Insulator pollution flashover voltage prediction method and system, storage medium and equipment

Abstract

The embodiment of the invention discloses an insulator pollution flashover voltage prediction method comprising the following steps: obtaining an insulator training data set which comprises insulator information of a pollution insulator and a corresponding flashover voltage; an LS-SVM model is constructed; the insulator information is input into the LS-SVM model, parameters of the LS-SVM model are optimized until the difference between the output of the LS-SVM model and the flashover voltage corresponding to the insulator information meets a threshold condition, and it is determined that training of the LS-SVM model is completed; and the trained LS-SVM model carries out pollution flashover voltage prediction on test data and obtains a prediction result. According to the method, the prediction model is constructed and optimized based on historical data, and accurate prediction of the pollution flashover voltage of the insulator is realized by using the model, so that support is provided for safe and stable operation of a power system.

Description

Technical Field

[0001] The present invention relates to the technical field of flashover voltage prediction, and in particular to a method, system, storage medium and device for predicting insulator contamination flashover voltage. Background Art

[0002] Pollution flashover is a serious threat to safe operation in power systems. Pollution flashover accidents occur mainly because a large amount of dirty substances are deposited on the surface of insulators running in the power grid. When these dirt layers are wetted, the external insulation capacity of the insulators will drop sharply, which will lead to flashover phenomenon and flashover voltage. Pollution flashover will not only cause damage to power equipment, but may also cause power outages on a larger scale. Therefore, the effective prevention and control of pollution flashover is of great practical significance.

[0003] As the power system develops towards high voltage, large capacity and long distance, the performance requirements for insulators are getting higher and higher. The accurate prediction of insulator flashover voltage becomes particularly important. Accurate flashover voltage prediction methods can realize real-time monitoring and early warning of insulator status, which helps power system operation and maintenance personnel to take measures in advance, reduce the risk of flashover accidents, and ensure the safe and stable operation of the power system. Summary of the invention

[0004] Based on this, it is necessary to propose a method for predicting insulator pollution flashover voltage to address the above problems.

[0005] A method for predicting insulator contamination flashover voltage, the method comprising the following steps:

[0006] Acquire an insulator training data set, wherein the insulator training data set includes insulator information of contaminated insulators and corresponding flashover voltages;

[0007] Construct LS-SVM model;

[0008] Inputting the insulator information into the LS-SVM model, optimizing the parameters of the LS-SVM model until the difference between the output of the LS-SVM model and the flashover voltage corresponding to the insulator information meets a threshold condition, and then determining that the LS-SVM model training is completed;

[0009] According to the trained LS-SVM model, the pollution flashover voltage of the test data is predicted and the prediction results are obtained.

[0010] In the above solution, the step of obtaining an insulator training data set specifically includes:

[0011] Determine the pollution level, construct a number of corresponding test insulators according to the pollution level, test the number of test insulators, and obtain the flashover voltage of the test insulators;

[0012] Collecting geometric parameters and environmental parameters of the plurality of test insulators;

[0013] The insulator information is cleaned and normalized to obtain an insulator training data set.

[0014] In the above scheme, the determining of the pollution level and constructing a number of corresponding test insulators according to the pollution level specifically include:

[0015] The pollution level is determined according to the salt density value of the insulator, and the pollution level includes level 0, level I, level II, level III, and level IV.

[0016] In the above scheme, the pollution level is determined according to the salt density value of the insulator, specifically including:

[0017] When the salt density of the insulator is in the range of (0,0.03], the corresponding pollution level is 0;

[0018] When the salt density of the insulator is in the range of (0.03, 0.06], the corresponding pollution level is level I;

[0019] When the salt density of the insulator is in the range of (0.06, 0.10], the corresponding pollution level is level II;

[0020] When the salt density of the insulator is in the range of (0.10, 0.25], the corresponding pollution level is level III;

[0021] When the salt density of the insulator is in the range of (0.25, 0.35], the corresponding pollution level is level IV.

