Insulating layer eccentricity correction method for cable

Abstract

The invention relates to the technical field of cable production, and particularly discloses a cable insulation layer eccentricity calibration method, which comprises the following steps: S1, on a cable production line, multiple groups of optical sensors are uniformly arranged around a cable at certain angle intervals, and the optical sensors are used for measuring distance information of different positions on the surface of the cable; s2, acquiring data of the sensor in real time by using a high-speed data acquisition card, and transmitting the data to a computer for processing; and S3, constructing a three-dimensional shape model of the cable insulation layer by using the collected distance data through a triangulation method. According to the cable insulation layer eccentricity correction method, eccentricity calculation under different scales, including local eccentricity and overall eccentricity, is considered, the non-uniformity of the insulation layer can be better reflected, more comprehensive information is provided for quality control, and compared with the prior art, the cable insulation layer eccentricity correction method has the advantage that the calculation efficiency is improved. The system can accurately position a local problem area so as to adjust production parameters in a targeted manner, and can be used as an early warning system to timely discover potential problems.

Description

Technical Field

[0001] The invention relates to the technical field of cable production, and in particular to a method for calibrating the eccentricity of an insulating layer of a cable. Background Art

[0002] Cables are widely used in power station lines, industrial and coal mine power supply, urban and rural underground power grids, and underwater transmission lines such as rivers and seas. The qualified eccentricity of cables is an important link in ensuring the quality of cables.

[0003] In cable production, after the conductor is extruded with an insulating layer, the eccentricity of the extruded insulating layer needs to be measured and adjusted to ensure that it meets production requirements. However, there is a large deviation between the machine-measured eccentricity automatically collected by the eccentricity instrument currently used and the measured eccentricity measured by workers using manual tools. Therefore, when each batch of cable production is started, workers must compare the machine-measured eccentricity with the measured eccentricity, and adjust the actual eccentricity to meet the production standards. At the same time, the actual eccentricity cannot be too small, because this will increase the working hours, significantly increase the cable consumption, and increase the cost. Therefore, the actual eccentricity is usually kept within a range slightly smaller than the standard value, which makes the measured eccentricity often close to the standard value, and then a large amount of manual work is required to correct the machine-measured eccentricity to ensure that the actual eccentricity is less than the standard value. Under this situation, there is an urgent need for a more scientific and efficient cable insulation layer eccentricity correction method that can reduce the workload of manual correction and improve the accuracy and production efficiency of eccentricity measurement.

[0004] Therefore, it is necessary to propose a method for calibrating the eccentricity of the insulation layer of a cable to solve the above problems. Summary of the invention

[0005] The main purpose of the present invention is to provide a method for calibrating the eccentricity of an insulation layer for a cable, which can effectively solve the problems in the background technology.

[0006] To achieve the above object, the technical solution adopted by the present invention is: A method for calibrating the eccentricity of an insulation layer for a cable comprises the following steps: S1: On the cable production line, multiple groups of optical sensors are evenly arranged around the cable at certain angles. The optical sensors are used to measure the distance information at different positions on the cable surface. S2: Use a high-speed data acquisition card to collect sensor data in real time and transmit it to a computer for processing: Assume that n optical sensors are installed, and each sensor measures the distance to a specific point on the cable surface as , the cable radius is r, and the distance between the sensor and the cable center is , then: ,in It is the angle between the line connecting the sensor and the cable center and a certain reference direction; S3: Using the collected distance data, a three-dimensional shape model of the cable insulation layer is constructed by triangulation; S4: Using edge detection algorithm, the boundary between the insulation layer and the conductor is extracted, and then the approximate geometric shape of the insulation layer can be obtained by fitting multiple points on the boundary; S5: Eccentricity calculation, which includes local eccentricity calculation and overall eccentricity calculation. Local eccentricity calculation: for the point set in the local area , the local center coordinates of the conductor are , the local center coordinates of the insulating layer are , the local radius of the insulating layer is , then the local eccentricity , where j represents the point index in the local area; Calculation of overall eccentricity: The center coordinates of the known conductor are , for the geometric shape of the insulating layer obtained by fitting, determine its center coordinates and radius R, then the overall eccentricity ; The final eccentricity value is calculated by weighted average: the local eccentricity weight is set to α, the overall eccentricity weight is set to β, and (α+β=1), then the final eccentricity ; S6: Collect historical production data, including insulation layer eccentricity measurement results under different production conditions and corresponding production equipment parameters; S7: Using a machine learning algorithm to train an eccentricity prediction model, with the input being the current production equipment parameters and the output being the predicted eccentricity; S8: During the production process, the current equipment parameters are input into the model in real time to predict the eccentricity trend. If the predicted eccentricity exceeds the set threshold range, the production equipment parameters are automatically adjusted.

