A catenary virtual model optimization method based on scene compensation coefficient

Abstract

The present application relates to the technical field of contact network simulation, and discloses a contact network virtual model optimization method based on a scene compensation coefficient, comprising: establishing a contact network initial virtual model and binding a tension response table and a conductive response table for a virtual node; obtaining a measured height sequence and a virtual height sequence of four equidistant positions of a contact line, respectively making a three-order difference, normalizing a neighboring height difference weighted result, obtaining a four-point arc line deviation ratio and a virtual arc line deviation ratio, and determining an arc line coherence degree; querying a theoretical tension and a theoretical conductive rate according to a node temperature and a node humidity, combining a measured tension and a measured conductive rate to determine an original compensation coefficient, and contracting the original compensation coefficient according to the arc line coherence degree to obtain a scene compensation coefficient; updating the tension response table and the conductive response table by using the scene compensation coefficient, writing back a corrected tension table and a corrected conductive table to the virtual node, and obtaining an optimized contact network virtual model, which is used for contact network model correction and state early warning.

Description

Technical Field

[0001] This invention relates to the field of overhead contact line simulation technology, and more specifically, to an optimization method for a virtual model of an overhead contact line based on scene compensation coefficients. Background Technology

[0002] In the operation and maintenance of the overhead contact system of electrified railways, contact wire tension, conductivity, suspension point location, and contact wire height collectively affect the power supply status. Inspection vehicles, laser scanning equipment, temperature and humidity acquisition devices, tension acquisition devices, and conductivity acquisition devices can generate timestamped monitoring records according to line mileage. To facilitate status verification and maintenance plan development, the engineering system typically incorporates the line topology, contact wire material, contact wire cross-sectional area, and monitoring records into a virtual model of the overhead contact system, and configures tension response tables and conductivity response tables in the virtual nodes.

[0003] Existing virtual catenary models typically look up tension and conductivity response tables based on node temperature and humidity to obtain theoretical tension and conductivity, which are then compared with measured tension and conductivity. Some systems revise the response tables based on the comparison results, gradually bringing the output of virtual nodes closer to field monitoring results under corresponding temperature and humidity conditions. This type of processing is suitable for overall state shifts dominated by environmental changes and also facilitates the reuse of existing experimental data, design data, and accepted operational data.

[0004] When local dropper deviation, local bending, point cloud outliers, or time registration errors occur in a single-span contact wire, the differences between the measured tension and measured conductivity and the model output may not entirely stem from nodal temperature and humidity. If the tension response table and conductivity response table are still updated directly based on state differences, local anomalies may be written into the temperature and humidity response relationship and continue to affect the theoretical tension, theoretical conductivity, warning temperature, and warning humidity under subsequent similar temperature and humidity conditions.

[0005] The contact wire span itself also has differences in height between the two ends, design presag, and normal sag variation. It is difficult to reliably distinguish between overall uplift, track slope, normal sag variation and local uneven variation by comparing the height at a certain position, mid-span sag or average height alone. Summary of the Invention

[0006] This invention provides a method for optimizing a virtual catenary model based on scene compensation coefficients, which solves the technical problems mentioned in the background art.

[0007] To achieve the above-mentioned objectives, the present invention provides the following technical solution.

[0008] A method for optimizing a virtual catenary model based on scene compensation coefficients includes establishing an initial virtual catenary model and binding virtual nodes to tension response tables and conductivity response tables.

[0009] Obtain the measured height sequence and virtual height sequence at four equally spaced positions on the contact wire;

[0010] The two height sequences are subjected to third-order difference processing, and the measured total fluctuation and the virtual total fluctuation are determined by weighting the adjacent height differences at both ends equally and the adjacent height differences in the middle with double weights.

[0011] When the measured total fluctuation is non-zero, the result is processed by normalizing the measured difference; otherwise, it is zero, and the deviation ratio of the four-point arc is obtained.

[0012] When the virtual total fluctuation is non-zero, the result is processed by its normalized virtual difference; otherwise, it is zero, and the virtual arc deviation ratio is obtained.

[0013] The degree of coincidence of two arcs is defined as being in the range of zero to one and decreasing as the difference in the deviation ratio between the two arcs increases; when the difference is zero, the degree of coincidence of the arcs is one.

[0014] The theoretical tension and theoretical conductivity are obtained by querying two response tables using node temperature and node humidity. The tension error ratio is obtained by normalizing the absolute difference between the measured tension and the theoretical tension using the tension range. The conductivity error ratio is obtained by normalizing the absolute difference between the measured conductivity and the theoretical conductivity using the conductivity range. The larger one is limited to the range of zero to one to obtain the original compensation coefficient.

[0015] The original compensation coefficient is reduced by the arc matching degree to obtain the scene compensation coefficient that decreases as the difference increases. When the difference is zero, the scene compensation coefficient is equal to the original compensation coefficient.

[0016] The difference between measured tension and theoretical tension and the difference between measured conductivity and theoretical conductivity are scaled according to the scenario compensation coefficient. The tension response table and conductivity response table are updated to form a corrected tension table and a corrected conductivity table, which are then written back to the virtual node to obtain the optimized catenary virtual model.

[0017] Compared with the prior art, the present invention has the following substantial features and significant progress:

[0018] This invention obtains measured height sequences and virtual height sequences at four equally spaced locations within the contact wire span between adjacent suspension points. Through third-order difference and weighted normalization of adjacent height differences, it reduces the impact of overall uplift, linear slope, and approximate quadratic sag on local deviation judgment. The arc alignment is determined based on the four-point arc deviation ratio and the virtual arc deviation ratio. The original compensation coefficient is then reduced based on the arc alignment, ensuring that the original response table update amplitude is retained when the measured arc aligns with the virtual arc. When local suspension wire deviation, local bending, or point cloud outliers cause arc misalignment, the write amplitude of tension residuals and conductivity residuals is reduced, thereby reducing the possibility of local anomalies entering the tension and conductivity response tables. After correcting the tension and conductivity tables and writing them back to the virtual nodes, they can be used to update theoretical tension, theoretical conductivity, warning temperature, and warning humidity, improving the consistency between the updated parameters of the contact wire virtual model and the on-site monitoring status. Attached Figure Description

[0019] Figure 1 This is a diagram illustrating the optimization mechanism of the overhead contact line virtual model of the present invention. Detailed Implementation

[0020] The following description is provided in conjunction with the accompanying drawings. It should be understood that the following embodiments are for illustrative purposes only and are not intended to limit the invention; those skilled in the art can make equivalent substitutions or combinations for the specific implementations.

[0021] This implementation method addresses scenarios involving digital modeling, computer-aided modeling, and simulation parameter correction of overhead contact lines. It handles discrepancies between the tension and conductivity response tables in the initial virtual model of the overhead contact line and the actual field data. The initial virtual model provides theoretical tension and conductivity based on the node temperature and humidity corresponding to the virtual nodes. However, the measured tension and conductivity obtained in the field are affected not only by environmental conditions but also by local dropper deviations, local bends, or outlier points in the point cloud. If these effects are not distinguished and all virtual-to-real differences are directly written into the tension and conductivity response tables, local geometric anomalies will be incorrectly encoded as temperature and humidity response relationships, causing subsequent model outputs to continuously deviate from the actual state.