[0022] The above scheme also includes:

[0023] Optimizing the parameters of the LS-SVM model through an objective function;

[0024] The objective function is:

[0025]

[0026] Among them, ω is the weight vector, b is the bias term, ξ is the error term, and C is the penalty parameter.

[0027] In the above scheme, the objective function is optimized according to the insulator training data set until the LS-SVM model parameters corresponding to the current objective function meet the threshold condition, and then the LS-SVM model training is determined to be completed, which specifically includes:

[0028] The optimal parameter combination of the LS-SVM model is determined by optimizing the objective function, and the LS-SVM model parameters include a regularization parameter γ, a kernel parameter σ, and a penalty parameter C.

[0029] In the above scheme, determining the optimal parameter combination of the LS-SVM model by optimizing the objective function specifically includes:

[0030] Encoding the regularization parameter γ, the kernel parameter σ and the penalty parameter C and representing them by floating point numbers;

[0031] Randomly generate several different combinations of regularization parameters γ, kernel parameters σ and penalty parameters C to obtain multiple chromosomes;

[0032] Determining the fitness value of each chromosome through a fitness function, and selecting some chromosomes as candidate chromosomes according to the fitness value;

[0033] Performing a crossover operation on the candidate chromosomes to generate a new chromosome combination;

[0034] performing mutation operation on some chromosomes in the new chromosome combination;

[0035] The fitness value of the chromosome after the mutation operation is obtained. When the fitness value meets the threshold range, the LS-SVM model parameters corresponding to the chromosome are the optimal parameter combination of the LS-SVM model.

[0036] In the above scheme, the prediction of the pollution flashover voltage of the test data based on the regression function of the trained LS-SVM model and obtaining the prediction result specifically include:

[0037] Acquiring the test data;

[0038] The test data includes insulator geometric parameters and environmental parameters. The environmental parameters include salt density, ash density, and relative humidity. The geometric parameters include the height, diameter, and creepage distance of the insulator.

[0039] The present application also proposes an insulator pollution flashover voltage prediction system, the system comprising: a data set acquisition unit, a model construction unit, a model training unit and a prediction unit;

[0040] The data set acquisition unit is used to acquire an insulator training data set, wherein the insulator training data set includes insulator information of contaminated insulators and corresponding flashover voltages;

[0041] The model building unit is used to build a LS-SVM model;

[0042] The model training unit is used to input the insulator information into the LS-SVM model, and optimize the parameters of the LS-SVM model until the difference between the output of the LS-SVM model and the flashover voltage corresponding to the insulator information meets a threshold condition, and then determine that the LS-SVM model training is completed;

[0043] The prediction unit is used to predict the pollution flashover voltage of the test data according to the trained LS-SVM model and obtain the prediction result.

[0044] The present application also proposes a readable storage medium storing a computer program. When the computer program is executed by a processor, the processor executes the following steps:

[0045] Acquire an insulator training data set, wherein the insulator training data set includes insulator information of contaminated insulators and corresponding flashover voltages;

[0046] Construct LS-SVM model;

[0047] Inputting the insulator information into the LS-SVM model, optimizing the parameters of the LS-SVM model until the difference between the output of the LS-SVM model and the flashover voltage corresponding to the insulator information meets a threshold condition, and then determining that the LS-SVM model training is completed;

[0048] According to the trained LS-SVM model, the pollution flashover voltage of the test data is predicted and the prediction results are obtained.

[0049] The present application also proposes a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor executes the following steps of the computer program:

[0050] Acquire an insulator training data set, wherein the insulator training data set includes insulator information of contaminated insulators and corresponding flashover voltages;

[0051] Construct LS-SVM model;

[0052] Inputting the insulator information into the LS-SVM model, optimizing the parameters of the LS-SVM model until the difference between the output of the LS-SVM model and the flashover voltage corresponding to the insulator information meets a threshold condition, and then determining that the LS-SVM model training is completed;

[0053] According to the trained LS-SVM model, the pollution flashover voltage of the test data is predicted and the prediction results are obtained.