[0007] Preferably, in step S1, the optical sensor is a laser displacement sensor or a structured light sensor, and is arranged in a ring array or a spiral arrangement.

[0008] Preferably, in step S2, the optical sensor data is pre-processed in real time to remove outliers, filter, and calibrate before transmitting the data to the computer for processing.

[0009] Preferably, in step S3, the specific steps of constructing the three-dimensional shape model of the cable insulation layer using the triangulation method are as follows: S3.1: For each optical sensor, based on the measured distance and the known position of the optical sensor, the position of a point on the cable surface can be determined; S3.2: Construct multiple triangles using points measured by multiple optical sensors; S3.3: By solving the side lengths and angles of the triangle, the three-dimensional shape of the cable surface can be determined.

[0010] Preferably, in step S4, the edge detection algorithm is a Sobel operator, and the specific steps of extracting the boundary between the insulating layer and the conductor by using the Sobel operator are: S4.1: Calculate the gradient of each point, for each point in the 3D shape model Calculate its gradient in the x, y, and z directions; S4.2: Determine edge points. According to the magnitude and direction of the gradient, determine possible edge points. By setting a threshold, when the magnitude of the gradient exceeds the threshold, the point is considered to be an edge point. S4.3: Connect the edge points, connect the determined edge points to form a boundary between the insulation layer and the conductor, and use curve fitting to connect the edge points into a smooth curve or surface.

[0011] Preferably, when the approximate geometric shape of the insulating layer is a sphere, the specific fitting steps are as follows: (1) Establish the equation of the sphere and set the equation of the sphere to ,in is the coordinate of the center of the sphere, R is the radius; (2) Using the least squares method to fit, substitute the points on the boundary into the sphere equation, and then use the least squares method to solve the coordinates of the sphere center and radius; (3) Evaluate the fitting effect. Use the root mean square error to evaluate the fitting effect.

[0012] Preferably, when the approximate geometric shape of the insulating layer is a cylinder, the specific fitting method is as follows: (1) Assume that the equation of the cylinder is ,in is the center coordinate of the bottom surface of the cylinder, R is the radius of the bottom surface, is half the height of the cylinder, and k is a constant; (2) Fit the points on the boundary using the least squares method, substitute the points on the boundary into the equation of the cylinder and solve for the parameters of the cylinder.

[0013] Preferably, in step S7, the machine learning algorithm adopted is a support vector machine or a deep learning neural network.

[0014] Preferably, in step S3, for the constructed three-dimensional shape model of the cable insulation layer, the three-dimensional model needs to be surface processed and filtered to remove noise and abnormal points.

[0015] Preferably, in step S8, automatically adjusting the production equipment parameters includes but is not limited to adjusting the temperature, pressure and speed of the extruder.

