Power equipment state evaluation method and device

Through the processing of infrared images and environmental parameters, combined with intelligent models to calculate the insulator state membership degree and operating risk value, the problem of online real-time assessment of insulator surface pollution is solved, timely quantification of pollution degree and preventive maintenance guidance, and the grid stability is improved and operation and maintenance efficiency.

CN119046790APending Publication Date: 2024-11-29STATE GRID QINGHAI ELECTRIC POWER COMPANY +3
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411095927.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-12
Publication Date
2024-11-29

AI Technical Summary

Technical Problem

The prior art cannot realize the online real-time status evaluation of insulator surface filth, resulting in untimely detection of contaminated states, resulting in lagging maintenance work and unable to solve the problem in a timely manner.

Method used

通过获取绝缘子的红外图像和环境参数,进行预处理和特征提取,利用预训练的智能模型计算状态隶属度和运行风险值,实现绝缘子状态的实时监测和评估。

Benefits of technology

It has achieved timely quantification of the degree of insulator pollution, guided preventive maintenance, and improved the stability and operation and maintenance efficiency of the power grid.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119046790A_ABST
    Figure CN119046790A_ABST
Patent Text Reader

Abstract

The invention discloses a power equipment state evaluation method and device, and relates to the technical field of power equipment detection. The method comprises the following steps: acquiring an infrared image and environmental parameters of a target insulator, and processing the infrared image to obtain target temperature characterization data of the target insulator; according to the target temperature characterization data and the environment temperature, characteristic parameters are obtained, the characteristic parameters serve as input of a pre-trained intelligent model, and the state membership degree of the target insulator is obtained; and calculating an operation risk value of the target insulator according to the state membership degree. On-line real-time monitoring of the insulator can be realized through an infrared image method. Characteristic parameters closely related to state classification are extracted from the infrared image to serve as a classification basis, so that the calculation result of the state membership degree is more reliable. The operation risk value is calculated according to the state membership degree, the pollution degree of the insulator is quantified, and preventive maintenance work can be timely and effectively guided.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of power equipment detection, and in particular to a method and device for evaluating the state of power equipment. Background Art

[0002] Insulators are a special type of insulating control that can play an important role in overhead transmission lines. According to different materials, insulators can be divided into porcelain insulators, glass insulators, and composite insulators. As key stress carriers and insulating devices in transmission lines, the operating reliability of insulators is closely related to the safety of the transmission system. However, glass insulators work in outdoor environments and are affected by electric fields and climate for a long time, and dirt accumulation on their surfaces is inevitable. The presence of dirt causes the hydrophobicity of the insulator surface to decrease, resulting in discharge. At the same time, dirt reduces the surface resistance of the insulator and increases the leakage current. In severe cases, flashover accidents are triggered, causing large-scale power outages, greatly increasing the cost of power grid operation and maintenance, and seriously threatening the stability of the power grid. In view of the above problems, qualitative or quantitative evaluation of the surface contamination of insulators is of great significance for timely elimination of hidden dangers.

[0003] At present, the degree of contamination is determined by manual inspection and regular testing. However, due to limitations such as human factors and measurement conditions, online real-time status assessment has not been achieved, resulting in delayed detection of contamination status, delayed maintenance work, and inability to solve the problem in a timely manner. Summary of the invention

[0004] The purpose of the present invention is to solve the problem of lack of timeliness mentioned in the above background technology, and to propose a method and device for evaluating the status of power equipment.

[0005] A first aspect of the present invention provides a method for evaluating the state of an electric power device, the method comprising:

[0006] Acquire an infrared image and environmental parameters of a target insulator, and extract a temperature matrix of the infrared image; the environmental parameters include an ambient temperature;

[0007] Preprocessing the infrared image to obtain a target binary image, and extracting a binary matrix of the target binary image;

[0008] Obtaining target temperature characterization data of the target insulator according to the temperature matrix and the binary matrix;

[0009] Obtaining characteristic parameters according to the target temperature characterization data and the ambient temperature, and using the characteristic parameters as input of a pre-trained intelligent model to obtain a state membership of the target insulator;

[0010] According to the state membership, an operation risk value of the target insulator is calculated.