[0054] The embodiment of the present invention has the following beneficial effects: firstly, an insulator training data set is obtained, wherein the insulator training data set includes the insulator information of the contaminated insulator and the corresponding flashover voltage; then, an LS-SVM model is constructed; the insulator information is input into the LS-SVM model, and the parameters of the LS-SVM model are optimized until the difference between the output of the LS-SVM model and the flashover voltage corresponding to the insulator information meets the threshold condition, then it is determined that the LS-SVM model training is completed; the trained LS-SVM model predicts the contaminated flashover voltage of the test data and obtains the prediction result. The present invention constructs and optimizes the LS-SVM model based on historical data, and finally uses the prediction model after the training optimization to accurately predict the contaminated flashover voltage of the newly input insulator data, which helps to warn the flashover fault caused by the contamination of the insulator in advance, thereby ensuring the safe and stable operation of the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the embodiments of the present invention 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 invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0056] in:

[0057] Figure 1 The figure is a flow chart of a method for predicting flashover voltage of an insulator due to pollution in one embodiment. DETAILED DESCRIPTION

[0058] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0059] In the following description, a large number of specific details are given in order to provide a more thorough understanding of the present invention; however, it is obvious to those skilled in the art that the present invention can be implemented without one or more of these details; in other examples, in order to avoid confusion with the present invention, some technical features known in the art are not described, and it should be understood that the present invention can be implemented in different forms and should not be interpreted as limited to the embodiments presented here; on the contrary, providing these embodiments will make the disclosure thorough and complete and fully convey the scope of the present invention to those skilled in the art.

[0060] The purpose of the terms used herein is only to describe specific embodiments and is not intended to be limiting of the present invention. When used herein, the singular forms "a", "an" and "said / the" are also intended to include the plural forms, unless the context clearly indicates otherwise, and it should also be understood that the terms "consisting of" and / or "comprising", when used in this specification, determine the presence of the features, integers, steps, operations, elements and / or parts, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, parts and / or groups. When used herein, the term "and / or" includes any and all combinations of the relevant listed items.

[0061] In order to deal with the problem of insulator flashover, prediction models are disclosed in the prior art. However, the prediction of insulator flashover voltage is a complex process and faces many challenges. First, there are many factors that affect the flashover voltage, such as the degree of pollution, wetness conditions, insulator materials and environmental conditions, and there are complex interactions between these factors. Secondly, the influence of these factors on the flashover voltage shows significant randomness and nonlinear characteristics, making it difficult for traditional statistical methods to accurately describe the complex relationship between the flashover voltage and its influencing factors. Therefore, it is particularly important to find a prediction method that can handle these complexities.

[0062] This application combines LS-SVM and genetic algorithm, and by simulating natural selection and genetic mechanisms, it can find the optimal solution in a complex parameter space and accurately predict the insulator pollution flashover voltage, which helps to effectively prevent and control pollution flashover accidents and ensure the safe and stable operation of the power system.

[0063] In order to thoroughly understand the present invention, a detailed structure will be presented in the following description to illustrate the technical solution proposed by the present invention; optional embodiments of the present invention are described in detail as follows, but in addition to these detailed descriptions, the present invention may also have other implementation methods.

[0064] like Figure 1 As shown, in one embodiment, a method for predicting an insulator contaminated flashover voltage is provided, and the method for predicting an insulator contaminated flashover voltage includes steps S101 to S104, which are described in detail as follows:

[0065] S101, obtaining an insulator training data set, where the insulator training data set includes insulator information of contaminated insulators and corresponding flashover voltages;

[0066] This step can provide a basis for subsequent model training by collecting historical data. Ensuring the diversity and representativeness of the data set is crucial to improving the accuracy of model predictions.