[0022] In this embodiment, the line mileage, suspension point location, contact wire material, contact wire cross-sectional area, minimum tension, maximum tension, minimum conductivity, and maximum conductivity used are derived from existing design data, acceptance data, or the initial virtual model configuration of the contact network; the node temperature, node humidity, measured tension, and measured conductivity are derived from existing monitoring data; and the contact wire heights at the four equidistant locations are derived from point cloud data collected by the laser scanning equipment mounted on the inspection vehicle.

[0023] like Figure 1 As shown, this implementation can be achieved collaboratively by a measured height acquisition module 100, a node status acquisition module 200, an initial virtual model of the overhead contact system 300, an arc deviation ratio determination module 400, an arc compatibility determination module 500, a compensation coefficient determination module 600, a response table update module 700, and an optimized virtual model of the overhead contact system 800. The measured height acquisition module 100 outputs a measured height sequence; the initial virtual model of the overhead contact system 300 outputs a virtual height sequence, a tension response table, a conductivity response table, a tension range, and a conductivity range; the arc deviation ratio determination module 400 obtains the four-point arc deviation ratio and the virtual arc deviation ratio based on the measured height sequence and the virtual height sequence; the arc compatibility determination module 500 obtains the arc compatibility based on the difference between the two arc deviation ratios.

[0024] The node status acquisition module 200 outputs node temperature, node humidity, measured tension, and measured conductivity. The compensation coefficient determination module 600 queries the tension response table and conductivity response table using the node temperature and humidity to obtain the theoretical tension and theoretical conductivity. It then combines this with the measured tension, measured conductivity, tension range, and conductivity range to determine the original compensation coefficient. Finally, it shrinks the original compensation coefficient according to the arc compatibility to obtain the scene compensation coefficient. The response table update module 700 scales the tension residual and conductivity residual according to the scene compensation coefficient, and writes the scaling results into the tension response table and conductivity response table to form the corrected tension table and corrected conductivity table. After the corrected tension table and corrected conductivity table are written back to the virtual node, the optimized contact network virtual model 800 is formed.

[0025] Specifically, the tension error ratio and conductivity error ratio only indicate that the initial virtual model of the overhead contact system needs updating under the current node's temperature and humidity coordinates. The arc alignment degree is used to limit the extent to which this update requirement can be written into the tension response table and conductivity response table. When the span shape reflected by the measured height sequence matches the virtual height sequence, the scene compensation coefficient maintains the update range determined by the original compensation coefficient. When the local shapes of the measured height sequence and the virtual height sequence do not match, the scene compensation coefficient decreases as the difference between the four-point arc deviation ratio and the virtual arc deviation ratio increases. Therefore, local geometric anomalies will not be indiscriminately interpreted as changes in the temperature and humidity response relationship.

[0026] Furthermore, the initial virtual model of the overhead contact line is established as follows:

[0027] The initial virtual model 300 of the overhead contact system is a computer model that already contains line topology, span geometry, virtual nodes, and state response data before executing this method. This implementation reads the suspension point positions and line topology, and divides the contact line between two adjacent suspension points along the line direction into a contact line span unit. The contact line span unit is a model division of an existing contact line. For continuous lines, the contact line span units are divided sequentially according to the line topology, so that each suspension point can serve as an endpoint of an adjacent contact line span unit.

[0028] Each contact wire span unit is assigned a span number, and the corresponding virtual node is assigned a virtual node number. The span number identifies the contact wire span unit and its height data, while the virtual node number identifies the location where response tables and status data are written. Based on the span number, the contact wire span units are associated with the model span objects in the initial virtual model of the catenary. This association allows the measured height sequence under the same span number to be compared with the virtual height sequence output by the model span object, and ensures that the calculation results are returned to the correct virtual node.

[0029] Line mileage is used to help determine the location of the field data acquisition point in the initial virtual model of the overhead contact system. During implementation, the corresponding contact wire span unit can be determined first by the span number, and then the longitudinal position of the data within the model span object can be determined by the line mileage. Using both span number and line mileage together can avoid erroneous associations when there are similar spans or duplicate local coordinates in the line topology.

[0030] The tension response table and conductivity response table are selected based on the contact wire material and cross-sectional area. The tension response table is an existing mapping between node temperature, node humidity, and tension output, while the conductivity response table is an existing mapping between node temperature, node humidity, and conductivity output. Both response tables can be derived from existing test data, design manuals, or accepted operational data, and have predefined temperature and humidity grid points and defined domains.

[0031] The tension response table and conductivity response table are bound to the virtual nodes according to their virtual node numbers. Each bound virtual node uses two response tables that match its contact wire material and cross-sectional area. The response table of one contact wire span unit is not directly used for another contact wire span unit with a different contact wire material or cross-sectional area. If multiple virtual nodes have the same contact wire material and cross-sectional area, they can reference the same set of initial response data, but independent write-back results should be maintained for each corresponding virtual node during correction to prevent the update of one virtual node from overwriting the state of another.

[0032] Obtain the minimum tension, maximum tension, minimum conductivity, and maximum conductivity. The minimum and maximum tensions are derived from the two boundaries of the rated tension range, and the minimum and maximum conductivity are derived from the two boundaries of the standard conductivity range. The difference between the maximum and minimum tensions is defined as the tension range, and the difference between the maximum and minimum conductivity is defined as the conductivity range, respectively written as:

[0033]

[0034] In the formula, Indicates the span number is The tension range of the virtual node corresponding to the contact line span unit; Indicates the maximum tension; This indicates the minimum tension. The tension range is used to normalize the absolute difference between the measured tension and the theoretical tension, so that the tension error ratio under different rated tension configurations has a comparable scale of zero to one.

[0035]

[0036] In the formula, Indicates the range of conductivity; Indicates the highest conductivity; This represents the minimum conductivity. The conductivity range is used to normalize the absolute difference between the measured conductivity and the theoretical conductivity. Before performing the calculation, it is confirmed that the maximum tension is greater than the minimum tension and the maximum conductivity is greater than the minimum conductivity; if either condition is not met, the corresponding range cannot be used as a divisor, the virtual node is marked as having incomplete configuration data, the original response table is retained, and this round of updates is not performed.

[0037] The initial virtual model 300 of the overhead contact line also stores the virtual contact line heights of the model span objects at the left, one-third, two-thirds, and right ends. These four positions are determined based on the existing coordinates of the two suspension points in the model and are evenly distributed along the line direction of the contact line span unit. These four positions serve as software sampling coordinates. The model span objects output virtual left-end height, virtual left-middle height, virtual right-middle height, and virtual right-end height at the same coordinates, forming a virtual height sequence in positional order.