[0011] Optionally, preprocessing the infrared image to obtain a target binary image includes:

[0012] Performing median filtering on the infrared image to obtain a first target image;

[0013] enhancing the first target image by using a second-order difference method to obtain a second target image;

[0014] Performing morphological processing on the second target image to obtain a third target image;

[0015] Converting the third target image into a binary image to obtain a fourth target image;

[0016] The fourth target image is segmented using a watershed algorithm to obtain a target binary image.

[0017] Optionally, the environmental parameters further include environmental humidity and wind speed; and the characteristic parameters obtained according to the target temperature characterization data and the environmental temperature include:

[0018] Obtaining first temperature rise data according to the target temperature characterization data and the ambient temperature;

[0019] The first temperature rise data is corrected according to the ambient humidity and wind speed to obtain the target temperature rise data:

[0020]

[0021] Wherein, TTR is the target temperature rise data; TRD is the first temperature rise data; S is the wind speed; H is the ambient humidity; a, b, c, k are preset constants;

[0022] Calculate characteristic parameters of the target temperature rise data; the characteristic parameters include mean value, median, root mean square, maximum value, variance, standard deviation and kurtosis.

[0023] Optionally, the status of the insulator is divided into multiple levels; the intelligent model is a probabilistic neural network model; wherein:

[0024] In the model training stage, the structure of the first probabilistic neural network model is determined according to the number of preset characteristic parameters and the number of state levels; the first probabilistic neural network model includes an input layer, a pattern layer, a summation layer and an output layer; the number of neurons in the input layer is equal to the number of preset characteristic parameters; the number of neurons in the pattern layer is determined by the number of training samples; the number of neurons in the summation layer is determined by the number of state levels; the output layer includes a neuron for outputting a representative number of the state level;

[0025] During the model use phase, the output layer of the first probabilistic neural network model is deleted to obtain a target intelligent model; the target intelligent model is used to output the probability that the target insulator is in each state level as the state membership.

[0026] Optionally, calculating the operation risk value of the target insulator according to the state membership includes:

[0027] According to the state membership, multiple risk reference values ​​and corresponding weights are determined:

[0028]

[0029] Among them, w j is the weight of the j-th risk reference value; MD j is the state membership corresponding to the j-th risk reference value; n is the specific number of the multiple risk reference values;

[0030] Multiple risk reference values ​​are weighted and calculated to obtain the operational risk value.

[0031] A second aspect of the present invention provides a device for evaluating the state of an electric power device, the device comprising:

[0032] A data acquisition module, used to acquire an infrared image and environmental parameters of a target insulator, and extract a temperature matrix of the infrared image; the environmental parameters include an ambient temperature;

[0033] A preprocessing module, used for preprocessing the infrared image to obtain a target binary image and extracting a binary matrix of the target binary image;

[0034] An information extraction module, used for obtaining target temperature characterization data of the target insulator according to the temperature matrix and the binary matrix;

[0035] A state classification module, used to obtain characteristic parameters according to the target temperature characterization data and the ambient temperature, and use the characteristic parameters as input of a pre-trained intelligent model to obtain the state membership of the target insulator;

[0036] The risk calculation module is used to calculate the operation risk value of the target insulator according to the state membership.

[0037] Optionally, the preprocessing module includes:

[0038] An image denoising module, used for performing median filtering on the infrared image to obtain a first target image;

[0039] An image enhancement module, used for enhancing the first target image by using a second-order difference method to obtain a second target image;

[0040] an edge enhancement module, used for performing morphological processing on the second target image to obtain a third target image;

[0041] A binarization module, used for converting the third target image into a binary image to obtain a fourth target image;

[0042] The image segmentation module is used to segment the fourth target image using a watershed algorithm to obtain a target binary image.

[0043] Optionally, the environmental parameters also include environmental humidity and wind speed; the state classification module includes:

[0044] A first temperature rise calculation module, used for obtaining first temperature rise data according to the target temperature characterization data and the ambient temperature;

[0045] A temperature rise correction module is used to correct the first temperature rise data according to the ambient humidity and wind speed to obtain target temperature rise data:

[0046]

[0047] Wherein, TTR is the target temperature rise data; TRD is the first temperature rise data; S is the wind speed; H is the ambient humidity; a, b, c, k are preset constants;

[0048] The feature extraction module is used to calculate the feature parameters of the target temperature rise data; the feature parameters include mean value, median, root mean square, maximum value, variance, standard deviation and kurtosis.

