Composite insulator infrared temperature measurement compensation calibration method

Through the infrared temperature measurement calibration model combined with BP neural network and GA genetic algorithm, the problem of infrared temperature measurement error in drone inspection is solved, the temperature measurement accuracy and stability are improved, and the safety of transmission lines is ensured.

CN119063851APending Publication Date: 2024-12-03STATE GRID HUNAN ELECTRIC POWER COMPANY LIMITED +2
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411174637.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-26
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

Infrared temperature measurement technology is greatly affected by distance and viewing angle when the drone inspects composite insulators, resulting in large errors in temperature measurement results, affecting accuracy and fault judgment.

Method used

The infrared temperature measurement calibration model is constructed by BP neural network combined with GA genetic algorithm, and the model weight and bias are optimized through the training data set, and the distance view angle compensation calibration is performed to improve the temperature measurement accuracy.

Benefits of technology

It significantly improves the accuracy and stability of infrared temperature measurement, can quickly and accurately judge the operating status of composite insulators, and ensures the safe and stable operation of the transmission line.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119063851A_ABST
    Figure CN119063851A_ABST
Patent Text Reader

Abstract

The invention discloses a composite insulator infrared temperature measurement compensation calibration method comprising the following steps: obtaining composite insulator infrared temperature measurement calibration data, and obtaining a training data set; constructing a composite insulator infrared temperature measurement calibration initial model based on a BP neural network; training the composite insulator infrared temperature measurement calibration initial model by using the training data set in combination with a GA (Genetic Algorithm) to obtain a composite insulator infrared temperature measurement calibration model; and performing actual composite insulator infrared temperature measurement compensation calibration by using the composite insulator infrared temperature measurement calibration model. According to the method, the GA is adopted to optimize the BP neural network to construct the model, and the distance visual angle compensation calibration of infrared temperature measurement is performed, so that the temperature measurement precision is improved, the stability and reliability of the model are improved, the operation state of the composite insulator can be quickly and accurately judged during infrared inspection, and the safe and stable operation of a power transmission line is ensured.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the field of infrared temperature measurement, and in particular relates to an infrared temperature measurement compensation calibration method for a composite insulator. Background Art

[0002] With the development of the modern electric power industry, the stable operation of transmission lines is crucial to the safety and reliability of the power system. As an important part of the transmission line, the working state of the composite insulator directly affects the stability and safety of the transmission line. Therefore, the status monitoring of the composite insulator is particularly important. As a non-contact temperature detection method, infrared temperature measurement technology has been widely used in transmission line inspection due to its advantages such as high efficiency and convenience. In recent years, some scholars at home and abroad have studied the correction of infrared temperature measurement accuracy. In China, with the rapid development of drone technology, it has gradually become a trend to use drones equipped with infrared temperature measurement equipment to inspect transmission lines.

[0003] However, when infrared temperature measurement technology is applied to UAV inspection of composite insulators, it is greatly affected by distance and viewing angle, and the temperature measurement results may have large errors. This is mainly because infrared radiation is interfered by various factors during the transmission process, such as air absorption and scattering, and the measurement angle of infrared camera equipment. Therefore, how to effectively compensate for the distance and viewing angle errors in infrared temperature measurement has become a key issue to improve temperature measurement accuracy and accurately determine the fault level of insulators. Summary of the invention

[0004] The purpose of the present invention is to provide a composite insulator infrared temperature measurement compensation calibration method to improve the temperature measurement accuracy of unmanned aerial vehicle infrared inspection.

[0005] The present invention provides a composite insulator infrared temperature measurement compensation calibration method, comprising the following steps:

[0006] S1. Obtain infrared temperature measurement calibration data of composite insulators to obtain a training data set;

[0007] S2. Based on BP neural network, the initial model of infrared temperature measurement calibration of composite insulator is constructed;

[0008] S3. Using the training data set obtained in step S1, combined with the GA genetic algorithm, the initial model of the composite insulator infrared temperature measurement calibration obtained in step S2 is trained to obtain a composite insulator infrared temperature measurement calibration model;

[0009] S4. Using the composite insulator infrared temperature measurement calibration model obtained in step S3, actual composite insulator infrared temperature measurement compensation calibration is performed.

