Surface pretreatment device and method for insulating coating of electrified bare conductor

Abstract

The surface pretreatment device comprises a supporting frame, a plurality of layers of box bodies arranged at the lower end of the supporting frame and used for bearing treating agents with different concentrations, and a supporting mechanism arranged at the upper end of the supporting frame and used for being supported on a wire. The box body is connected with a nozzle arranged at the upper end of the support frame through a hose; during working, the device is supported on a wire through the supporting mechanism, and then treatment agents with different concentrations are sprayed through the spray heads according to the position of a pollution layer on the wire, so that the surface pretreatment of the electrified bare wire is realized. According to the position and the thickness of a polluted layer, treatment agents with different concentrations can be sprayed, an effective surface pretreatment quality real-time monitoring and dynamic deviation correction mechanism can be established in an electrified environment, and accurate recognition of surface states of different wire materials and self-adaptive adjustment of pretreatment parameters are achieved. And the uniformity and consistency of the pretreatment quality are ensured.

Description

Technical Field

[0001] This invention relates to the field of insulation coating for bare conductors, and more specifically to a surface pretreatment apparatus and method for insulation coating of live bare conductors. Background Technology

[0002] Insulation coating technology for live bare conductors is a core technology for ensuring the safe operation of power systems. It plays a vital role in improving the insulation performance of lines while ensuring continuous power supply to the power grid.

[0003] This technology enables insulation modifications and maintenance of operating transmission lines without interrupting power supply, which is of great significance for ensuring the stable operation of the power grid.

[0004] Traditional methods of conductor insulation treatment mainly rely on power outages, which not only cause huge economic losses, but also make it difficult to meet the high reliability power supply requirements of modern power grids.

[0005] While existing live-line working insulation coating technology can achieve uninterrupted construction, it has obvious defects in the surface pretreatment stage. The pretreatment quality control methods are limited, and there is a lack of effective real-time monitoring and adjustment mechanisms, resulting in unstable coating effects.

[0006] Pretreatment of conductor surfaces in energized environments presents complex technical challenges.

[0007] Different types of wires will form surface contamination layers of different types and thicknesses during long-term operation. The distribution of these contaminants is often extremely uneven, and traditional pretreatment methods are difficult to control precisely.

[0008] This inconsistency in pretreatment quality directly affects the adhesion and overall performance of the subsequent insulating coating, significantly reducing the reliability of the entire insulating coating system. Especially in complex electromagnetic environments, traditional monitoring equipment is difficult to operate normally, and operators cannot accurately judge the pretreatment effect.

[0009] How to establish an effective real-time monitoring and dynamic correction mechanism for surface pretreatment quality in a live environment, and achieve accurate identification of the surface state of different conductor materials and adaptive adjustment of pretreatment parameters to ensure the uniformity and consistency of pretreatment quality, has become a key issue in the development of insulation coating technology for live bare conductors. Summary of the Invention

[0010] The purpose of this invention is to solve the above-mentioned problems and provide a surface pretreatment device and method for insulating bare conductors, which can establish an effective real-time monitoring and dynamic correction mechanism for surface pretreatment quality in a live environment, realize accurate identification of the surface state of different conductor materials and adaptive adjustment of pretreatment parameters, and ensure the uniformity and consistency of pretreatment quality.

[0011] The technical solution adopted by this invention to solve its technical problem is: A surface pretreatment device for insulating and coating live bare wires includes a support frame, a box with several layers for carrying treatment agents of different concentrations disposed at the lower end of the support frame, and a support mechanism disposed at the upper end of the support frame for supporting the wire. The box is connected to a nozzle disposed at the upper end of the support frame via a hose. The support mechanism includes a support wheel disposed on the upper end of the support frame and a lifting frame disposed inside the support frame and moving up and down within the support frame. One end of the housing of the support wheel is rotatably connected to the support frame. The lifting frame is connected to one end of the housing of the support wheel via a connecting rod. The upper end of the connecting rod is rotatably connected to the housing of the support wheel, and the lower end of the connecting rod is rotatably connected to the lifting frame. The lower end of the support frame is provided with a connecting mechanism for connecting the lifting frame. The lifting frame is connected to the upper end of the support frame via a tension spring.