[0067] In some embodiments, obtaining an insulator training data set specifically includes:

[0068] Determine the pollution level, construct a number of corresponding test insulators according to the pollution level, test the number of test insulators, and obtain the flashover voltage of the test insulators;

[0069] Collect geometric parameters and environmental parameters of several test insulators;

[0070] The insulator information is cleaned and normalized to obtain the insulator information.

[0071] In the above scheme, several corresponding test insulators are constructed according to these pollution levels, which can simulate various pollution levels that may be encountered in actual operation, ensuring the breadth and representativeness of the training data set.

[0072] Specifically, the geometric parameters of the test insulator include the size, shape and structure of the insulator, while the environmental parameters involve factors such as temperature, humidity, wind speed, etc. that may affect the flashover voltage. The collection of these parameters helps to fully understand the relationship between insulator performance and external conditions.

[0073] In order to ensure the quality and consistency of the data, this application also performs data cleaning and normalization on the collected insulator information. The data cleaning process includes removing outliers, filling missing values ​​and correcting erroneous data. Normalization converts data of different dimensions and ranges into a unified standard form. These preprocessing steps can improve the quality of the training data set and lay a solid foundation for subsequent model training and prediction.

[0074] Preferably, the finite element method is used to simulate the electric field and potential distribution of the insulator to generate a training data set. The training data set is used for model training and includes insulator data under different geometric shapes and pollution conditions. The basic process includes: (1) using a triangular mesh to divide the surface of the test insulator; (2) calculating the potential distribution on the surface of the insulator to determine the electric field strength and pollution flashover voltage; (3) based on the simulation results, generating 200 training sample data sets.

[0075] Preferably, the data set is properly normalized before training. Normalization can scale the data to the same range, which helps to speed up the convergence of the model and improve the prediction accuracy. The normalization formula is as follows:

[0076]

[0077] Among them, x' is the pollution parameter (such as salt density, ash density and relative humidity), x max and x min are the minimum and maximum values ​​in the sample, respectively.

[0078] When processing multidimensional data, redundant variables may be generated. Therefore, PCA (principal component analysis) is used in the normalization process to reduce the dimension of input variables and further improve the efficiency of model training.

[0079] In some embodiments, determining the pollution level and constructing a number of corresponding test insulators according to the pollution level specifically includes:

[0080] The pollution level is determined according to the salt density of the insulator, and the pollution levels include level 0, level I, level II, level III, and level IV.

[0081] In some embodiments, the pollution level is determined according to the salt density of the insulator, specifically including:

[0082] When the salt density of the insulator is in the range of (0,0.03], the corresponding pollution level is 0;

[0083] When the salt density of the insulator is in the range of (0.03, 0.06], the corresponding pollution level is level I;

[0084] When the salt density of the insulator is in the range of (0.06, 0.10], the corresponding pollution level is level II;

[0085] When the salt density of the insulator is in the range of (0.10, 0.25], the corresponding pollution level is level III;

[0086] When the salt density of the insulator is in the range of (0.25, 0.35], the corresponding pollution level is level IV.

[0087] Specifically, by making the salt density value correspond to the pollution level, the test insulator can be constructed according to a unified standard, ensuring the repeatability and comparability of the test; among them, the division of different pollution levels enables the test insulator to better simulate the different pollution conditions in the real environment, thereby improving the applicability of the prediction model in practical applications. The test insulator data set constructed based on the pollution level can provide rich training samples for the subsequent machine learning model, which helps the model to learn the flashover voltage characteristics under different pollution levels. Through precise pollution level division, the prediction model can more accurately identify and predict the flashover voltage of the insulator, providing a strong guarantee for the safe operation of the power system.

[0088] Preferably, by selecting insulator samples of different geometric shapes, different materials and surface coatings to simulate the common insulator materials in actual power lines, it can be ensured that the test results have wide applicability.

[0089] S102, constructing LS-SVM model;

[0090] LS-SVM is an improved support vector machine. It introduces the least squares linear system to replace the traditional quadratic programming problem, thereby improving the computational efficiency and prediction accuracy. It is suitable for dealing with complex problems such as nonlinearity, high dimension and local minima, and is suitable for predicting insulator contamination flashover voltage. The construction of the LS-SVM model lays the foundation for subsequent parameter optimization and prediction.