[0010] Step S1 specifically includes: constructing a composite insulator temperature rise experimental platform, conducting experiments on composite insulators to obtain measured values ​​and true values ​​of the composite insulators at different measurement distances and angles by an infrared thermal imager, and arranging them to obtain a training data set;

[0011] The experiment was conducted indoors with a wind speed of 0 m / s and constant temperature and humidity;

[0012] The composite insulator temperature rise test platform includes a high-voltage power supply system, a high-voltage measurement system, a pressurizing device, a composite insulator suspension vehicle, a composite insulator, an optical fiber thermometer, an infrared thermometer and an infrared thermal imager;

[0013] The experiment includes the following steps:

[0014] First, hang the composite insulator vertically on the composite insulator hanging vehicle, fix its top, measure the vertical distance between the composite insulator and the ground, and ensure that the distance between any part of the composite insulator and the ground object is not less than 0.7m;

[0015] Install the fiber optic thermometer, stick the three probes of the fiber optic thermometer on the surface of the composite insulator with insulating thermal double-sided adhesive, and place the fourth probe in the air to measure the ambient temperature;

[0016] Connect the high-voltage end of the composite insulator to the live wire end of the pressurizing device, and the top end to the bottom line to simulate the actual operating state; set the voltage to 1.1 times the actual operating voltage, and confirm that the safety guardrail and grounding rod are correctly placed before pressurizing;

[0017] After pressurization, use an optical fiber thermometer to measure the actual temperature of the composite insulator, and obtain the infrared temperature through an infrared thermometer and an infrared thermal imager; according to the requirements of the UAV infrared inspection, change the distance and viewing angle between the infrared thermometer and the infrared thermal imager to perform measurements respectively, and record the temperature data of the infrared thermometer, infrared thermal imager and optical fiber thermometer;

[0018] Replace the new composite insulator and repeat the above steps, and organize the temperature data obtained by the infrared thermometer, infrared thermal imager and optical fiber thermometer to obtain a training data set.

[0019] Step S2 is specifically as follows: based on the BP neural network, an initial model for infrared temperature measurement calibration of composite insulators is constructed, including an input layer, one or several hidden layers and an output layer connected in series in sequence; in the input layer, one or several hidden layers and the output layer, each layer includes several neurons, and the layers are fully connected; the input data is transmitted layer by layer starting from the input layer, and the prediction results are finally generated in the output layer through the calculation of the neurons in each layer.

[0020] Step S3 specifically includes: firstly using GA genetic algorithm to optimize the weight and bias of the neural network, and then using forward propagation and back propagation to train the initial model of infrared temperature measurement calibration of composite insulators to obtain the infrared temperature measurement calibration model of composite insulators;

[0021] Among them, the optimization of the weights and biases of the neural network using the GA genetic algorithm includes the following steps:

[0022] 1) Set genetic algorithm parameters: set population size, crossover probability p_c and mutation probability p_m;

[0023] 2) Population initialization: Randomly generate an initial population, where each individual represents a set of weights and bias combinations of a neural network; the initial weights and biases are generated through uniform distribution or normal distribution;

[0024] 3) Define the fitness function and calculate the fitness value: The fitness function is used to evaluate the performance of each individual. The fitness is defined based on the error of the neural network on the validation set. The higher the fitness value, the smaller the model error of the individual, and therefore the better the performance.

[0025] 4) Selection operation: using roulette wheel selection and tournament selection methods, select individuals with fitness higher than the preset value from the current population to enter the next generation;

[0026] 5) Crossover operation: According to the crossover probability p_c obtained in step 1, randomly pair the selected individuals and use single-point crossover, two-point crossover or uniform crossover method to exchange genes to generate new individuals;

[0027] 6) Mutation operation: According to the mutation probability p_m obtained in step 1, the genes (i.e. weights or biases) in the new individuals are randomly adjusted to increase the diversity of the population and prevent the algorithm from falling into a local optimal solution;

[0028] 7) Fitness evaluation and selection: The fitness of the newly generated individuals is evaluated again, and the individuals of the parent generation are selected. The elite retention strategy is used to retain the individuals with the highest fitness to ensure the quality of the solution;

[0029] 8) Stopping of genetic algorithm: When the set maximum number of generations is reached or the fitness change within a number of generations is less than the preset threshold, the genetic algorithm stops and the individual with the highest fitness is selected as the optimal solution, representing the final optimized neural network weights and biases.

