Pantograph carbon slide plate transverse deviation prediction method, device, equipment, medium and product

Abstract

The application discloses a pantograph carbon slide plate transverse deviation prediction method, device, equipment, medium and product. The pantograph carbon slide plate transverse deviation prediction method comprises the following steps: obtaining an input parameter set of a to-be-predicted train running condition, and preprocessing the input parameter set to obtain an input feature vector; based on the input feature vector and a pre-constructed training sample set, a K nearest neighbor regression method is used for transverse deviation prediction to obtain a dynamic transverse deviation prediction value; the training sample set is obtained through a multi-rigid-body dynamics simulation model of a wheel rail and a vehicle body under a plurality of preset conditions; a static transverse deviation value is obtained, the static transverse deviation value is a fixed deviation determined based on a wheel rail gap and a construction deviation; and based on the dynamic transverse deviation prediction value and the static transverse deviation value, a total transverse deviation prediction value is determined. The technical scheme of the embodiment of the application can accurately predict the pantograph carbon slide plate transverse deviation based on the train running condition.

Description

Technical Field

[0001] This invention relates to the field of rail transit technology, and in particular to a method, apparatus, equipment, medium, and product for predicting the lateral deviation of a pantograph carbon sliding plate. Background Technology

[0002] With the rapid development of urban rail transit and the increase in operating speed, the current collection quality of the pantograph and overhead contact system has become one of the key factors affecting the safe operation of trains.

[0003] As the direct contact component of the pantograph-catenary system, the lateral deviation (i.e., pull-out value) of the carbon contactor directly affects the dynamic matching performance of the pantograph-catenary system and the service life of the carbon contactor. If the pull-out value is designed to be too large, it can easily lead to risks such as uneven wear of the carbon contactor and detachment from the catenary; if it is designed to be too small, it may cause problems such as accelerated local wear of the contact wire and unstable current collection.

[0004] In existing technologies, the design of pantograph carbon skid plate deviation often relies on standards or given data, resulting in design redundancy and affecting the dynamic performance of pantograph-catenary matching. Therefore, accurately predicting the lateral deviation of the pantograph carbon skid plate is of great significance for optimizing pantograph-catenary parameters and improving the reliability of locomotive operation. Summary of the Invention

[0005] This invention provides a method, device, equipment, medium, and product for predicting the lateral deviation of the pantograph carbon sliding plate, in order to solve the problem that the deviation design of the existing pantograph carbon sliding plate does not take into account the actual train running conditions, thus affecting the dynamic performance of pantograph-catenary matching.

[0006] According to one aspect of the present invention, a method for predicting the lateral deviation of a pantograph carbon sliding plate is provided, comprising: Obtain the input parameter set of the train's operating conditions to be predicted, and preprocess the input parameter set to obtain the input feature vector; Based on the input feature vector and the pre-constructed training sample set, the K-nearest neighbor regression method is used to predict the lateral deviation and obtain the dynamic lateral deviation prediction value; the training sample set is obtained by simulating multiple preset working conditions through a multi-rigid-body dynamics simulation model of wheel-rail and vehicle body. Obtain the static lateral deviation value, which is a fixed deviation determined based on wheel-rail clearance and construction deviation; The total lateral deviation prediction value is determined based on the dynamic lateral deviation prediction value and the static lateral deviation value.

[0007] According to another aspect of the present invention, a pantograph carbon slider lateral deviation prediction device is provided, comprising: The input parameter acquisition module is used to acquire the set of input parameters for the train's operating conditions to be predicted, and to preprocess the set of input parameters to obtain the input feature vector. The dynamic deviation prediction module is used to predict the lateral deviation based on the input feature vector and the pre-constructed training sample set, using the K-nearest neighbor regression method to obtain the dynamic lateral deviation prediction value; the training sample set is obtained by simulating multiple preset working conditions through a multi-rigid-body dynamics simulation model of wheel-rail and vehicle body. The static deviation acquisition module is used to acquire the static lateral deviation value, which is a fixed deviation determined based on the wheel-rail clearance and construction deviation. The total deviation determination module is used to determine the total lateral deviation prediction value based on the dynamic lateral deviation prediction value and the static lateral deviation value.

[0008] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the pantograph carbon skid plate lateral deviation prediction method according to any embodiment of the present invention.

[0009] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions for causing a processor to execute and implement the pantograph carbon skid plate lateral deviation prediction method according to any embodiment of the present invention.

[0010] According to another aspect of the present invention, a computer program product is provided, comprising a computer program that, when executed by a processor, implements the pantograph carbon skid plate lateral deviation prediction method of any embodiment of the present invention.