[0012] Furthermore, the lower end of the support frame is provided with a connecting frame for supporting the box body.

[0013] Furthermore, the uppermost box is provided with a collection trough.

[0014] Furthermore, the lifting frame is provided with a groove that mates with the wire.

[0015] Furthermore, the connecting mechanism includes a crossbar disposed on the side of the lifting frame, an elastic plate is provided on the support frame, and a connector head connected to the crossbar is provided at the upper end of the elastic plate.

[0016] Furthermore, a surface pretreatment method for insulating bare conductors, utilizing the aforementioned apparatus, includes the following steps: S101 Obtain the spectral feature dataset of the contamination layer on the conductor surface, extract the thickness distribution and material identification information of the contamination layer using wavelength analysis technology, and generate contamination layer feature data; S102 Construct a three-dimensional model of the conductor surface based on the contamination layer thickness distribution data, segment the region using a mesh generation algorithm, calculate the contamination density value, and determine the baseline values ​​of the preprocessing parameters; S103 Query the pre-established mapping database based on the material identification information, obtain the initial preprocessing parameters, and correct them according to the contamination density value to generate the final preprocessing parameter configuration; S104 Monitor the impedance change during the preprocessing process using electrochemical impedance spectroscopy to determine the stable state of the preprocessing; S105 Establish a quality assessment model based on the impedance change trend, calculate the quality assessment value, and dynamically adjust the parameters; S106 Use laser interferometry to detect surface roughness, determine the preprocessing quality, and record the parameter combination; S107 Train historical data using machine learning algorithms to predict the optimal preprocessing parameters; S108 Generate a conductor quality distribution map, and adjust the deviation areas until the quality uniformity meets the standard.

[0017] Further, step S101 includes the following steps: The surface of the conductor is scanned by a multispectral imaging sensor to obtain reflectance spectral data at different wavelengths, generating a spectral feature dataset. Wavelength analysis technology is used to extract the contamination layer thickness distribution and material identification information from the spectral feature dataset, generating contamination layer feature data. If there are areas in the contamination layer thickness distribution data that exceed a preset threshold, then the marking data of key treatment areas is generated based on the thickness distribution data.

[0018] Further, step S103 includes the following steps: Material information of the conductor is extracted from the material identification system to determine the material type and obtain the material classification result. Based on the material classification result, the pre-established material-preprocessing parameter mapping database is queried to obtain the initial preprocessing parameters. The influence of pollution density on the initial preprocessing parameters is calculated using a preset pollution density-parameter correction model to obtain the correction coefficient. The initial preprocessing parameters are adjusted by combining the correction coefficient with a linear regression algorithm to generate the final preprocessing parameter configuration.

[0019] Further, step S104 includes the following steps: Impedance signals from the surface of a conductor are collected using an electrochemical impedance spectroscopy device to generate an original impedance dataset. The original impedance dataset is then denoised using a Kalman filter algorithm to obtain a smoothed impedance dataset. The rate of change of impedance within a continuous time window is calculated based on the smoothed impedance dataset. If the rate of change sequence is lower than a preset threshold within the continuous time window, the preprocessing is considered to have reached a stable state.

[0020] The beneficial effects of this invention are: 1. This invention includes a support frame, a multi-layered housing at the lower end of the support frame for holding treatment agents of different concentrations, and a support mechanism at the upper end of the support frame for supporting the conductor. The housing is connected to a nozzle at the upper end of the support frame via a flexible hose. During operation, the conductor is supported by the support mechanism, and then different concentrations of treatment agent are sprayed through the nozzle according to the location of the contamination layer on the conductor, thereby achieving surface pretreatment of the live bare conductor. Furthermore, it can spray different concentrations of treatment agent according to the location and thickness of the contamination layer, and can establish an effective real-time monitoring and dynamic correction mechanism for surface pretreatment quality in a live environment. This enables accurate identification of the surface condition of different conductor materials and adaptive adjustment of pretreatment parameters, ensuring the uniformity and consistency of pretreatment quality.

