A contact line self-adaptive polishing method based on profile measurement

Abstract

The application provides a contact line adaptive polishing method based on profile measurement, relates to the technical field of electrified railway component maintenance, and solves the problems of low precision and efficiency of the existing contact line polishing operation. In the method, the cross-sectional profile data of the contact line is collected, the wear range is determined, the edge of the wear surface is fitted with a circular arc, and the fillet parameter representing the wear state is extracted; the contact line is tested and polished at a constant pressure, and the response coefficient between the constant pressure of each measurement point and the change amount of the fillet parameter is calculated; then in the normal polishing process, the deviation of the measured fillet parameter and the preset target threshold value is calculated, and the deviation and the corresponding response coefficient are used to calculate the optimal polishing pressure of each point in the next polishing; the polishing device is controlled to perform the polishing operation on the contact line, the fillet parameter is updated, and the polishing pressure calculation and the polishing operation are continuously performed. The application realizes the differentiated and accurate adaptive polishing of each point of the contact line in the longitudinal direction, and improves the overall efficiency.

Description

Technical Field

[0001] This invention relates to the field of maintenance technology for electrified railway components, and specifically to an adaptive grinding method for contact wires based on profile measurement. Background Technology

[0002] The contact wire is a core component of the electrified railway power supply system, transmitting electrical energy through direct contact between the pantograph and the locomotive. During long-term operation, the contact wire surface gradually wears down due to the combined effects of mechanical friction, arc erosion, and environmental factors, causing changes in its cross-sectional profile. The degree of wear directly affects the quality of current collection; excessive wear leads to a reduction in cross-sectional area and mechanical strength, potentially causing wire breakage accidents in severe cases. Simultaneously, changes in the radius of the transition fillet at the edge of the worn surface affect the smooth transition of the pantograph's contact wire, exacerbating localized wear and reducing current collection stability. Therefore, regular grinding and maintenance of the contact wire to restore its ideal cross-sectional profile is a crucial aspect of ensuring the safe operation of electrified railways.

[0003] Current contact wire grinding operations mainly rely on manual experience or fixed parameters. Traditional manual grinding methods depend heavily on the practical experience of on-site operators, who judge the degree of wear by visual inspection or simple measuring tools and set grinding parameters based on experience. This method lacks scientific basis, cannot perform differentiated and precise grinding according to the actual wear condition of different locations on the contact wire, and is prone to over-grinding or under-grinding. Furthermore, the quality of the work is greatly affected by the skill level of the personnel, resulting in poor stability.

[0004] Although some existing equipment uses preset fixed grinding parameters to perform grinding with the same parameters throughout the entire section, this method fails to establish a dynamic mapping relationship between the actual profile of the contact line and the grinding parameters. It cannot adaptively adjust the grinding pressure according to the actual wear level of each measurement point, resulting in wasted materials and time in areas with shallow wear, and insufficient grinding in areas with deep wear. Overall, the grinding accuracy and efficiency are still low.

[0005] In summary, current contact line grinding operations face numerous technical bottlenecks. Existing equipment generally lacks point-to-point precision grinding capabilities, fails to establish a quantitative mapping relationship between profile measurements and grinding pressure, and cannot perform differentiated pressure calculations based on the actual conditions of each longitudinal measurement point. The determination of key grinding parameters lacks automated algorithm support, resulting in low standardization of operations.

[0006] Existing methods typically separate the measurement and grinding processes, failing to establish a control mechanism that integrates both. This prevents dynamic strategy adjustments based on real-time profile data during operation, resulting in a lack of quantitative feedback and iterative optimization of grinding effects. Furthermore, existing equipment often lacks a systematic calibration process before formal operation, hindering the automatic learning and establishment of pressure-wear change response relationships at various points through trial grinding. This leads to a lack of scientific basis for subsequent pressure calculations. The inability to automatically determine whether a section needs further grinding and to automatically calculate the optimal pressure distribution results in low efficiency in generating grinding plans and redundant grinding passes.

[0007] Therefore, there is an urgent need for an adaptive grinding method that can achieve data-driven decision-making, iterative control, and automatically generate the optimal grinding scheme, so as to improve the automation level, accuracy, and efficiency of contact line grinding operations. Summary of the Invention

[0008] The purpose of this invention is to address the problems of low grinding accuracy and efficiency in existing contact wire grinding operations, such as the lack of point-to-point precision grinding capabilities, the heavy reliance on manual experience in parameter setting, and the inability to generate feedback. This invention proposes an adaptive contact wire grinding method based on profile measurement. By extracting contact wire wear parameters in real time and learning the pressure response coefficient, it automatically calculates the optimal grinding pressure and planned number of grinding passes for each point, constructing a control scheme that coordinates measurement and grinding. This invention achieves differentiated, precise, and adaptive grinding at various points along the longitudinal direction of the contact wire, effectively reducing manual intervention and significantly improving the automation level, quality consistency, and overall efficiency of the grinding operation.

[0009] The present invention employs the following technical solutions to achieve its objective: An adaptive grinding method for contact lines based on profile measurement, comprising the following steps: S1. Collect cross-sectional profile data of each measurement point of the contact line, preprocess and identify the cross-sectional profile data to determine the wear range; based on the wear range, perform arc fitting on the edge of the wear surface to extract the rounded corner parameters characterizing the wear state. S2. Control the grinding device to perform test grinding on the contact line with a set constant pressure, and obtain the change in the fillet parameter before and after the test grinding; based on the constant pressure and the change in the fillet parameter, calculate the response coefficient between the constant pressure and the change in the fillet parameter at each measurement point, and complete the calibration; S3. During the normal polishing process after calibration, obtain the currently measured fillet parameters and calculate the deviation between the currently measured fillet parameters and the preset target threshold; use the response coefficient and the deviation to calculate the optimal polishing pressure for each measurement point in the next polishing pass. S4. Based on the optimal grinding pressure, control the grinding device to perform grinding operations on the contact line, and after grinding is completed, re-collect the cross-sectional profile data and update the fillet parameters; based on the updated fillet parameters, determine whether the preset grinding end condition is met. If not, use the updated fillet parameters as the currently measured fillet parameters, and repeat steps S3 to S4 until the grinding end condition is met, thus completing the adaptive grinding of the contact line.

