A composite insulator electric field distortion detection method based on electroluminescence effect

Abstract

The application provides a composite insulator electric field distortion detection method based on electroluminescence effect, first, silicon rubber is used as a base material, a formula is optimized, and a preparation process is prepared, an electroluminescence composite coating material which is highly sensitive to electric field intensity and has excellent adhesion and environmental stability is prepared. Secondly, an unmanned aerial vehicle is used to carry a high-resolution camera, and comprehensive and clear insulator surface images are obtained by combining optimized flight route design. Then, the image data is predicted by multivariate regression algorithm, the local and global characteristics of the electric field distribution are accurately captured, and the modeling error is reduced. Finally, by comparing the actual measured electric field distribution curve with the theoretical curve, the electric field distortion is analyzed, and the potential defects are diagnosed by using the feature matching technology. The application can maintain high detection accuracy in different environments, greatly improves the efficiency and precision of electric field distribution detection, and provides reliable technical support for the maintenance and safe operation of power equipment.

Description

Technical Field

[0001] The present invention relates to the technical field of power equipment monitoring, and in particular to a composite insulator electric field distortion detection system and an intelligent diagnosis method thereof based on electroluminescence effect and intelligent algorithm. Background Art

[0002] With the further development of my country's ultra-high voltage (UHV) power grid, high-voltage transmission lines will utilize an increasing number of composite insulators. Due to design flaws, material degradation, and environmental factors, composite insulators may develop defects, causing localized electric field distortion and threatening the safe operation of the power system. Traditional electric field distortion detection technology primarily relies on sensors to measure electric field strength and then analyze the electric field distribution using complex mathematical algorithms. However, the complex structure of composite insulators is affected by various factors, such as surface contamination, mechanical stress, and temperature variations. This makes traditional technology often unable to accurately capture changes in electric field distortion, resulting in low monitoring accuracy. Furthermore, technologies that rely on electric field sensors to detect electric field distortion in composite insulators are labor-intensive and require a high level of operator skill. Therefore, a simple, intuitive, and labor-saving self-display technology for composite insulator electric field distortion is urgently needed. Summary of the Invention

[0003] To address the shortcomings of the aforementioned existing technologies, the present invention introduces electroluminescence and white display technology to achieve real-time monitoring of electric field distortion. This allows for the timely detection and resolution of potential composite insulator faults, enabling appropriate maintenance measures to be implemented, reducing the risk of further failures and improving the reliability and safety of the power system. Furthermore, intelligent monitoring methods enable better remote monitoring and fault warning of the power system, enhancing the intelligence level of the power system and promoting the development and progress of the power industry.

[0004] The technical solution of the present invention is: a method for detecting electric field distortion of a composite insulator based on the electroluminescence effect, comprising the steps of:

[0005] Step 1, preparing an electroluminescent composite coating material, including a base material and a fluorescent material, using silicone rubber as the base material and ZnS:Cu as the fluorescent material;

[0006] Step 2: Utilize a drone equipped with a high-resolution camera, combined with optimized flight path design, to obtain comprehensive and clear images of the insulator surface. Algorithms are used to predict the image data, accurately capturing the local and global characteristics of the electric field distribution and reducing modeling errors.

[0007] Step 3: By comparing the actual measured electric field distribution curve with the theoretical curve, analyze the electric field distortion and use feature matching technology to diagnose potential defects.

[0008] Further, in step 1, ZnS:Cu is used as the fluorescent material, and the specific preparation and coating process of the ZnS:Cu composite coating material is as follows: first, weigh the required mass of silicone rubber material, pour it into a flask and heat it to 60°C to improve the flowability of the matrix, add the curing agent in proportion, then add the ZnS:Cu fluorescent material and barium titanate filler, then keep the heating temperature at 60°C unchanged, continue stirring for 1 h, then perform vacuum operation in a vacuum drying box until there are no bubbles on the surface, finally, use spray coating to coat the ZnS:Cu composite coating material on the surface of the composite insulator that has been cleaned and treated, ensure that the coating thickness is uniform and complete, avoid the generation of bubbles, void defects during the coating process, use appropriate coating speed and coating thickness control methods to ensure the quality of the coating, and after high-temperature curing, the final coated electroluminescent composite coating material composite insulator is obtained.

[0009] Further, the specific process of step 2 is as follows:

[0010] Image acquisition part: use a drone to carry a camera for shooting, choose an appropriate model of drone, carry a high-resolution camera, ensure that clear and high-quality images can be obtained during flight, and the camera can dynamically adjust the shooting angle, control the drone to fly according to the pre-designed route, adjust the distance between the drone and the composite insulator surface, control the attitude and camera angle, and shoot the image of the composite insulator surface;

[0011] Image processing and analysis part: mainly based on the luminescent image to deduce the insulator surface electric field distribution to provide basis for subsequent defect diagnosis, based on the principle that the stronger the field strength, the greater the luminescent intensity of the electroluminescent material, the relationship between field strength and luminescent intensity can be further obtained; first, denoising the collected image to eliminate possible interference and improve image quality and clarity, then using image processing software to convert the preprocessed image to a grayscale image and extract the grayscale value information of each pixel point; finally, establish an electric field size model based on grayscale value, convert the grayscale value of each pixel point according to the measured light intensity-field relationship to obtain the corresponding electric field size value.