[0021] 2. This invention employs a multispectral imaging sensor to scan the surface of the conductor, analyzes the reflectance spectral characteristics to obtain the thickness distribution and material information of the contamination layer, establishes a three-dimensional surface model, and performs region segmentation. Initial processing parameters are extracted from the database based on the material information and corrected according to the contamination density. The pretreatment quality is assessed by real-time monitoring of surface impedance changes, and processing parameters are dynamically adjusted. Laser interferometry is used to detect surface roughness and determine pretreatment quality. Machine learning algorithms are used to optimize pretreatment parameters and improve processing efficiency. Finally, an interpolation algorithm is used to generate a quality distribution map to ensure the uniformity of pretreatment quality across the entire conductor. This invention achieves intelligent and precise pretreatment of the surface of charged conductors, improving pretreatment efficiency and quality consistency, and providing a reliable guarantee for subsequent conductor maintenance. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a schematic diagram of the structure of the present invention. Figure 1 ; Figure 2 This is a schematic diagram of the structure of the present invention. Figure 2 ; Figure 3 This is a schematic diagram of the structure of the present invention. Figure 3 ; Figure 4 This is the front view of the present invention; Figure 5 This is a flowchart of the present invention.

[0024] In the diagram: 1. Support frame; 2. Box body; 3. Hose; 4. Support wheel; 5. Lifting frame; 6. Connecting rod; 7. Tension spring; 8. Connecting frame; 9. Collection trough; 10. Crossbar; 11. Elastic plate; 12. Connector. Detailed Implementation

[0025] like Figure 1As shown, a surface pretreatment device for insulating bare conductors includes a support frame 1, a box 2 with several layers for carrying different concentrations of treatment agent disposed at the lower end of the support frame 1, and a support mechanism disposed at the upper end of the support frame 1 for supporting the conductor. The box 2 is connected to a nozzle disposed at the upper end of the support frame 1 via a hose 3. During operation, the device is supported on the conductor by the support mechanism, and then different concentrations of treatment agent are sprayed through the nozzle according to the position of the contamination layer on the conductor, thereby achieving surface pretreatment of the bare conductor. Furthermore, it can spray different concentrations of treatment agent according to the position and thickness of the contamination layer, and can establish an effective real-time monitoring and dynamic correction mechanism for surface pretreatment quality in a live environment, achieving accurate identification of the surface state of different conductor materials and adaptive adjustment of pretreatment parameters, ensuring the uniformity and consistency of pretreatment quality.

[0026] like Figure 1 , Figure 2 and Figure 4 As shown, the support mechanism includes a support wheel 4 mounted on the upper end of the support frame 1 and a lifting frame 5 mounted and lifted within the support frame 1. The lifting frame 5 and the support wheel 4 are connected by a guide rail slider. One end of the housing of the support wheel 4 is rotatably connected to the support frame 1. The lifting frame 5 and one end of the housing of the support wheel 4 are connected by a connecting rod 6. The upper end of the connecting rod 6 is rotatably connected to the housing of the support wheel 4, and the lower end of the connecting rod 6 is rotatably connected to the lifting frame 5. The lower end of the support frame 1 is provided with a connecting mechanism for connecting the lifting frame 5. The lifting frame 5 is connected to the upper end of the support frame 1 by a tension spring 7. The support frame 1 is supported on the drone and raised to the position of the guide wire. Then, the support frame 1 continues to rise. The guide wire presses down on the housing of the support wheel 4, causing it to swing downwards. After the housing of the support wheel 4 separates from the guide wire, the lifting frame 5 moves upwards under the action of the tension spring 7. Under the action of the connecting rod 6, the housing of the support wheel 4 rotates upwards, the drone separates from the support frame 1, and the support wheel 4 is supported on the guide wire and can move on the guide wire. After the cleaning is completed, the drone continues to rise, driving the support frame 1 to rise. The wire presses down on the lifting frame 5. Under the action of the connecting rod, the lifting frame 5 descends and is fixed to the bottom through the connecting mechanism. At this time, the shaft of the support wheel 4 is in a vertical position, and the wire can leave the support frame 1.

[0027] like Figure 3 As shown, the lower end of the support frame 1 is provided with a connecting frame 8 for supporting the box 2, and the bottom of the connecting frame 8 is provided with a connecting plate for connecting to the drone.