[0010] Furthermore, in step S1, the collected original cross-sectional profile data is cleaned, invalid marker points are removed, and the cleaned profile coordinates are normalized, scaled, and sorted; wherein, the cross-sections corresponding to the original cross-sectional profile data include the contact line cross-section and the clamping component cross-section. Using the central axis of the original cross-sectional profile data as the dividing axis, the sliding window width is set for one side of the dividing axis, and the normalized data is divided into multiple windows. The point with the smallest ordinate is found in each window to form a sequence of minimum points. Search for a transition pattern of first decreasing and then increasing in the minimum point sequence, and record the corresponding transition point as the characteristic valley position of the single-sided profile. The characteristic valley position is located on the cross section of the clamping member. Find the point with the largest ordinate in the region from the characteristic valley position toward the contact line cross section profile, and use it as the characteristic peak position of the single-sided profile. Flip the data on the other side of the profile along the dividing axis and repeat the above process to obtain the corresponding feature valley position and feature peak position on the other side, thus completing the feature identification of the left and right sides of the profile, which serves as the basis for determining the wear range.

[0011] Furthermore, in step S1, determining the wear range includes: dividing the data between the characteristic peak positions on the left and right sides of the profile into windows again with a set wear window width, and taking the point with the largest ordinate in each window to form a sequence of maximum points; Calculate the first-order slope between adjacent points in the maximum point sequence and the second-order difference of the first-order slope; Find a continuous interval where the absolute value of the second-order difference is less than a preset difference threshold and the absolute value of the first-order slope is less than a preset slope threshold, and define the continuous interval as a flat area in the contact line cross-sectional profile corresponding to the wear surface. The starting point and ending point of the flat area are respectively taken as the left and right boundary points of the wear range.

[0012] Furthermore, in step S1, based on the wear range, the edge of the wear surface is fitted with an arc, including: selecting the boundary point with the smaller ordinate from the left and right boundary points of the wear range as the basis for arc fitting; taking the selected boundary point as the center point, selecting a preset number of data points on both sides of the center point, and using the least squares method to fit these data points into a circle, and the radius value of the fitted circle is the fillet parameter of the corresponding measurement point.

[0013] Preferably, after selecting a preset number of data points, an outlier filtering algorithm based on the distance between points is used to remove outliers from the selected data points to obtain a precise point set for the transition arc; the precise point set for the transition arc is fitted using the least squares method, and the radius value of the fitted circle is the fillet parameter.

[0014] Specifically, in step S2, the response coefficient between the constant pressure and the change in the fillet parameter at each measurement point is calculated, including: obtaining the initial fillet parameter of the measurement point before the test grinding, and the calibrated fillet parameter of the measurement point after the test grinding; calculating the difference between the initial fillet parameter and the calibrated fillet parameter to obtain the change in the fillet parameter; and dividing the change in the fillet parameter by the constant pressure to calculate the response coefficient.

[0015] Furthermore, in step S3, the optimal polishing pressure for each measurement point in the next polishing pass is calculated using the response coefficient and the deviation. This includes: subtracting the currently measured fillet parameter from the preset target threshold to obtain the deviation; dividing the deviation by the response coefficient to obtain the theoretical polishing pressure for each measurement point in the next polishing pass; smoothing and limiting the theoretical polishing pressure to obtain the optimal polishing pressure; and simulating the change in fillet parameter after a single polishing pass based on the optimal polishing pressure and the response coefficient, and planning the number of subsequent polishing passes required to reach the preset target threshold.

[0016] Preferably, the optimal grinding pressure is obtained by smoothing and limiting the theoretical grinding pressure, including: performing a moving average filtering process on the theoretical grinding pressure of adjacent measurement points along the longitudinal direction of the contact line; and comparing the processed grinding pressure with preset minimum grinding pressure threshold and maximum grinding pressure threshold. If the processed grinding pressure is less than the minimum grinding pressure threshold, then the optimal grinding pressure is set to the minimum grinding pressure threshold; if the processed grinding pressure is greater than the maximum grinding pressure threshold, then the optimal grinding pressure is set to the maximum grinding pressure threshold.

[0017] Specifically, in step S4, based on the updated fillet parameters, it is determined whether the preset grinding end condition is met, including: determining whether the updated fillet parameters fall within a preset tolerance range centered on the preset target threshold. If the updated fillet parameters fall within the preset tolerance range, the polishing end condition is determined to be met; if the updated fillet parameters do not fall within the preset tolerance range, it is further determined whether the current cumulative polishing passes have reached the preset maximum polishing pass threshold. If the maximum polishing pass threshold is reached, the polishing end condition is determined to be met.

[0018] Specifically, in step S1, the cross-sectional profile data of each measurement point of the contact line is collected. Specifically, the profile measurement unit set in front of the grinding device is used to collect a two-dimensional coordinate point array of each measurement point in the longitudinal direction of the contact line, including the contact line and the clamping part, in real time as the grinding device moves along the longitudinal direction of the contact line, as the cross-sectional profile data. While extracting the fillet parameters that characterize the wear state, the wear surface width, wear angle, remaining height, and left and right wear surface heights are also extracted as auxiliary wear parameters based on the cross-sectional profile data. The auxiliary wear parameters are associated with the corresponding measurement points and stored for real-time monitoring of the grinding state and historical data tracing.