[0091] Preferably, the present application uses insulator height (Hi-mm), diameter (D-mm), creepage distance (L-mm), surface conductivity (σ-μS), number of strings (n) and number of sheds (d) as input variables of LS-SVM, dynamically models the insulator, and the pollution flashover voltage prediction value is the output variable.

[0092] S103, inputting the insulator information into the LS-SVM model, optimizing the parameters of the LS-SVM model, until the difference between the output of the LS-SVM model and the flashover voltage corresponding to the insulator information meets the threshold condition, then determining that the LS-SVM model training is completed;

[0093] By inputting insulator information, the parameters of the LS-SVM model are optimized so that the difference between the model output (predicted flashover voltage) and the actual flashover voltage is as small as possible until the set threshold condition is met. This process ensures the accuracy and generalization ability of the model. Parameter optimization is an important task in machine learning and directly affects the prediction performance of the model.

[0094] Specifically, the preprocessed insulator information, including but not limited to pollution level, geometric parameters, environmental parameters, etc., is input into the LS-SVM model as a feature vector. After the data is input, the LS-SVM model is initialized and the initial model parameters are set. These parameters include regularization parameters, kernel function parameters, etc., and then the parameters of the LS-SVM model are adjusted through an iterative optimization process until the model meets the training completion conditions.

[0095] In some embodiments, it also includes:

[0096] The parameters of the LS-SVM model are optimized through the objective function;

[0097] The objective function is:

[0098]

[0099] Among them, ω is the weight vector, b is the bias term, ξ is the error term, and C is the penalty parameter.

[0100] In some embodiments, the objective function is optimized according to the insulator training data set until the LS-SVM model parameters corresponding to the current objective function meet the threshold condition, and then the LS-SVM model training is determined to be completed, specifically including:

[0101] The optimal parameter combination of the LS-SVM model is determined by optimizing the objective function. The LS-SVM model parameters include the regularization parameter γ, the kernel parameter σ and the penalty parameter C.

[0102] In some embodiments, the optimal parameter combination of the LS-SVM model is determined by optimizing the objective function, specifically including:

[0103] Encode the regularization parameter γ, kernel parameter σ and penalty parameter C and represent them by floating point numbers;

[0104] Randomly generate several different combinations of regularization parameters γ, kernel parameters σ and penalty parameters C to obtain multiple chromosomes;

[0105] The fitness value of each chromosome is determined by the fitness function, and some chromosomes are selected as candidate chromosomes according to the fitness value;

[0106] Perform crossover operation on candidate chromosomes to generate new chromosome combinations;

[0107] Perform mutation operations on some chromosomes in the new chromosome combination;

[0108] The fitness value of the chromosome after the mutation operation is obtained. When the fitness value meets the threshold range, the LS-SVM model parameters corresponding to the chromosome are the optimal parameter combination of the LS-SVM model.

[0109] In some embodiments, the minimum RMSE value (fitness value) corresponds to the optimal LS-SVM parameter combination.

[0110] Specifically, fitness function: Set the fitness function to the root mean square error (RMSE) to evaluate the quality of each chromosome. The formula is:

[0111]

[0112] Where N is the number of data patterns in the dataset, y pre,k is the predicted value, u ac,k is the test value of a data point k.

[0113] Furthermore, the optimal parameter combination of the LS-SVM model can be determined by setting a threshold condition, which defines an acceptable range of difference between the model output and the actual flashover voltage. This threshold can be the mean square error (MSE), root mean square error or other applicable performance indicators. If the square error or root mean square error of the output of the current parameter combination meets the difference range, the current parameter combination is the optimal parameter combination.