[0028] like Figure 2 As shown, the uppermost box 2 is provided with a collection tank 9, which is used to collect the fallen treatment agent and reduce waste.

[0029] like Figure 2 As shown, the lifting frame 5 is provided with a groove that mates with the wire.

[0030] like Figure 2 As shown, the connecting mechanism includes a crossbar 10 disposed on the side of the lifting frame 5, an elastic plate 11 disposed on the support frame 1, a connector 12 disposed at the upper end of the elastic plate 11 and connected to the crossbar 10, the upper end of the connector 12 is provided with an inclined surface, and the lower end of the connector 12 is provided with a plane that cooperates with the crossbar.

[0031] like Figure 5 As shown, a surface pretreatment method for insulating bare conductors using the aforementioned apparatus includes the following steps: S101 Obtain the spectral feature dataset of the contamination layer on the conductor surface, extract the thickness distribution and material identification information of the contamination layer using wavelength analysis technology, and generate contamination layer feature data; S102 Construct a three-dimensional model of the conductor surface based on the contamination layer thickness distribution data, segment the region using a mesh generation algorithm, calculate the contamination density value, and determine the baseline values ​​of the preprocessing parameters; S103 Query the pre-established mapping database based on the material identification information, obtain the initial preprocessing parameters, and correct them according to the contamination density value to generate the final preprocessing parameter configuration; S104 Monitor the impedance change during the preprocessing process using electrochemical impedance spectroscopy to determine the stable state of the preprocessing; S105 Establish a quality assessment model based on the impedance change trend, calculate the quality assessment value, and dynamically adjust the parameters; S106 Use laser interferometry to detect surface roughness, determine the preprocessing quality, and record the parameter combination; S107 Train historical data using machine learning algorithms to predict the optimal preprocessing parameters; S108 Generate a conductor quality distribution map, and adjust the deviation areas until the quality uniformity meets the standard.

[0032] Step S101 includes the following steps: scanning the surface of the conductor with a multispectral imaging sensor to obtain reflectance spectral data at different wavelengths and generating a spectral feature dataset; extracting contamination layer thickness distribution and material identification information from the spectral feature dataset using wavelength analysis technology to generate contamination layer feature data; if there are areas in the contamination layer thickness distribution data that exceed a preset threshold, generating marker data for key processing areas based on the thickness distribution data.

[0033] By scanning the surface of a charged conductor with a multispectral imaging sensor, reflectance spectral data at different wavelengths can be acquired, forming a spectral feature dataset. This process is based on multispectral imaging technology, utilizing the interaction between light of different wavelengths and the contamination layer on the conductor surface to generate feature data reflecting material and thickness information.

[0034] Wavelength analysis techniques extract information on contaminant layer thickness distribution and material identification from spectral feature datasets. For example, principal component analysis can be used to separate key features from spectral data, and combined with a pre-trained material spectral database, the main components in the contaminant layer, such as grease, dust, or salts, can be identified.

[0035] If any area in the contamination layer thickness distribution data exceeds a preset threshold, then marker data for key treatment areas is generated. For example, if the thickness threshold is set to 0.5 mm, when the contamination layer thickness on a certain section of the conductor reaches 0.7 mm, the system automatically marks that area as a key treatment area.

[0036] Step S102 includes the following steps: acquiring the contamination layer thickness distribution data on the conductor surface; constructing a three-dimensional model of the conductor surface using three-dimensional scanning technology to obtain the contamination layer thickness distribution; using an adaptive mesh generation algorithm to divide the three-dimensional model into meshes, adjusting the mesh density according to the contamination layer thickness distribution to obtain a set of mesh cells; calculating the area A of each mesh cell using the set of mesh cells, and combining it with the contaminant mass M, calculating the contamination density using the formula D=M / A to obtain the contamination density value of each mesh cell; if the contamination density value is greater than a preset threshold, the corresponding mesh cell is marked as a high-contamination area, resulting in a set of high-contamination areas.

[0037] Based on the set of highly polluted areas, the average pollution density of each area is obtained, and the baseline values ​​of preprocessing parameters are calculated using a linear interpolation method to obtain the baseline values ​​of preprocessing parameters for each area.