[0019] In summary, due to the adoption of this technical solution, the beneficial effects of this invention are as follows: This invention automatically learns the response coefficient of each measurement point through a calibration algorithm, establishing a quantitative mapping between pressure and fillet parameter changes. In application, based on this mapping, the grinding pressure is calculated independently for each point, truly realizing differentiated and precise grinding according to specific conditions, effectively avoiding over-grinding and under-grinding.

[0020] The method of this invention covers a complete decision chain from profile measurement → feature extraction → calibration learning → pressure calculation → effect simulation → feedback iteration. Operators do not need to judge the polishing parameters; the method automatically completes all judgments and calculations regarding whether the fillet parameters meet the standards, whether polishing is necessary, the required pressure, and the number of polishing passes, achieving complete automation in the generation of polishing plans.

[0021] This invention can simulate the polishing effect before each actual polishing, predict the distribution of fillet parameters after polishing, and intelligently determine whether subsequent polishing passes are needed. After the actual polishing, the effect is remeasured and evaluated, and the pressure for the next round is dynamically updated, forming a cyclical process of "measurement → calculation → simulation → polishing → remeasurement → recalculation" to ensure that the polishing effect is controllable throughout the process.

[0022] The calibration operation in this invention only requires the first round trip to learn the response coefficient curve of the entire line. The response coefficient contains complex information such as material properties and geometric response characteristics of each point. All subsequent grinding operations can be directly reused without repeating the calibration process, which not only ensures the consistency of pressure calculation, but also greatly improves the work efficiency.

[0023] This invention uses a "calculation → simulation → judgment" predictive mechanism to automatically determine the minimum necessary number of polishing passes, avoiding redundant polishing passes in the traditional fixed mode, and maximizing polishing efficiency while ensuring quality.

[0024] The pressure smoothing operation used in this invention eliminates pressure jumps caused by measurement noise, allowing the grinding force to transition smoothly in a step-like manner along the longitudinal direction of the contact line, thus ensuring the uniformity of the surface quality after grinding. Attached Figure Description

[0025] The technical solution of this invention is described in detail with reference to the following figures, specifically including two figures as follows: Figure 1 This is a schematic diagram illustrating the overall process of the adaptive grinding method for contact lines of the present invention. Figure 2 This is a schematic diagram of the cross-sectional profile of the contact wire and its clamping member in the method of the present invention. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The parts of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0027] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0028] An adaptive grinding method for contact lines based on profile measurement. Figure 1 The overall process of this method is briefly described below and can be viewed concurrently; the key steps of this method can be summarized as follows: S1. Collect cross-sectional profile data of each measurement point of the contact line, preprocess and identify the features of the cross-sectional profile data to determine the wear range; based on the wear range, perform arc fitting on the edge of the wear surface to extract the fillet parameters characterizing the wear state. S2. Control the grinding device to perform test grinding on the contact line with a set constant pressure, and obtain the change in fillet parameters before and after test grinding; based on the constant pressure and the change in fillet parameters, calculate the response coefficient between the constant pressure and the change in fillet parameters at each measurement point, and complete the calibration. S3. During the normal polishing process after calibration, obtain the currently measured fillet parameters and calculate the deviation between the currently measured fillet parameters and the preset target threshold; use the response coefficient and deviation to calculate the optimal polishing pressure for each measurement point in the next polishing pass. S4. Based on the optimal grinding pressure, control the grinding device to perform grinding operations on the contact line, and re-collect the cross-sectional profile data after grinding and update the fillet parameters. Based on the updated fillet parameters, determine whether the preset grinding end condition is met. If not, use the updated fillet parameters as the current measured fillet parameters and repeat steps S3 to S4 until the grinding end condition is met, thus completing the adaptive grinding of the contact line.

[0029] In this embodiment, a profile measurement unit located in front of the grinding device is used to collect cross-sectional profile data of each measurement point on the contact line in real time, including the profile of the contact line and the clamping parts on both sides. The key parameter characterizing the wear state, namely the fillet parameter, is extracted through a measurement algorithm. Then, in the calibration stage, a constant pressure test grinding is conducted to establish the "pressure-fillet parameter change" response coefficient for each measurement point. Finally, in the application stage, the optimal grinding pressure for each measurement point is automatically calculated based on the deviation between the current measured fillet parameter and the target threshold using the calibrated response coefficient, and the grinding effect is simulated to plan the number of subsequent grinding passes, thereby forming an adaptive grinding control process of "measurement-grinding-measurement".

[0030] The method steps of this implementation can be summarized into three main algorithms: a profile measurement algorithm, a calibration algorithm, and an application algorithm. The profile measurement algorithm automatically extracts fillet parameters characterizing the wear state from the cross-sectional profile data; the calibration algorithm learns the pressure-fillet parameter response coefficients at each point through a single constant pressure grinding test; and the application algorithm calculates the grinding pressure and plans the number of grinding passes using the calibration coefficients. After these three algorithms work together, a complete adaptive grinding control process of "measurement → calculation → grinding → re-measurement" is formed.

[0031] This implementation will now elaborate on the algorithm logic, processing steps, and input-output relationships of the three algorithms mentioned above.

[0032] I. Profile Measurement Algorithm Before introducing the profile measurement algorithm, this implementation method first explains the evolution law of contact line wear state and fillet parameter.