[0012] Further, the specific steps of step 2 are as follows:

[0013] Step 2.1, drone setting

[0014] Choosing the right model of drone is the key to ensuring image acquisition quality, the following are the specific selection criteria:

[0015] Flight stability: the drone should have good flight stability to reduce image blurring caused by vibration or wind;

[0016] Load capacity: The UAV needs to have sufficient load capacity to carry high-resolution cameras and other sensors;

[0017] Endurance time: Choose a UAV with longer endurance time to ensure that it can complete the image acquisition task of the entire route;

[0018] Therefore, the selected UAV has high load capacity and long endurance time, and is equipped with an RTK module, which can achieve centimeter-level flight accuracy;

[0019] Next, fix the high-resolution camera on the UAV and ensure that the camera can dynamically adjust the shooting angle. The specific installation steps are as follows:

[0020] Choose a high-resolution camera and use a shockproof fixing bracket to securely install the camera at the bottom of the UAV, ensuring that it does not shake during flight. To ensure that the camera can dynamically adjust the shooting angle to adapt to different shooting needs, use a gimbal stabilizer to achieve this function;

[0021] Finally, the route is designed. According to the pre-designed route, control the UAV to fly autonomously, adjust the distance between the UAV and the composite insulator surface, control the attitude and camera angle, and take pictures of the composite insulator surface. The specific steps are as follows:

[0022] Use route planning software (DJI Terra) to pre-set the flight path to ensure that the UAV can cover all the composite insulator surfaces to be detected;

[0023] Flight parameter settings: Flight height (H): Set the appropriate flight height according to the height of the insulator and the focal length of the camera; Shooting interval (I): Set the shooting interval to ensure sufficient overlap between images, with a recommended value of 3 seconds;

[0024] Autonomous flight: Start the autonomous flight mode of the UAV and fly according to the predetermined route, while dynamically adjusting the camera angle to achieve the best shooting effect;

[0025] Data monitoring: Monitor the status of the UAV and the image acquisition during flight in real time to ensure the accuracy and completeness of the data;

[0026] Step 2.2, image acquisition and preprocessing;

[0027] The UAV flies along the predetermined route and takes pictures of the composite insulator surface. To ensure that the collected images can fully and accurately reflect the information of the composite insulator surface, the following settings are required:

[0028] Shooting angle: Set the pitch angle (θ) and yaw angle (φ) of the camera to obtain images at different angles;

[0029] Flight speed: Set the appropriate flight speed (v) to ensure image clarity;

[0030] Image capture: The drone flies according to the designed route and continuously captures images of the composite insulator surface during the flight;

[0031] Coverage: To ensure complete surface information, the image capture should cover all areas of the composite insulator, including the upper, middle and lower parts;

[0032] Image preprocessing includes denoising and grayscale conversion, aiming to improve image quality and extract useful information;

[0033] Denoising: Denoising is performed on the collected images to eliminate possible interference;

[0034] Among them, Gaussian filtering is a linear smoothing filtering method, which is often used to remove high-frequency noise in images. The formula of Gaussian filtering is:

[0035]

[0036] Where G(x, y) is the value of the Gaussian function at point (x, y), which usually represents the value of the pixel value of the image at point (x, y) after Gaussian filtering. x is the horizontal offset from the center of the Gaussian function, in pixels. y is the vertical offset from the center of the Gaussian function, in pixels. σ is the standard deviation, which determines the width of the Gaussian distribution (i.e., the degree of smoothness). e is the base of the natural logarithm.

[0037] Grayscale conversion: Convert the preprocessed image into a grayscale image and extract the grayscale value information of each pixel. The grayscale value calculation formula is:

[0038] G(x,y)=0.299·R(x,y)+0.587·G(x,y)+0.114·B(x,y)

[0039] Where R(x, y), G(x, y), and B(x, y) are the red, green, and blue component values ​​of the pixel point (x, y), respectively.

[0040] Step,2.3, electric field distribution modeling;

[0041] The relationship between the grayscale value at different positions and the applied voltage is shown below:

[0042]

[0043] Among them, I, R, G, and B are the grayscale value, red value, green value, and blue value of different positions (x, y) under voltage U, and all have values ​​0-1.

[0044] The relationship between the electric field strength and the applied voltage at the calibration point is as follows:

[0045] E(x0, y0) = kU (2)

[0046] Among them, k is the corresponding coefficient, which is obtained from simulation and experimental data.

[0047] The relationship between the electric field intensity and the grayscale value at the calibration point is as follows:

[0048] E(x0,y0)=f(I(x0,y0,U)) (3)

[0049] Where f is the relationship function.

[0050] Therefore, the actual electric field strength can be inferred based on the grayscale value, as shown below:

[0051] E(x,y,U1)=f(I(x,y,U1)) (4)

[0052] Wherein, U1 is the corresponding voltage value at the position (x, y).

[0053] Furthermore, the specific process of step 3 is:

[0054] Step 3.1: First, the electric field value of each pixel is plotted into an electric field distribution map to intuitively display the electric field distribution on the surface of the composite insulator;

[0055] Step 3.2: The actual predicted electric field distribution is then compared with the theoretical electric field distribution. Based on the difference analysis, the presence of electric field distortion, the degree and range of the distortion are determined. The location of the electric field distortion is matched with the known electric field distortion characteristics to identify the possible defect types and locations, including internal pores and partial discharge, and thus perform defect diagnosis.

[0056] Step 3.3, actuator operation: When electric field distortion or defects are detected, the system sends an alarm signal, records the location and type of the defect, takes corresponding treatment measures according to the preset alarm level, and notifies maintenance personnel to carry out inspection and replacement.

[0057] Further, the specific steps of steps 3.2-3.3 are as follows:

[0058] Step 3.2, data preparation: grayscale image conversion to graph structure. Convert the preprocessed grayscale image data into a graph structure. A node is each pixel in the image, and an edge is the connection between adjacent pixels. If the image size is M×N, the number of nodes is M×N.

[0059] Graph structure representation:

[0060] G=(V,E)

[0061] Among them, V represents the node set, that is, the pixel points; E represents the edge set, that is, the relationship between adjacent pixel points;

[0062] Grayscale value conversion: In order to convert the grayscale value in the image into the corresponding electric field intensity value, it is necessary to use the experimentally measured light intensity-electric field relationship:

[0063] E(x0,y0)=f(I(x0,y0,U))

[0064] Next, the LSSVM-Adaboost model is used to train the above light intensity-electric field relationship;

[0065] Input feature, grayscale value: represents the brightness information of the pixel, denoted as I (x，y) ;

[0066] Pixel coordinates: represents the position of the pixel, recorded as (x, y);

[0067] Output data: target variable electric field intensity E;

[0068] LSSVM-Adaboost model introduction:

[0069] The objective function of the least squares support vector machine (LSSVM) is:

[0070]

[0071] Among them, w is the weight vector, b is the bias term, e is the error vector, γ is the regularization parameter, is the square of the deviation between the true value and the predicted value of sample i, N is the number of samples, and min is the minimum function;