[0038] By preprocessing the baseline parameter values, the k-means clustering algorithm is used to classify the grid cells, resulting in a set of classified grid cells.

[0039] Based on the classified grid cell set, a preprocessing parameter distribution map of each region is generated to determine the final preprocessing scheme.

[0040] For example, by constructing a 3D model of the conductor surface using 3D scanning technology, precise data on the thickness distribution of the contamination layer can be obtained. The scanning equipment uses lidar or structured light scanning to cover the conductor surface, generating high-resolution point cloud data. This point cloud data is then processed to form a 3D model that reflects the geometric features of the conductor surface and variations in the contamination layer thickness.

[0041] When calculating pollution density, the pollution density value is obtained using the formula D=M / A, which combines the area of ​​a grid cell and the mass of the pollutant. Assuming a grid cell has an area of ​​100 square centimeters and a pollutant mass of 5 grams, the calculated pollution density is 0.05 grams per square centimeter. If the preset threshold is 0.04 grams per square centimeter, this grid cell is marked as a high-pollution area.

[0042] Assuming the average pollution density of a highly polluted area is 0.06 g / cm², the baseline values ​​of pretreatment parameters for each grid cell, such as cleaning intensity or reagent dosage, are obtained through interpolation.

[0043] The k-means clustering algorithm classifies grid cells into high, medium, and low pollution levels based on their pollution density values.

[0044] Step S103 includes the following steps: Obtaining the conductor material identification result; extracting conductor material information from the material identification system; determining the material type as aluminum alloy or copper conductor; obtaining the material classification result; querying the pre-established material-pretreatment parameter mapping database based on the material classification result; extracting the first concentration value and first action time if the material is aluminum alloy, and the second concentration value and second action time if the material is copper conductor; obtaining the initial pretreatment parameters; obtaining the contamination density value of the conductor surface from the contamination density detection system; calculating the impact of contamination density on the initial pretreatment parameters using a preset contamination density-parameter correction model; obtaining the correction coefficient; adjusting the treatment agent concentration and action time in the initial pretreatment parameters using a linear regression algorithm combined with the correction coefficient; obtaining the first adjustment parameter; using the first adjustment parameter, determining whether the adjusted parameters meet the preset process constraints using a decision tree algorithm; if not, iteratively optimizing the parameters to obtain the second adjustment parameter; generating the final pretreatment parameter configuration using the second adjustment parameter, including the adjusted treatment agent concentration and action time, and outputting it to the pretreatment control system. Based on the final pretreatment parameter configuration, update the parameter settings of the pretreatment control system to complete the automated adjustment of the wire pretreatment process.

[0045] When retrieving parameters from the material-pretreatment parameter mapping database, if the wire is aluminum alloy, the database may return a first concentration value of 20% and a first treatment time of 60 seconds; if the wire is copper, it will return a second concentration value of 15% and a second treatment time of 45 seconds. This difference stems from the thicker oxide layer on the surface of aluminum alloy, requiring a higher concentration and longer treatment time to ensure cleaning effectiveness.

[0046] A linear regression algorithm is used to adjust the initial preprocessing parameters. Assuming the initial parameters are a concentration of 20% and a time of 60 seconds, with a correction factor of 1.2, the adjusted parameters may be a concentration of 24% and a time of 72 seconds.

[0047] The decision tree algorithm is used to verify whether the adjusted parameters meet the process constraints. Constraints may include a concentration not exceeding 30% or a time not exceeding 90 seconds. If the first adjusted parameter has a concentration of 24% and a time of 72 seconds, the conditions are met, and the algorithm passes directly. If the concentration exceeds the limit, iterative optimization is used to adjust the concentration to 28% and the time to 80 seconds. The decision tree ensures that the parameters meet the process requirements through multi-level conditional judgments.

