and penalty parameters of the target support vector machine are derived based on the cross-validation algorithm, the detection accuracy of the target support vector machine is greatly improved.

[0054] In one embodiment, Figure 3 As shown, it shows a flow chart of determining kernel function parameters and penalty parameters provided by an embodiment of the present application; the process of determining the kernel function parameters and penalty parameters of the target support vector machine includes:

[0055] Step 301, construct K initial support vector machines;

[0056] The initial vector machine refers to the constructed candidate support vector machines. From the initial vector machines, a support vector machine with the most accurate judgment result can be selected as the target support vector machine for detecting the insulation defect type of the target switchgear.

[0057] Please refer to Figure 4, which shows a schematic diagram of a process for constructing K initial support vector machines provided in an embodiment of the present application; the process for constructing K initial support vector machines includes:

[0058] Step 401 : Acquire experimental data, which includes the concentration of the target gas in the test switch cabinet under different insulation defect types, the production rate of the target gas in the test switch cabinet, and the temperature of the test switch cabinet.

[0059] Before determining the target support vector machine, it is first necessary to obtain a large amount of experimental data related to insulation defect types. Based on the experimental data, a target support vector machine is obtained for subsequent detection of insulation defect types of the target switchgear; the switchgear used to obtain the experimental data is used as a test switchgear. The switchgear can have different insulation defect types. Therefore, different insulation defect types can be manufactured for the test switchgear to obtain experimental data, wherein the insulation defect types include solid metal protrusion, overheating fault, insulator air gap, metal contamination on the insulator surface and free metal particle defects.

[0060] Exemplarily, the process of obtaining experimental data can be as follows: under different types of insulation defects, the degree of insulation degradation of the test switch cabinet is continuously changed. For example, under solid metal protrusion defects, different degrees of insulation degradation can be set for the test switch cabinet by changing the size of the solid metal protrusion. The discharge amount and voltage value of the test switch cabinet are gradually changed until obvious partial discharge or overheating failure is observed in the test switch cabinet. The experiment is stopped, and the time when partial discharge occurs in the test switch cabinet under different degrees of solid metal protrusion defects is obtained, and the concentrations of carbon monoxide, nitric oxide and nitrogen dioxide gases in the test switch cabinet and the temperature of the test switch cabinet are collected; the insulation defect type of the test switch cabinet is changed, and the above experiment is repeated to obtain the concentrations of carbon monoxide, nitric oxide and nitrogen dioxide gases under different insulation defect types. The concentration of the gas is obtained by dividing the net increase in the concentration of each gas by the time when the partial discharge occurs, and the generation rate of carbon monoxide, nitric oxide and nitrogen dioxide gases is obtained. The net increase in the concentration of each gas is the concentration difference between the start and end of the gas collection. Thus, the concentration of the target gas in the test switch cabinet, the production rate of the target gas in the test switch cabinet and the temperature of the test switch cabinet are obtained. The obtained data are discrete data, and each data is normalized and stored in the database. It should be noted that this application does not limit the process of obtaining the insulation degradation test data, as long as the experimental method can obtain the insulation degradation test data; optionally, the normalization process can use the min-max normalization formula to convert each data into a value between [0,1]. The min-max normalization formula is as follows:

[0061]

[0062] In the above formula, is the value of the normalized experimental data, x i is the experimental data value before normalization, x min is the minimum value of this type of experimental data, for example, normalizing the carbon monoxide concentration, x min is the minimum value in the carbon monoxide concentration data; x max is the maximum value of this type of experimental data, for example, normalizing the carbon monoxide concentration, x max is the maximum value in the carbon monoxide concentration data.

[0063] Step 402: Divide the experimental data into K experimental data sets.

[0064] Among them, the experimental data stored in the database can be divided, so as to perform cross-validation based on the divided data; the experimental data is divided into K experimental data sets, each experimental data set is divided into the concentrations and generation rates of carbon monoxide, nitric oxide and nitrogen dioxide gases of the test switch cabinet obtained under different insulation defect types, as well as the temperature of the test switch cabinet; thereby obtaining K experimental data sets, wherein K is a positive integer greater than zero, and the value of K can be set according to actual needs.