[0049] Optionally, the status of the insulator is divided into multiple levels; the intelligent model is a probabilistic neural network model; the device also includes:

[0050] A model training module, used for determining the structure of a first probabilistic neural network model according to the number of preset characteristic parameters and the number of state levels during the model training phase; the first probabilistic neural network model comprises an input layer, a pattern layer, a summation layer and an output layer; the number of neurons in the input layer is equal to the number of preset characteristic parameters; the number of neurons in the pattern layer is determined by the number of training samples; the number of neurons in the summation layer is determined by the number of state levels; the output layer comprises a neuron for outputting a representative number of the state level;

[0051] The model modification module is used to delete the output layer of the first probabilistic neural network model during the model use phase to obtain a target intelligent model; the target intelligent model is used to output the probability of the target insulator being in each state level as the state membership.

[0052] Optionally, the risk calculation module includes:

[0053] The weight calculation module is used to determine multiple risk reference values ​​and corresponding weights according to the state membership:

[0054]

[0055] Among them, w j is the weight of the j-th risk reference value; MD j is the state membership corresponding to the j-th risk reference value; n is the specific number of the multiple risk reference values;

[0056] The operation risk determination module is used to perform weighted calculation on multiple risk reference values ​​to obtain an operation risk value.

[0057] Beneficial effects of the present invention:

[0058] The present invention proposes a method for evaluating the state of electric power equipment, which includes: acquiring an infrared image and environmental parameters of a target insulator, and extracting a temperature matrix of the infrared image; the environmental parameters include ambient temperature; preprocessing the infrared image to obtain a target binary image, and extracting a binary matrix of the target binary image; obtaining target temperature characterization data of the target insulator according to the temperature matrix and the binary matrix; obtaining characteristic parameters according to the target temperature characterization data and the ambient temperature, and using the characteristic parameters as input of a pre-trained intelligent model to obtain a state membership of the target insulator; and calculating an operation risk value of the target insulator according to the state membership.

[0059] Through the infrared imaging method, the online real-time monitoring of insulators can be realized. The characteristic parameters closely related to the state classification are extracted from the infrared image as the classification basis, so that the calculation result of the state membership is more reliable. The operation risk value is calculated according to the state membership, and the pollution degree of the insulator is quantified, which can guide the preventive maintenance work in a timely and effective manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] The present invention will be further described below in conjunction with the accompanying drawings.

[0061] Figure 1 A flowchart of a method for evaluating the state of an electric device is provided for an embodiment of the present invention;

[0062] Figure 2 A structural diagram of a power equipment status assessment device is provided for an embodiment of the present invention. DETAILED DESCRIPTION

[0063] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0064] The embodiment of the present invention provides a method for evaluating the state of an electric power device. Figure 1 , Figure 1 A flowchart of a method for evaluating the state of an electric device provided by an embodiment of the present invention. The method comprises the following steps:

[0065] S101, acquiring an infrared image and environmental parameters of a target insulator, and extracting a temperature matrix of the infrared image.

[0066] S102, preprocessing the infrared image to obtain a target binary image, and extracting a binary matrix of the target binary image.

[0067] S103, obtaining target temperature characterization data of the target insulator according to the temperature matrix and the binary matrix.

[0068] S104, obtaining characteristic parameters according to the target temperature characterization data and the ambient temperature, and using the characteristic parameters as input of a pre-trained intelligent model to obtain the state membership of the target insulator.

[0069] S105, calculating the operation risk value of the target insulator according to the state membership.

[0070] Among them, the environmental parameters include ambient temperature.

[0071] Based on a power equipment status assessment method provided by an embodiment of the present invention, online real-time monitoring of insulators can be achieved through infrared imaging. Feature parameters closely related to status classification are extracted from infrared images as a basis for classification, thereby making the calculation result of status membership more reliable. The operation risk value is calculated based on the status membership, and the degree of contamination of the insulator is quantified, which can guide preventive maintenance work in a timely and effective manner.