[0072] Constraints:

[0073] y i =w T φ(x i )+b+e i , i=1,...,N

[0074] Among them, y i is the actual output value (true value) of the i-th sample, φ(x i ) represents the i-th sample x i The feature map of , i = 1, ..., N is the index of the sample;

[0075] Adaboost algorithm goal: iteratively adjust weights to enhance the model's attention to difficult-to-classify samples;

[0076] Weight update:

[0077]

[0078] Among them, w i is the sample weight at the i-th iteration, α i is the model weight at the i-th iteration, I is the indicator function, if the weak classifier h i The prediction result h for sample x i (x) is not equal to the true label y, then the function takes the value of 1, otherwise it takes the value of 0, h i is the weak classifier of the i-th iteration, y represents the true label of sample x,

[0079] LSSVM-Adaboost combined model, objective function and constraints:

[0080]

[0081] y i =w T φ(x i )+b+e i , i=1,…,N

[0082] Implementation steps:

[0083] Initialize weights:

[0084]

[0085] Training weak classifiers: For each round of iteration, use LSSVM to train a weak classifier and calculate the error e i And update the weights;

[0086] Model fusion: Use weighted voting to fuse the outputs of multiple weak classifiers;

[0087] Iterative update: Repeat the above process until the preset number of iterations or error threshold is reached;

[0088] After training, the model can predict the electric field intensity value of each pixel in the image;

[0089] Distribution comparison: The actual measured electric field distribution is compared with the theoretical electric field distribution. The comparison method uses image difference calculation. This patent uses the mean square error (MSE):

[0090]

[0091] Where N is the total number of pixels in the image, E measured (i) is the actual measured electric field strength value of the i-th pixel, E theoretical (i) is the electric field intensity value of the i-th pixel in the theoretical model;

[0092] Defect identification: Match the electric field distortion location with known electric field distortion features to identify possible defect types and locations, including internal pores and partial discharges.

[0093] Step 3.2, actuator operation;

[0094] Alarm signal: When electric field distortion or defects are detected, the system will issue an alarm signal and record the location and type of the defect. According to the preset alarm level, appropriate treatment measures will be taken;

[0095] Actuator operation is a crucial part of the system, ensuring the precise execution of drones in image acquisition and electric field distribution detection through automated control systems;

[0096] Actuators include motors, servos, and camera gimbals, while sensors include GPS modules, gyroscopes, accelerometers, and altimeters.

[0097] The formulas for adjusting the flight path and camera parameters are as follows:

[0098] v=v base +Δv

[0099] h=h base +Δh

[0100] θ=θ base +Δθ

[0101] ψ=ψ base +Δψ

[0102] Among them, v is the adjusted flight speed, h is the adjusted flight altitude, θ is the adjusted camera pitch angle, ψ is the adjusted heading angle, and v base 、h base ,θ base , ψ base are the basic flight data of flight speed, flight altitude, camera pitch angle, and heading angle, respectively. Δv, Δh, Δθ, and Δψ are the flight speed, flight altitude, camera pitch angle, and heading angle adjusted according to the prediction results.

[0103] Real-time monitoring: Real-time monitoring of the drone’s flight status and image acquisition to ensure the accuracy and completeness of the data.

[0104] Beneficial effects:

[0105] 1. The present invention comprehensively utilizes the electroluminescent effect and self-display technology, and realizes real-time monitoring of electric field distortion and defect diagnosis by preparing specific luminescent composite coating materials. At the same time, by using a camera mounted on an unmanned aerial vehicle for image acquisition, combined with image processing technology, it is possible to quickly and accurately obtain the electric field distribution information on the surface of the composite insulator. Ultimately, through the comparison of electric field distribution curves and defect diagnosis algorithms, automatic detection and alarm of electric field distortion of composite insulators are achieved, providing reliable protection for the safe and stable operation of the power system. The present invention not only simplifies the operating procedures of electric field detection of composite insulators, lowers the technical threshold, saves human resources, and provides important support for the intelligent management of power systems, promoting the development and progress of the power industry.

[0106] 2. By introducing the LSSVM-Adaboost algorithm and combining it with the automated control system of drones and high-resolution cameras, the present invention significantly improves the accuracy and stability of modeling the electric field distribution on the surface of composite insulators. First, through the flexible route design and multi-angle shooting of drones, comprehensive image data of the insulator surface is obtained to ensure the diversity and representativeness of the data. Secondly, the LSSVM-Adaboost algorithm is used to perform multivariate regression prediction on the image data, effectively capturing the local and global features in the electric field distribution and significantly reducing the modeling error. Experimental verification shows that the model can maintain a high prediction accuracy under different environmental conditions. In addition, the electric field distribution diagram after model optimization intuitively shows the electric field changes on the insulator surface, which helps to quickly identify electric field distortion and potential defects. Overall, the improvements of the present invention have greatly improved the efficiency and accuracy of electric field distribution detection, providing reliable technical support for the maintenance and safe operation of power equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0107] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without paying any creative work.

[0108] Figure 1 The specific preparation and coating process of ZnS:Cu composite coating material;

[0109] Figure 2 This is the flow chart of the intelligent detection system for composite insulator distorted electric field;

[0110] Figure 3 This is a flow chart of an intelligent feedback algorithm for smart power station data information according to the present invention. DETAILED DESCRIPTION

[0111] In the following, only certain exemplary embodiments are simply described. As those skilled in the art can recognize, the described embodiments can be modified in various different ways without departing from the spirit or scope of the embodiments of the present application. Therefore, the drawings and description are considered to be exemplary in nature and not limiting.

[0112] The disclosure below provides many different embodiments or examples for implementing different structures of the embodiments of the present application. In order to simplify the disclosure of the embodiments of the present application, the components and arrangements of specific examples are described below. Of course, they are only examples and the purpose is not to limit the embodiments of the present application. In addition, the embodiments of the present application can refer to the same reference numerals and / or reference letters in different examples, and such repetition is for the purpose of simplification and clarity, which does not indicate the relationship between the various embodiments and / or arrangements discussed.

[0113] In order to better understand the purpose, structure and function of the present application, the following describes in further detail an electric field distortion detection technology for composite insulators based on electroluminescence effect, in combination with the accompanying drawings.