[0065] Step 403: Based on the K experimental data sets, perform K model training processes to obtain K initial support vector machines; wherein, the model training process includes: selecting K-1 experimental data sets from the K experimental data sets as training sets, and using the unselected experimental data sets as test sets; performing model training based on the training sets and the test sets to obtain the initial support vector machine.

[0066] Among them, according to the obtained K experimental data sets, a cross-validation algorithm is used to train the support vector machine to obtain K initial support vector machines; the specific training process can be: select an experimental data set from the K experimental data sets as a test set, and use the remaining K-1 experimental data sets as training sets, input the training set and the test set into the support vector machine, and train the support vector machine to perform data classification to obtain an initial support vector machine; each experimental data set is used as a test set, and the remaining K-1 experimental data sets are used as training sets, and are input into the support vector machine for data classification training, and a total of K training times are performed to obtain K initial support vector machines.

[0067] Step 302: Acquire multiple parameter pairs, each of which includes different candidate kernel function parameters and candidate penalty parameters.

[0068] Among them, the traversal algorithm is used to perform multiple traversal searches on the predetermined preset support vector machine within a certain range. Since the kernel function parameters and penalty parameters generated by the support vector machine are different each time it performs classification, each traversal search obtains a kernel function parameter and penalty parameter, which are used as candidate kernel function parameters and candidate penalty parameters to form a parameter pair; multiple traversal searches obtain multiple parameter pairs, each parameter pair includes different candidate kernel function parameters and candidate penalty parameters, and the traversal search range can be 2 -10 -2 10 Specifically, the decision function of the preset support vector machine is:

[0069]

[0070] In the above formula, ω is the weight vector in the preset support vector machine, b is the bias value in the preset support vector machine, is the kernel function, m is the number of the five types of insulation defects, specifically, overheating fault defect, solid metal protrusion defect, insulator air gap defect, insulator surface metal contamination defect and free metal particle defect are numbered 1, 2, 3, 4, 5 respectively; x i The experimental data are input, specifically, the concentration of the target gas, the generation rate of the target gas, and the temperature of the target gas are i=1, 2, and 3 respectively; that is, in the preset support vector machine, according to the difference of m, each insulation defect type corresponds to a decision function, and the experimental data corresponding to each insulation defect type in the database can be input into each decision function accordingly, so that the preset support vector machine performs five binary classifications on each experimental data; and each time the five binary classifications are performed, one insulation defect type data is divided into +1 class, and the other four insulation defect type data are divided into -1 class, and the five insulation defect types are respectively treated as +1 classes for five binary classifications; and the above The kernel function may specifically be a Gaussian radial basis kernel function, and the Gaussian radial basis kernel function expression is as follows:

[0071] K(x,x i )=exp(-g||xx i || 2 ), g>0

[0072] In the above formula, g is the kernel function parameter, g = 1 / 2σ 2 , σ is the kernel function width parameter, x is the data that needs to be classified by the preset support vector machine, x i The data used for preset support vector machine training.

[0073] In this process, the preset support vector machine optimizes each decision function according to the objective function. When the objective function value is minimized, the parameters of the obtained decision function are optimal. The objective function is as follows:

[0074]

[0075] In the above formula, C is the penalty parameter, C>0; ξ is the relaxation factor, ξ≥0; is the relaxation factor of the t-th type of data input in the m-th type of insulation defect data; N is the sample size; ω m is the weight vector in the preset support vector machine; is the transpose of the weight vector; b is the bias value in the preset support vector machine; m = 1, 2, 3, 4, 5; y t ∈{1,2,...,m},y t is the data type label, t is the index of the sample, t = 1, 2, ..., N; when the support vector machine is preset to perform data classification, the objective function value changes continuously. When the objective function is minimized, its corresponding ω and b make the decision function optimal.