[0072] In one implementation, conventional classification methods use thresholds to classify or categorize different pollution levels. Data near the thresholds are uncertain, which has a negative impact on pollution treatment decisions. Based on the principle of fuzzy algorithms, using state membership as an evaluation indicator for insulators and then calculating risk values ​​can more accurately identify pollution levels.

[0073] In one embodiment, step S102 includes:

[0074] Step 1: Perform median filtering on the infrared image to obtain a first target image.

[0075] Step 2: Use the second-order difference method to enhance the first target image to obtain the second target image.

[0076] Step three, performing morphological processing on the second target image to obtain a third target image.

[0077] Step 4: Convert the third target image into a binary image to obtain a fourth target image.

[0078] Step 5: Use the watershed algorithm to segment the fourth target image to obtain a target binary image.

[0079] In one implementation, the target insulator can be accurately extracted by filtering, enhancing, reconstructing and segmenting the infrared image, so that the temperature information of the target insulator can be extracted from the pixel temperature of the infrared image, providing a data basis for subsequent classification.

[0080] In one implementation, a median filter can be used to smooth the image, thereby effectively eliminating salt and pepper noise in the image. A double-pass filter method can also be used, which can achieve parallel processing, thereby improving the efficiency and speed of data processing.

[0081] In one implementation, the second-order difference of the image can be calculated using a Sobel operator to adjust pixels and enhance edge detail information.

[0082] In one implementation, the image can be smoothed by the opening and closing operations in the morphological reconstruction to obtain a reconstructed image, and the reconstructed image can be converted into a binary image. The vector distance of each pixel in the binary image can be obtained by the Euclidean distance transformation, and the foreground and background can be separated by the watershed transformation of the vector distance matrix to achieve image segmentation. Thus, the pixel information of the insulator in the image can be obtained, and the temperature information of the insulator can be extracted from the infrared image. Through this method, the pixel position information of the insulator in the image can be accurately identified, so as to obtain comprehensive and accurate insulator temperature information.

[0083] In one embodiment, step S103 may be implemented by performing element-by-element multiplication of the binary matrix and the temperature matrix to obtain temperature characterization data of the insulator.

[0084] Among them, each element of the temperature matrix represents the temperature value of each pixel in the infrared image; the binary matrix only has 0 and 1. By multiplying the corresponding elements, only the temperature information of the insulator pixels can be retained, and the insulator temperature characterization data can be extracted, providing a data basis for subsequent classification.

[0085] In one embodiment, the environmental parameters also include environmental humidity and wind speed; step S104 includes:

[0086] Step 1: Obtain first temperature rise data according to target temperature characterization data and ambient temperature.

[0087] Step 2: Correct the first temperature rise data according to the ambient humidity and wind speed to obtain the target temperature rise data:

[0088]

[0089] Among them, TTR is the target temperature rise data; TRD is the first temperature rise data; S is the wind speed; H is the ambient humidity; a, b, c, k are preset constants determined by experimental data.

[0090] Step three, calculate the characteristic parameters of the target temperature rise data.

[0091] Among them, the characteristic parameters include mean, median, root mean square, maximum value, variance, standard deviation and kurtosis.

[0092] In one implementation, the target temperature rise data is obtained by correcting the first temperature rise data. This method can eliminate the influence of environmental factors on the temperature data, thereby improving the accuracy of the assessment.

[0093] In one implementation, the mean, median, root mean square, maximum value, variance, standard deviation and kurtosis of the temperature rise data are closely related to the pollution of the insulator surface. By extracting these characteristic parameters, the accuracy of the evaluation can be improved.

[0094] In one embodiment, the status of the insulator is divided into multiple levels; the intelligent model is a probabilistic neural network model; wherein:

[0095] During the model training phase, the structure of the first probabilistic neural network model is determined according to the number of preset feature parameters and the number of state levels; the first probabilistic neural network model includes an input layer, a pattern layer, a summation layer and an output layer; the number of neurons in the input layer is equal to the number of preset feature parameters; the number of neurons in the pattern layer is determined by the number of training samples; the number of neurons in the summation layer is determined by the number of state levels; the output layer includes a neuron for outputting a representative number of the state level.