[0038] After completing the above processing, the initial virtual model 300 of the overhead contact line can at least provide the virtual height sequence, tension response table, conductivity response table, tension range and conductivity range for subsequent processing, and can locate the model span objects and virtual nodes according to the span number and virtual node number.

[0039] Furthermore, the implementation methods for acquiring, correlating, and preprocessing on-site data are as follows:

[0040] The node status acquisition module 200 associates the temperature and humidity acquisition locations, tension acquisition locations, and conductivity acquisition locations with the corresponding virtual nodes according to the line mileage. The temperature and humidity acquisition locations are used to obtain the node temperature and humidity, the tension acquisition locations are used to obtain the measured tension, and the conductivity acquisition locations are used to obtain the measured conductivity. When the acquisition locations from different data sources cannot completely coincide with the coordinates of the virtual nodes, they can be associated with the corresponding virtual nodes based on the existing line mileage index.

[0041] Node temperature represents the temperature reading of the corresponding virtual node at the current acquisition time; node humidity represents the relative humidity reading of the corresponding virtual node at the current acquisition time; measured tension represents the current mechanical state of the contact wire; and measured conductivity represents the current electrical state of the contact wire. Node temperature and node humidity are used to look up the tension response table and conductivity response table, while measured tension and measured conductivity are used to determine the tension error ratio and conductivity error ratio.

[0042] Before determining the original compensation coefficients, node temperature, node humidity, measured tension, and measured conductivity are aligned according to timestamps. The timestamp of the measured height sequence used to calculate the four-point arc deviation ratio in this round is taken as the target time. For data with the same sampling frequency and consistent timestamps, they are directly grouped into state data sets for the same time. For data with different sampling frequencies, if the state channel has been configured with a synchronization tolerance, the nearest time matching only selects the record with the smallest time difference from the target time that does not exceed the synchronization tolerance; if the state channel has not been configured with a separate synchronization tolerance, half of the adjacent valid sampling interval of the state channel is used as the synchronization tolerance. When there are valid records before and after the target time and the target time is located between the two record timestamps, linear interpolation is used to form the state data set for the target time; linear interpolation only uses the valid record before and after the target time and does not cross missing segments or equipment stop sampling segments. State channels that do not meet the conditions for nearest time matching or linear interpolation are marked as having incomplete state data in this round.

[0043] The status data set undergoes range checking, duplicate value checking, and median filtering. Range checking is based on the equipment range or the minimum and maximum tension, minimum and maximum conductivity configured in the initial virtual model of the overhead contact line, as well as the monitoring system's configuration boundaries. Records exceeding these boundaries are excluded from the status data set. Duplicate value checking uses identical acquisition locations, timestamps, data channels, and numerical values ​​as criteria for duplicate records, retaining only records entering the data sequence for the first time. Median filtering utilizes an odd-numbered window length already configured for the status channel; if no window length is configured for the status channel, three valid recording windows centered on the target time are used; if there are fewer than three valid records at the start or end of the sequence, existing valid records are used to form a shortened window.

[0044] The measured height acquisition module 100 determines the left end, one-third position, two-thirds position, and right end within the contact wire span unit based on the suspension point location. It acquires the left end height, left-middle height, right-middle height, and right end height at these four positions, forming a measured height sequence strictly following the order from left to right. The left end height corresponds to the contact wire height at the left suspension point, the left-middle height corresponds to the contact wire height at the one-third position, the right-middle height corresponds to the contact wire height at the two-thirds position, and the right end height corresponds to the contact wire height at the right suspension point.

[0045] This embodiment uses a laser scanning device mounted on a testing vehicle to collect point cloud data and extracts the measured height sequence from the point cloud data. When the testing vehicle passes through the target contact line span unit along the line, the laser scanning device acquires point cloud data containing spatial points of the contact line; based on existing point cloud classification or contact line identification results, the contact line points are selected, and then the vertical coordinates are read at four equally spaced positions.

[0046] After the point cloud data has completed contact line point identification, the measured height acquisition module 100 sorts the contact line points within the same contact line span unit according to the line mileage direction, and determines the line mileage corresponding to the left end, one-third position, two-thirds position, and right end based on the suspension point position. If the detection device has already output the guide height results for the above four positions, then the guide height result is directly used as the left end height, left-middle height, right-middle height, and right end height; if the point cloud data only provides discrete contact line points, then two contact line points with adjacent line mileages are selected on both sides of each equidistant position, and the vertical coordinates of the corresponding positions are obtained by linear interpolation according to the line mileage. When the equidistant position coincides with the line mileage of the identified contact line point, the vertical coordinates of the contact line point are directly read. If adjacent contact line points cannot be obtained on both sides of the equidistant position and the detection device does not output the corresponding guide height result, the height is recorded as missing.

[0047] The measured height sequence is associated with the corresponding contact wire span unit using span number and line mileage. The span number determines the contact wire span unit to which the measured height sequence belongs, and the line mileage is used to verify the order and spatial range of the four positions along the line direction. If the line mileage direction, vertical reference, or length unit used in the point cloud data is inconsistent with the initial virtual model of the catenary, existing calibration relationships can be used to convert it to the same coordinate reference.

[0048] Use timestamps to confirm that the measured height sequence comes from the same inspection stroke or the same time-registered data frame. Before determining the four-point arc deviation ratio, reconfirm that the four heights in the measured height sequence have the same inspection stroke identifier or the same time registration relationship. Data from different inspection strokes or incomplete time registration cannot be spliced ​​into a single measured height sequence because the contact line state may be different at different times, and the spliced ​​result cannot represent the arc shape of the same span under the same state.

[0049] If any of the four heights is missing, the order of the four positions cannot be confirmed, or the measured height sequence cannot be associated with the model span object, the four-point arc deviation ratio will not be calculated in this round, nor will the two response tables be updated based on the measured height sequence. The system can retain the original data in the initial virtual model of the overhead contact line and record the missing status, waiting for a complete measured height sequence before execution.

[0050] After the on-site data preprocessing is completed, the measured height acquisition module 100 transmits the measured height sequence to the arc deviation ratio determination module 400, and the node status acquisition module 200 transmits the node temperature, node humidity, measured tension, and measured conductivity to the compensation coefficient determination module 600. The measured height sequence and the virtual height sequence correspond to the same span number and the same four positions, and the status data group corresponds to the two response tables according to the same virtual node number.

[0051] Furthermore, the methods for determining the theoretical state, state error, and original compensation coefficients are as follows:

[0052] The compensation coefficient determination module 600 determines the current node temperature and humidity coordinates based on the node temperature and humidity. These coordinates are jointly determined by the node temperature and humidity data from the same state data group and used as query conditions for the tension response table and the conductivity response table. Before querying, it is confirmed that the current node temperature and humidity coordinates are within the respective domains of the two response tables. If the coordinates exceed the domain of either response table, the extrapolation is stopped using the existing boundary processing rules of the initial virtual model of the contact network, boundary grid points are used, or the virtual node is marked as not updating in this round.