[0033] In electrified railway systems, the contact wire and the pantograph slide plate are in a state of high-speed sliding friction for a long time. This continuous mechanical wear will cause a significant change in the geometry of the bottom cross-section of the contact wire. Under the repeated friction of the pantograph, the bottom of the contact wire will gradually be "flattened" to form a relatively flat wear surface. At the junction of this flat wear surface and the original contours on both sides of the contact wire, due to the non-uniform removal of material, a transitional, seemingly arc-shaped contour will naturally form, which is referred to as the wear surface edge fillet in this embodiment.

[0034] From a geometric evolution perspective, the degree of wear on the contact wire is related to the size of the edge fillet. As wear intensifies, the worn surface gradually widens and deepens, and the material in the edge transition area is further cut away, leading to a continuous increase in the curvature of the approximate transition arc. Reflected in geometric parameters, the radius of the fitted circle corresponding to this arc-shaped profile, i.e., the fillet parameter, gradually decreases. A smaller fillet parameter means a sharper edge and a more abrupt transition. This sharp edge not only easily causes localized stress concentration on the contact wire, accelerating conductor fatigue and increasing the risk of wire breakage, but also leads to mechanical impact, hard spots, and even arcing when the pantograph slide passes, severely deteriorating the dynamic current collection quality of the pantograph-catenary system.

[0035] To eliminate this undesirable geometric condition, the contact line must be ground and repaired. One of the core objectives of this grinding process is not only to eliminate localized unevenness on the worn surface, but more importantly, to smooth out the sharp transitions at the edges of the worn surface. After proper grinding, the interface between the worn surface and the side contours will be smoothed and leveled. Geometrically, this physical repair process manifests as a reduction in the curvature of the transition arc, meaning the "corner radius" of the edge is increased again, restoring it to a level close to the original design or meeting the ideal threshold range for safe operation.

[0036] The main function of the profile measurement algorithm is to input the two-dimensional profile coordinate data of a measurement point on the contact line, automatically identify the geometric features of the contact line, locate the wear area, fit the transition arc of the wear surface edge, and output its fillet parameters. The algorithm outputs wear parameters, which characterize the degree of wear. It can also output other wear parameters that are only used for data recording, such as wear surface width, wear angle, remaining height, and height of the left and right wear surfaces, which facilitates data recording, analysis and traceability.

[0037] The input to the profile measurement algorithm is the original cross-sectional profile data of a single measurement point. Since there are clamping elements on both sides of the contact line, they provide obvious analytical features, facilitating algorithm analysis and processing. Therefore, the measurement points selected by the profile measurement unit of the grinding device all correspond to clamping elements. The cross-sections corresponding to the original cross-sectional profile data it collects include the contact line cross-section and the clamping element cross-section, such as... Figure 2As shown. The original cross-sectional profile data is input in the form of a two-dimensional coordinate point array, i.e. .

[0038] An exemplary processing step of the profile measurement algorithm is as follows: The input raw cross-sectional profile data is cleaned, and points marked as invalid values ​​are removed, such as those recorded by the laser profile sensor of the profile measurement unit in its coordinate records. Points with coordinates equal to specific values ​​such as 32768 are designated as sensor blind zone markers and are therefore discarded. The profile coordinates are then normalized and scaled. Divide the coordinates by 1000. Divide the coordinates by 2000 to standardize the units and reduce scale differences in subsequent numerical calculations; finally, according to... Sort the coordinates from smallest to largest, and arrange the outline coordinates from left to right, thus completing the preprocessing.

[0039] Next, using the centerline of the original cross-sectional profile data as the dividing axis, the width of the data coverage is determined for one side of the dividing axis. Set the width of the sliding window The normalized data is divided into There are several windows; for each window, find the point with the smallest ordinate to form a sequence of minimum points; search for a transition pattern of first decreasing and then increasing in the sequence of minimum points, and record the corresponding transition point as the feature valley position of the single-sided profile. This feature valley position is located on the cross-section of the clamping part, as shown in [reference]. Figure 2 The illustration.

[0040] Continue searching for the point with the largest ordinate within the region from the feature valley bottom towards the contact line cross-sectional profile. This point will be used as the feature peak location for that side profile. (See also...) Figure 2 The characteristic peak can be understood as the intersection of the contact line section and the clamping part section.

[0041] Flip the data on the other side of the profile along the dividing axis and repeat the above process to obtain the corresponding feature valley and feature peak positions on the other side, thus completing the feature identification of the left and right sides of the profile and serving as the basis for determining the wear range. The contact line section is located between the feature peaks on both sides, and the wear range is located within the contact line section, so it can serve as the basis for further processing of the original data collected by the profile measurement unit.

[0042] Begin by determining the wear range, using data between the feature peak positions on the left and right sides of the profile, with a set wear window width. The window is divided again, without considering the center boundary. Within each window, the point with the largest ordinate is selected to form a sequence of maximum points, which will cover the entire contact line cross section. Then, the first-order slope between adjacent points in the sequence of maximum points and the second-order difference of the first-order slope are calculated.

[0043] In this embodiment, a continuous interval is found where the absolute value of the second-order difference is less than a preset difference threshold, and the absolute value of the first-order slope is less than a preset slope threshold. This continuous interval is defined as a flat region in the contact line cross-sectional profile corresponding to the wear surface. The start and end points of the flat region are respectively taken as the left and right boundary points of the wear range. (See [reference needed]). Figure 2 The annotations in the text are shown.

[0044] Once the left and right boundary points of the wear range are determined, circular arc fitting can begin. However, when processing the data for the left and right boundary points, to facilitate the recording of other auxiliary parameters, this implementation method uses the contact line profile data (excluding the data for the entire wear range) as a basis and employs the least squares method to fit a large circle. This large circle is located within... Figure 2 The coordinates of its center and radius are not shown in the diagram. All values ​​are recorded as data, which characterizes the condition of the entire contact line. This can be understood as the fitted large circle essentially encompassing the profile of the entire contact line under its clamping element. When subsequently fitting the smaller circles around the wear range of the contact line, the least squares method is also used to fit the relevant data point set.