[0096] In the model use stage, the output layer of the first probabilistic neural network model is deleted to obtain the target intelligent model; the target intelligent model is used to output the probability of the target insulator being in each state level as the state membership.

[0097] In one implementation, a probabilistic neural network model (PNN) is a feedforward neural network based on Bayesian decision theory, which is suitable for pattern classification problems. For multi-level classification of insulator states, PNN can use its unique structure to quickly and accurately determine the state level of the insulator.

[0098] In one implementation, during the model use phase, by deleting the output layer, the target intelligent model can output the probability of the target insulator being in each state level, that is, the state membership. The pollution degree can be quantified more precisely based on these state memberships.

[0099] In one embodiment, step S105 includes:

[0100] Step 1: Determine multiple risk reference values ​​and corresponding weights based on the state membership:

[0101]

[0102] Among them, w j is the weight of the j-th risk reference value; MD j is the state membership corresponding to the j-th risk reference value; n is the specific number of the multiple risk reference values.

[0103] Step 2: Perform weighted calculation on multiple risk reference values ​​to obtain the operational risk value.

[0104] In one implementation, the state membership is an important indicator for insulator state assessment, which can reflect the state level of the insulator, but the level classification is uncertain. Combining multiple risk reference values ​​and corresponding weights for weighted calculation can be finely quantified, providing an important reference for maintenance decision-making.

[0105] In one implementation, each state level corresponds to a risk reference value. For example, the state levels are divided into four categories: A, B, C, and D, and their risk reference values ​​are 1, 2, 3, and 4, respectively. The larger the risk reference value, the more serious the surface pollution of the insulator. In the case of four categories, the risk reference values ​​of the states corresponding to the two largest state memberships can be selected. For example, the output results of the classification are A: 0.12, B: 0.10, C: 0.37, and D: 0.41. The weight calculation is w1 = 0.37 / (0.37+0.41) = 0.474, w2 = 0.42 / (0.37+0.41) = 0.538; the operation risk value is rv = 0.474*3+0.538*4 = 3.674. The operation and maintenance personnel can determine the maintenance strategy according to the bias of the operation risk value.

[0106] The embodiment of the present invention provides a device for evaluating the state of an electric power device. Figure 2 , Figure 2 A structural diagram of a power equipment status assessment device provided by an embodiment of the present invention. The device includes:

[0107] The data acquisition module is used to acquire the infrared image and environmental parameters of the target insulator and extract the temperature matrix of the infrared image.

[0108] The preprocessing module is used to preprocess the infrared image to obtain a target binary image and extract a binary matrix of the target binary image.

[0109] The information extraction module is used to obtain target temperature characterization data of the target insulator according to the temperature matrix and the binary matrix.

[0110] The state classification module is used to obtain characteristic parameters according to the target temperature characterization data and the ambient temperature, and use the characteristic parameters as the input of the pre-trained intelligent model to obtain the state membership of the target insulator.

[0111] The risk calculation module is used to calculate the operation risk value of the target insulator according to the state membership.

[0112] Among them, the environmental parameters include ambient temperature.

[0113] Based on an electric power equipment status assessment device provided by an embodiment of the present invention, online real-time monitoring of insulators can be achieved through infrared imaging. Feature parameters closely related to status classification are extracted from infrared images as classification basis, so that the calculation result of status membership is more reliable. The operation risk value is calculated based on the status membership, and the degree of contamination of the insulator is quantified, which can guide preventive maintenance work in a timely and effective manner.

[0114] In one embodiment, the pre-processing module includes:

[0115] The image denoising module is used to perform neighborhood mean filtering on the infrared image to obtain a first target image.

[0116] The image enhancement module is used to enhance the first target image by using a second-order difference method to obtain a second target image.

[0117] The edge enhancement module is used to perform morphological processing on the second target image to obtain a third target image.

[0118] The image segmentation module is used to segment the third target image using a watershed algorithm to obtain a fourth target image.

[0119] The binarization module is used to convert the fourth target image into a binary image to obtain a target binary image.