[0048] Step S104 includes the following steps: Impedance signals from the conductor surface are acquired using an electrochemical impedance spectroscopy device to generate an original impedance dataset. The original impedance dataset is then denoised using a Kalman filter algorithm to obtain a smoothed impedance dataset. Based on the smoothed impedance dataset, the rate of change of impedance within a continuous time window is calculated to obtain a rate of change sequence. If the rate of change sequence is below a preset threshold for three consecutive seconds, the conductor surface pretreatment is deemed to have reached a stable state, and a stable state flag is generated. Based on the stable state flag, the corresponding timestamp and impedance value are extracted to generate a preprocessing completion dataset. Using the preprocessing completion dataset, the statistical characteristics of the conductor surface impedance are calculated to obtain a feature parameter set. The feature parameter set is then classified using a support vector machine algorithm to determine the stability quality level of the conductor surface pretreatment, resulting in a quality assessment result.

[0049] In the scenario of wire surface pretreatment, an electrochemical impedance spectroscopy (EIS) device measures the impedance response of the wire surface at different frequencies by applying an AC signal. The device can acquire data within the frequency range of 0.1 Hz to 100 kHz, generating a raw dataset containing information such as complex impedance and phase angle.

[0050] Assuming that there is random noise in the original data, the variance of the impedance value can be significantly reduced after filtering, and the smoothed dataset can more accurately reflect the true state of the conductor surface.

[0051] Based on the smoothed impedance dataset, the rate of change of impedance within a continuous time window is calculated to obtain a rate of change sequence, which aims to quantify the dynamic changes in the surface state of the conductor. For example, a 3-second time window is selected to calculate the relative rate of change of impedance value per second. If the rate of change sequence is lower than a preset threshold (e.g., 0.5%) for three consecutive seconds, the conductor surface preprocessing is determined to have reached a stable state, and a stable state indicator is generated.

[0052] Step S105 includes the following steps: acquiring impedance change, action time, and reagent concentration data, forming an initial dataset through real-time data acquisition via sensors. Based on the initial dataset, a linear regression algorithm is used to determine the weighting coefficients α, β, and γ, resulting in a quality assessment model Q = α × ΔZ + β × T + γ × C. The quality assessment model is used to calculate the quality assessment value of the current treatment process. If the quality assessment value is lower than a preset threshold, a dynamic adjustment mechanism is triggered, generating an adjustment command. Based on the adjustment command, an optimization algorithm is used to adjust the reagent concentration or action time, obtaining updated parameter values. Using the updated parameter values, impedance change, action time, and reagent concentration data are reacquired to generate a new dataset. Based on the new dataset, the quality assessment value is recalculated to determine whether it meets the preset threshold, obtaining the final treatment parameters.

[0053] The change in impedance reflects the dynamic changes on the wire surface during the electrochemical reaction, usually measured in ohms. For example, in a certain experiment, the change in impedance decreased from an initial 50 ohms to 45 ohms.

[0054] The action time refers to the duration when the treatment agent contacts the wire surface, usually measured in seconds, such as 60 seconds. The concentration of the treatment agent is expressed in moles per liter, such as 0.1 mol / L.

[0055] Based on the initial data set, the linear regression algorithm is used to determine the weight coefficients α, β, and γ. For example, by analyzing multiple sets of experimental data, it is assumed that α = 0.4, β = 0.2, and γ = 0.3, forming a quality assessment model Q = 0.4×ΔZ + 0.2×T + 0.3×C. This model comprehensively considers the effects of the change in impedance ΔZ, the action time T, and the concentration of the treatment agent C on the pretreatment quality.

[0056] Suppose in a certain experiment, the change in impedance ΔZ is 5 ohms, the action time T is 60 seconds, and the concentration of the treatment agent C is 0.1 mol / L. Substituting into the model gives Q = 0.4×5 + 0.2×60 + 0.3×0.1 = 14.03. If the preset threshold is 15, then the current quality assessment value is lower than the threshold, indicating that the pretreatment quality does not meet the standard and needs to be further optimized.

[0057] Step S106 includes the following steps: Use laser interferometry to scan the wire surface to obtain surface profile data. Process the surface profile data through the fast Fourier transform algorithm to obtain the profile height distribution. According to the profile height distribution, calculate the root mean square deviation Ra value to determine the surface roughness characteristics. If the root mean square deviation Ra value is within the preset range, it is determined that the pretreatment quality is qualified, and a quality determination result is generated. According to the quality determination result, extract the combination of treatment parameters for the qualified area and store it as a parameter record. For the parameter record, use the K-means clustering algorithm to analyze the parameter distribution characteristics to obtain an optimized parameter set. Through the optimized parameter set, update the pretreatment process database and record the correspondence between the detection area and the parameters.