[0030] The forward propagation is specifically to input the data in the training data set into the initial model of infrared temperature measurement calibration of composite insulators to obtain the output value, and then calculate the error between the output value and the corresponding true value in the training data set;

[0031] Back propagation calculates the error between the predicted result and the actual value, and adjusts the weights and biases of the model neural network through the gradient descent method to minimize the output error. The weight change of back propagation is expressed by the following formula:

[0032]

[0033] Among them, Δw kj is the weight change of the back-propagation output layer; η is the learning rate; E is the total error; w kj is the weight from the jth hidden layer neuron to the kth output layer neuron; u′ k is the output of the kth neuron in the output layer;

[0034] The change in the output layer neuron bias is expressed using the following formula:

[0035]

[0036] Where Δb k is the change in the output layer neuron bias; b k is the bias of the kth neuron in the output layer;

[0037] The transformation of the input layer neuron weights is expressed using the following formula:

[0038]

[0039] Among them, Δw ji is the weight change from the input layer neurons to the hidden layer neurons; w ji is the weight from the jth hidden layer neuron to the ith output layer neuron; h′ j is the output of the jth hidden layer neuron;

[0040] The change in the bias of the hidden layer neurons is expressed using the following formula:

[0041]

[0042] Where Δb j is the transformation amount of the hidden layer neuron bias; b j is the bias of the jth neuron in the hidden layer;

[0043] The invention discloses a method for infrared temperature measurement compensation calibration of a composite insulator. The method adopts GA genetic algorithm to optimize BP neural network to construct a model, and performs distance and viewing angle compensation calibration of infrared temperature measurement, thereby improving the temperature measurement accuracy, increasing the stability and reliability of the model, and enabling the operating status of the composite insulator to be judged quickly and accurately during infrared inspection, thereby ensuring the safe and stable operation of the transmission line. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 It is a schematic diagram of the process of the present invention;

[0045] Figure 2 is the temperature measurement result of the infrared thermometer;

[0046] Figure 3 The temperature measurement result of the infrared thermal imager

[0047] Figure 4 This is a comparison chart of the model correction effect;

[0048] Figure 5 This is a comparison chart of the correction results of the infrared thermometer;

[0049] Figure 6 Comparison chart of correction results for infrared imager;

[0050] Figure 7 This is the infrared temperature measurement correction diagram. DETAILED DESCRIPTION

[0051] The present invention provides a composite insulator infrared temperature measurement compensation calibration method, the flow diagram of which is as follows Figure 1 As shown, the following steps are included:

[0052] S1. Obtain the infrared temperature measurement calibration data of composite insulators to obtain the training data set, specifically:

[0053] A temperature rise experimental platform for composite insulators was built, and experiments were conducted on composite insulators to obtain the measured values ​​and true values ​​of the composite insulators at different measurement distances and angles by the infrared thermal imager, and the values ​​were collated to obtain a training data set. The experiment was conducted indoors with a wind speed of 0 m / s and constant temperature and humidity.

[0054] The composite insulator temperature rise test platform includes a high-voltage power supply system, a high-voltage measurement system, a pressurizing device, a composite insulator suspension vehicle, a composite insulator, an optical fiber thermometer, an infrared thermometer and an infrared thermal imager;

[0055] The experiment includes the following steps:

[0056] First, hang the composite insulator vertically on the composite insulator hanging vehicle, fix its top, measure the vertical distance between the composite insulator and the ground, and ensure that the distance between any part of the composite insulator and the ground object is not less than 0.7m;

[0057] An optical fiber thermometer is installed outside, and three probes of the optical fiber thermometer are glued to the surface of the composite insulator with insulating thermal conductive double-sided adhesive, and the fourth probe is placed in the air to measure the ambient temperature;

[0058] Connect the high-voltage end of the composite insulator to the live wire end of the pressurizing device, and the top end to the bottom line to simulate the actual operating state; set the voltage to 1.1 times the actual operating voltage, and confirm that the safety guardrail and grounding rod are correctly placed before pressurizing;

[0059] After pressurization, use an optical fiber thermometer to measure the actual temperature of the composite insulator, and obtain the infrared temperature through an infrared thermometer and an infrared thermal imager; according to the requirements of the UAV infrared inspection, change the distance and viewing angle between the infrared thermometer and the infrared thermal imager to perform measurements respectively, and record the temperature data of the infrared thermometer, infrared thermal imager and optical fiber thermometer;

[0060] Replace the new composite insulator and repeat the above steps, and organize the temperature data obtained by the infrared thermometer, infrared thermal imager and optical fiber thermometer to obtain a training data set.