[0045] Furthermore, after fitting the large circle that represents the entire contact line, the distance deviation of each data point in the contact line profile data to the large circle can be calculated. Data points with distance deviations exceeding the threshold are marked as anomalies and removed. Then, the large circle is refitted to make the large circle a more realistic and accurate representation of the entire contact line.

[0046] Next, based on the wear range, a circular arc fitting is performed on the edge of the wear surface. From the left and right boundary points of the wear range, the boundary point with the smaller ordinate is selected as the basis for the circular arc fitting. Normally, the wear surface of the contact line is close to horizontal, so the ordinates of the left and right boundary points are almost the same. However, during the actual wear process of the contact line, there may be various abnormal contact reasons, causing the wear surface to be inclined. At this time, the ordinate of a certain boundary point will be lower, and the bending angle may even tend to be acute. The fitted circle will be smaller, and a larger degree of grinding is needed to use this circle as the basis for the fillet parameters of this measurement point.

[0047] Therefore, using the selected boundary point as the center point, a predetermined number of data points are selected on both sides of the center point. The least squares method is used to fit these data points into a circle. The radius of the fitted circle is the fillet parameter for the corresponding measurement point. (See [reference needed]). Figure 2 This is an illustration. As a preferred embodiment, after selecting a preset number of data points, an outlier filtering algorithm based on inter-point distance statistics is used to remove outliers from the selected data points, obtaining a precise point set for the transition arc. The precise point set for the transition arc is then fitted using the least squares method, and the radius value of the fitted circle is the fillet parameter. value.

[0048] The final output of the profile measurement algorithm includes: the fillet parameters corresponding to each measurement point. The values, as well as auxiliary wear parameters at each measurement point. The auxiliary wear parameters are mainly obtained by conventional processing of relevant parameters characterizing the large fitted circle of the entire contact line, and by combining data such as the left and right positions and heights determined by the position coordinates of characteristic peaks.

[0049] II. Calibration Algorithm The main function of the calibration algorithm is to learn the pressure and fillet parameters point by point using two measurement data: before grinding and after grinding with constant pressure. Linear response coefficient between values ​​of change This allows for the establishment of a comprehensive personalized polishing response model, based on the current... Value status, pre-calculate the pressure distribution for the next two polishing passes.

[0050] The input to the calibration algorithm includes: all measurements taken along the longitudinal direction of the contact line during the initial measurement. List of cross-sectional profile data for each point; during the second measurement after constant pressure grinding, also along the longitudinal direction of the contact line. A list of cross-sectional profile data for each point. The configuration parameters for the calibration algorithm include the calibration constant pressure. Maximum grinding pressure Preset target threshold Allowable error range and pressure smoothing group size Among them, the pressure smoothing group size These are specific parameters for preferred operation and will be explained in detail later when introducing the pressure smoothing algorithm; these configuration parameters will also be partially used in subsequent application algorithms. In this embodiment, a preset target threshold is used. The tolerance is 4mm, and the allowable error range is... 0.02mm; fillet radius measured at a typical wear point. The value is 3.6 mm or smaller.

[0051] The exemplary processing steps of the calibration algorithm are as follows: The first and second measurements respectively The cross-sectional profile data at each point is fed into the profile measurement algorithm one by one to obtain the fillet parameters before grinding. value list and the rounded corner parameters after polishing value list .

[0052] Perform point-by-point Parameter calculation, for each point Utilizing the before and after polishing Value change and calibrated constant pressure Calculate the response coefficient at this point. As shown in the following formula:

[0053] Response coefficient The physical significance lies in the fact that, at this point, under a calibrated constant unit grinding pressure, The increase in value can be expressed in mm / pressure unit; The specific value will be related to factors such as the material of the contact line, the geometry of the point, and the contact state of the grinding tool. It will be automatically learned through actual measurement data and will not require manual preset.

[0054] This allows all longitudinal sections of the contact wire to be... Each point The values ​​form a configuration list, which serves as the grinding response model for the entire contact line.

[0055] Next, this implementation method can be based on the polished material already prepared at the current calibration time. First, a check is performed to determine if each point meets the standard. For each point... The judgment method is as follows: like If the target is reached, the grinding pressure will be set to 0 when the grinding device reaches the target. like and If the wear level is too high, it means that the wear at that point is excessive and needs to be polished.

[0056] If the above judgments are made, all If all points meet the standards, no further polishing is required, and the algorithm returns and ends directly. If there are areas where wear exceeds the standard, the first grinding operation needs to be carried out.

[0057] For each point exceeding the standard The calibration algorithm calculates the required grinding pressure as follows:

[0058] In the formula, This represents the calculated locations that exceed the standard. The grinding pressure required during the first grinding operation; This represents taking the minimum value; the core logic of this formula is: the point that exceeds the standard. Needs improvement Value, before and after polishing Subtract the values ​​and divide by the number of points exceeding the standard. response coefficient This means obtaining the required polishing pressure; simultaneously, it is subjected to the maximum polishing pressure. Constraints are used to prevent over-wearing.

[0059] This calculation will yield all locations that exceed the standard. The corresponding grinding pressure is listed in a smooth group by pressure size. We then perform grouping and smoothing processes, which will be detailed later, to obtain the smoothed first grinding pressure distribution.