[0076] Among them, the constraints of the objective function are:

[0077]

[0078] In the above formula, is the relaxation factor of the t-th type of data input in the m-th type of insulation defect data; is the transpose of the weight vector corresponding to the insulation defect type corresponding to the t-th type of data; b m are the deviation values ​​corresponding to the five insulation defect types in the preset support vector machine; b yt is the deviation value corresponding to the insulation defect type corresponding to the t-th type of data; x i The data used for training in the preset support vector machine training sample set is y t is the data type label, y t ∈{1,2,...,m}.

[0079] The final decision function for the preset support vector machine that can perform multiple data classifications is as follows:

[0080]

[0081] In the above formula, is the weight vector in the preset support vector machine; b is the deviation value, and m is the insulation defect type; is the kernel function.

[0082] Step 303 : For each parameter pair, assign the parameter pair to the K initial vector machines respectively, and determine the insulation defect classification quality of the K support vector machines after the assignment.

[0083] Among them, the obtained multiple parameter pairs are assigned to each initial vector machine respectively, so that each initial vector machine can use the candidate penalty parameters and candidate kernel function parameters contained in the assigned parameter pairs as candidate penalty parameters and kernel function parameters, thereby classifying the experimental data; according to the classification results of the K support vector machines after the assignment, the insulation defect classification quality of each of the K support vector machines after the assignment is determined, wherein, the more accurate the classification of the experimental data by each of the K support vector machines after the assignment is, the higher the insulation defect classification quality of the K support vector machines is.

[0084] Step 304 : Obtain a target parameter pair corresponding to the highest insulation defect classification quality, and use the candidate kernel function parameters and candidate penalty parameters included in the target parameter pair as kernel function parameters and penalty parameters of the target support vector machine, respectively.

[0085] Among them, according to the insulation defect classification quality of the K support vector machines after assignment, the parameter pair used by the support vector machine with the highest insulation defect classification quality is obtained as the target parameter pair, and the candidate kernel function parameters and candidate penalty parameters included in the target parameter pair are used as the kernel function parameters and penalty parameters of the target support vector machine that finally performs the target switchgear insulation defect type detection.

[0086] By constructing K initial vector machines and assigning the obtained multiple candidate penalty parameters and candidate kernel function parameters to each initial vector machine respectively, the classification quality of the insulation defect type of each initial vector machine with different assignments can be obtained, so as to help determine the optimal penalty parameters and kernel function parameters according to the classification quality of the insulation defect type, so that the detection result of the target support vector machine for insulation defect type detection based on the optimal penalty parameters and kernel function parameters is more accurate.

[0087] In one embodiment, Figure 5 As shown, it shows a flow chart of determining the insulation defect classification quality provided by an embodiment of the present application, wherein the insulation defect classification quality of the K support vector machines after assignment is determined, including:

[0088] Step 501: For each assigned support vector machine, obtain a target test set corresponding to the assigned support vector machine, and use the target test set as the input of the assigned support vector machine to obtain the output of the assigned support vector machine.

[0089] Among them, the divided K experimental data sets are respectively used as target test sets and input into each assigned support vector machine. Each assigned support vector machine classifies the data in the input target test set to obtain the classification result, which is the output of the assigned support vector machine.

[0090] Step 502: Obtain the mean square error values ​​of the outputs of the K support vector machines after the assignment.

[0091] Among them, after each assigned K support vector machine classifies and outputs the data in each target test set, its mean square error value is calculated. The mean square error value can represent the accuracy of the assigned support vector machine in classifying the data in the target test set.

[0092] Step 503: Determine the insulation defect classification quality of the K support vector machines after the assignment based on the mean square error value.

[0093] The method of obtaining the target parameter pair corresponding to the highest insulation defect classification quality includes: obtaining the target parameter pair corresponding to the K support vector machines after the assignment with the minimum mean square error.