[0011] The technical solution of this invention involves obtaining a set of input parameters for the train's operating conditions to be predicted, preprocessing the input parameter set to obtain an input feature vector, using the K-nearest neighbor regression method to predict lateral deviation based on the input feature vector and a pre-constructed training sample set, obtaining a dynamic lateral deviation prediction value, and obtaining a static lateral deviation value. The static lateral deviation value is a fixed deviation determined based on wheel-rail clearance and construction deviation. Based on the dynamic lateral deviation prediction value and the static lateral deviation value, a total lateral deviation prediction value is determined. This method can predict the lateral deviation of the pantograph's carbon sliding plate under different train operating conditions, providing a reliable reference for pantograph selection, installation location optimization, and vehicle clearance design.

[0012] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0013] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1a This is a flowchart of a method for predicting the lateral deviation of a pantograph carbon sliding plate according to Embodiment 1 of the present invention; Figure 1b This is a schematic diagram of the vehicle system motion coordinate system provided in Embodiment 1 of the present invention; Figure 1c This is a wind speed and wind pressure fitting curve provided according to Embodiment 1 of the present invention; Figure 2a This is a flowchart of a method for predicting the lateral deviation of a pantograph carbon sliding plate according to Embodiment 2 of the present invention; Figure 2b This is a technical roadmap of the pantograph carbon sliding plate lateral deviation prediction method provided in Embodiment 2 of the present invention; Figure 2c This is a schematic diagram of the pantograph carbon sliding plate lateral deviation prediction interface provided in Embodiment 2 of the present invention; Figure 2d This is a schematic diagram of a warning pop-up window provided according to Embodiment 2 of the present invention; Figure 3 This is a schematic diagram of the structure of a pantograph carbon sliding plate lateral deviation prediction device according to Embodiment 3 of the present invention; Figure 4 This is a schematic diagram of the structure of an electronic device that implements the pantograph carbon slider lateral deviation prediction method according to an embodiment of the present invention. Detailed Implementation

[0015] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0016] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0017] Example 1 Figure 1a This invention provides a flowchart of a method for predicting the lateral deviation of a pantograph carbon sliding plate according to Embodiment 1. This embodiment is applicable to situations where the lateral deviation of the pantograph carbon sliding plate is predicted based on train operating conditions. This method can be executed by a pantograph carbon sliding plate lateral deviation prediction device, which can be implemented in hardware and / or software and can be configured in various general-purpose computing devices. Figure 1a As shown, the method includes: S110. Obtain the input parameter set of the train's operating conditions to be predicted, and preprocess the input parameter set to obtain the input feature vector.

[0018] The input parameter set is a multi-dimensional set of parameters used to characterize the track conditions, train operating status, environmental information, and pantograph installation status under the predicted operating condition. For example, the input parameter set includes parameters characterizing track conditions such as track type (straight or curved), curve radius, and superelevation status (whether superelevation exists); vehicle speed parameters characterizing train operating status; lateral wind pressure status (no wind, inner wind, and outer wind) parameters characterizing environmental information; and pantograph installation position parameters on the car body characterizing pantograph installation status.

[0019] In this embodiment of the invention, multi-dimensional input parameters for the train's operating conditions to be predicted are first obtained. For example, the input parameters include vehicle speed, track type, curve radius, superelevation, lateral wind pressure, and the pantograph's mounting position on the vehicle body.

[0020] Furthermore, the input parameters are preprocessed to unify all input parameters into a dimensionless numerical format, facilitating subsequent deviation prediction. For example, categorical parameters are converted into numerical values ​​according to a preset ordered encoding rule to obtain the feature values ​​of the categorical input parameters. For instance, the line type "straight line" is encoded as 0, and "curve" is encoded as 1.

[0021] Simultaneously, the numerical input parameters are normalized to obtain the feature values ​​of the numerical input parameters. For example, Min-Max normalization is used to map the vehicle speed to the [0,1] interval. Finally, the feature values ​​of the preprocessed input parameters are combined in a preset order to form the input feature vector.

[0022] S120. Based on the input feature vector and the pre-constructed training sample set, the K-nearest neighbor regression method is used to predict the lateral deviation and obtain the dynamic lateral deviation prediction value.

[0023] The training sample set is a dataset obtained through simulations using a multi-rigid-body dynamics simulation model of the wheel-rail and vehicle body under multiple preset working conditions. Each training sample in the set contains a set of sample feature vectors and their corresponding true values ​​of lateral deviations (including positive and negative lateral deviations). The training sample set serves as a knowledge base for K-nearest neighbor regression, providing historical data for prediction.

[0024] In this embodiment of the invention, based on the input feature vector and a pre-constructed training sample set, the K-nearest neighbor regression method is used to predict lateral bias, obtaining a dynamic lateral bias prediction value. Specifically, the input feature vector is compared with the sample feature vector of each sample in the training sample set, and the spatial distance between them is calculated. For example, the spatial distance is Euclidean distance. Then, based on the calculated spatial distance, all training samples are sorted in ascending order of distance, and the K samples with the smallest distance are selected as neighboring samples, where the value of K is predetermined through cross-validation.