[0114] As Figure 1-3 , in combination with actual needs, the invention achieves the following points: First, a material with high sensitivity and stability is prepared, which can produce obvious luminescence effect under the action of electric field, has good adhesion, can be firmly coated on the surface of composite insulators, and has the characteristics of high temperature resistance, cold resistance, moisture resistance, etc. to meet the application needs of composite insulators in different environmental conditions. Second, a suitable image processing technology needs to be developed, which can accurately back-solve the electric field distribution according to the luminescence signal generated by the material when the electric field is distorted, so as to realize the defect diagnosis of composite insulators. Finally, an online monitoring and alarm system is established, which can respond to the changes of the electric field distortion on the surface of composite insulators in time. When the monitoring device detects that the electric field distortion exceeds the preset threshold, the alarm mechanism can be automatically triggered to inform the relevant personnel for repair and maintenance in time, so as to avoid potential failure risks.

[0115] 1. The electric field distortion self-display technology for composite insulators based on electroluminescence effect mainly includes two parts of electroluminescence composite coating material preparation and coating and composite insulator distortion electric field intelligent detection system.

[0116] 2. Furthermore, the electroluminescent composite coating material mainly includes a matrix material and a fluorescent material. To ensure that the coating has good adhesion and environmental stability, the present invention uses a silicone rubber material as the matrix material; the fluorescent material is the core of the electroluminescent composite coating material, which is mostly a sulfide doped with metal ions. The present invention uses ZnS:Cu as the fluorescent material. Under different conditions, the ZnS:Cu material can emit two different colors of light. When copper ions replace zinc ions in the ZnS lattice, the material emits green light, and when copper ions are located in the lattice gap, the light emission color is blue. To ensure the luminous intensity and dielectric properties of the material, in the preparation of the electroluminescent composite coating material, During the process, barium titanate is added as a supplementary filler. The essence of the luminescence phenomenon of the ZnS:Cu composite coating material is that the electric field affects the distribution of electron energy levels inside the material. Under the action of the electric field, some electrons will be excited to a higher energy level, forming an excited state. When these excited electrons return to the ground state, they release energy and generate photons, thereby producing luminescence. Depending on the carrier, the luminescence phenomenon is divided into two types: one is charge polarization luminescence, in which the external electric field is lower than the discharge starting field strength, and the other is gas discharge induced luminescence, in which the external electric field is higher than the discharge starting field strength, and the charge generated by the gas discharge accumulates on the coating surface, further emitting light in the form of radiation transition.

[0117] The specific preparation and coating process of ZnS:Cu composite coating material is as follows: First, weigh the required mass of silicone rubber material, pour it into a flask and heat it to 60°C to improve the fluidity of the matrix, add curing agent in proportion, and then add ZnS:Cu fluorescent material and barium titanate filler. Then, keep heating at 60°C and continue stirring for 1 hour. After that, enter the vacuum drying oven for vacuum operation until there are no bubbles on the surface. Finally, use spraying to coat the ZnS:Cu composite coating material on the surface of the cleaned and treated composite insulator to ensure that the coating thickness is uniform and complete. During the coating process, defects such as bubbles and gaps should be avoided. Appropriate coating speed and coating thickness control methods can be used to ensure the coating quality. After high-temperature curing, the composite insulator finally coated with electroluminescent composite coating material is obtained.

[0118] 3. Further development of composite insulator distorted electric field intelligent detection system The composite insulator distorted electric field intelligent detection system includes three parts: image acquisition, image processing and analysis, and electric field distortion diagnosis;

[0119] The image acquisition part adopts the solution of using a drone equipped with a camera for shooting. The appropriate model of drone is selected, equipped with a high-resolution camera to ensure that clear and high-quality images can be obtained during the flight. At the same time, the camera can dynamically adjust the shooting angle. According to the pre-designed route, the drone is controlled to fly autonomously, adjust the distance between the drone and the surface of the composite insulator, and control the posture and camera angle to capture images of the composite insulator surface.

[0120] The image processing and analysis part mainly derives the electric field distribution on the surface of the insulator based on the luminescence image, thereby providing a basis for subsequent defect diagnosis. Based on the principle that the greater the field intensity, the greater the luminescence intensity of the electroluminescent material, the relationship between field intensity and luminescence intensity can be further obtained. First, the collected image is denoised to eliminate possible interference and improve image quality and clarity. Then, using image processing software, the pre-processed image is converted into a grayscale image, and the grayscale value information of each pixel is extracted. Finally, an electric field size model based on grayscale value is established. According to the experimentally measured light intensity-electric field relationship, the grayscale value of each pixel is converted to obtain the corresponding electric field size value.

[0121] The relationship between the grayscale value at different positions and the applied voltage is shown below:

[0122]

[0123] Where I, R, G, and B are the grayscale value, red value, green value, and blue value at different positions (x, y) under voltage U, respectively, and all range from 0 to 1.

[0124] The relationship between the electric field strength and the applied voltage at the calibration point is as follows:

[0125] E(x0, y0) = kU

[0126] Where k is the corresponding coefficient obtained from simulation and experimental data, and E is the electric field intensity at the calibration point;

[0127] The relationship between the electric field intensity and the grayscale value at the calibration point is as follows:

[0128] E(x0,y0)=f(I(x0,y0,U))

[0129] Where, f is the relationship function;

[0130] Therefore, the actual electric field strength can be inferred based on the grayscale value, as shown below:

[0131] E(x,y,U1)=f(I(x,y,U1))

[0132] The electric field distortion diagnosis part mainly includes the drawing of the electric field distribution curve on the insulator surface, curve analysis and comparison, and defect diagnosis. First, the electric field magnitude value of each pixel point is drawn into an electric field distribution map to intuitively display the electric field distribution on the composite insulator surface. Then, the actual measured electric field distribution map is compared with the theoretical electric field distribution. Based on the difference analysis, it is determined whether there is electric field distortion, as well as the degree and range of the distortion. The electric field distortion position is matched with the known electric field distortion characteristics to identify possible defect types and locations, including internal pores and partial discharges, so as to perform defect diagnosis. When electric field distortion or defects are detected, the system sends an alarm signal, records the location and type of the defect, and takes corresponding treatment measures according to the preset alarm level, notifying maintenance personnel to carry out inspection and replacement.