[0120] In one embodiment, the environmental parameters further include environmental humidity and wind speed; and the state classification module includes:

[0121] The first temperature rise calculation module is used to obtain first temperature rise data according to the target temperature characterization data and the ambient temperature.

[0122] The temperature rise correction module is used to correct the first temperature rise data according to the ambient humidity and wind speed to obtain the target temperature rise data:

[0123]

[0124] Among them, TTR is the target temperature rise data; TRD is the first temperature rise data; S is the wind speed; H is the ambient humidity; a, b, c, k are constants determined by experimental data.

[0125] The feature extraction module is used to calculate the characteristic parameters of the target temperature rise data.

[0126] Among them, the characteristic parameters include mean, median, root mean square, maximum value, variance, standard deviation and kurtosis.

[0127] In one embodiment, the status of the insulator is divided into multiple levels; the intelligent model is a probabilistic neural network model; the device also includes:

[0128] The model training module is used to determine the structure of the first probabilistic neural network model according to the number of preset feature parameters and the number of state levels during the model training stage; the first probabilistic neural network model includes an input layer, a pattern layer, a summation layer and an output layer; the number of neurons in the input layer is equal to the number of preset feature parameters; the number of neurons in the pattern layer is determined by the number of training samples; the number of neurons in the summation layer is determined by the number of state levels; the output layer includes a neuron for outputting a representative number of the state level.

[0129] The model modification module is used to delete the output layer of the first probabilistic neural network model during the model use phase to obtain a target intelligent model; the target intelligent model is used to output the probability of the target insulator being in each state level as the state membership.

[0130] In one embodiment, the risk calculation module includes:

[0131] The weight calculation module is used to determine multiple risk reference values ​​and corresponding weights according to the state membership:

[0132]

[0133] Among them, w j is the weight of the j-th risk reference value; MD j is the state membership corresponding to the j-th risk reference value; n is the specific number of the multiple risk reference values.

[0134] The operation risk determination module is used to perform weighted calculation on multiple risk reference values ​​to obtain an operation risk value.

[0135] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A method for evaluating the state of electric power equipment, characterized in that: The method comprises: Acquire an infrared image and environmental parameters of a target insulator, and extract a temperature matrix of the infrared image; the environmental parameters include an ambient temperature; Preprocessing the infrared image to obtain a target binary image, and extracting a binary matrix of the target binary image; Obtaining target temperature characterization data of the target insulator according to the temperature matrix and the binary matrix; Obtaining characteristic parameters according to the target temperature characterization data and the ambient temperature, and using the characteristic parameters as input of a pre-trained intelligent model to obtain a state membership of the target insulator; According to the state membership, an operation risk value of the target insulator is calculated.

2. A method for evaluating the state of electric power equipment according to claim 1, characterized in that: The preprocessing of the infrared image to obtain a target binary image comprises: Performing median filtering on the infrared image to obtain a first target image; enhancing the first target image by using a second-order difference method to obtain a second target image; Performing morphological processing on the second target image to obtain a third target image; Converting the third target image into a binary image to obtain a fourth target image; The fourth target image is segmented using a watershed algorithm to obtain a target binary image.

3. A method for evaluating the state of electric power equipment according to claim 1, characterized in that: The environmental parameters also include environmental humidity and wind speed; the characteristic parameters obtained according to the target temperature characterization data and the environmental temperature include: Obtaining first temperature rise data according to the target temperature characterization data and the ambient temperature; The first temperature rise data is corrected according to the ambient humidity and wind speed to obtain the target temperature rise data: Wherein, TTR is the target temperature rise data; TRD is the first temperature rise data; S is the wind speed; H is the ambient humidity; a, b, c, k are preset constants; Calculate characteristic parameters of the target temperature rise data; the characteristic parameters include mean value, median, root mean square, maximum value, variance, standard deviation and kurtosis.