[0114] Preferably, determining the optimal parameter combination of the LS-SVM model by optimizing the objective function also includes:

[0115] Encode the regularization parameter γ, kernel parameter σ and penalty parameter C and represent them by floating point numbers;

[0116] Randomly generate several different combinations of regularization parameters γ, kernel parameters σ and penalty parameters C to obtain multiple chromosomes;

[0117] The fitness value of each chromosome is determined by the fitness function, and some chromosomes are selected as candidate chromosomes according to the fitness value;

[0118] Perform crossover operation on candidate chromosomes to generate new chromosome combinations;

[0119] Perform mutation operations on some chromosomes in the new chromosome combination;

[0120] The above screening, crossover and mutation operations are regarded as an iterative process. When the iterative process meets a certain number of requirements, the LS-SVM model parameters corresponding to the current chromosome are output as the optimal parameter combination of the LS-SVM model.

[0121] Specifically, in the above screening step, a roulette wheel or tournament selection strategy may be used according to the fitness value to select chromosomes with better performance as candidates for the next generation.

[0122] Furthermore, the above crossover operation is to perform a crossover operation on the selected chromosomes with a certain probability to generate a new chromosome combination. The crossover formula is:

[0123] C new =λC1+(1-λ)C2

[0124] C new =λC1+(1-λ)σ2

[0125] Among them, λ is a random number between 0 and 1, the crossover probability is generally set to 0.8-0.9, C1 and C2 are two parent chromosomes, C new To generate chromosomes for offspring.

[0126] Furthermore, the above crossover operation is to perform mutation operation on some chromosomes to prevent falling into local optimum. The mutation formula is:

[0127] C new =C old +δ×randn()

[0128] C new =σ old +δ×randn()

[0129] Among them, δ is the variation range, the variation rate is set to 0.01-0.1, randn() is a normally distributed random number, C new To generate offspring chromosomes, C old The parent chromosome.

[0130] By setting threshold conditions and optimizing parameters, the risk of model overfitting is effectively reduced, ensuring the robustness of the model in practical applications. The trained LS-SVM model is faster and more efficient in predicting new data, providing support for real-time monitoring and preventive maintenance of power systems.

[0131] S104, predicting the pollution flashover voltage of the test data according to the trained LS-SVM model and obtaining a prediction result.

[0132] After completing the model training, the trained LS-SVM model is used to predict the test data to obtain the flashover voltage prediction results of the contaminated insulator. This step is the model application stage. By comparing the predicted results with the actual results, the prediction performance and generalization ability of the model can be evaluated. If the prediction results are accurate, it means that the model can effectively predict the contaminated flashover voltage of the insulator, which is of great significance to the safe and stable operation of the power system.

[0133] In some embodiments, the pollution flashover voltage is predicted on the test data according to the regression function of the trained LS-SVM model and a prediction result is obtained, specifically including:

[0134] Get test data;

[0135] The test data includes the insulator geometric parameters and environmental parameters. The environmental parameters include salt density, ash density, and relative humidity. The geometric parameters include the insulator height, diameter, and creepage distance.

[0136] Specifically, the regression function of the above LS-SVM model is:

[0137]

[0138] Kernel function K(x i ,x) usually uses the radial basis kernel function (RBF):

[0139]

[0140] Among them, f(x) is the predicted pollution flashover voltage value, b is the bias term, a i is the Lagrange multiplier, K(x i ,x) is the kernel function, and σ is the width parameter of the kernel function.

[0141] Through the kernel function, the LS-SVM model can handle nonlinear problems, enabling it to efficiently capture the complex relationship between insulator pollution parameters and flashover voltage.

[0142] In some embodiments, the absolute variance score (R 2 ), mean error function (MEF), and coefficient of variation (COV) to compare the predicted and tested values ​​and calculate the model validation. The absolute variance score (R 2 ), mean error function (MEF) and coefficient of variation (COV) are calculated as follows:

[0143]

[0144] Among them, u ac,k is the actual value of the kth data point, y pre,k is the predicted value of the kth data point, n is the total number of data points, |u ac,k -y pre,k | is the absolute value of the prediction error of the kth data point, max(u ac,k ) is the maximum actual value of the kth data point, min(u ac,k ) is the minimum actual value of the kth data point, N is the total number of data points, RMSE is the fitness value, is the average of all data points.