[0058] Step S107 includes the following steps: Obtain the surface state features of the conductor and the corresponding treatment agent concentration and treatment time from historical preprocessed data to construct a training dataset. Train the training dataset using the random forest algorithm to generate a parameter optimization model, obtaining the mapping relationship between the conductor surface state features and the treatment parameters. Obtain newly detected conductor surface state features, input them into the parameter optimization model, predict the treatment agent concentration and treatment time, and obtain the predicted parameters. Evaluate the accuracy of the predicted parameters using cross-validation to obtain the prediction accuracy. If the prediction accuracy is higher than a preset threshold, the predicted parameters are used for processing to obtain the processing result; if the prediction accuracy is lower than or equal to the preset threshold, standard parameters are used for processing to obtain the processing result. Update the historical preprocessed data based on the processing result to generate an updated training dataset. Retrain the parameter optimization model using the updated training dataset to obtain the optimized mapping relationship.

[0059] Assuming a conductor surface has a peak-to-valley height of 2.5 micrometers and a corrugation spacing of 0.1 millimeters, and the treatment agent concentration is recorded at 5% and the treatment time at 30 seconds, this constitutes one training dataset. Historical data may contain hundreds of similar records, covering different surface conditions and parameter combinations, forming a training dataset.

[0060] When training the random forest algorithm, surface state features can be used as input, and treatment concentration and treatment time can be used as output to construct multiple decision trees.

[0061] After training, the model can capture the complex nonlinear relationship between features and parameters.

[0062] Specifically, after the newly detected surface features of the conductor are input into the model, suitable processing parameters can be predicted. Assuming the peak-valley height of the new conductor is 2.3 micrometers and the corrugation spacing is 0.12 millimeters, the model may output a concentration of 5.2% and a time of 31 seconds.

[0063] When evaluating prediction accuracy through cross-validation, ten-fold cross-validation can be used, which divides the dataset into ten parts: nine for training and one for testing, and repeats this process ten times.

[0064] Suppose that in a certain verification, the predicted concentration deviation is less than 0.2%, the time deviation is less than 2 seconds, and the accuracy reaches 95%. If the preset threshold is 90%, the prediction parameters are used; if the accuracy is only 85%, the standard parameters are switched, such as a concentration of 5% and a time of 30 seconds.

[0065] Acquire preprocessed conductor quality data for each region, and obtain initial quality datasets by collecting quality parameters of each conductor segment through sensors.

[0066] Step S108 includes the following steps: Using the Lagrange interpolation algorithm, a quality distribution map of the entire conductor segment is generated based on the initial quality dataset, resulting in a continuous quality distribution curve. Quality values ​​of adjacent regions are extracted from the quality distribution curve. If the quality deviation between adjacent regions exceeds a preset threshold, regions with larger deviations are marked, resulting in a list of regions to be processed. For the list of regions to be processed, a local weighted regression algorithm is used to predict the quality of the marked regions, obtaining a parameter set for supplementary processing. Based on the parameter set for supplementary processing, the preprocessing parameters of the regions to be processed are adjusted to generate an updated quality dataset. The conductor quality distribution map is regenerated from the updated quality dataset. If the uniformity of the quality distribution across the entire segment meets a preset standard, the final quality distribution map is output; otherwise, the process returns to the step of marking regions with larger deviations, and iterative processing continues.

Claims

1. A surface pretreatment apparatus for insulating bare conductors, characterized in that, It includes a support frame (1), a box (2) with several layers at the lower end of the support frame (1) for carrying different concentrations of treatment agents, and a support mechanism at the upper end of the support frame (1) for supporting the wire. The box (2) is connected to the nozzle at the upper end of the support frame (1) through a hose (3). The support mechanism includes a support wheel (4) set on the upper end of the support frame (1) and a lifting frame (5) set inside the support frame (1) and raised and lowered inside the support frame (1). One end of the housing of the support wheel (4) is rotatably connected to the support frame (1). The lifting frame (5) is connected to one end of the housing of the support wheel (4) through a connecting rod (6). The upper end of the connecting rod (6) is rotatably connected to the housing of the support wheel (4), and the lower end of the connecting rod (6) is rotatably connected to the lifting frame (5). The lower end of the support frame (1) is provided with a connecting mechanism for connecting the lifting frame (5). The lifting frame (5) is connected to the upper end of the support frame (1) through a tension spring (7).