[0061] S2. Based on BP neural network, the initial model of infrared temperature measurement calibration of composite insulator is constructed, specifically:

[0062] Based on the BP neural network, an initial model for infrared temperature measurement and calibration of composite insulators is constructed, including an input layer, one or several hidden layers and an output layer connected in series in sequence; each layer of the input layer, one or several hidden layers and the output layer includes several neurons, and the layers are fully connected; the input data is transmitted layer by layer starting from the input layer, and the prediction results are finally generated in the output layer through the calculation of the neurons in each layer.

[0063] S3. Using the training data set obtained in step S1, combined with the GA genetic algorithm, deep learning training is performed on the initial model of composite insulator infrared temperature measurement calibration obtained in step S2 to obtain a composite insulator infrared temperature measurement calibration model, specifically:

[0064] Firstly, the GA genetic algorithm is used to optimize the weights and biases of the neural network, and then the initial model of infrared temperature measurement calibration of composite insulators is trained by forward propagation and back propagation to obtain the infrared temperature measurement calibration model of composite insulators.

[0065] Among them, the optimization of the weights and biases of the neural network using the GA genetic algorithm includes the following steps:

[0066] 1) Set genetic algorithm parameters: set population size, crossover probability p_c and mutation probability p_m;

[0067] 2) Population initialization: Randomly generate the initial population, where each individual represents a set of weights and bias combinations of the neural network; the initial weights and biases are generated through uniform distribution or normal distribution to ensure that the population is diverse and avoid premature convergence.

[0068] 3) Define the fitness function and calculate the fitness value: The fitness function is used to evaluate the performance of each individual. The fitness is defined based on the error of the neural network on the validation set. The higher the fitness value, the smaller the model error of the individual, and therefore the better the performance.

[0069] 4) Selection operation: Use roulette selection and tournament selection methods to select individuals with higher fitness from the current population to enter the next generation;

[0070] 5) Crossover operation: According to the crossover probability p_c obtained in step 1, randomly pair the selected individuals and use single-point crossover, double-point crossover or uniform crossover methods to exchange genes and generate new individuals;

[0071] 6) Mutation operation: According to the mutation probability p_m obtained in step 1, some genes (i.e. weights or biases) in the new individuals are randomly adjusted slightly to increase the diversity of the population and prevent the algorithm from falling into a local optimal solution;

[0072] 7) Fitness evaluation and selection: The fitness of the newly generated individuals is evaluated again, and the individuals of the parent generation are selected. Usually, the elite retention strategy is used to retain the individuals with the highest fitness to ensure the quality of the solution;

[0073] 8) Stopping of genetic algorithm: When the set maximum number of generations is reached or the fitness change within a number of generations is less than the preset threshold, the genetic algorithm stops and the individual with the highest fitness is selected as the optimal solution, representing the final optimized neural network weights and biases.

[0074] The forward propagation is specifically to input the data in the training data set into the initial model of infrared temperature measurement calibration of composite insulators to obtain the output value, and then calculate the error between the output value and the corresponding true value in the training data set;

[0075] Back propagation calculates the error between the predicted result and the actual value, and adjusts the weights and biases of the model neural network through the gradient descent method to minimize the output error. The weight change of back propagation is expressed by the following formula:

[0076]

[0077] Among them, Δw kj is the weight change of the back-propagation output layer; η is the learning rate; E is the total error; w kj is the weight from the jth hidden layer neuron to the kth output layer neuron; u′ k is the output of the kth neuron in the output layer;

[0078] The change in the output layer neuron bias is expressed using the following formula:

[0079]