[0060] The effect of the first polishing operation was simulated, and the results were analyzed at each point. Simulate the corresponding distribution of the first grinding pressure. The value is as follows:

[0061] In the formula, To simulate polishing Value. If the point Since no polishing is required, the corresponding first polishing pressure distribution does not include polishing pressure at that point, and is therefore set to 0. Based on This allows us to predict the polishing effect without actually performing the polishing, and then determine the need for a second polishing and calculate the pressure.

[0062] In simulation Based on this, if all If all points meet the standard, there is no need for a second polishing process, and the algorithm will set the mark for the second polishing to 0. If there are still points where wear exceeds the standard, the same method is used to calculate the required grinding pressure. Each point is also grouped and smoothed to obtain the second grinding pressure distribution.

[0063] This implementation assumes that only two actual grinding operations are performed. Therefore, the final output of the calibration algorithm includes: a marker indicating whether grinding is needed for the first grinding operation, a list of pressure points for each point during the first grinding operation (an empty list if grinding is not needed), a marker indicating whether grinding is needed for the second grinding operation, and a list of pressure points for each point during the second grinding operation (an empty list if grinding is not needed). The calibration algorithm also provides the linear response coefficients for the entire contact wire. This is used in all subsequent application stages. During the calibration phase, the grinding device can be driven by the calibration algorithm and perform grinding directly based on its initial grinding scheme.

[0064] III. Application Algorithm The main function of the application algorithm is to receive the currently measured cross-sectional profile data during normal grinding operation after calibration, and to utilize the linear response coefficients of the entire contact line obtained during calibration. The algorithm, along with relevant configuration parameters, automatically calculates the grinding pressure distribution at each point and predicts subsequent grinding requirements through simulation. In the example steps of the calibration algorithm, the pressure calculation and simulation logic it calls is essentially the main function of the application algorithm, but it is used to output the initial grinding plan during the calibration stage. The most crucial aspect of the calibration stage is outputting the linear response coefficient of the entire contact line. This is so that it can be used during multiple polishing processes in the application phase.

[0065] The input to the applied algorithm includes: all longitudinal lines along the contact line during the actual grinding operation. List of cross-sectional profile data for each point; linear response coefficients of the entire contact wire obtained during the calibration phase. The configuration parameters for the applied algorithm include maximum grinding pressure. Preset target threshold Allowable error range and pressure smoothing group size .

[0066] The exemplary processing steps of the applied algorithm are as follows: Will The cross-sectional profile data of each point is fed into the profile measurement algorithm one by one to obtain the current cross-sectional profile data. value list For each point Execute the following judgment logic: Scenario 1: If the target point has been met, then the corresponding grinding pressure will be applied when the grinding device reaches that point. Set to 0; Scenario 2: and If the wear at that point exceeds the standard, it indicates that the point needs to be polished. The polishing pressure should be calculated using the following formula. That's all:

[0067] Scenario 3: This situation represents The value is abnormally high, which should not occur in theory. It may be caused by measurement noise, and this situation will also affect the corresponding grinding pressure. Set it to 0.

[0068] In addition, there is a special case where the linear response coefficient of the entire contact wire returned during the calibration phase is... In the equation, there exists a response coefficient corresponding to a certain point. That is, after grinding at this point If the value does not increase or even decreases, then it is directly taken in the application phase. .

[0069] Based on the assessment results, determine whether the first polishing is necessary, and if the polishing pressure at all points is... If all values ​​are 0, it means all requirements are met, and no further polishing is needed; the algorithm returns directly. Otherwise, the calculated original polishing pressure list is grouped by pressure smoothing size. We then perform grouping and smoothing to obtain the smoothed first grinding pressure distribution; the logic here is the same as that described in the calibration algorithm.

[0070] For each point In the application phase, the polishing effect is also simulated, and the simulation is obtained after applying the first polishing pressure distribution. The value is as follows:

[0071] In the formula, It was obtained after simulation. value; It is the smoothed grinding pressure distribution. Based on this... The judgment logic and polishing pressure calculation in the repeated application stage are repeated; if all points meet the standard after polishing, no further polishing is required; if there are still points that exceed the standard, the second and third calculations and polishing processes are continued.

[0072] The final output of the applied algorithm includes: a marker indicating whether grinding is needed for the first grinding cycle, a list of pressure points for the first grinding cycle (an empty list if grinding is not needed), a marker indicating whether grinding is needed for the second grinding cycle, a list of pressure points for the second grinding cycle (an empty list if grinding is not needed), and whether a third or fourth grinding cycle is required. Driven by the applied algorithm, the grinding device can then begin performing the actual grinding operation.

[0073] As a preferred embodiment of this method, the pressure smoothing algorithm smooths the grinding pressure calculated independently at each point between adjacent points, avoiding drastic changes in grinding pressure caused by single-point measurement noise or local profile abnormalities, so that the grinding force transitions smoothly along the longitudinal direction of the contact line, ensuring uniform surface quality after grinding.

[0074] Assuming that the calculated grinding pressure distribution during the calibration or application phase is denoted as follows: It also has a pressure smoothing group size configured. First, divide the pressure list into indexed sections. Group , for the th group Group, calculate the average pressure within that group. As shown in the following formula:

[0075] In the formula, Representing the The number of grinding pressures in the group; set the grinding pressure of all points in the group to [value]. This means that the pressure within a group is uniform, while the pressure varies stepwise between groups.

[0076] After smoothing, adjacent Each measurement point (corresponding to the longitudinal direction of the contact line, i.e.) (Multiplied by the length of the sampling interval) using the same polishing pressure, polishing pressure and The value changes gradually in a stepped manner along the longitudinal direction of the contact line, avoiding grinding marks caused by high-frequency pressure fluctuations.