[0053] When the current node's temperature and humidity coordinates coincide with a tension temperature and humidity grid point, the response value of that tension temperature and humidity grid point is directly read as the theoretical tension. When the current node's temperature and humidity coordinates do not coincide with a tension temperature and humidity grid point but are within the domain defined by the tension response table, the theoretical tension is obtained based on the tension-adjacent temperature and humidity grid points surrounding the current node's temperature and humidity coordinates, according to the existing interpolation method of the tension response table. When the tension response table has an existing interpolation method configuration, the compensation coefficient determination module 600 reads this configuration and retains the tension-adjacent temperature and humidity grid points involved in the theoretical tension calculation and their weights. When the tension response table does not have an interpolation method configured and the current node's temperature and humidity coordinates are within a rectangular tension temperature and humidity grid cell, ordinary bilinear interpolation is used; if the current node's temperature and humidity coordinates are on the grid line between two tension temperature and humidity grid points, linear interpolation is used; if they coincide with a tension temperature and humidity grid point, the tension temperature and humidity grid point is directly read. When the current node's temperature and humidity coordinates cannot be surrounded by the grid cells of the tension response table and are not on the boundary grid line, the virtual node is marked as not updating in this round. This process can be represented as:

[0054]

[0055] In the formula, Indicates theoretical tension; This represents the tension response table for virtual node bindings; Indicates node temperature; Indicates node humidity; subscript Corresponding span number. If existing bilinear interpolation is used, only the tension-adjacent temperature and humidity grid points enclosing the current node's temperature and humidity coordinates and their existing weights are used from the tension response table.

[0056] When the current node's temperature and humidity coordinates coincide with a conductive temperature and humidity grid point, the response value of that conductive temperature and humidity grid point is directly read as the theoretical conductivity. When the current node's temperature and humidity coordinates do not coincide with a conductive temperature and humidity grid point but are within the domain defined by the conductive response table, the theoretical conductivity is obtained based on adjacent conductive temperature and humidity grid points according to the existing interpolation method of the conductive response table. When the conductive response table has an existing interpolation method configuration, the compensation coefficient determination module 600 reads this configuration and retains the adjacent conductive temperature and humidity grid points and their weights involved in the theoretical conductivity calculation. When the conductive response table does not have an interpolation method configured and the current node's temperature and humidity coordinates are within a rectangular conductive temperature and humidity grid cell, ordinary bilinear interpolation is used; if the current node's temperature and humidity coordinates are on the grid line between two conductive temperature and humidity grid points, linear interpolation is used; if they coincide with a conductive temperature and humidity grid point, the conductive temperature and humidity grid point is directly read. When the current node's temperature and humidity coordinates cannot be surrounded by the grid cells of the conductive response table and are not on the boundary grid line, the virtual node is marked as not updating in this round. This process can be represented as:

[0057]

[0058] In the formula, Indicates theoretical conductivity; This represents the conductive response table to which the virtual node is bound; and These represent node temperature and node humidity, respectively. The tension temperature and humidity grid points and the conductivity temperature and humidity grid points can have different grid configurations; therefore, when querying the two response tables, their corresponding grid points are used respectively.

[0059] The tension error ratio is determined based on the percentage of the absolute difference between the measured tension and the theoretical tension relative to the tension range. The calculation formula is as follows:

[0060]

[0061] In the formula, Indicates the tension error ratio; Indicates the measured tension; Indicates theoretical tension; This indicates the tension range. The tension error ratio is a dimensionless proportion used to represent the magnitude of the current tension difference relative to the allowable tension span.

[0062] The conductivity error ratio is determined based on the percentage of the absolute difference between the measured conductivity and the theoretical conductivity relative to the conductivity range. The calculation formula is as follows:

[0063]

[0064] In the formula, Indicates the conductivity error ratio; Indicates the measured conductivity; Indicates theoretical conductivity; Indicates the range of conductivity.

[0065] The larger of the tension error ratio and the conductivity error ratio is limited to the range of zero to one to obtain the original compensation coefficient, calculated as follows:

[0066]

[0067] In the formula, This represents the original compensation coefficient. First, the larger of the tension error ratio and the conductivity error ratio is selected, ensuring that the channel with the larger deviation between the two state channels can determine the update requirement. Then, an outer constraint ensures that the original compensation coefficient is neither less than zero nor greater than one. Since both the tension error ratio and the conductivity error ratio are calculated from their absolute differences, negative value constraints are usually not required. The original compensation coefficient reflects the update magnitude proposed by the tension error ratio and the conductivity error ratio.

[0068] Simultaneously, the tension residual and conductivity residual are determined for use by the response table update module 700. The tension residual and conductivity residual are written as follows:

[0069]

[0070] In the formula, This represents the tension residual. The tension residual retains the directional information of the measured tension relative to the theoretical tension; a positive tension residual indicates that the tension response near the current node's temperature and humidity coordinates needs to be increased, while a negative tension residual indicates that the corresponding tension response needs to be decreased.

[0071]

[0072] In the formula, This represents the conductivity residual. A positive conductivity residual indicates that the conductivity response near the current node's temperature and humidity coordinates needs to be increased, while a negative conductivity residual indicates that the corresponding conductivity response needs to be decreased. The original compensation coefficient determines the base update magnitude corresponding to the tension error ratio and conductivity error ratio. The tension residual and conductivity residual provide the update direction and difference for their respective response tables.

[0073] Furthermore, the third-order difference processing and the determination method for the deviation ratio of the two arcs are as follows:

[0074] The arc deviation ratio determination module 400 receives the measured height sequence and the virtual height sequence of the same contact wire span unit. The measured height sequence is processed in the order of left end height, left center height, right center height, and right end height, while the virtual height sequence is processed in the order of virtual left end height, virtual left center height, virtual right center height, and virtual right end height. The four positions in the two height sequences correspond one-to-one, and both are arranged from left to right.

[0075] The measured height sequence was subjected to third-order differencing to obtain the measured differencing result. The measured differencing result is the absolute value of four equally spaced heights combined with coefficients 1, -3, 3, and -1, calculated as follows:

[0076]

[0077] In the formula, This represents the measured difference processing result; Indicates the height of the left end; Indicates the height of the left center; Indicates the height of the right center; The height at the right end is indicated. All four heights use the same unit of length. For an equidistant height sequence consisting of a constant term, a first-order term, and a second-order term, the above third-order difference processing can eliminate the effects of overall height translation, linear slope, and approximate second-order sag, so that higher-order local non-uniform variations are reflected in the measured difference processing results.