[0080] Where Δb k is the change in the output layer neuron bias; b k is the bias of the kth neuron in the output layer;

[0081] The transformation of the input layer neuron weights is expressed using the following formula:

[0082]

[0083] Among them, Δw ji is the weight change from the input layer neurons to the hidden layer neurons; w ji is the weight from the jth hidden layer neuron to the ith output layer neuron; h′ j is the output of the jth hidden layer neuron;

[0084] The change in the bias of the hidden layer neurons is expressed using the following formula:

[0085]

[0086] Where Δb j is the transformation amount of the hidden layer neuron bias; b j is the bias of the jth neuron in the hidden layer;

[0087] S4. Using the composite insulator infrared temperature measurement calibration model obtained in step S3, actual composite insulator infrared temperature measurement compensation calibration is performed.

[0088] The method of the present invention is further described below in conjunction with an embodiment:

[0089] According to the method of the present invention, an experimental platform is built to carry out infrared temperature measurement experiments. The WT-YH-400 optical fiber temperature measuring device is used to measure the temperature of the surface of the composite insulator. The temperature measurement accuracy is ±0.5°C. The temperature value obtained by the optical fiber temperature measuring device can be used as an accurate value to judge the measurement accuracy of other temperature measurement methods, and as a data set input for subsequent infrared temperature measurement correction methods. This embodiment uses two types of infrared temperature measuring instruments, namely infrared laser thermometers and infrared thermal imagers. The infrared laser thermometer mainly performs point temperature measurement through lasers and has no image display function, while the infrared thermal imager can generate infrared images to display the specific temperature distribution of the measured object. The infrared temperature measurement results of these two types will be corrected later. The specific parameters of the three temperature measuring instruments are detailed in Table 1.

[0090] Table 1 Infrared temperature measuring instrument parameters

[0091]

[0092] The composite insulator in this embodiment is a 110kV FXBW4-110 / 70 composite insulator with a voltage equalizing ring. According to the requirements of GB / T 4585-2004, the surfaces of three normal insulators are cleaned with tap water (heated to 50°C) with trisodium phosphate or other cleaning agents, and then rinsed with tap water. Then, two composite insulators with surface dirt are artificially constructed by the solid coating method. 40g of loess, 1000g of tap water and an appropriate amount of NaCl are mixed, and the dirty liquid is covered on the two insulators after cleaning and drying by the pouring method, and placed in the environment to dry naturally for 24 hours.

[0093] According to the steps of the infrared temperature measurement experiment in the method of the present invention, the temperature measurement results of the composite insulator at different distances and angles by the infrared thermometer and the infrared thermal imager are obtained. The results are as follows Figure 2 , Figure 3 As shown. From the temperature measurement results, it can be seen that as the temperature measurement distance of the infrared thermometer and the infrared thermal imager increases, the temperature difference between the measured temperature and the accurate temperature basically shows a trend of increasing, and the temperature measurement accuracy will also be affected by changing the pitch angle of the temperature measurement. The larger the pitch angle, the greater the temperature measurement error. The maximum temperature difference of the infrared thermometer is 1.6°C. The infrared thermal imager is 1.1°C. From the infrared temperature measurement principle and experiments, it can be seen that distance and viewing angle are important factors affecting the accuracy of infrared temperature measurement. In order to ensure the effectiveness of the infrared temperature measurement value, a certain temperature compensation method must be used to correct the influence of these influencing factors on the infrared temperature measurement.

[0094] The infrared temperature measurement calibration model of composite insulators obtained by the method of the present invention is used together with the statistical regression correction methods commonly used in the prior art, including linear, nonlinear and exponential function fitting methods, to calibrate the results of the infrared thermometer and infrared thermal imager, and compare the correction effects of different methods on these two infrared temperature measurement devices.

[0095] The commonly used statistical regression models currently include linear, nonlinear, exponential and logarithmic function fitting. However, since the input viewing angle value has negative values, the logarithmic and power function fitting is not considered in the present invention. The present invention uses the linear, nonlinear and exponential function fitting methods to perform fitting analysis on the infrared temperature measurement data set. The fitting results are shown in Table 2. 2 represents the correlation coefficient of the fitting curve. The closer it is to 1, the stronger the model explanatory ability and the better the fitting effect. r1 is the corrected temperature value of the infrared thermal imager, d is the distance, θ is the viewing angle, T m is the temperature value measured by the infrared thermal imager, T r2 is the corrected temperature value of the infrared thermometer, T m2 It is the temperature value measured by infrared thermometer.