2. The surface pretreatment apparatus for insulating bare conductors as described in claim 1, characterized in that, The lower end of the support frame (1) is provided with a connecting frame (8) for supporting the box body (2).

3. The surface pretreatment apparatus for insulating bare conductors as described in claim 1, characterized in that, The uppermost box (2) is provided with a collection trough (9).

4. The surface pretreatment apparatus for insulating bare conductors as described in claim 1, characterized in that, The lifting frame (5) is provided with a groove that cooperates with the guide wire.

5. The surface pretreatment apparatus for insulating bare conductors as described in claim 1, characterized in that, The connecting mechanism includes a crossbar (10) disposed on the side of the lifting frame (5), an elastic plate (11) is provided on the support frame (1), and a connector (12) connected to the crossbar (10) is provided at the upper end of the elastic plate (11).

6. A surface pretreatment method for insulating bare conductors, utilizing the apparatus according to any one of claims 1 to 5, characterized in that, Includes the following steps: S101 Obtain the spectral feature dataset of the contamination layer on the conductor surface, extract the contamination layer thickness distribution and material identification information through wavelength analysis technology, and generate contamination layer feature data; S102 Construct a three-dimensional model of the conductor surface based on the contamination layer thickness distribution data, divide the region using a mesh generation algorithm, calculate the contamination density value, and determine the baseline values ​​of the preprocessing parameters. S103 queries the pre-established mapping database based on the material identification information, obtains the initial preprocessing parameters, corrects them according to the pollution density value, and generates the final preprocessing parameter configuration. S104 Monitor the impedance change during the pretreatment process using electrochemical impedance spectroscopy to determine the stable state of the pretreatment; S105 Establish a quality assessment model based on the impedance change trend, calculate the quality assessment value, and dynamically adjust the parameters; S106 Use laser interferometry to detect surface roughness, determine the pretreatment quality, and record the parameter combination; S107 Train historical data using machine learning algorithms to predict the optimal pretreatment parameters; S108 Generate a conductor quality distribution map and adjust the deviation area until the quality uniformity meets the standard.

7. The surface pretreatment method for insulating bare conductors as described in claim 6, characterized in that, Step S101 includes the following steps: The surface of the conductor is scanned by a multispectral imaging sensor to obtain reflectance spectral data at different wavelengths, generating a spectral feature dataset. Wavelength analysis technology is used to extract the contamination layer thickness distribution and material identification information from the spectral feature dataset, generating contamination layer feature data. If there are areas in the contamination layer thickness distribution data that exceed a preset threshold, then the marking data of key treatment areas is generated based on the thickness distribution data.

8. The surface pretreatment method for insulating bare conductors as described in claim 6, characterized in that, Step S103 includes the following steps: Material information of the conductor is extracted from the material identification system to determine the material type and obtain the material classification result. Based on the material classification result, the pre-established material-preprocessing parameter mapping database is queried to obtain the initial preprocessing parameters. The influence of pollution density on the initial preprocessing parameters is calculated using a preset pollution density-parameter correction model to obtain the correction coefficient. The initial preprocessing parameters are adjusted by combining the correction coefficient with a linear regression algorithm to generate the final preprocessing parameter configuration.

9. The surface pretreatment method for insulating bare conductors as described in claim 6, characterized in that, Step S104 includes the following steps: Impedance signals from the surface of a conductor are collected using an electrochemical impedance spectroscopy device to generate an original impedance dataset. The original impedance dataset is then denoised using a Kalman filter algorithm to obtain a smoothed impedance dataset. The rate of change of impedance within a continuous time window is calculated based on the smoothed impedance dataset. If the rate of change sequence is lower than a preset threshold within the continuous time window, the preprocessing is considered to have reached a stable state.

Images