[0016] Compared with the prior art, the present invention provides a method for calibrating the eccentricity of the insulation layer of a cable, which has the following beneficial effects: 1. The cable insulation layer eccentricity calibration method considers the eccentricity calculation at different scales, including local and overall eccentricity, which can better reflect the non-uniformity of the insulation layer and provide more comprehensive information for quality control. Compared with the existing technology, it can accurately locate the local problem area so as to adjust the production parameters in a targeted manner, and timely discover potential problems as an early warning system, evaluate the overall quality of the cable from a macro perspective, guide the overall production optimization and verify with the local eccentricity, realize comprehensive quality control, provide a scientific basis for production decisions and support continuous quality improvement, and can effectively overcome the shortcomings of the existing technology in quality control, such as the difficulty of comprehensive coverage and poor accuracy of manual detection, the lack of scientificity of experience adjustment, the lack of multi-scale analysis of non-uniformity, and the inability to predict the trend of eccentricity for automatic adjustment. 2. The cable insulation layer eccentricity calibration method, the cable insulation layer eccentricity calibration method, through multiple groups of optical sensors evenly arranged at a certain angle interval, combined with triangulation and three-dimensional reconstruction algorithms, can more accurately construct a three-dimensional shape model of the cable insulation layer, and can more accurately measure and calculate eccentricity than the existing technology. 3. The cable insulation layer eccentricity calibration method uses a machine learning algorithm to train an eccentricity prediction model, which can predict the eccentricity trend based on real-time equipment parameters and automatically adjust production equipment parameters to improve production efficiency and product quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0018] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the present invention is further explained below in conjunction with specific implementation methods.

[0019] like Figure 1As shown, a method for calibrating the eccentricity of the insulation layer of a cable is as follows: first, on a cable production line, multiple groups of optical sensors, such as laser displacement sensors or structured light sensors, are evenly arranged around the cable at a certain angle interval. These sensors can emit specific light and receive reflected signals to measure the distance information of different positions on the cable surface; wherein the laser displacement sensor: has the advantages of high precision, high resolution and non-contact measurement, and is suitable for measuring small changes on the cable surface. Laser displacement sensors with different measurement ranges and accuracies can be selected to meet the needs of different cable specifications. During installation, ensure that the laser beam of the sensor is perpendicular to the cable surface to improve the measurement accuracy. Structured light sensor: by projecting a specific light pattern onto the cable surface and then analyzing the deformation of the reflected light to obtain three-dimensional shape information, the structured light sensor can provide richer surface details, but may be more sensitive to ambient light and surface reflectivity. During installation, it is necessary to consider the lighting conditions and cable surface characteristics, and adjust the angle and distance of the sensor to obtain the best measurement effect. In addition, the arrangement of the sensor can be arranged in a ring array or a spiral to ensure that the entire circumference of the cable can be covered. At the same time, according to the cable diameter and production speed, the number and spacing of sensors can be adjusted to ensure sufficient measurement density and real-time performance.

[0020] Furthermore, a high-speed data acquisition card is used to collect sensor data in real time and transmit it to a computer for processing. Before the data is transmitted to the computer for processing, the sensor data is preprocessed in real time, such as removing outliers, filtering, calibration, etc. A field programmable gate array (FPGA) or a dedicated data preprocessing chip can be used to increase the processing speed. Among them, the high-speed data acquisition card selects a high-speed data acquisition card with a high sampling rate and multi-channel input to meet the needs of real-time acquisition of multiple sensor data. The resolution and accuracy of the acquisition card should match the sensor to ensure the accuracy of the data. In addition, the data transmission adopts a high-speed data transmission interface, such as USB3.0, Ethernet or PCIExpress, to transmit the collected data to the computer in real time for processing. In order to improve the stability and reliability of data transmission, data caching and error checking mechanisms can be adopted. During the data acquisition process, a trigger signal can be set, such as a cable position sensor or a synchronization signal on a production line, to ensure the synchronization and accuracy of the data. At the same time, the collected data can be monitored and displayed in real time to detect abnormal conditions in time.

[0021] Based on the above, assume that n optical sensors are installed and each sensor measures the distance to a specific point on the cable surface as , the cable radius is r, and the distance between the sensor and the cable center is , then: ,in The angle between the line connecting the sensor and the center of the cable and a certain reference direction (assuming that the cable moves from left to right on the production line, the horizontal right direction can be defined as the reference direction. At this time, for each sensor, measure the angle between the line connecting the sensor and the center of the cable and the horizontal right direction ).