[0133] 4. To more accurately diagnose electric field distortion and defects, drones equipped with high-resolution cameras were used, combined with optimized flight paths, to obtain comprehensive and clear images of the insulator surface. Algorithms were used to predict the image data, accurately capturing both local and global characteristics of the electric field distribution and reducing modeling errors. The specific steps are as follows:

[0134] Step 1: Drone settings;

[0135] Choosing the right drone model is key to ensuring image acquisition quality. The following are specific selection criteria and models:

[0136] Flight stability: The drone should have good flight stability to reduce image blur caused by vibration or wind;

[0137] Payload capacity: The drone needs to have sufficient payload capacity to carry high-resolution cameras and other sensors;

[0138] Flight time: Choose a drone with a long flight time to ensure it can complete the image acquisition mission for the entire route;

[0139] Therefore, the drone model chosen is the DJI Matrice 300RTK, which has a high payload capacity and long flight time, and is equipped with an RTK module, which can achieve centimeter-level flight accuracy;

[0140] Next, fix the high-resolution camera on the drone and ensure that the camera can dynamically adjust the shooting angle. The specific installation steps are as follows:

[0141] The high-resolution camera chosen was the Sony Alpha 7R IV, which boasts a 61-megapixel resolution and provides clear images. A shock-resistant mounting bracket secures the camera to the bottom of the drone, preventing it from shaking during flight. A gimbal stabilizer was used to ensure the camera's ability to dynamically adjust shooting angles to suit different shooting requirements.

[0142] The final step is to design the route. According to the pre-designed route, the drone is controlled to fly autonomously, the distance between the drone and the composite insulator surface is adjusted, the attitude and camera angle are controlled, and images of the composite insulator surface are captured. The specific steps are as follows:

[0143] Use flight planning software (DJI Terra) to pre-set the flight path to ensure that the drone can cover all composite insulator surfaces to be inspected;

[0144] Flight parameter settings: Flight altitude (H): Set the appropriate flight altitude based on the height of the insulator and the focal length of the camera; Shooting interval (I): Set the shooting interval to ensure sufficient overlap between images. The recommended value is 3 seconds.

[0145] Autonomous flight: Activate the drone's autonomous flight mode and fly along a predetermined route, dynamically adjusting the camera angle for optimal filming.

[0146] Data monitoring: During the flight, the drone's status and image acquisition are monitored in real time to ensure the accuracy and integrity of the data;

[0147] Step 2: Image acquisition and preprocessing;

[0148] The drone flies along a predetermined route to capture images of the composite insulator surface. To ensure that the captured images fully and accurately reflect the composite insulator surface information, the following settings are required:

[0149] Shooting angle: Set the camera's pitch angle (θ) and yaw angle (φ) to obtain images at different angles;

[0150] Flight speed: Set the appropriate flight speed (v) to ensure image clarity;

[0151] Image capture: The drone flies according to the designed route and continuously captures images of the composite insulator surface during the flight;

[0152] Coverage: To ensure complete surface information, the image capture should cover all areas of the composite insulator, including the upper, middle and lower parts;

[0153] Image preprocessing includes denoising and grayscale conversion, aiming to improve image quality and extract useful information;

[0154] Denoising: Denoising is performed on the collected images to eliminate possible interference;

[0155] Among them, Gaussian filtering is a linear smoothing filtering method, which is often used to remove high-frequency noise in images. The formula of Gaussian filtering is:

[0156]

[0157] Where G(x, y) is the value of the Gaussian function at point (x, y), which usually represents the value of the pixel value of the image at point (x, y) after Gaussian filtering. x is the horizontal offset from the center of the Gaussian function, in pixels. y is the vertical offset from the center of the Gaussian function, in pixels. σ is the standard deviation, which determines the width of the Gaussian distribution (i.e., the degree of smoothness). e is the base of the natural logarithm.

[0158] Grayscale conversion: Convert the preprocessed image into a grayscale image and extract the grayscale value information of each pixel. The grayscale value calculation formula is:

[0159] G(x,y)=0.299·R(x,y)+0.587·G(x,y)+0.114·B(x,y)

[0160] Where R(x, y), G(x, y), and B(x, y) are the red, green, and blue component values ​​of the pixel point (x, y), respectively;

[0161] Step 3. First, data preparation: grayscale image is converted into a graph structure. The preprocessed grayscale image data is converted into a graph structure. A node is each pixel in the image, and an edge is the connection between adjacent pixels. If the image size is M×N, the number of nodes is M×N.

[0162] Graph structure representation:

[0163] G=(V,E)

[0164] Among them, V represents the node set, that is, the pixel points; E represents the edge set, that is, the relationship between adjacent pixel points;

[0165] Grayscale value conversion: In order to convert the grayscale value in the image into the corresponding electric field intensity value, it is necessary to use the experimentally measured light intensity-electric field relationship:

[0166] E(x0,y0)=f(I(x0,y0,U))

[0167] Then the LSSVM-Adaboost model is used to train the above light intensity-electric field relationship;

[0168] Input feature, grayscale value: represents the brightness information of the pixel, denoted as I (x，y) ;

[0169] Pixel coordinates: represents the position of the pixel, recorded as (x, y);

[0170] LSSVM-Adaboost model introduction:

[0171] The objective function of the least squares support vector machine (LSSVM) is:

[0172]

[0173] Among them, w is the weight vector, b is the bias term, e is the error vector, γ is the regularization parameter, is the square of the deviation between the true value and the predicted value of sample i, N is the number of samples, and min is the minimum function;

[0174] Constraints:

[0175] y i =w T φ(x i )+b+e i , i=1,…,N

[0176] Among them, y i is the actual output value (true value) of the i-th sample, φ(x i ) represents the i-th sample x i The feature map of , i = 1, ..., N is the index of the sample;

[0177] Adaboost algorithm goal: iteratively adjust weights to enhance the model's attention to difficult-to-classify samples;

[0178] Weight update:

[0179]