[0094] Among them, the mean square error value of each assigned support vector machine is calculated accordingly. The smaller the mean square error value, the more accurate the insulation defect classification result of the assigned support vector machine is, and thus the candidate penalty parameters and candidate kernel function parameters contained in the target parameter pair used are more suitable for use in the target support vector machine.

[0095] In one embodiment, the method further includes: after determining the kernel function parameters and penalty parameters of the target support vector machine based on a cross-validation algorithm, training using the experimental data based on the determined kernel function parameters and penalty parameters to obtain the target support vector machine.

[0096] Among them, after the kernel function parameters and penalty parameters of the target support vector machine are determined, the experimental data in the database are input into the target support vector machine. The target support vector machine predicts and classifies the experimental data and continuously adjusts the error to obtain a target support vector machine whose input is the concentration of the target gas, the generation rate of the target gas and the temperature of the target switchgear, and whose output is the insulation defect type of the target switchgear.

[0097] In one embodiment, Figure 6 As shown, it shows a schematic diagram of a process for determining a target support vector machine and performing insulation defect type detection provided by an embodiment of the present application, including:

[0098] Step 601: Acquire experimental data.

[0099] The experimental data includes the concentration of the target gas in the test switchgear under different insulation defect types, the target gas production rate in the test switchgear, and the temperature of the test switchgear. The target gases are nitrogen dioxide, nitric oxide, and carbon monoxide. The process of obtaining the experimental data is as follows: under different insulation defect types, the insulation degradation degree of the test switchgear is continuously changed, and the discharge amount and voltage value of the test switchgear are gradually changed until obvious partial discharge or overheating fault is observed in the test switchgear. The time when partial discharge occurs in the test switchgear under different insulation defect types is determined. The concentrations of carbon monoxide, nitric oxide, and nitrogen dioxide in the test switchgear and the temperature of the test switchgear are collected, the obtained data are normalized, and stored in a database.

[0100] Step 602: Divide the experimental data into K experimental data sets, perform K model training processes, and obtain K initial support vector machines.

[0101] The experimental data is divided into K experimental data sets, each of which is divided into the concentrations and generation rates of carbon monoxide, nitric oxide and nitrogen dioxide gases of the test switchgear obtained under different insulation defect types, as well as the temperature of the test switchgear; K-1 experimental data sets are selected from the K experimental data sets as training sets, and the unselected experimental data sets are used as test sets; each experimental data set is used as a test set, and the remaining K-1 experimental data sets are used as training sets. The training set and the test set are input into a support vector machine, and the support vector machine is trained for data classification. A total of K training times are performed to obtain K initial support vector machines.

[0102] Step 603: Acquire multiple parameter pairs, each of which includes different candidate kernel function parameters and candidate penalty parameters.

[0103] The traversal algorithm is used to perform multiple traversal searches on the predetermined support vector machine within a certain range. Since the kernel function parameters and penalty parameters generated by the support vector machine are different each time it performs classification, a kernel function parameter and penalty parameter are obtained in each traversal search, which are used as candidate kernel function parameters and candidate penalty parameters to form a parameter pair; multiple traversal searches obtain multiple parameter pairs, each parameter pair includes different candidate kernel function parameters and candidate penalty parameters, and the traversal search range can be 2 -10 -2 10 ; The preset support vector machine is a support vector machine that pre-classifies experimental data of different insulation defect types.

[0104] Step 604: For each parameter pair, assign the parameter pair to the K initial vector machines respectively, and obtain the mean square error values ​​of the K support vector machines after the assignment.

[0105] The obtained multiple parameter pairs are assigned to each initial vector machine respectively, so that each initial vector machine can use the candidate penalty parameters and candidate kernel function parameters contained in the assigned parameter pairs as candidate penalty parameters and kernel function parameters, thereby classifying the experimental data; after each assigned K support vector machine classifies and outputs the data in each target test set, its mean square error value is calculated to obtain the mean square error value of the data classification of each assigned support vector machine.

[0106] Step 605: The parameter pair used by the support vector machine after the assignment with the smallest mean square error value is used as the target parameter pair.