[0077] In summary, in the method of this embodiment, the profile measurement algorithm is the foundational layer, providing a unified framework for the calibration algorithm and the application algorithm. Value extraction capability; whether it's the profile before and after polishing in the calibration stage, or the current profile in the application stage, it is all converted into the same measurement algorithm. Values ​​to ensure consistent data standards.

[0078] The calibration algorithm is executed only once on the first run, relying on the profile measurement algorithm to provide before and after polishing. The value is obtained by learning and outputting the linear response coefficient of the entire contact wire. The calibration algorithm also calls the same pressure calculation and simulation logic as the subsequent application algorithm, so that the first polishing solution can be output during the calibration stage.

[0079] The application algorithm is executed on every run after calibration, directly using the linear response coefficients output by the calibration algorithm. The relevant pressure calculations can be performed without recalibration. The pressure calculation formula of the applied algorithm is essentially the same as the pressure calculation formula inside the calibration algorithm; the only difference lies in the formula used. The values ​​come from different sources.

[0080] Finally, this embodiment describes the overall workflow of the method, which can be summarized as a calibration stage and an application stage. The calibration stage only requires the initial measurement run from the contact wire start point A to the end point B and the calibration run from the end point B to the start point A; the application stage can continuously perform reciprocating runs between the contact wire start point A and the end point B.

[0081] During the initial measurement run from start point A to end point B in the calibration phase, the profile measurement unit of the grinding device collects data along the entire line at the front end. The cross-sectional profile data of each measurement point is used as the initial state before grinding; at this time, the grinding execution unit of the device is not working. The data processing and control unit calls the profile measurement algorithm to extract the profile data point by point. The value is obtained to determine the entire contact line before polishing. The value distribution establishes a baseline wear condition for the entire line.

[0082] The device then returns from endpoint B to starting point A, reversing its direction of travel. At this point, the grinding execution unit is in front, and the profile measurement unit is behind. The grinding execution unit operates at a preset constant pressure. The contact line is uniformly polished at all points. Immediately after polishing, the profile measurement unit measures the profile data at each point and uses the profile measurement algorithm to extract the data. The value is obtained by polishing the entire contact line. Value distribution.

[0083] The data processing and control unit runs a calibration algorithm for each point. The response coefficient is calculated to form a corresponding configuration list. Then, it is determined whether there are any points exceeding the standard. If so, the pressure distribution of the first grinding is calculated and smoothed by grouping. Then, the process after the first grinding is simulated. The value is checked again to see if it exceeds the standard. If it does, the pressure distribution for the second grinding is calculated. Finally, the calibration algorithm outputs the linear response coefficient obtained in the calibration stage. List, the requirement for two polishing stages, and the corresponding polishing pressure distribution.

[0084] Once in the application phase, the grinding device reciprocates along the contact line. During each run, the data processing and control unit executes the application algorithm, eliminating the need for repeated calibration. The following operations are automatically performed during each run: Call the profile measurement algorithm to convert the current measured profile into List of values; Perform a compliance assessment for each location; For locations exceeding the standard, the response coefficient is used. Calculate the grinding pressure and be subject to... constraint; The calculated grinding pressure distribution is grouped and smoothed. Simulate the effect of the first round of sanding to determine if a second round of sanding is still needed. If so, continue to calculate the pressure distribution for the second round of sanding in the same way.

[0085] Based on the calculated pressure distribution, the grinding execution unit applies a specified grinding pressure to each point on the contact line during the device's movement. After one round of grinding, the profile is measured and extracted again in the next round of operation. If there are still points that exceed the standard, the algorithm is used to recalculate the pressure based on the latest measurement data and the grinding continues; if all points meet the standard, the grinding operation is completed.

[0086] Practical verification has shown that the method of this embodiment can effectively clean the contact wire after grinding. The value deviation is controlled within ±0.02mm. Compared to the traditional fixed-parameter grinding method that typically requires 5 to 8 passes, the method in this embodiment can optimize the number of grinding passes to 2 to 4 passes depending on the initial wear level. The grinding pressure is along the longitudinal direction of the contact line. The points with the number of values ​​are grouped together for a smooth transition, and the pressure changes between adjacent groups are gradual. After grinding, the profile surface has no obvious fluctuations or abrupt changes, thus achieving a highly efficient and precise adaptive grinding operation.

Claims

1. A contact line adaptive grinding method based on profile measurement, characterized in that, The method includes the following steps: S1. Collect cross-sectional profile data of each measurement point of the contact line, preprocess and identify the cross-sectional profile data to determine the wear range; based on the wear range, perform arc fitting on the edge of the wear surface to extract the rounded corner parameters characterizing the wear state. S2. Control the grinding device to perform test grinding on the contact line with a set constant pressure, and obtain the change in the fillet parameter before and after the test grinding; based on the constant pressure and the change in the fillet parameter, calculate the response coefficient between the constant pressure and the change in the fillet parameter at each measurement point, and complete the calibration; S3. During the normal polishing process after calibration, obtain the currently measured fillet parameters and calculate the deviation between the currently measured fillet parameters and the preset target threshold; use the response coefficient and the deviation to calculate the optimal polishing pressure for each measurement point in the next polishing pass. S4. Based on the optimal grinding pressure, control the grinding device to perform grinding operations on the contact line, and after grinding is completed, re-collect the cross-sectional profile data and update the fillet parameters; based on the updated fillet parameters, determine whether the preset grinding end condition is met. If not, use the updated fillet parameters as the currently measured fillet parameters, and repeat steps S3 to S4 until the grinding end condition is met, thus completing the adaptive grinding of the contact line.