[0180] Among them, w i is the sample weight at the i-th iteration, α i is the model weight at the i-th iteration, I is the indicator function, if the weak classifier h i The prediction result h for sample x i (x) is not equal to the true label y, then the function takes the value of 1, otherwise it takes the value of 0, h i is the weak classifier of the i-th iteration, y represents the true label of sample x,

[0181] LSSVM-Adaboost combined model, objective function and constraints:

[0182]

[0183] y i =w T φ(x i )+b+e i , i=1,...,N

[0184] Implementation steps:

[0185] Initialize weights:

[0186]

[0187] Training weak classifiers: For each round of iteration, use LSSVM to train a weak classifier and calculate the error e i And update the weights;

[0188] Model fusion: Use weighted voting to fuse the outputs of multiple weak classifiers;

[0189] Iterative update: Repeat the above process until the preset number of iterations or error threshold is reached;

[0190] After training, the model can predict the electric field intensity value of each pixel in the image;

[0191] Step 4: Generation and diagnosis of electric field distribution map;

[0192] Distribution diagram comparison: The actual measured electric field distribution diagram is compared with the theoretical electric field distribution. The comparison method uses image difference calculation. The present invention selects the mean square error (MSE):

[0193]

[0194] Where N is the total number of pixels in the image, E measured (i) is the actual measured electric field strength value of the i-th pixel, E theoretical (i) is the electric field intensity value of the i-th pixel in the theoretical model;

[0195] Defect identification: Match the electric field distortion location with known electric field distortion features to identify possible defect types and locations, including internal pores and partial discharges.

[0196] Alarm signal: When electric field distortion or defects are detected, the system will issue an alarm signal and record the location and type of the defect. According to the preset alarm level, appropriate treatment measures will be taken;

[0197] Step 5: Actuator operation;

[0198] Actuator operation is a crucial part of the system, ensuring the precise execution of drones in image acquisition and electric field distribution detection through automated control systems;

[0199] Actuators include motors, servos, and camera gimbals, while sensors include GPS modules, gyroscopes, accelerometers, and altimeters.

[0200] The formulas for adjusting the flight path and camera parameters are as follows:

[0201] v=v base +Δv

[0202] h=h base +Δh

[0203] θ=θ base +Δθ

[0204] ψ=ψ base +Δψ

[0205] Among them, v is the adjusted flight speed, h is the adjusted flight altitude, θ is the adjusted camera pitch angle, ψ is the adjusted heading angle, and v base 、h base ,θ base , ψ base are the basic flight data of flight speed, flight altitude, camera pitch angle, and heading angle, respectively. Δv, Δh, Δθ, and Δψ are the flight speed, flight altitude, camera pitch angle, and heading angle adjusted according to the prediction results.

[0206] Real-time monitoring: Real-time monitoring of the drone’s flight status and image acquisition to ensure the accuracy and completeness of the data.

[0207] Specific numerical examples:

[0208] Material preparation and coating process:

[0209] Base material: Weigh 50g of silicone rubber material, put it into a flask and heat it to 60℃ to improve fluidity;

[0210] Additives: Add curing agent in the ratio (10:1) and add 5g ZnS;

[0211] Fluorescent material and 2g barium titanate filler to ensure the material's luminescence performance and dielectric properties;

[0212] Stirring and vacuuming: Maintain heating at 60°C and continue stirring for 1 hour. Then place the mixed material in a vacuum drying oven and vacuum until there are no bubbles on the surface.

[0213] Coating operation:

[0214] Surface Treatment: Clean the surface of the composite insulator to ensure no stains and grease;

[0215] Spraying: Use spraying method to evenly coat ZnS composite coating material on the surface of the insulator, ensuring uniform thickness of the coating. Control the coating speed during spraying to avoid bubbles or voids;

[0216] Curing: After coating, place the insulator in a high-temperature environment of 150°C for 2 hours to ensure firm bonding of the coating with the substrate and stable luminescence performance.

[0217] Electric Field Distortion Diagnosis Case:

[0218] In some locations on the surface of the insulator, the predicted electric field strength is 130 kV / m, while the theoretical electric field strength should be 100 kV / m. By comparison, it is found that the electric field distortion amplitude is 30%. This area shows an abnormally high brightness on the electric field distribution map, suggesting that there may be internal porosity defects at this location.

[0219] The electric field distortion information detected by the system is as follows:

[0220] Location: 10 cm from the top of the insulator, left offset 2 cm;

[0221] Electric field distortion degree: 30%;

[0222] Possible defect type: internal porosity;

[0223] Alarm signal: yellow alert, prompting further inspection;

[0224] Using LSSVM-Adaboost algorithm to predict the electric field strength distribution on the surface of the composite insulator, experimental design:

[0225] Equipment: UAV: DJI Phantom 4;

[0226] Camera: Canon EOS R5 (high-resolution camera, 45MP);

[0227] Data collection:

[0228] Flight route design: set the flight height of the UAV to 50 meters;

[0229] Scheduled route: the UAV flies within a 30-meter range along the surface of the insulator, taking pictures covering the entire surface;

[0230] Image acquisition: take a picture every 1 meter, a total of 50 pictures;

[0231] The resolution of the collected images is 8192x5460 pixels;

[0232] Data preprocessing;

[0233] Image preprocessing: denoise each image and smooth it using a Gaussian filter;

[0234] Convert each image into a grayscale image and extract the grayscale value of each pixel;

[0235] LSSVM-Adaboost model training;

[0236] Input data: Feature selection: Gray value (G): the gray value of each pixel in the image; Pixel coordinates (x, y): the spatial position of each pixel;

[0237] Training data set: Each image is converted into a set of feature vectors (G, x, y);

[0238] There are 250,000 samples in total (50 images, 5000 pixels each);

[0239] Output data: Target variable: Electric field intensity (E): A linear regression model is established based on the relationship between light intensity and electric field measured experimentally. The electric field intensity range is 0-10kV / m.