[0078] The absolute values ​​of adjacent height differences between the left end height and the left center height, the left center height and the right center height, and the right center height and the right end height are obtained respectively. The absolute values ​​of adjacent height differences at both ends are assigned a single weight, and the absolute values ​​of adjacent height differences in the middle are assigned double the weight. The sum of the weighted results is determined as the measured total fluctuation, calculated as follows:

[0079]

[0080] In the formula, This represents the measured total fluctuation. The measured total fluctuation and the measured difference processing result are obtained from the same measured height sequence and have the same length dimension. The absolute value of the intermediate adjacent height difference is doubled to ensure that the measured total fluctuation corresponds to the coefficient structure in the third-order difference processing and that the measured difference processing result is no greater than the measured total fluctuation. The measured total fluctuation is used to eliminate the influence of the overall vertical variation of the span on the normalization result.

[0081] When the measured total fluctuation is non-zero, the measured difference processing result is normalized to the measured total fluctuation to obtain the four-point arc deviation ratio; when the measured total fluctuation is zero, the four-point arc deviation ratio is set to zero. The piecewise expression of the four-point arc deviation ratio is as follows:

[0082]

[0083] In the formula, This represents the deviation ratio of the four-point arc. When the measured total fluctuation is non-zero, the deviation ratio of the four-point arc is obtained by dividing the measured differential processing result by the measured total fluctuation; when the measured total fluctuation is zero, it means that there is no change in adjacent heights among the four measured heights, so it is directly taken as zero, without setting a separate minimum constant. The deviation ratio of the four-point arc is located in the range of zero to one.

[0084] Following the same processing method as the measured altitude sequence, the virtual altitude sequence is subjected to third-order difference processing to obtain the virtual difference processing result, calculated as follows:

[0085]

[0086] In the formula, This represents the result of virtual difference processing; Indicates the virtual left height; Indicates the virtual left-center height; Indicates the virtual right-center height; This indicates the virtual right-hand height.

[0087] The absolute values ​​of adjacent height differences in the virtual height sequence are weighted equally at both ends and twice in the middle. The sum of the weighted results is determined as the virtual total fluctuation, calculated as follows:

[0088]

[0089] In the formula, This represents the virtual total fluctuation. The virtual total fluctuation and the virtual difference processing result are obtained from the same virtual height sequence and use the same weighting configuration as the measured total fluctuation. Applying the same processing to the two height sequences can avoid misjudging existing endpoint height differences, design pre-slack, or inherent span morphology in the model as local anomalies on site.

[0090] When the virtual total fluctuation is non-zero, the virtual arc deviation ratio is obtained by normalizing the virtual difference processing result based on the virtual total fluctuation; when the virtual total fluctuation is zero, the virtual arc deviation ratio is set to zero. The piecewise expression of the virtual arc deviation ratio is as follows:

[0091]

[0092] In the formula, This represents the virtual arc deviation ratio. The virtual arc deviation ratio is in the range of zero to one and has no length dimension. When the total virtual ripple is zero, it is directly taken as zero.

[0093] After completing the above processing, the arc deviation ratio determination module 400 simultaneously outputs the four-point arc deviation ratio and the virtual arc deviation ratio to the arc compatibility determination module 500.

[0094] Furthermore, the determination methods for the arc matching degree and scene compensation coefficient are as follows:

[0095] The arc matching degree determination module 500 determines the arc matching degree based on the difference between the four-point arc deviation ratio and the virtual arc deviation ratio. The difference is taken as the absolute value of the difference between the two arc deviation ratios, and can be expressed as:

[0096]

[0097] In the formula, Indicate the difference; This indicates the deviation ratio of the four-point arc; This represents the virtual arc deviation ratio. Since both the four-point arc deviation ratio and the virtual arc deviation ratio are within the range of zero to one, the difference is also within the range of zero to one. A difference of zero indicates that after the measured height sequence and the virtual height sequence undergo the same third-order differencing and normalization processing, the four-point arc deviation ratio and the virtual arc deviation ratio are the same.

[0098] The arc coincidence degree is defined as being in the range of zero to one, decreasing as the difference in the deviation ratio between the two arcs increases. This implementation uses a linear correspondence:

[0099]

[0100] In the formula, The degree of arc alignment is represented by the following formula: When the difference is zero, the degree of arc alignment is one; when the difference reaches the maximum difference between the deviation ratios of the two arcs, the degree of arc alignment is zero; when the difference is between the two endpoints, the degree of arc alignment continuously decreases as the difference increases. The degree of arc alignment is used to constrain the original compensation coefficient.

[0101] A congruence of one indicates that, based on the comparison between the four-point arc deviation ratio and the virtual arc deviation ratio, there is no need to reduce the update magnitude of the tension error ratio and conductivity error ratio due to local arc mismatch. A congruence close to zero indicates a significant difference between the local morphology reflected by the measured height sequence and the virtual height sequence. In this case, even if there is a difference between the measured tension or measured conductivity and the theoretical output, the magnitude of this difference recorded in the tension response table and conductivity response table should be reduced.

[0102] The compensation coefficient determination module 600 receives the arc coincidence degree output by the arc coincidence degree determination module 500, and shrinks the original compensation coefficient according to the arc coincidence degree to obtain the scene compensation coefficient. The calculation formula is:

[0103]

[0104] In the formula, Indicates the scene compensation coefficient; Indicates the degree of coincidence of the arcs; This represents the original compensation coefficient. Both the original compensation coefficient and the arc coincidence are within the range of zero to one; therefore, the scene compensation coefficient is also within the range of zero to one and is not greater than the original compensation coefficient. The scene compensation coefficient is multiplicatively contracted compared to the original compensation coefficient.

[0105] When the difference is zero, the arc matching degree is one, and the scene compensation coefficient is equal to the original compensation coefficient; wherein, when the difference increases, the arc matching degree decreases, and the scene compensation coefficient decreases accordingly; when the difference reaches the maximum difference, both the arc matching degree and the scene compensation coefficient are zero.

[0106] Furthermore, the update methods for the tension response table and the conductivity response table are as follows:

[0107] The response table update module 700 receives scene compensation coefficients, tension residuals, conductivity residuals, tension response tables, and conductivity response tables. The update objects are the two response tables bound to the virtual nodes. The response table update module 700 first selects whether to directly overlay or allocate according to an existing interpolation method based on the relationship between the current node's temperature and humidity coordinates and the corresponding temperature and humidity grid points, and then forms the corrected tension table and the corrected conductivity table.

[0108] When the current node's temperature and humidity coordinates coincide with the tension temperature and humidity grid point, the tension residual, scaled according to the scene compensation coefficient, will be superimposed onto that tension temperature and humidity grid point. The update can be represented as:

[0109]

[0110] In the formula, Indicates correction of the tension table; Represents the tension response table; Indicates the scene compensation coefficient; Indicates tension residual; and These represent node temperature and node humidity, respectively. When the tension residual is positive, the tension response of the corresponding grid point increases; when the tension residual is negative, the tension response of the corresponding grid point decreases; the scene compensation coefficient determines the stacking ratio for this round.