[0107] Step 606 : The candidate kernel function parameters and candidate penalty parameters included in the target parameter pair are used as kernel function parameters and penalty parameters of a target support vector machine, and are trained using experimental data to obtain a target support vector machine.

[0108] Among them, after the kernel function parameters and penalty parameters of the target support vector machine are determined, the experimental data in the database are input into the target support vector machine. The target support vector machine predicts and classifies the experimental data and continuously adjusts the error to obtain a target support vector machine whose input is the concentration of the target gas, the generation rate of the target gas and the temperature of the target switchgear, and whose output is the insulation defect type of the target switchgear.

[0109] Step 607 : Acquire detection data of the target switchgear and input it into the target support vector machine to obtain the detection result of the target support vector machine.

[0110] For a target switchgear whose insulation defect type needs to be detected, detection data of the target switchgear is obtained, and the detection data includes the gas concentrations of nitrogen dioxide, nitric oxide, and carbon monoxide in the target switchgear, the generation rates of nitrogen dioxide, nitric oxide, and carbon monoxide gases, and the temperature of the target switchgear; the detection data is input into a target support vector machine to obtain a detection result of the target support vector machine, and the detection result is the insulation defect type of the target switchgear.

[0111] It should be understood that, although the various steps in the flowcharts involved in the various embodiments described above are displayed in sequence according to the instructions of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowcharts involved in the various embodiments described above can include multiple steps or multiple stages, and these steps or stages are not necessarily executed and completed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of steps or stages in other steps.

[0112] Based on the same inventive concept, embodiments of the present application also provide a switchgear insulation defect type detection device for implementing the aforementioned switchgear insulation defect type detection method. The solution provided by this device is similar to the solution described in the aforementioned method. Therefore, the specific limitations of one or more switchgear insulation defect type detection device embodiments provided below can be found in the aforementioned limitations of the switchgear insulation defect type detection method and will not be further elaborated here.

[0113] In one embodiment, Figure 7 As shown, a switch cabinet insulation defect type detection device is provided. The device 700 includes: an acquisition module 701 and a detection module 702, wherein:

[0114] An acquisition module 701 is configured to acquire detection data of a target gas in a target switch cabinet, wherein the detection data includes a concentration of the target gas, a generation rate of the target gas, and a temperature of the target switch cabinet;

[0115] The detection module 702 is used to input the detection data into the target support vector machine to obtain the detection result of the target support vector machine; the detection result is the insulation defect type of the target switchgear; wherein the kernel function parameters and penalty parameters of the target support vector machine are obtained based on the cross-validation algorithm.

[0116] In one embodiment, Figure 8 As shown, a second switch cabinet insulation defect type detection device is provided, and the device 800 further includes:

[0117] A construction module 703 is used to construct K initial support vector machines;

[0118] A second acquisition module 704 is configured to acquire a plurality of parameter pairs, each of which includes a different candidate kernel function parameter and a candidate penalty parameter;

[0119] A determination module 705 is configured to assign, for each parameter pair, the parameter pair to the K initial vector machines, and determine the insulation defect classification quality of the K support vector machines after the assignment;

[0120] The third acquisition module 706 is configured to acquire a target parameter pair corresponding to the highest insulation defect classification quality, and use the candidate kernel function parameters and candidate penalty parameters included in the target parameter pair as kernel function parameters and penalty parameters of the target support vector machine, respectively.

[0121] In one embodiment, the construction module 703 is specifically used to: obtain experimental data, which includes the concentration of the target gas in the test switch cabinet under different insulation defect types, the production rate of the target gas in the test switch cabinet, and the temperature of the test switch cabinet; divide the experimental data into K experimental data sets; based on the K experimental data sets, perform K model training processes to obtain K initial support vector machines; wherein, the model training process includes: selecting K-1 experimental data sets from the K experimental data sets as training sets, and using the unselected experimental data sets as test sets; performing model training based on the training set and the test set to obtain the initial support vector machine.