2. The adaptive grinding method for contact lines according to claim 1, characterized in that, In step S1, the collected original cross-sectional profile data is cleaned, invalid marker points are removed, and the cleaned profile coordinates are normalized, scaled, and sorted; wherein, the cross-sections corresponding to the original cross-sectional profile data include the contact line cross-section and the clamping component cross-section; Using the central axis of the original cross-sectional profile data as the dividing axis, the sliding window width is set for one side of the dividing axis, and the normalized data is divided into multiple windows. The point with the smallest ordinate is found in each window to form a sequence of minimum points. Search for a transition pattern of first decreasing and then increasing in the minimum point sequence, and record the corresponding transition point as the characteristic valley position of the single-sided profile. The characteristic valley position is located on the cross section of the clamping member. Find the point with the largest ordinate in the region from the characteristic valley position toward the contact line cross section profile, and use it as the characteristic peak position of the single-sided profile. Flip the data on the other side of the profile along the dividing axis and repeat the above process to obtain the corresponding feature valley position and feature peak position on the other side, thus completing the feature identification of the left and right sides of the profile, which serves as the basis for determining the wear range.

3. The adaptive grinding method for contact lines according to claim 2, characterized in that, In step S1, determining the wear range includes: dividing the data between the characteristic peak positions on the left and right sides of the profile into windows again with a set wear window width, and taking the point with the largest ordinate in each window to form a sequence of maximum points; Calculate the first-order slope between adjacent points in the maximum point sequence and the second-order difference of the first-order slope; Find a continuous interval where the absolute value of the second-order difference is less than a preset difference threshold and the absolute value of the first-order slope is less than a preset slope threshold, and define the continuous interval as a flat area in the contact line cross-sectional profile corresponding to the wear surface. The starting point and ending point of the flat area are respectively taken as the left and right boundary points of the wear range.

4. The adaptive grinding method for contact lines according to claim 3, characterized in that, In step S1, based on the wear range, the edge of the wear surface is fitted with an arc, including: selecting the boundary point with the smaller ordinate from the left and right boundary points of the wear range as the basis for arc fitting; taking the selected boundary point as the center point, selecting a preset number of data points on both sides of the center point, and using the least squares method to fit these data points into a circle, and the radius value of the fitted circle is the fillet parameter of the corresponding measurement point.

5. The adaptive grinding method for contact lines according to claim 4, characterized in that: After selecting a preset number of data points, an outlier filtering algorithm based on the statistical distance between points is used to remove outliers from the selected data points, resulting in a precise set of transition arc points. The precise set of transition arc points is then fitted using the least squares method, and the radius value of the fitted circle is the fillet parameter.

6. The adaptive grinding method for contact lines according to claim 1, characterized in that, In step S2, the response coefficient between the constant pressure and the change in the fillet parameter at each measurement point is calculated, including: obtaining the initial fillet parameter of the measurement point before the test grinding, and the calibrated fillet parameter of the measurement point after the test grinding; calculating the difference between the initial fillet parameter and the calibrated fillet parameter to obtain the change in the fillet parameter; dividing the change in the fillet parameter by the constant pressure to calculate the response coefficient.

7. The adaptive grinding method for contact lines according to claim 6, characterized in that, In step S3, the optimal polishing pressure for each measurement point in the next polishing pass is calculated using the response coefficient and the deviation. This includes: subtracting the currently measured fillet parameter from the preset target threshold to obtain the deviation; dividing the deviation by the response coefficient to obtain the theoretical polishing pressure for each measurement point in the next polishing pass; smoothing and limiting the theoretical polishing pressure to obtain the optimal polishing pressure; and simulating the change in fillet parameter after a single polishing pass based on the optimal polishing pressure and the response coefficient, and planning the number of subsequent polishing passes required to reach the preset target threshold.

8. The adaptive grinding method for contact lines according to claim 7, characterized in that, The optimal grinding pressure is obtained by smoothing and limiting the theoretical grinding pressure, including: performing a moving average filtering process on the theoretical grinding pressure of adjacent measurement points along the longitudinal direction of the contact line; and comparing the processed grinding pressure with preset minimum grinding pressure threshold and maximum grinding pressure threshold. If the processed grinding pressure is less than the minimum grinding pressure threshold, then the optimal grinding pressure is set to the minimum grinding pressure threshold; if the processed grinding pressure is greater than the maximum grinding pressure threshold, then the optimal grinding pressure is set to the maximum grinding pressure threshold.

9. The adaptive grinding method for contact lines according to claim 1, characterized in that, In step S4, based on the updated fillet parameters, it is determined whether the preset grinding end condition is met, including: determining whether the updated fillet parameters fall within a preset tolerance range centered on the preset target threshold. If the updated fillet parameters fall within the preset tolerance range, the polishing end condition is determined to be met; if the updated fillet parameters do not fall within the preset tolerance range, it is further determined whether the current cumulative polishing passes have reached the preset maximum polishing pass threshold. If the maximum polishing pass threshold is reached, the polishing end condition is determined to be met.

10. The adaptive grinding method for contact lines according to claim 1, characterized in that, In step S1, the cross-sectional profile data of each measurement point of the contact line is collected. Specifically, the profile measurement unit set in front of the grinding device is used to collect a two-dimensional coordinate point array of each measurement point in the longitudinal direction of the contact line, including the contact line and the clamping part, in real time as the grinding device moves along the longitudinal direction of the contact line, as the cross-sectional profile data. While extracting the fillet parameters that characterize the wear state, the wear surface width, wear angle, remaining height, and left and right wear surface heights are also extracted as auxiliary wear parameters based on the cross-sectional profile data. The auxiliary wear parameters are associated with the corresponding measurement points and stored for real-time monitoring of the grinding state and historical data tracing.

Images