[0022] This formula is derived based on the cosine theorem. Specifically, in the triangle OAB, O is the cable center, A is the sensor position, and B is the point on the cable surface measured by the sensor. , OB=r, , According to the law of cosines, ,Right now , we can get ,Through this formula, the position of the sensor in the three-dimensional space can be determined ,according to the sensor’s measurement data and the known cable radius, ,which provides a basis for the subsequent three-dimensional shape reconstruction.

[0023] Furthermore, the collected distance data is used to construct a three-dimensional shape model of the cable insulation layer through triangulation or other three-dimensional reconstruction algorithms.

[0024] Specifically, triangulation is a 3D reconstruction method based on geometric principles. The distance information measured by multiple sensors can be used to determine the positions of multiple points on the cable surface, and then the geometric relationship of triangles can be used to construct a 3D shape model. The specific steps are as follows: 1. For each sensor, based on the measured distance 1. With the known sensor positions, the position of a point on the cable surface can be determined. 2. Using the points measured by multiple sensors, multiple triangles can be constructed. 3. By solving the side lengths and angles of the triangles, the three-dimensional shape of the cable surface can be determined.

[0025] In addition to triangulation, other 3D reconstruction algorithms can be used, such as stereo vision, time of flight (ToF), etc. You can choose the appropriate method according to the actual situation. For example, stereo vision requires two or more cameras to shoot the cable surface from different angles, and then reconstruct the 3D shape through image matching and triangulation. The time of flight method determines the distance by measuring the time from the emission to the reception of the light signal, thereby constructing the 3D shape.

[0026] Further, the three-dimensional model is subjected to surface processing and filtering to remove noise and outliers and improve the accuracy of the model: Among them, denoising, the collected distance data may be affected by noise, such as sensor noise, environmental noise, etc. In order to improve the accuracy of the model, the data needs to be denoised. Mean filtering, median filtering, Gaussian filtering and other methods can be used to remove noise points in the data. Outlier removal: There may be some outliers in the data, such as sensor failures, defects on the cable surface, etc. These outliers will affect the accuracy of the three-dimensional shape model and need to be removed. Statistical methods, such as standard deviation method, box plot method, etc., can be used to identify and remove outliers. Surface smoothing: In order to improve the visualization effect of the model and the accuracy of subsequent boundary extraction, the three-dimensional model can be surface smoothed. Laplace smoothing, bilateral filtering and other methods can be used to smooth the model surface.

[0027] Furthermore, an edge detection algorithm, such as the Sobel operator or the Laplacian operator, is used to extract the boundary between the insulating layer and the conductor. The Sobel operator is a commonly used edge detection algorithm that detects edges by calculating the gradient of an image. In a three-dimensional shape model, the coordinates of each point can be regarded as the pixel value of the image, and then the Sobel operator is applied to detect the boundary between the insulating layer and the conductor. The specific steps are as follows: 1. Calculate the gradient of each point: For each point in the 3D shape model , calculate its gradient in the x, y, and z directions. The central difference method or other gradient calculation methods can be used. 2. Determine edge points: Determine possible edge points based on the size and direction of the gradient. A threshold can be set. When the size of the gradient exceeds the threshold, the point is considered to be an edge point. 3. Connect edge points: Connect the determined edge points to form a boundary between the insulation layer and the conductor. Curve fitting or other methods can be used to connect the edge points into a smooth curve or surface.

[0028] The Laplacian operator is also a commonly used edge detection algorithm that detects edges by calculating the second-order derivative of the image. In 3D shape models, a similar method can be used to apply the Laplacian operator to detect boundaries. Compared with the Sobel operator, the Laplacian operator is more sensitive to noise and requires more pre-processing and post-processing.

[0029] In the process of edge extraction, multiple algorithms can be combined, such as using the Sobel operator for preliminary edge detection, and then using the Laplacian operator for refinement and optimization. At the same time, the parameters of the algorithm, such as threshold, filter size, etc., can be adjusted according to actual conditions to obtain the best edge extraction effect.