[0096] Table 2 Fitting correction results

[0097]

[0098] The comparison diagram of the correction effect of the correction method of the present invention in this embodiment is as follows: Figure 4 The comparison of the correction effect of the method of the present invention and the commonly used statistical regression model used in this embodiment is shown in the figure below. Figure 5 , Figure 6 The specific correction data are shown in Table 3.

[0099] Table 3 Comparison of correction effect data

[0100]

[0101] It can be seen from the correction results that before correction, the temperature difference between the infrared thermometer and the infrared thermal imager was large. After correction by different algorithms, the temperature difference of infrared temperature measurement was improved. Different algorithms have different correction effects on different temperature measuring instruments. Linear, nonlinear and exponential statistical regression correction methods have certain correction effects on infrared temperature measurement, but the overall effect is not good, especially when processing complex data, the generalization ability of statistical regression correction is weak. The correction effect of the method of the present invention is significantly better than the traditional statistical regression correction method. Specifically, after the temperature measurement accuracy of the infrared thermometer is corrected by the method of the present invention, the average error is 1.23%. The correction effect of the method of the present invention is relatively stable. Whether it is an infrared thermometer or an infrared thermal imager, the correction accuracy of the deep learning model is significantly higher than other methods. Therefore, among all the correction methods, the model proposed by the method of the present invention performs the best and is suitable for high-precision correction applications of infrared temperature measurement. The specific correction effects are as follows: Figure 7 shown.

Claims

1. A composite insulator infrared temperature measurement compensation calibration method, characterized in that: The following steps are involved: S1. Obtain infrared temperature measurement calibration data of composite insulators to obtain a training data set; S2. Based on BP neural network, the initial model of infrared temperature measurement calibration of composite insulator is constructed; S3. Using the training data set obtained in step S1, combined with the GA genetic algorithm, the initial model of the composite insulator infrared temperature measurement calibration obtained in step S2 is trained to obtain a composite insulator infrared temperature measurement calibration model; S4. Using the composite insulator infrared temperature measurement calibration model obtained in step S3, actual composite insulator infrared temperature measurement compensation calibration is performed.

2. The infrared temperature measurement compensation calibration method for composite insulators according to claim 1 is characterized in that: Step S1 specifically includes: constructing a composite insulator temperature rise experimental platform, conducting experiments on composite insulators to obtain measured values ​​and true values ​​of the composite insulators at different measurement distances and angles by an infrared thermal imager, and arranging them to obtain a training data set.

3. The infrared temperature measurement compensation calibration method for composite insulators according to claim 2 is characterized in that: The experiment was carried out indoors with a wind speed of 0m / s and constant temperature and humidity. The composite insulator temperature rise experimental platform includes a high-voltage power supply system, a high-voltage measurement system, a pressurizing device, a composite insulator suspension vehicle, a composite insulator, an optical fiber thermometer, an infrared thermometer and an infrared thermal imager. The experiment includes the following steps: First, hang the composite insulator vertically on the composite insulator hanging vehicle, fix its top, measure the vertical distance between the composite insulator and the ground, and ensure that the distance between any part of the composite insulator and the ground object is not less than 0.7m; An optical fiber thermometer is installed outside, and three probes of the optical fiber thermometer are glued to the surface of the composite insulator with insulating thermal conductive double-sided adhesive, and the fourth probe is placed in the air to measure the ambient temperature; Connect the high-voltage end of the composite insulator to the live wire end of the pressurizing device, and the top end to the bottom line to simulate the actual operating state; set the voltage to 1.1 times the actual operating voltage, and confirm that the safety guardrail and grounding rod are correctly placed before pressurizing; After pressurization, use an optical fiber thermometer to measure the actual temperature of the composite insulator, and obtain the infrared temperature through an infrared thermometer and an infrared thermal imager; according to the requirements of the UAV infrared inspection, change the distance and viewing angle between the infrared thermometer and the infrared thermal imager to perform measurements respectively, and record the temperature data of the infrared thermometer, infrared thermal imager and optical fiber thermometer; Replace the new composite insulator and repeat the above steps, and organize the temperature data obtained by the infrared thermometer, infrared thermal imager and optical fiber thermometer to obtain a training data set.