[0145] When the fitting effect is good, R 2 A value of 1 is obtained, while a poor fit results in values ​​close to zero.

[0146] The present application also proposes an insulator pollution flashover voltage prediction system, the system comprising: a data set acquisition unit, a model building unit, a model training unit and a prediction unit;

[0147] A data set acquisition unit, used for acquiring an insulator training data set, wherein the insulator training data set includes insulator information of contaminated insulators and corresponding flashover voltages;

[0148] A model building unit, used to build a LS-SVM model;

[0149] The model training unit is used to input the insulator information into the LS-SVM model and optimize the parameters of the LS-SVM model until the difference between the output of the LS-SVM model and the flashover voltage corresponding to the insulator information meets the threshold condition, and then it is determined that the LS-SVM model training is completed;

[0150] The prediction unit is used to predict the pollution flashover voltage of the test data according to the trained LS-SVM model and obtain the prediction result.

[0151] The present application also proposes a readable storage medium storing a computer program. When the computer program is executed by a processor, the processor executes the following steps:

[0152] Acquire an insulator training data set, where the insulator training data set includes insulator information of contaminated insulators and corresponding flashover voltages;

[0153] Construct LS-SVM model;

[0154] The insulator information is input into the LS-SVM model, and the parameters of the LS-SVM model are optimized until the difference between the output of the LS-SVM model and the flashover voltage corresponding to the insulator information meets the threshold condition, and then the LS-SVM model training is determined to be completed;

[0155] According to the trained LS-SVM model, the pollution flashover voltage of the test data is predicted and the prediction results are obtained.

[0156] The present application also proposes a computer device, including a memory and a processor, wherein the memory stores a computer program, and the computer program is executed by the processor to perform the following steps:

[0157] Acquire an insulator training data set, where the insulator training data set includes insulator information of contaminated insulators and corresponding flashover voltages;

[0158] Construct LS-SVM model;

[0159] The insulator information is input into the LS-SVM model, and the parameters of the LS-SVM model are optimized until the difference between the output of the LS-SVM model and the flashover voltage corresponding to the insulator information meets the threshold condition, and then the LS-SVM model training is determined to be completed;

[0160] According to the trained LS-SVM model, the pollution flashover voltage of the test data is predicted and the prediction results are obtained.

[0161] Those skilled in the art will understand that all or part of the processes in the above-mentioned embodiments can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in the present application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0162] The technical features of the above embodiments can be combined arbitrarily. 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.

[0163] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the patent of the present application. It should be pointed out that, for ordinary technicians in this field, several modifications and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. The above disclosure is only a preferred embodiment of the present invention, and of course it cannot be used to limit the scope of the rights of the present invention. Therefore, equivalent changes made according to the claims of the present invention are still within the scope covered by the present invention.

Claims

1. A method for predicting insulator contamination flashover voltage, characterized in that: include: Acquire an insulator training data set, wherein the insulator training data set includes insulator information of contaminated insulators and corresponding flashover voltages; Construct LS-SVM model; Inputting the insulator information into the LS-SVM model, optimizing the parameters of the LS-SVM model until the difference between the output of the LS-SVM model and the flashover voltage corresponding to the insulator information meets a threshold condition, and then determining that the LS-SVM model training is completed; According to the trained LS-SVM model, the pollution flashover voltage of the test data is predicted and the prediction results are obtained.

2. The method for predicting insulator pollution flashover voltage according to claim 1, characterized in that: The obtaining of the insulator training data set specifically includes: Determine the pollution level, construct a number of corresponding test insulators according to the pollution level, test the number of test insulators, and obtain the flashover voltage of the test insulators; Collecting geometric parameters and environmental parameters of the plurality of test insulators; The insulator information is cleaned and normalized to obtain an insulator training data set.