[0030] Assume that the coordinates of the point on the boundary of the insulation layer are By fitting these points, we can get the approximate geometric shape of the insulating layer, such as a sphere or a cylinder. If the shape of the insulating layer is approximately a sphere, we can use the least squares method or other fitting algorithms to fit the points on the boundary into a sphere. The specific steps are as follows: 1. Establish the equation of the sphere: Assume that the equation of the sphere is ,in 2. Least squares fitting: Substitute the points on the boundary into the sphere equation, and then solve the sphere center coordinates and radius by the least squares method. You can use iterative algorithms, such as Newton's method, gradient descent method, etc., to solve the least squares problem. 3. Evaluate the fitting effect: You can use the root mean square error (RMSE) or other indicators to evaluate the fitting effect. If the fitting effect is not ideal, you can adjust the parameters of the fitting algorithm or try other fitting methods.

[0031] If the shape of the insulating layer is approximately a cylinder, a similar method can be used for fitting. Assume that the equation of the cylinder is , ,in is the center coordinate of the bottom surface of the cylinder, R is the radius of the bottom surface, is half the height of the cylinder, k is a constant, and the parameters of the cylinder can be solved by fitting the points on the boundary using the least squares method.

[0032] For the acquisition of k, specifically, k is mainly used to determine the position of the cylinder along the axis in space. Specifically, when the center coordinates of the bottom surface of the cylinder are given Parameters related to bottom radius R and height Finally, by adjusting the value of k, the cylinder can be placed in different spatial locations. If k is a positive number, then the cylinder will be If k is a negative number, the cylinder will move upward by the corresponding distance. If it is known that the cylinder needs to be placed at a specific position in space, and the coordinate value of the position in the z-axis direction is known, and the coordinates of the center of the bottom of the cylinder are also known Bottom radius R and height related parameters At this point, the z coordinate of the specific position can be calculated and For example, if a cylinder needs to be placed at a point (x, y, z) in space, and it is known that , R and .if ,So .

[0033] In the process of geometric shape fitting, the appropriate fitting method and model can be selected according to the actual situation. If the shape of the insulating layer is more complex, more advanced fitting algorithms can be used, such as polynomial surface fitting, spline surface fitting, etc. At the same time, visualization tools such as 3D graphics software can be combined to intuitively display and evaluate the fitting results.

[0034] Furthermore, eccentricity calculation includes local eccentricity calculation and overall eccentricity calculation, where local eccentricity calculation: for the point set in the local area , the local center coordinates of the conductor are , the local center coordinates of the insulating layer are , the local radius of the insulating layer is , then the local eccentricity: , where j represents the point index in the local area.

[0035] For the above point set generation, for each distance data measured by the sensor, the coordinates of a corresponding point on the cable surface can be determined according to the position and measurement angle of the sensor. Assume that the position coordinates of the sensor are , the measured distance is , based on the triangulation principle or other 3D reconstruction algorithms, the coordinates of the corresponding points on the cable surface can be calculated , for example, for triangulation, calculate the distance between the sensor and the cable center Then, the coordinates of the corresponding points on the cable surface are further calculated based on the geometric relationship. Repeat the above steps to calculate the corresponding cable surface point coordinates for each sensor measured data, thus obtaining the point set .

[0036] For the local center coordinates of the conductor The local area can be divided according to the cable geometry, the characteristics of the insulation layer or specific requirements. For example, the cable surface can be divided into several small rectangular or circular areas, or the local area can be determined according to the thickness change of the insulation layer, material characteristics and other factors.

[0037] If the conductor has a regular geometric shape in a local area, such as a circle or a cylinder, the local center coordinates of the conductor can be determined by calculating the geometric center. For example, for a circular conductor, the local center coordinates of the conductor can be the coordinates of the center of the circle.

[0038] For the local center coordinates of the conductor If the insulating layer has a regular geometric shape in the local area, such as a sphere or a cylinder, the local center coordinates of the insulating layer can be determined by calculating the geometric center; for the local radius of the insulating layer If the insulating layer has a regular geometric shape in the local area, such as a sphere or a cylinder, the local radius of the insulating layer can be determined directly according to the definition of the geometric shape. For example, for a spherical insulating layer, the local radius can be the distance from any point in the local area to the local center of the insulating layer.