[0111] When the current node's temperature and humidity coordinates do not coincide with the tension temperature and humidity grid points, the response table update module 700 uses the same set of adjacent tension temperature and humidity grid points and the same set of weights retained when obtaining the theoretical tension. It then distributes the tension residuals scaled according to the scene compensation coefficient to the adjacent tension temperature and humidity grid points according to their weights. If the theoretical tension is obtained by ordinary bilinear interpolation, the four adjacent tension temperature and humidity grid points each acquire their corresponding bilinear weights; if the theoretical tension is obtained by linear interpolation, the two adjacent tension temperature and humidity grid points each acquire their corresponding linear weights; if the theoretical tension is obtained by direct reading, all scaled tension residuals are superimposed on the corresponding tension temperature and humidity grid points. The sum of the weights remains one.

[0112] When the current node's temperature and humidity coordinates coincide with a conductive temperature and humidity grid point, the conductivity residual scaled according to the scene compensation coefficient will be superimposed onto that conductive temperature and humidity grid point. The update can be represented as:

[0113]

[0114] In the formula, Indicates correction of conductivity meter; Represents the conductivity response table; Indicates the scene compensation coefficient; This represents the conductivity residual. The sign of the conductivity residual determines the update direction of the conductivity response at the corresponding grid point, and the scene compensation coefficient determines the proportion of the conductivity residual that can be written into the conductivity response table.

[0115] When the current node's temperature and humidity coordinates do not coincide with the conductive temperature and humidity grid points, the response table update module 700 uses the same set of adjacent conductive temperature and humidity grid points and the same set of weights retained when obtaining the theoretical conductivity. It then distributes the conductivity residuals scaled according to the scene compensation coefficient to the adjacent conductive temperature and humidity grid points according to their weights. If the theoretical conductivity is obtained by ordinary bilinear interpolation, the four adjacent conductive temperature and humidity grid points each receive their corresponding bilinear weights; if the theoretical conductivity is obtained by linear interpolation, the two adjacent conductive temperature and humidity grid points each receive their corresponding linear weights; if the theoretical conductivity is obtained by direct reading, all scaled conductivity residuals are superimposed on the corresponding conductive temperature and humidity grid points. The sum of the weights remains one.

[0116] The revised tension table and tension response table are two data objects, one before and one after the update; similarly, the revised conductivity table and conductivity response table are also two data objects, one before and one after the update. During execution, the revised tension table and revised conductivity table can be generated in the temporary data area first. After both results pass the data integrity check, they are then written back to the corresponding virtual nodes according to their virtual node numbers.

[0117] Furthermore, the implementation methods for early warning temperature, early warning humidity, and the optimized virtual model of the overhead contact line are as follows:

[0118] After writing the corrected tension table and corrected conductivity table back to the corresponding virtual nodes, check the model output starting from the current node's temperature and humidity coordinates. First, generate candidate boundary grid points in the tension temperature and humidity grid points of the corrected tension table and the conductivity temperature and humidity grid points of the corrected conductivity table, respectively: tension temperature and humidity grid points where the tension reaches the minimum or maximum tension are added to the candidate boundary grid point set, and conductivity temperature and humidity grid points where the conductivity reaches the minimum or maximum conductivity are added to the candidate boundary grid point set. All candidate boundary grid points originate from existing grids in the corrected tension table or corrected conductivity table.

[0119] The candidate boundary grid points are sorted in a fixed order. First, the distance between each candidate boundary grid point and the current node's temperature and humidity coordinates is calculated; the distance is the sum of the absolute values ​​of the temperature coordinate difference and the humidity coordinate difference. Then, they are sorted in ascending order of distance; when the distances are the same, tension temperature and humidity grid points are sorted first, followed by conductivity temperature and humidity grid points; when they belong to the same response table and the distances are still the same, they are sorted in ascending order of temperature grid number, followed by humidity grid number. When checking grid by grid according to the above fixed order, the first candidate boundary grid point after sorting is the first boundary grid point to appear.

[0120] When only one boundary grid point appears initially, this boundary grid point is designated as the warning temperature and humidity grid point. The temperature of the warning temperature and humidity grid point is designated as the warning temperature, and the humidity of the warning temperature and humidity grid point is designated as the warning humidity. The warning temperature and warning humidity are then written back to the corresponding virtual nodes. The warning temperature and warning humidity are the corrected model output boundaries, used for virtual model display and risk warning.

[0121] If the candidate boundary grid point set is empty, no warning temperature and humidity grid points will be generated in this round, and the optimized catenary virtual model 800 will retain the existing warning temperature and humidity before being written back. If the same candidate boundary grid point satisfies multiple boundary conditions, it will still participate in the fixed sorting as a candidate boundary grid point; if there are multiple candidate boundary grid points after the fixed sorting, the candidate boundary grid point ranked first will be determined as the warning temperature and humidity grid point.

[0122] The optimized catenary virtual model 800 includes corrected tension and conductivity tables for the virtual nodes that have been written back, and may include warning temperature and humidity data obtained from subordinate processing. For a given node temperature and humidity, the optimized catenary virtual model 800 provides tension and conductivity outputs using the corrected tension and conductivity tables. After all contact wire span units have been written back, a global virtual model version of the catenary corresponding one-to-one with the physical line locations can be generated.

[0123] As an optional implementation, a contact wire span unit with complete data is selected. This contact wire span unit has a valid span number and a corresponding virtual node number. The contact wire material, contact wire cross-sectional area, minimum tension, maximum tension, minimum conductivity, and maximum conductivity are all configured. The tension response table and conductivity response table are bound to the virtual node. The laser scanning equipment on the inspection vehicle acquires point cloud data for the same inspection journey. The measured height acquisition module 100 sorts the contact wire points according to the line mileage direction and reads the left end height, left-middle height, right-middle height, and right end height through existing conductivity results from the inspection equipment or linear interpolation of adjacent contact wire points, thereby forming a measured height sequence. The node status acquisition module 200 acquires the node temperature, node humidity, measured tension, and measured conductivity of the same virtual node. The four status data are aligned using the timestamp of the measured height sequence as the target time and are checked for range, duplicate values, and median filtering. The compensation coefficient determination module 600 queries the two response tables based on the current node temperature and humidity coordinates to obtain the theoretical tension and theoretical conductivity. There is a small difference between the measured tension and the theoretical tension, and there is also a small difference between the measured conductivity and the theoretical conductivity. Therefore, the tension error ratio and the conductivity error ratio are both small, and the original compensation coefficient is in the lower range of zero to one. The arc deviation ratio determination module 400 processes the measured height sequence and the virtual height sequence respectively. Since the measured span shape and the model span shape are basically consistent under normal steady-state conditions, the four-point arc deviation ratio and the virtual arc deviation ratio are close, and the difference is close to zero, and the arc consistency is close to one. The scene compensation coefficient is therefore close to the original compensation coefficient. The response table update module 700 scales the small tension residual and conductivity residual according to the scene compensation coefficient and makes a small correction to the grid point corresponding to the current node temperature and humidity coordinate or the adjacent temperature and humidity grid point. At this time, the arc consistency will not unnecessarily lower the original compensation coefficient, and the initial virtual model of the contact network can absorb the tension residual and conductivity residual consistent with the field according to the tension error ratio and conductivity error ratio. After correcting the tension table and the conductivity table and writing them back to the virtual node, the output of the optimized catenary virtual model 800 under the same current node temperature and humidity coordinates is closer to the measured tension and conductivity, while maintaining the existing mesh structure and interpolation mechanism.