[0122] In one embodiment, the determination module 705 is specifically used to: for each assigned support vector machine, obtain a target test set corresponding to the assigned support vector machine, and use the target test set as the input of the assigned support vector machine to obtain the output of the assigned support vector machine; obtain the mean square error value of the outputs of the K assigned support vector machines; and determine the insulation defect classification quality of the K assigned support vector machines based on the mean square error value.

[0123] In one embodiment, the third acquisition module 706 is specifically configured to obtain the target parameter pairs corresponding to the K assigned support vector machines with the smallest mean square error.

[0124] In one embodiment, Figure 9 As shown, a third switch cabinet insulation defect type detection device is provided, and the device 900 further includes:

[0125] The training module 707 is used to determine the kernel function parameters and penalty parameters of the target support vector machine based on the cross-validation algorithm, and then perform training using the experimental data based on the determined kernel function parameters and penalty parameters to obtain the target support vector machine.

[0126] Each module in the switchgear insulation defect type detection device described above can be implemented in whole or in part through software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in hardware form, or can be stored in a computer device memory in software form, so that the processor can call and execute the corresponding operations of each module.

[0127] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 10 As shown. The computer device includes a processor, memory, and a network interface connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device is used to store switchgear insulation defect type detection data. The network interface of the computer device is used to communicate with an external terminal via a network connection. When executed by the processor, the computer program implements a switchgear insulation defect type detection method.

[0128] Those skilled in the art will understand that Figure 10 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0129] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:

[0130] Acquire detection data of a target gas in a target switchgear, the detection data including the concentration of the target gas, the generation rate of the target gas, and the temperature of the target switchgear; input the detection data into a target support vector machine to obtain a detection result of the target support vector machine; the detection result is the insulation defect type of the target switchgear; wherein the kernel function parameters and penalty parameters of the target support vector machine are derived based on a cross-validation algorithm.

[0131] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0132] Construct K initial support vector machines; obtain multiple parameter pairs, each parameter pair including different candidate kernel function parameters and candidate penalty parameters; for each parameter pair, assign the parameter pair to the K initial vector machines respectively, and determine the insulation defect classification quality of the K support vector machines after the assignment; obtain a target parameter pair corresponding to the highest insulation defect classification quality, and use the candidate kernel function parameters and candidate penalty parameters included in the target parameter pair as the kernel function parameters and penalty parameters of the target support vector machine respectively.

[0133] The K initial support vector machines are constructed, including: obtaining experimental data, the experimental data including the concentration of the target gas in the test switch cabinet under different insulation defect types, the production rate of the target gas in the test switch cabinet, and the temperature of the test switch cabinet; dividing the experimental data into K experimental data sets; performing K model training processes based on the K experimental data sets to obtain K initial support vector machines; wherein the model training process includes: selecting K-1 experimental data sets from the K experimental data sets as training sets, and using unselected experimental data sets as test sets; performing model training based on the training sets and the test sets to obtain the initial support vector machines;

[0134] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0135] For each assigned support vector machine, a target test set corresponding to the assigned support vector machine is obtained, and the target test set is used as the input of the assigned support vector machine to obtain the output of the assigned support vector machine; the mean square error value of the output of the K assigned support vector machines is obtained; and the insulation defect classification quality of the K assigned support vector machines is determined based on the mean square error value.

[0136] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0137] After the kernel function parameters and penalty parameters of the target support vector machine are determined based on the cross-validation algorithm, the target support vector machine is trained using the experimental data based on the determined kernel function parameters and penalty parameters.

[0138] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0139] Acquire detection data of a target gas in a target switchgear, the detection data including the concentration of the target gas, the generation rate of the target gas, and the temperature of the target switchgear; input the detection data into a target support vector machine to obtain a detection result of the target support vector machine; the detection result is the insulation defect type of the target switchgear; wherein the kernel function parameters and penalty parameters of the target support vector machine are derived based on a cross-validation algorithm.