[0039] Further, the overall eccentricity calculation: the center coordinates of the known conductor are , for the geometric shape of the insulating layer obtained by fitting, determine its center coordinates and radius R, then the overall eccentricity The final eccentricity value is calculated by weighted average and other methods: the local eccentricity weight is set to α, the overall eccentricity weight is set to β, and (α+β=1), then the final eccentricity At the same time, considering the measurement error and model uncertainty, Monte Carlo simulation or Bayesian method is used to estimate the uncertainty range of eccentricity; when used, the eccentricity calculation at different scales, including local and overall eccentricity, can better reflect the non-uniformity of the insulation layer and provide more comprehensive information for quality control. Compared with the existing technology, it can accurately locate the local problem area so as to adjust the production parameters in a targeted manner, and timely discover potential problems as an early warning system, evaluate the overall quality of the cable from a macro perspective, guide the overall production optimization and verify with the local eccentricity, realize comprehensive quality control, provide a scientific basis for production decisions and support continuous quality improvement, which can effectively overcome the shortcomings of the existing technology in quality control, such as the difficulty of comprehensive coverage and poor accuracy of manual inspection, the lack of scientificity of empirical adjustment, the lack of multi-scale analysis of non-uniformity and the inability to predict the trend of eccentricity for automatic adjustment.

[0040] Furthermore, a large amount of historical production data is collected, including the measurement results of the eccentricity of the insulation layer under different production conditions and the corresponding production equipment parameters (such as temperature, pressure, speed, etc.). These data can come from the monitoring system, quality inspection report or experimental data in the actual production process. The collected data is preprocessed, including data cleaning, denoising, normalization and other operations. Data cleaning can remove outliers and erroneous data, denoising can reduce noise interference in the data, and normalization can convert the data to the same scale for subsequent analysis and processing.

[0041] Furthermore, a machine learning algorithm, such as support vector machine (SVM) or deep learning neural network, is used to train an eccentricity prediction model. The input is the current production equipment parameters, and the output is the predicted eccentricity.

[0042] Among them, support vector machine (SVM): SVM is a commonly used machine learning algorithm that classifies or regresses data by finding an optimal hyperplane. In eccentricity prediction, production equipment parameters can be used as input features and eccentricity as output labels to train an SVM regression model. SVM has good generalization ability and robustness, and can handle high-dimensional data and small sample problems.

[0043] Deep learning neural network is a powerful machine learning algorithm that can automatically learn features and patterns in data. In eccentricity prediction, a deep learning model can be trained using architectures such as convolutional neural network (CNN), recurrent neural network (RNN) or deep neural network (DNN). Deep learning models can handle large-scale data and complex nonlinear relationships, but require a lot of computing resources and time to train.

[0044] When choosing a machine learning algorithm, you need to consider the characteristics of the data, the complexity of the problem, and the limitations of computing resources. You can choose the most suitable algorithm for training by experimenting and comparing the performance of different algorithms.

[0045] Finally, during the production process, the current equipment parameters are input into the model in real time to predict the eccentricity trend. If the predicted eccentricity exceeds the set threshold range, the production equipment parameters are automatically adjusted to reduce the eccentricity.

[0046] For example, the thickness distribution of the insulation layer can be changed by adjusting the temperature, pressure or speed of the extruder, thereby reducing the eccentricity. The specific adjustment strategy can be optimized according to the actual situation, such as using optimization algorithms such as gradient descent or genetic algorithm to find the best equipment parameter adjustment solution.

[0047] It should be noted that for the distribution of local eccentricity weight α and overall eccentricity weight β, a large number of experiments are required to collect eccentricity data of different cable specifications under different temperature, pressure, rotation speed, material type and other conditions, and record the eccentricity adjustment results under different local and overall insulation layer weight combinations; analyze the experimental data to observe the influence of different weight combinations on the eccentricity reduction. For example, for 110kV high-voltage cross-linked polyethylene insulated power cable (240mm²), its initial eccentricity and adjusted eccentricity will change under different local and overall insulation layer weights. By comparing the eccentricity reduction under different weight combinations, a relatively optimal weight combination range can be found.