4. The infrared temperature measurement compensation calibration method for composite insulators according to claim 3 is characterized in that: Step S2 is specifically as follows: based on the BP neural network, an initial model for infrared temperature measurement calibration of composite insulators is constructed, including an input layer, one or several hidden layers and an output layer connected in series in sequence; in the input layer, one or several hidden layers and the output layer, each layer includes several neurons, and the layers are fully connected; the input data is transmitted layer by layer starting from the input layer, and the prediction results are finally generated in the output layer through the calculation of the neurons in each layer.

5. The infrared temperature measurement compensation calibration method for composite insulators according to claim 4 is characterized in that: Step S3 is specifically as follows: firstly, the weight and bias of the neural network are optimized by using the GA genetic algorithm, and then the initial model of the composite insulator infrared temperature measurement calibration is trained by using forward propagation and back propagation to obtain the composite insulator infrared temperature measurement calibration model.

6. The infrared temperature measurement compensation calibration method for composite insulators according to claim 5, characterized in that: The optimization of the weights and biases of the neural network using the GA genetic algorithm includes the following steps: 1) Set genetic algorithm parameters: set population size, crossover probability p_c and mutation probability p_m; 2) Population initialization: Randomly generate an initial population, where each individual represents a set of weights and bias combinations of a neural network; the initial weights and biases are generated through uniform distribution or normal distribution; 3) Define the fitness function and calculate the fitness value: The fitness function is used to evaluate the performance of each individual. The fitness is defined based on the error of the neural network on the validation set. The higher the fitness value, the smaller the model error of the individual, and therefore the better the performance. 4) Selection operation: using roulette wheel selection and tournament selection methods, select individuals with fitness higher than the preset value from the current population to enter the next generation; 5) Crossover operation: According to the crossover probability p_c obtained in step 1, randomly pair the selected individuals and use single-point crossover, two-point crossover or uniform crossover method to exchange genes to generate new individuals; 6) Mutation operation: According to the mutation probability p_m obtained in step 1, the genes in the new individuals are randomly adjusted to increase the diversity of the population and prevent the algorithm from falling into a local optimal solution; 7) Fitness evaluation and selection: The fitness of the newly generated individuals is evaluated again, and the individuals of the parent generation are selected. The elite retention strategy is used to retain the individuals with the highest fitness to ensure the quality of the solution; 8) Stopping of genetic algorithm: When the set maximum number of generations is reached or the fitness change within a number of generations is less than the preset threshold, the genetic algorithm stops and the individual with the highest fitness is selected as the optimal solution, representing the final optimized neural network weights and biases.

7. The infrared temperature measurement compensation calibration method for composite insulators according to claim 6 is characterized in that: The forward propagation is specifically to input the data in the training data set into the initial model of infrared temperature measurement calibration of composite insulators to obtain the output value, and then calculate the error between the output value and the corresponding true value in the training data set; Back propagation calculates the error between the predicted result and the actual value, and adjusts the weights and biases of the model neural network through the gradient descent method to minimize the output error. The weight change of back propagation is expressed by the following formula: Among them, Δw kj is the weight change of the back-propagation output layer; η is the learning rate; E is the total error; w kj is the weight from the jth hidden layer neuron to the kth output layer neuron; u′ k is the output of the kth neuron in the output layer; The change in the output layer neuron bias is expressed using the following formula: Where Δb k is the change in the output layer neuron bias; b k is the bias of the kth neuron in the output layer; The transformation of the input layer neuron weights is expressed using the following formula: Among them, Δw ji is the weight change from the input layer neurons to the hidden layer neurons; w ji is the weight from the jth hidden layer neuron to the ith output layer neuron; h′ j is the output of the jth hidden layer neuron; The change in the bias of the hidden layer neurons is expressed using the following formula: Where Δb j is the transformation amount of the hidden layer neuron bias; b j is the bias of the jth neuron in the hidden layer.