[0140] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0141] Construct K initial support vector machines; obtain multiple parameter pairs, each parameter pair including different candidate kernel function parameters and candidate penalty parameters; for each parameter pair, assign the parameter pair to the K initial vector machines respectively, and determine the insulation defect classification quality of the K support vector machines after the assignment; obtain a target parameter pair corresponding to the highest insulation defect classification quality, and use the candidate kernel function parameters and candidate penalty parameters included in the target parameter pair as the kernel function parameters and penalty parameters of the target support vector machine respectively.

[0142] The K initial support vector machines are constructed, including: obtaining experimental data, the experimental data including the concentration of the target gas in the test switch cabinet under different insulation defect types, the production rate of the target gas in the test switch cabinet, and the temperature of the test switch cabinet; dividing the experimental data into K experimental data sets; performing K model training processes based on the K experimental data sets to obtain K initial support vector machines; wherein the model training process includes: selecting K-1 experimental data sets from the K experimental data sets as training sets, and using unselected experimental data sets as test sets; performing model training based on the training sets and the test sets to obtain the initial support vector machines;

[0143] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0144] For each assigned support vector machine, a target test set corresponding to the assigned support vector machine is obtained, and the target test set is used as the input of the assigned support vector machine to obtain the output of the assigned support vector machine; the mean square error value of the output of the K assigned support vector machines is obtained; and the insulation defect classification quality of the K assigned support vector machines is determined based on the mean square error value.

[0145] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0146] After the kernel function parameters and penalty parameters of the target support vector machine are determined based on the cross-validation algorithm, the target support vector machine is trained using the experimental data based on the determined kernel function parameters and penalty parameters.

[0147] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0148] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, database or other media used in the embodiments provided in this application may include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory may include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchains. The processor involved in the various embodiments provided herein may be, but are not limited to, a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic unit, a data processing logic unit based on quantum computing, and the like.

[0149] The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0150] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

Claims

1. A method for detecting insulation defect types in a switch cabinet, characterized in that: The method comprises: Acquiring detection data of a target gas in a target switch cabinet, the detection data including a concentration of the target gas, a generation rate of the target gas, and a temperature of the target switch cabinet; Inputting the detection data into a target support vector machine to obtain a detection result of the target support vector machine; the detection result is the insulation defect type of the target switchgear; The kernel function parameters and penalty parameters of the target support vector machine are obtained based on a cross-validation algorithm; the process of determining the kernel function parameters and penalty parameters includes: constructing K initial support vector machines; obtaining multiple parameter pairs, each parameter pair including different candidate kernel function parameters and candidate penalty parameters; for each parameter pair, assigning the parameter pair to the K initial support vector machines, and determining the insulation defect classification quality of the K support vector machines after the assignment; obtaining a target parameter pair corresponding to the highest insulation defect classification quality, and using the candidate kernel function parameters and candidate penalty parameters included in the target parameter pair as the kernel function parameters and penalty parameters of the target support vector machine, respectively; A traversal algorithm is used to perform multiple traversal searches on a predetermined support vector machine within a preset range. Each traversal search obtains a kernel function parameter and a penalty parameter, which are used as candidate kernel function parameters and candidate penalty parameters. The preset support vector machine optimizes each decision function according to the objective function. When the objective function value is minimized, the parameters of the obtained decision function are optimal. The objective function is: The constraints of the objective function are: ; C is the penalty parameter; ξ is the relaxation factor; is the relaxation factor of the t-th type of data input in the m-th type of insulation defect data; N is the sample size; ω m is the weight vector in the preset support vector machine; is the transpose of the weight vector; b is the bias value in the preset support vector machine; m = 1, 2, 3, 4, 5; y t is the data type label, t is the index of the sample; is the transpose of the weight vector corresponding to the insulation defect type corresponding to the t-th type of data; b m are the deviation values ​​corresponding to the five insulation defect types in the preset support vector machine; b yt is the deviation value corresponding to the insulation defect type corresponding to the t-th type of data; x i is the training data in the preset support vector machine training sample set; t The data type label.