[0048] Based on the structure and performance characteristics of the cable insulation layer, a theoretical model for eccentricity calculation is established. In the model, the characteristics of the local and overall insulation layer and their influence mechanism on eccentricity are considered. Through theoretical analysis, the approximate range of the weights of the local and overall insulation layer is determined. For example, according to the material properties and electric field distribution of the insulation layer, the influence of local eccentricity on the overall electric field distribution is analyzed, so as to determine the upper and lower limits of the weight of the local insulation layer.

[0049] Combined with the experience in the actual production process, we can understand the influence of different factors on the cable eccentricity. Production personnel can evaluate and provide feedback on the effects of different weight combinations based on previous production experience. For example, in actual production, if it is found that a certain cable specification has a more prominent local eccentricity problem under specific production conditions, the weight of the local insulation layer can be appropriately increased. At the same time, according to the quality control requirements and customer feedback in the production process, the weight combination is continuously adjusted to achieve the best eccentricity control effect.

[0050] Cables of different specifications and types may have different requirements for the weight of local and overall insulation layers. Generally speaking, high-voltage cables have higher requirements for overall insulation performance due to their complex structure and high electric field strength, so the overall insulation layer weight may be relatively large, while low-voltage cables may pay more attention to local insulation performance, and the weight of the local insulation layer can be appropriately increased. For example, for 110kV high-voltage cross-linked polyethylene insulated power cables (240mm²), due to their high voltage level, the stability of the overall insulation performance needs to be ensured, and the overall insulation layer weight can be set to 0.6 and the local insulation layer weight to 0.4. For 0.6 / 1kV low-voltage polyvinyl chloride insulated power cables (50mm²), the local insulation layer weight can be set to 0.5 and the overall insulation layer weight can be set to 0.5 according to actual conditions.

[0051] In addition, production conditions and processes will also affect the distribution of weights. For example, when production parameters such as temperature, pressure, and rotation speed fluctuate greatly, more attention should be paid to the stability of the overall insulation performance, and the weight of the overall insulation layer can be appropriately increased. When the production process is relatively stable, the weight of the local insulation layer can be appropriately adjusted according to the local eccentricity of the cable.

[0052] This method can be applied to the production of cables of different types and specifications. By adjusting the sensor layout and algorithm parameters, it can meet the needs of different production environments. In addition, the intelligent algorithm can learn the relationship between eccentricity and production equipment parameters based on historical data, and automatically adjust the equipment parameters to correct eccentricity, thereby reducing manual intervention and improving production efficiency.

[0053] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited by the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, and these changes and improvements all fall within the scope of the present invention to be protected.

Claims

1. A method for calibrating the eccentricity of an insulation layer of a cable, characterized in that: The following steps are involved: S1: On the cable production line, multiple groups of optical sensors are evenly arranged around the cable at certain angles. The optical sensors are used to measure the distance information at different positions on the cable surface. S2: Use a high-speed data acquisition card to collect sensor data in real time and transmit it to a computer for processing: Assume that n optical sensors are installed, and each sensor measures the distance to a specific point on the cable surface as , the cable radius is r, and the distance between the sensor and the cable center is , then: ,in It is the angle between the line connecting the sensor and the cable center and a certain reference direction; S3: Using the collected distance data, a three-dimensional shape model of the cable insulation layer is constructed by triangulation; S4: Using edge detection algorithm, the boundary between the insulation layer and the conductor is extracted, and then the approximate geometric shape of the insulation layer can be obtained by fitting multiple points on the boundary; S5: Eccentricity calculation, which includes local eccentricity calculation and overall eccentricity calculation. The local eccentricity calculation is: for the point set in the local area , the local center coordinates of the conductor are , the local center coordinates of the insulating layer are , the local radius of the insulating layer is , then the local eccentricity , where j represents the point index in the local area; Calculation of overall eccentricity: The center coordinates of the known conductor are , for the geometric shape of the insulating layer obtained by fitting, determine its center coordinates and radius R, then the overall eccentricity ; The final eccentricity value is calculated by weighted average: the local eccentricity weight is set to α, the overall eccentricity weight is set to β, and (α+β=1), then the final eccentricity ; S6: Collect historical production data, including insulation layer eccentricity measurement results under different production conditions and corresponding production equipment parameters; S7: Using a machine learning algorithm to train an eccentricity prediction model, with the input being the current production equipment parameters and the output being the predicted eccentricity; S8: During the production process, the current equipment parameters are input into the model in real time to predict the eccentricity trend. If the predicted eccentricity exceeds the set threshold range, the production equipment parameters are automatically adjusted.