3. The method for predicting insulator pollution flashover voltage according to claim 2, characterized in that: The determining of the pollution level and constructing a number of corresponding test insulators according to the pollution level specifically includes: The pollution level is determined according to the salt density value of the insulator, and the pollution level includes level 0, level I, level II, level III, and level IV.

4. The method for predicting insulator pollution flashover voltage according to claim 3, characterized in that: Determining the pollution level according to the salt density value of the insulator specifically includes: When the salt density of the insulator is in the range of (0,0.03], the corresponding pollution level is 0; When the salt density of the insulator is in the range of (0.03, 0.06], the corresponding pollution level is level I; When the salt density of the insulator is in the range of (0.06, 0.10], the corresponding pollution level is level II; When the salt density of the insulator is in the range of (0.10, 0.25], the corresponding pollution level is level III; When the salt density of the insulator is in the range of (0.25, 0.35], the corresponding pollution level is level IV.

5. The method for predicting insulator pollution flashover voltage according to claim 1, characterized in that: Also includes: Optimizing the parameters of the LS-SVM model through an objective function; The objective function is: Among them, ω is the weight vector, b is the bias term, ξ is the error term, and C is the penalty parameter.

6. The method for predicting insulator pollution flashover voltage according to claim 1, characterized in that: The step of optimizing the objective function according to the insulator training data set until the LS-SVM model parameter corresponding to the current objective function meets a threshold condition, and then determining that the LS-SVM model training is completed, specifically includes: The optimal parameter combination of the LS-SVM model is determined by optimizing the objective function, and the LS-SVM model parameters include a regularization parameter γ, a kernel parameter σ, and a penalty parameter C.

7. The method for predicting insulator pollution flashover voltage according to claim 6, characterized in that: Determining the optimal parameter combination of the LS-SVM model by optimizing the objective function specifically includes: Encoding the regularization parameter γ, the kernel parameter σ and the penalty parameter C and representing them by floating point numbers; Randomly generate several different combinations of regularization parameters γ, kernel parameters σ and penalty parameters C to obtain multiple chromosomes; Determining the fitness value of each chromosome through a fitness function, and selecting some chromosomes as candidate chromosomes according to the fitness value; Performing a crossover operation on the candidate chromosomes to generate a new chromosome combination; performing mutation operation on some chromosomes in the new chromosome combination; The fitness value of the chromosome after the mutation operation is obtained. When the fitness value meets the threshold range, the LS-SVM model parameters corresponding to the chromosome are the optimal parameter combination of the LS-SVM model.

8. The method for predicting insulator pollution flashover voltage according to claim 1, characterized in that: The method of predicting the pollution flashover voltage of the test data according to the regression function of the trained LS-SVM model and obtaining the prediction result specifically includes: Acquiring the test data; The test data includes insulator geometric parameters and environmental parameters. The environmental parameters include salt density, ash density, and relative humidity. The geometric parameters include the height, diameter, and creepage distance of the insulator.

9. An insulator pollution flashover voltage prediction system, characterized in that: The system comprises: a data set acquisition unit, a model construction unit, a model training unit and a prediction unit; The data set acquisition unit is used to acquire an insulator training data set, wherein the insulator training data set includes insulator information of contaminated insulators and corresponding flashover voltages; The model building unit is used to build a LS-SVM model; The model training unit is used to input the insulator information into the LS-SVM model, and optimize the parameters of the LS-SVM model until the difference between the output of the LS-SVM model and the flashover voltage corresponding to the insulator information meets a threshold condition, and then determine that the LS-SVM model training is completed; The prediction unit is used to predict the pollution flashover voltage of the test data according to the trained LS-SVM model and obtain the prediction result.

10. A readable storage medium storing a computer program, wherein when the computer program is executed by a processor, the processor is caused to perform the steps of the method according to any one of claims 1 to 8.

11. A computer device comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the method according to any one of claims 1 to 8.

Images