[0240] Model training and validation:

[0241] LSSVM-Adaboost model: uses the least squares support vector machine (LSSVM) as the base learner, and Adaboost is used to weightedly combine multiple LSSVM models;

[0242] Use 10-fold cross validation to evaluate model performance;

[0243] Specific parameters:

[0244] LSSVM parameters: C = 1 (penalty parameter);

[0245] γ = 0.1 (kernel function parameter);

[0246] Adaboost parameters: number of base learners: 50;

[0247] Results and effect analysis;

[0248] Model output: Electric field distribution map generated by electric field strength prediction results. The predicted electric field strength in a specific area (x=25, y=30) is 6.5 kV / m. Accuracy assessment: By comparing with the actual measured electric field strength (6.3 kV / m), the prediction accuracy is 96%;

[0249] Effect summary:

[0250] Improved modeling accuracy: The LSSVM-Adaboost algorithm can effectively capture local and global features in the electric field distribution, significantly improving the accuracy of electric field distribution modeling;

[0251] Real-time monitoring capabilities:

[0252] Through the automated control system, the model can adjust the drone's flight path and camera angle in real time to obtain optimal image data, reducing errors caused by manual adjustments.

[0253] Rapid defect identification:

[0254] The generated electric field distribution map can intuitively display the electric field distribution, quickly identify the location of electric field distortion and potential defects, and reduce the risk of equipment failure.

[0255] It will be understood that the present invention is described by way of some embodiments, and it will be appreciated by those skilled in the art that various changes or equivalent substitutions may be made to these features and embodiments without departing from the spirit and scope of the present invention. In addition, under the teachings of the present invention, these features and embodiments may be modified to adapt to specific circumstances and materials without departing from the spirit and scope of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of this application are intended to be protected by the present invention.

Claims

1. A method for detecting electric field distortion of composite insulators based on electroluminescence effect, characterized in that: Including steps: Step 1, preparing an electroluminescent composite coating material, including a base material and a fluorescent material, using silicone rubber as the base material and ZnS:Cu as the fluorescent material; Step 2: Utilize a drone equipped with a high-resolution camera, combined with optimized flight path design, to obtain comprehensive and clear images of the insulator surface. Algorithms are used to predict the image data, accurately capturing the local and global characteristics of the electric field distribution and reducing modeling errors. Step 3: By comparing the measured electric field distribution curve with the theoretical curve, the electric field distortion is analyzed and potential defects are diagnosed using feature matching technology; The specific process of step 3 is: Step 3.1: First, the electric field value of each pixel is plotted into an electric field distribution map to intuitively display the electric field distribution on the surface of the composite insulator; Step 3.2: The actual predicted electric field distribution is then compared with the theoretical electric field distribution. Based on the difference analysis, the presence of electric field distortion, the degree and range of the distortion are determined. The location of the electric field distortion is matched with the known electric field distortion characteristics to identify the possible defect types and locations, including internal pores and partial discharge, and thus perform defect diagnosis. Step 3.3, actuator operation: When electric field distortion or defects are detected, the system issues an alarm signal, records the location and type of the defect, takes appropriate action based on the preset alarm level, and notifies maintenance personnel to perform inspection and replacement. The specific steps of step 3.2 are as follows: Step 3.2, data preparation: Convert the grayscale image to a graph structure. Convert the preprocessed grayscale image data to a graph structure, where a node is each pixel in the image and an edge is the connection between adjacent pixels. If the image size is M×N, then the number of nodes is M×N. The graph structure G represents: G=(V,E) Among them, V represents the node set, that is, the pixel points; E represents the edge set, that is, the relationship between adjacent pixel points; Grayscale value conversion: In order to convert the grayscale value in the image into the corresponding electric field intensity value, it is necessary to use the experimentally measured light intensity-electric field relationship: E(x0,y0)=f(I(x0,y0,U)) Then, the least squares support vector machine-adaptive boosting algorithm LSSVM-Adaboost model is used to train the above light intensity-electric field relationship; Input feature, grayscale value: represents the brightness information of the pixel, denoted as I (x,y) ; Pixel coordinates: represents the position of the pixel, recorded as (x, y); Output data: target variable electric field intensity E; Training weak classifiers: For each round of iteration, use LSSVM to train a weak classifier and calculate the error e i And update the weights; Model fusion: Use weighted voting to fuse the outputs of multiple weak classifiers; Iterative update: Repeat the above process until the preset number of iterations or error threshold is reached; After training, the model can predict the electric field intensity value of each pixel in the image; Distribution map comparison: The actual predicted electric field distribution map is compared with the theoretical electric field distribution. The comparison method uses image difference calculation. The present invention selects the mean square error (MSE): Where N is the total number of pixels in the image, E measured (i) is the actual measured electric field strength value of the i-th pixel, E theoretical (i) is the electric field intensity value of the i-th pixel in the theoretical model; Defect identification: Match the electric field distortion position with known electric field distortion characteristics to identify possible defect types and locations, including internal pores and partial discharges, thereby performing defect diagnosis.

2. The method for detecting electric field distortion of composite insulators based on electroluminescence effect according to claim 1, characterized in that: In step 1, ZnS:Cu is used as the fluorescent material. The specific preparation and coating process of the ZnS:Cu composite coating material is as follows: first, the required mass of silicone rubber material is weighed, poured into a flask and heated to 60°C to improve the fluidity of the matrix, a curing agent is added in proportion, and then the ZnS:Cu fluorescent material and barium titanate filler are added. Then, the heating at 60°C is maintained constant and stirring is continued for 1 hour. After that, the mixture is placed in a vacuum drying oven for vacuuming until there are no bubbles on the surface. Finally, the ZnS:Cu composite coating material is sprayed on the surface of the cleaned and treated composite insulator to ensure that the coating thickness is uniform and complete. During the coating process, bubbles and voids should be avoided. The coating speed and coating thickness are controlled to ensure the coating quality. After high-temperature curing, the composite insulator finally coated with the electroluminescent composite coating material is obtained.