2. The method according to claim 1, characterized in that The net increase in target gas concentration is the difference between the concentrations at the start and end of gas acquisition.

3. The method according to claim 1, characterized in that The step of constructing K initial support vector machines includes: Acquiring experimental data, the experimental data including the concentration of the target gas in the test switch cabinet under different insulation defect types, the production rate of the target gas in the test switch cabinet, and the temperature of the test switch cabinet; Dividing the experimental data into K experimental data sets; Based on the K experimental data sets, K model training processes are performed to obtain K initial support vector machines; wherein the model training process includes: Selecting K-1 experimental data sets from the K experimental data sets as training sets, and using the unselected experimental data sets as test sets; Model training is performed based on the training set and the test set to obtain the initial support vector machine.

4. The method according to claim 3, characterized in that Determining the insulation defect classification quality of the K support vector machines after the assignment includes: For each assigned support vector machine, obtaining a target test set corresponding to the assigned support vector machine, and using the target test set as the input of the assigned support vector machine to obtain the output of the assigned support vector machine; Obtaining mean square error values ​​of the outputs of the K support vector machines after the assignment; The insulation defect classification quality of the K support vector machines after the assignment is determined based on the mean square error value.

5. The method according to claim 4, characterized in that The target parameter pair corresponding to the highest insulation defect classification quality is obtained, including: Obtain the target parameter pairs corresponding to the K assigned support vector machines with the smallest corresponding mean square error.

6. The method according to claim 3, characterized in that The method further comprises: After the kernel function parameters and penalty parameters of the target support vector machine are determined based on a cross-validation algorithm, the target support vector machine is trained using the experimental data based on the determined kernel function parameters and penalty parameters to obtain the target support vector machine.

7. A switch cabinet insulation defect type detection device, characterized in that: The device comprises: an acquisition module, configured to acquire detection data of a target gas in a target switch cabinet, the detection data including a concentration of the target gas, a generation rate of the target gas, and a temperature of the target switch cabinet; A detection module, configured to input the detection data into a target support vector machine to obtain a detection result of the target support vector machine; the detection result is the insulation defect type of the target switchgear; Wherein, the kernel function parameters and penalty parameters of the target support vector machine are obtained based on a cross-validation algorithm; Construction module, used to construct K initial support vector machines; A second acquisition module is used to acquire multiple parameter pairs, each parameter pair including different candidate kernel function parameters and candidate penalty parameters; A determination module is configured to assign, for each parameter pair, the parameter pair to the K initial support vector machines, and determine the insulation defect classification quality of the K support vector machines after the assignment; a third acquisition module, configured to acquire a target parameter pair corresponding to the highest insulation defect classification quality, and use the candidate kernel function parameters and the candidate penalty parameters included in the target parameter pair as the kernel function parameters and the penalty parameters of the target support vector machine, respectively; A traversal algorithm is used to perform multiple traversal searches on a predetermined support vector machine within a preset range. Each traversal search obtains a kernel function parameter and a penalty parameter, which are used as candidate kernel function parameters and candidate penalty parameters. The preset support vector machine optimizes each decision function according to the objective function. When the objective function value is minimized, the parameters of the obtained decision function are optimal. The objective function is: The constraints of the objective function are: C is the penalty parameter; ξ is the relaxation factor; is the relaxation factor of the t-th type of data input in the m-th type of insulation defect data; N is the sample size; ω m is the weight vector in the preset support vector machine; is the transpose of the weight vector; b is the bias value in the preset support vector machine; m = 1, 2, 3, 4, 5; y t is the data type label, t is the index of the sample; is the transpose of the weight vector corresponding to the insulation defect type corresponding to the t-th type of data; b m are the deviation values ​​corresponding to the five insulation defect types in the preset support vector machine; b yt is the deviation value corresponding to the insulation defect type corresponding to the t-th type of data; x i is the training data in the preset support vector machine training sample set; t The data type label.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images