[0124] As an optional implementation, there is a significant difference between the measured tension or conductivity and the theoretical output. The larger of the tension error ratio and conductivity error ratio places the original compensation coefficient in a higher range. Simultaneously, the point cloud data reflects local dropper deviations, local bends, or isolated geometric outliers in the target contact line span unit, causing a higher-order local non-uniformity in the measured height sequence (left end height, left center height, right center height, and right end height) compared to the virtual height sequence. After third-order differential processing eliminates the combined effects of overall height translation, linear slope, and approximate quadratic sag, the proportion of the measured differential processing result relative to the measured total fluctuation is no longer close to the proportion of the virtual differential processing result relative to the virtual total fluctuation. The difference between the four-point arc deviation ratio and the virtual arc deviation ratio increases, and the arc matching degree decreases. The compensation coefficient determination module 600 multiplies the reduced arc matching degree by the original compensation coefficient to obtain a scene compensation coefficient smaller than the original compensation coefficient. The response table update module 700 still uses the tension residual and conductivity residual to determine the update direction, but only allows the portion determined by the scene compensation coefficient to be written into the two response tables. Compared to the treatment without setting arc compatibility constraints, the variation amplitude of the current round in the corrected tension table and corrected conductivity table is reduced. If the difference reaches the maximum difference, the scene compensation coefficient is zero, and the current round tension residual and conductivity residual are not written into the two response tables.

[0125] As an optional implementation, if any height in the measured height sequence is missing, or if the four heights come from different detection strokes and time registration cannot be completed, the existing heights are not stitched into a measured height sequence, and the four-point arc deviation ratio is not calculated. If the virtual height sequence is missing any corresponding height, other heights are not used as substitutes. The contact wire span unit maintains the original tension response table and conductivity response table in this round and records the incomplete input data state. If the measured total fluctuation is zero, the four-point arc deviation ratio is directly set to zero; if the virtual total fluctuation is zero, the virtual arc deviation ratio is directly set to zero. Neither of the above two boundaries introduces a minimum constant. If both the measured total fluctuation and the virtual total fluctuation are zero, the four-point arc deviation ratio and the virtual arc deviation ratio are both zero, the difference is zero, the arc matching degree is one, and then the original compensation coefficient and the update amplitude of this round are still determined by the tension error ratio and the conductivity error ratio. If the highest tension is not greater than the lowest tension or the highest conductivity is not greater than the lowest conductivity, the corresponding tension range or conductivity range cannot be used as the divisor of the error ratio. The system will mark the corresponding configuration as invalid and stop the current round of updates for that virtual node. If the current node's temperature and humidity coordinates are outside the domain defined by the response table, the existing boundary processing rules of the initial virtual model of the overhead contact system will be used. If the response table is missing, the boundary range is missing, the interpolation method configuration is missing and adjacent temperature and humidity grid points cannot be determined by ordinary bilinear interpolation, the virtual height sequence is missing, or the associated field is missing, the corresponding virtual node will not be updated in this round. The tension response table and conductivity response table before the write-back will be retained, and the span number, virtual node number, timestamp, and missing items will be recorded. If the temporary data area has already formed a corrected tension table but the corrected conductivity table calculation fails, or any data object fails verification during the write-back process, not only one data object will be written back, but both response tables before the write-back will be retained.

[0126] As an optional implementation, the corrected tension table and corrected conductivity table have been successfully written back to the virtual node. The system generates a set of candidate boundary grid points from the current node's temperature and humidity coordinates, and sorts them in a fixed order according to the distance between the candidate boundary grid points and the current node's temperature and humidity coordinates, the response table type, the temperature grid number, and the humidity grid number. The first candidate boundary grid point to appear during the grid-by-grid check in the fixed order is identified as the warning temperature and humidity grid point, and the warning temperature and humidity are obtained respectively. If the warning temperature and humidity grid point comes from the corrected tension table, its tension reaches the minimum or maximum tension; if the warning temperature and humidity grid point comes from the corrected conductivity table, its conductivity reaches the minimum or maximum conductivity. After the warning temperature and humidity are written back to the corresponding virtual node, the optimized catenary virtual model 800 can output this boundary information in the model display or risk warning. This allows the response table correction results to continue to be used for the state assessment of the catenary virtual model.

[0127] Although embodiments of this application have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of this application, and all such modifications and variations fall within the scope defined by the appended claims.

Claims

1. A method for optimizing a virtual catenary model based on scene compensation coefficients, comprising establishing an initial virtual model of the catenary and binding virtual nodes to tension response tables and conductivity response tables, characterized in that: Obtain the measured height sequence and virtual height sequence at four equally spaced positions on the contact wire; The two height sequences are subjected to third-order difference processing, and the measured total fluctuation and the virtual total fluctuation are determined by weighting the adjacent height differences at both ends equally and the adjacent height differences in the middle with double weights. When the measured total fluctuation is non-zero, the result is processed by normalizing the measured difference; otherwise, it is zero, and the deviation ratio of the four-point arc is obtained. When the virtual total fluctuation is non-zero, the result is processed by its normalized virtual difference; otherwise, it is zero, and the virtual arc deviation ratio is obtained. The degree of coincidence of two arcs is defined as being in the range of zero to one and decreasing as the difference in the deviation ratio between the two arcs increases; when the difference is zero, the degree of coincidence of the arcs is one. The theoretical tension and theoretical conductivity are obtained by querying two response tables using node temperature and node humidity. The tension error ratio is obtained by normalizing the absolute difference between the measured tension and the theoretical tension using the tension range. The conductivity error ratio is obtained by normalizing the absolute difference between the measured conductivity and the theoretical conductivity using the conductivity range. The larger one is limited to the range of zero to one to obtain the original compensation coefficient. The original compensation coefficient is reduced by the arc matching degree to obtain the scene compensation coefficient that decreases as the difference increases. When the difference is zero, the scene compensation coefficient is equal to the original compensation coefficient. The difference between measured tension and theoretical tension and the difference between measured conductivity and theoretical conductivity are scaled according to the scenario compensation coefficient. The tension response table and conductivity response table are updated to form a corrected tension table and a corrected conductivity table, which are then written back to the virtual node to obtain the optimized catenary virtual model.