4. A method for evaluating the state of electric power equipment according to claim 3, characterized in that: The states of the insulators are divided into multiple levels; the intelligent model is a probabilistic neural network model; wherein: In the model training stage, the structure of the first probabilistic neural network model is determined according to the number of preset characteristic parameters and the number of state levels; the first probabilistic neural network model includes an input layer, a pattern layer, a summation layer and an output layer; the number of neurons in the input layer is equal to the number of preset characteristic parameters; the number of neurons in the pattern layer is determined by the number of training samples; the number of neurons in the summation layer is determined by the number of state levels; the output layer includes a neuron for outputting a representative number of the state level; During the model use phase, the output layer of the first probabilistic neural network model is deleted to obtain a target intelligent model; the target intelligent model is used to output the probability that the target insulator is in each state level as the state membership.

5. A method for evaluating the state of electric power equipment according to claim 1, characterized in that: Calculating the operation risk value of the target insulator according to the state membership comprises: According to the state membership, multiple risk reference values ​​and corresponding weights are determined: Among them, w j is the weight of the j-th risk reference value; MD j is the state membership corresponding to the j-th risk reference value; n is the specific number of the multiple risk reference values; Multiple risk reference values ​​are weighted and calculated to obtain the operational risk value.

6. A device for evaluating the state of electric power equipment, characterized in that: The device comprises: A data acquisition module, used to acquire an infrared image and environmental parameters of a target insulator, and extract a temperature matrix of the infrared image; the environmental parameters include an ambient temperature; A preprocessing module, used for preprocessing the infrared image to obtain a target binary image and extracting a binary matrix of the target binary image; An information extraction module, used for obtaining target temperature characterization data of the target insulator according to the temperature matrix and the binary matrix; A state classification module, used to obtain characteristic parameters according to the target temperature characterization data and the ambient temperature, and use the characteristic parameters as input of a pre-trained intelligent model to obtain the state membership of the target insulator; The risk calculation module is used to calculate the operation risk value of the target insulator according to the state membership.

7. The device for evaluating the state of electric power equipment according to claim 6, characterized in that: The preprocessing module comprises: An image denoising module, used for performing median filtering on the infrared image to obtain a first target image; An image enhancement module, used for enhancing the first target image by using a second-order difference method to obtain a second target image; an edge enhancement module, configured to perform morphological processing on the second target image to obtain a third target image; A binarization module, used for converting the third target image into a binary image to obtain a fourth target image; The image segmentation module is used to segment the fourth target image using a watershed algorithm to obtain a target binary image.

8. The device for evaluating the state of electric power equipment according to claim 6, characterized in that: The environmental parameters also include environmental humidity and wind speed; the state classification module includes: A first temperature rise calculation module, used for obtaining first temperature rise data according to the target temperature characterization data and the ambient temperature; A temperature rise correction module is used to correct the first temperature rise data according to the ambient humidity and wind speed to obtain target temperature rise data: Wherein, TTR is the target temperature rise data; TRD is the first temperature rise data; S is the wind speed; H is the ambient humidity; a, b, c, k are preset constants; The feature extraction module is used to calculate the feature parameters of the target temperature rise data; the feature parameters include mean value, median, root mean square, maximum value, variance, standard deviation and kurtosis.

9. The power equipment status assessment device according to claim 8, characterized in that: The states of the insulators are divided into multiple levels; the intelligent model is a probabilistic neural network model; the device also includes: A model training module, used for determining the structure of a first probabilistic neural network model according to the number of preset characteristic parameters and the number of state levels during the model training phase; the first probabilistic neural network model comprises an input layer, a pattern layer, a summation layer and an output layer; the number of neurons in the input layer is equal to the number of preset characteristic parameters; the number of neurons in the pattern layer is determined by the number of training samples; the number of neurons in the summation layer is determined by the number of state levels; the output layer comprises a neuron for outputting a representative number of the state level; The model modification module is used to delete the output layer of the first probabilistic neural network model during the model use phase to obtain a target intelligent model; the target intelligent model is used to output the probability of the target insulator being in each state level as the state membership.

10. The power equipment status assessment device according to claim 1, characterized in that: The risk calculation module comprises: The weight calculation module is used to determine multiple risk reference values ​​and corresponding weights according to the state membership: Among them, w j is the weight of the j-th risk reference value; MD j is the state membership corresponding to the j-th risk reference value; n is the specific number of the multiple risk reference values; The operation risk determination module is used to perform weighted calculation on multiple risk reference values ​​to obtain an operation risk value.