3. The method for detecting electric field distortion of composite insulators based on electroluminescence effect according to claim 1, characterized in that: The specific process of step 2 is: Image acquisition: Use a drone equipped with a camera for shooting. Choose a suitable drone model equipped with a high-resolution camera to ensure that clear, high-quality images can be obtained during flight. At the same time, the camera can dynamically adjust the shooting angle. According to the pre-designed route, control the drone to fly autonomously, adjust the distance between the drone and the surface of the composite insulator, control the posture and camera angle, and capture images of the composite insulator surface. Image processing and analysis part: First, the collected image is denoised to eliminate possible interference and improve image quality and clarity. Then, image processing software is used to convert the pre-processed image into a grayscale image, and the grayscale value information of each pixel is extracted; finally, an electric field size model based on grayscale value is established. According to the experimentally measured light intensity-electric field relationship, the grayscale value of each pixel is converted to obtain the corresponding electric field size value.

4. The method for detecting electric field distortion of composite insulators based on electroluminescence effect according to claim 3, characterized in that: The specific steps of step 2 are as follows: Step 2.1, Drone Settings Choosing the right drone model is key to ensuring image quality. The following are specific selection criteria: Flight stability: The drone should have good flight stability to reduce image blur caused by vibration or wind; Payload capacity: The drone needs to have sufficient payload capacity to carry high-resolution cameras and other sensors; Flight time: Choose a drone with a long flight time to ensure it can complete the image acquisition mission for the entire route; Therefore, the drones selected have high payload capacity and long flight time, and are equipped with RTK modules that can achieve centimeter-level flight accuracy; Next, fix the high-resolution camera on the drone and ensure that the camera can dynamically adjust the shooting angle. The specific installation steps are as follows: Choose a high-resolution camera and use a shock-proof mounting bracket to securely mount the camera on the bottom of the drone to prevent shaking during flight. Use a gimbal stabilizer to ensure the camera can dynamically adjust the shooting angle to suit different shooting needs. The final step is to design the route. According to the pre-designed route, the drone is controlled to fly autonomously, the distance between the drone and the composite insulator surface is adjusted, the attitude and camera angle are controlled, and images of the composite insulator surface are captured. The specific steps are as follows: Use flight planning software to pre-set the flight path to ensure that the drone can cover all composite insulator surfaces to be inspected; Flight parameter settings: Flight altitude H: Set the appropriate flight altitude based on the height of the insulator and the focal length of the camera; Shooting interval I: Set the shooting interval to ensure sufficient overlap between images; Autonomous flight: Activate the drone's autonomous flight mode and fly along a predetermined route, dynamically adjusting the camera angle for optimal filming. Data monitoring: During the flight, the drone's status and image acquisition are monitored in real time to ensure the accuracy and integrity of the data; Step 2.2, image acquisition and preprocessing; The drone flies along a predetermined route to capture images of the composite insulator surface. To ensure that the captured images fully and accurately reflect the composite insulator surface information, the following settings are required: Shooting angle: Set the camera's pitch angle θ and yaw angle φ to obtain images at different angles; Flight speed: Set an appropriate flight speed v to ensure image clarity; Image capture: The drone flies according to the designed route and continuously captures images of the composite insulator surface during the flight; Coverage: To ensure complete surface information, the image capture should cover all areas of the composite insulator, including the upper, middle and lower parts; Image preprocessing includes denoising and grayscale conversion, aiming to improve image quality and extract useful information; Denoising: Denoising is performed on the collected images to eliminate possible interference; Among them, Gaussian filtering is a linear smoothing filtering method, which is often used to remove high-frequency noise in images. The formula of Gaussian filtering is: Where G(x,y) is the value of the Gaussian function at point (x,y), which usually represents the value of the pixel value of the image at point (x,y) after Gaussian filtering. x is the horizontal offset from the center of the Gaussian function in pixels. y is the vertical offset from the center of the Gaussian function in pixels. σ is the standard deviation, which determines the width of the Gaussian distribution. e is the base of the natural logarithm. Grayscale conversion: Convert the preprocessed image into a grayscale image and extract the grayscale value information of each pixel. The grayscale value calculation formula is: I(x,y)=0.299·R(x,y)+0.587·G(x,y)+0.114·B(x,y) Among them, R(x,y), G(x,y), and B(x,y) are the red, green, and blue component values ​​of the pixel point (x,y) respectively; Step,2.3, electric field distribution modeling; The relationship between the grayscale value at different positions and the applied voltage is shown below: Where I, R, G, and B are the grayscale value, red value, green value, and blue value at different positions (x, y) under voltage U, respectively, and all range from 0 to 1. The relationship between the electric field strength and the applied voltage at the calibration point (x0, y0) is as follows: E(x0,y0)=kU Wherein, k is the corresponding coefficient, obtained from simulation and test data; The relationship between the electric field intensity and the grayscale value at the calibration point is as follows: E(x0,y0)=f(I(x0,y0,U)) Where, f is the relationship function; Therefore, the actual electric field strength can be inferred based on the grayscale value, as shown below: E(x,y,U1)=f(I(x,y,U1)) Wherein, U1 is the corresponding voltage value at the position (x, y).

5. The method for detecting electric field distortion of composite insulators based on electroluminescence effect according to claim 1, characterized in that: The specific steps of step 3.3 are as follows: Step 3.3, Actuator Operation Alarm signal: When electric field distortion or defects are detected, the system will issue an alarm signal and record the location and type of the defect. According to the preset alarm level, appropriate treatment measures will be taken; Actuator operation is a crucial part of the system, ensuring the precise execution of drones in image acquisition and electric field distribution detection through automated control systems; Actuators include motors, servos, and camera gimbals, while sensors include GPS modules, gyroscopes, accelerometers, and altimeters. The formulas for adjusting the flight path and camera parameters are as follows: v=v base +Δv h=h base +Δh θ=θ base +Δθ ψ=ψ base +Dψ Among them, v is the adjusted flight speed, h is the adjusted flight altitude, θ is the adjusted camera pitch angle, ψ is the adjusted heading angle, and v base 、h base ,θ base , ψ base are the basic flight data of flight speed, flight altitude, camera pitch angle, and heading angle, respectively. Δv, Δh, Δθ, and Δψ are the flight speed, flight altitude, camera pitch angle, and heading angle adjusted according to the prediction results. Real-time monitoring and processing: Real-time monitoring of the drone's flight status and image acquisition to ensure data accuracy and completeness, record the location and type of defects, take appropriate measures based on the preset alarm level, and notify maintenance personnel to carry out inspection and replacement.

Images