2. The method for optimizing a virtual catenary model based on scene compensation coefficients according to claim 1, characterized in that, The establishment of the initial virtual model of the overhead contact line includes: Based on the location of the suspension points and the line topology, the contact wire between two adjacent suspension points is divided into contact wire span units. Span numbers are assigned to the contact wire span units and virtual node numbers are assigned to the virtual nodes. The contact wire span units are associated with the model span objects in the initial virtual model of the contact network according to the span numbers. Select the tension response table and the conductivity response table according to the contact wire material and the contact wire cross-sectional area, and bind the tension response table and the conductivity response table to the virtual node according to the virtual node number; Obtain the minimum tension, maximum tension, minimum conductivity, and maximum conductivity. Determine the tension range by the difference between the maximum tension and the minimum tension, and determine the conductivity range by the difference between the maximum conductivity and the minimum conductivity.

3. The method for optimizing a virtual catenary model based on scene compensation coefficients according to claim 2, characterized in that, According to the line mileage, the temperature and humidity acquisition locations, tension acquisition locations, and conductivity acquisition locations are associated with the corresponding virtual nodes. The association results are used to obtain the node temperature, node humidity, measured tension, and measured conductivity of the corresponding virtual nodes. The node temperature and node humidity are used to query the tension response table and the conductivity response table, and the measured tension and measured conductivity are used to determine the tension error ratio and the conductivity error ratio.

4. The method for optimizing a virtual catenary model based on scene compensation coefficients according to claim 3, characterized in that, Based on the suspension point location, the left end, one-third position, two-thirds position, and right end are determined in the contact wire span unit. The left end height, left middle height, right middle height, and right end height are obtained at the four positions to form a measured height sequence. Virtual left end height, virtual left middle height, virtual right middle height, and virtual right end height are also obtained to form a virtual height sequence. Point cloud data is collected by the laser scanning equipment mounted on the inspection vehicle, and the measured height sequence is extracted from the point cloud data. The measured height sequence is associated with the corresponding contact wire span unit using the span number and line mileage, and the timestamp is used to confirm that the measured height sequence comes from the same inspection trip or the same time registration data frame.

5. The method for optimizing a virtual catenary model based on scene compensation coefficients according to claim 4, characterized in that, Before determining the original compensation coefficients, the node temperature, node humidity, measured tension, and measured conductivity are aligned according to the timestamp. For node temperature, node humidity, measured tension, and measured conductivity with different sampling frequencies, the most recent time matching or linear interpolation is used to form a state data group at the same time. Abnormal measurement values ​​in the state data group are removed by range checking, duplicate value checking, and median filtering, and the cleaned node temperature, node humidity, measured tension, and measured conductivity are used for subsequent processing. Before determining the four-point arc deviation ratio, it is confirmed that the four heights in the measured height sequence come from the same detection stroke or the same time registration data frame.

6. The method for optimizing a virtual catenary model based on scene compensation coefficients according to claim 5, characterized in that, The current node temperature and humidity coordinates are determined based on the node temperature and humidity. When the current node temperature and humidity coordinates are within the domain of the tension response table and the conductivity response table, the two response tables are queried respectively. When the current node temperature and humidity coordinates coincide with the tension temperature and humidity grid points, the theoretical tension is read directly. Otherwise, the theoretical tension is obtained based on the tension adjacent temperature and humidity grid points according to the existing interpolation method of the tension response table. When the current node's temperature and humidity coordinates coincide with the conductive temperature and humidity grid points, the theoretical conductivity is read directly; otherwise, the theoretical conductivity is obtained based on the existing interpolation method of the conductivity response table using adjacent conductive temperature and humidity grid points. The tension error ratio is determined based on the ratio of the absolute difference between the measured tension and the theoretical tension to the tension range, and the conductivity error ratio is determined based on the ratio of the absolute difference between the measured conductivity and the theoretical conductivity to the conductivity range. The larger of the tension error ratio and the conductivity error ratio is limited to the range of zero to one to obtain the original compensation coefficient.

7. The method for optimizing a virtual catenary model based on scene compensation coefficients according to claim 6, characterized in that, The third-order difference processing processes the measured height sequence in the order of left-end height, left-middle height, right-middle height, and right-end height to obtain the measured difference processing result. The absolute values ​​of adjacent height differences between the left-end height and left-middle height, the left-middle height and right-middle height, and the right-middle height and right-end height are obtained respectively. The absolute values ​​of adjacent height differences at both ends are assigned a unit weight, and the absolute values ​​of adjacent height differences in the middle are assigned double weight. The sum of the weighted results is determined as the measured total fluctuation. The virtual height sequence is processed in the same way as the order of virtual left-end height, virtual left-middle height, virtual right-middle height, and virtual right-end height to obtain the virtual difference processing result and the virtual total fluctuation.

8. The method for optimizing a virtual catenary model based on scene compensation coefficients according to claim 7, characterized in that, The deviation ratios of the four-point arcs obtained from the measured differential processing results and the measured total fluctuations, as well as the deviation ratios of the virtual arcs obtained from the virtual differential processing results and the virtual total fluctuations, are all within the range of zero to one. When the difference between the four-point arc deviation ratio and the virtual arc deviation ratio is zero, the arc matching degree is set to one and the scene compensation coefficient is equal to the original compensation coefficient. When the difference reaches the maximum difference between the two arc deviation ratios, the arc matching degree and the scene compensation coefficient are set to zero.

9. The method for optimizing a virtual catenary model based on scene compensation coefficients according to claim 8, characterized in that, The difference between the measured tension and the theoretical tension is defined as the tension residual, and the difference between the measured conductivity and the theoretical conductivity is defined as the conductivity residual. When the temperature and humidity coordinates of the current node coincide with the tension temperature and humidity grid point, the tension residual scaled according to the scene compensation coefficient is superimposed on the tension temperature and humidity grid point; otherwise, the tension residual scaled according to the scene compensation coefficient is distributed to the adjacent tension temperature and humidity grid points according to the existing interpolation method of the tension response table, forming a corrected tension table. When the temperature and humidity coordinates of the current node coincide with the conductive temperature and humidity grid point, the conductivity residual scaled according to the scene compensation coefficient is superimposed on the conductive temperature and humidity grid point; otherwise, the conductivity residual scaled according to the scene compensation coefficient is distributed to the adjacent conductive temperature and humidity grid points according to the existing interpolation method of the conductive response table, forming a corrected conductivity table. The corrected tension table and the corrected conductivity table are written back to the corresponding virtual nodes.

10. The method for optimizing a virtual catenary model based on scene compensation coefficients according to claim 9, characterized in that, The model output is checked grid by grid along the tension temperature and humidity grid points of the corrected tension table and the conductivity temperature and humidity grid points of the corrected conductivity table from the current node temperature and humidity coordinates. When there is only one boundary grid point when it first appears, the boundary grid point is determined as the warning temperature and humidity grid point. The boundary grid point is the tension temperature and humidity grid point where the tension reaches the minimum tension or the maximum tension, or the conductivity temperature and humidity grid point where the conductivity reaches the minimum conductivity or the maximum conductivity. The temperature and humidity of the early warning temperature and humidity grid points are determined as early warning temperature and early warning humidity, respectively, and the early warning temperature and early warning humidity are written back to the corresponding virtual nodes.

Images