2. A method for calibrating the eccentricity of an insulation layer for a cable according to claim 1, characterized in that: In step S1, the optical sensor is a laser displacement sensor or a structured light sensor, and is arranged in a ring array or a spiral arrangement.

3. The method for calibrating the eccentricity of the insulation layer of a cable according to claim 1, characterized in that: In step S2, the optical sensor data is pre-processed in real time to remove outliers, filter, and calibrate before being transferred to a computer for processing.

4. The method for calibrating the eccentricity of the insulation layer of a cable according to claim 1, characterized in that: In step S3, the specific steps of constructing the three-dimensional shape model of the cable insulation layer by triangulation are as follows: S3.1: For each optical sensor, based on the measured distance and the known position of the optical sensor, the position of a point on the cable surface can be determined; S3.2: Construct multiple triangles using points measured by multiple optical sensors; S3.3: By solving the side lengths and angles of the triangle, the three-dimensional shape of the cable surface can be determined.

5. The method for calibrating the eccentricity of the insulation layer of a cable according to claim 1, characterized in that: In step S4, the edge detection algorithm is the Sobel operator, and the specific steps of extracting the boundary between the insulating layer and the conductor by the Sobel operator are: S4.1: Calculate the gradient of each point, for each point in the 3D shape model , calculate its gradient in the x, y, and z directions; S4.2: Determine edge points. According to the magnitude and direction of the gradient, determine possible edge points. By setting a threshold, when the magnitude of the gradient exceeds the threshold, the point is considered to be an edge point. S4.3: Connect the edge points, connect the determined edge points to form a boundary between the insulation layer and the conductor, and use curve fitting to connect the edge points into a smooth curve or surface.

6. A method for calibrating the eccentricity of an insulation layer for a cable according to claim 5, characterized in that: When the approximate geometric shape of the insulating layer is a sphere, the specific fitting steps are as follows: (1) Establish the equation of the sphere and set the equation of the sphere to ,in is the coordinate of the center of the sphere, R is the radius; (2) Using the least squares method to fit, substitute the points on the boundary into the sphere equation, and then use the least squares method to solve the coordinates of the sphere center and radius; (3) Evaluate the fitting effect. Use the root mean square error to evaluate the fitting effect.

7. A method for calibrating the eccentricity of an insulation layer for a cable according to claim 5, characterized in that: When the approximate geometric shape of the insulating layer is a cylinder, the specific fitting method is as follows: (1) Assume that the equation of the cylinder is ,in is the center coordinate of the bottom surface of the cylinder, R is the radius of the bottom surface, is half the height of the cylinder, and k is a constant; (2) Fit the points on the boundary using the least squares method, substitute the points on the boundary into the equation of the cylinder and solve for the parameters of the cylinder.

8. A method for calibrating the eccentricity of an insulation layer for a cable according to claim 1, characterized in that: In step S7, the machine learning algorithm adopted is a support vector machine or a deep learning neural network.

9. A method for calibrating the eccentricity of an insulation layer for a cable according to claim 1, characterized in that: In step S3, for the constructed three-dimensional shape model of the cable insulation layer, the three-dimensional model needs to be surface processed and filtered to remove noise and abnormal points.

10. A method for calibrating the eccentricity of an insulation layer for a cable according to claim 1, characterized in that: In step S8, automatically adjusting the production equipment parameters includes but is not limited to adjusting the temperature, pressure and speed of the extruder.

Images