[0025] Each neighboring sample is assigned a weight based on its spatial distance from the input feature vector. Samples with smaller spatial distances contribute more to the subsequent weighted summation, i.e., they have larger weights. For example, the reciprocal of the spatial distance between each neighboring sample is used as the initial weight; the smaller the distance, the larger the reciprocal, and the higher the weight. Then, the initial weights of all neighboring samples are summed, and each initial weight is divided by the sum of the weights to obtain a normalized weight value, ensuring that the sum of the weights of all neighboring samples equals 1.

[0026] Finally, the positive lateral deviation values ​​of each neighboring sample are multiplied by their respective weights and then summed to obtain the predicted dynamic lateral positive deviation value under the current operating condition. Similarly, the negative lateral deviation values ​​of each neighboring sample are multiplied by their respective weights and then summed to obtain the predicted dynamic lateral negative deviation value under the current operating condition. Based on this, the absolute values ​​of the predicted dynamic lateral positive deviation value and the predicted dynamic lateral negative deviation value are summed to obtain the dynamic deviation range.

[0027] Optionally, the training sample set can be constructed as follows: A multi-rigid-body dynamics simulation model of wheel-rail and vehicle body is established; the multi-rigid-body dynamics simulation model includes multiple vehicle body sections, multiple frames, multiple wheelsets and multiple axle box swing arms, and has multi-degree-of-freedom motion capability; Based on the multi-rigid-body dynamics simulation model, dynamics simulations are performed on various preset working conditions to obtain the lateral positive and negative deviation values ​​of the pantograph carbon slide plate center relative to the track center. The sample feature vector of the input parameter set associated with each preset working condition, along with the corresponding lateral positive and negative deviation values, are stored together as training samples to obtain the training sample set.

[0028] In this optional embodiment, a method for constructing the training sample set is provided, as follows: First, a multi-rigid-body dynamics simulation model of the wheel-rail system and the vehicle body is established. The multi-rigid-body dynamics simulation model includes multiple vehicle body sections, multiple frames, multiple wheelsets, and multiple axle box swing arms, and has multi-degree-of-freedom motion capabilities.

[0029] In a specific example, vehicle vibration affects the lateral pull-out value of the pantograph carbon sliding plate. Therefore, a vehicle model is built on the basis of the pantograph carbon sliding plate model to add vehicle vibration simulation function.

[0030] According to vehicle dynamics theory, the train consists of 4 bogies and 3 car bodies. The bogies are mainly composed of wheelsets, axle boxes, frames and other components. The car body and bogies transmit vertical force through air springs. The weight of the entire car body is borne by 8 running wheels on the two bogies.

[0031] A multi-rigid-body dynamics simulation model of the vehicle system was established using multibody dynamics software. This model mainly consists of 3 vehicle bodies, 4 frames, 8 wheelsets, and 16 axle box swingarms. The vehicle bodies, frames, and wheelsets each have 6 degrees of freedom (longitudinal, lateral, vertical, roll, pitch, and yaw), while the axle box swingarms have 1 degree of freedom (pitch). The entire vehicle system has 106 degrees of freedom, as shown in Table 1.

[0032] Table 1 like Figure 1bThe vehicle system motion coordinate system shown defines the first wheelset in the vehicle's forward direction as the first wheelset. The motion coordinate system is a right-handed coordinate system, with the forward direction as the positive x-axis and the downward direction of the vehicle as the positive z-axis.

[0033] Before establishing the multi-rigid-body dynamics simulation model, the following assumptions are made about the vehicle model: (1) The car body, frame, wheelset, and axle box are simplified as non-deformable rigid bodies, and their structural elastic deformation is not considered; (2) The suspension of the car body and bogie is symmetrically arranged; (3) The track beam is regarded as a rigid body, and the wheels do not leave the track during operation.

[0034] Based on the multi-rigid-body dynamics simulation model, dynamic simulations are performed on various preset working conditions to obtain the lateral positive and negative deviation values ​​of the pantograph carbon sliding plate center relative to the track center. The sample feature vectors of the input parameter set associated with each preset working condition, along with the corresponding lateral positive and negative deviation values, are stored together as training samples to obtain the training sample set.

[0035] In a specific example, during the design phase, track parameters are selected for straight and curved sections. The curved section parameters include curve radii of 50m and 300m. The maximum operating speed of the vehicle is 70km / h, and the track geometric irregularity spectrum is used as the input for track geometric irregularity. The influence of lateral wind speed and crosswind load calculations are shown in Table 2.

[0036] Table 2 It is important to note that trains should slow down during gales of force 8; and should be suspended immediately during gales of force 9 or above, or in severe weather conditions such as heavy fog, heavy snow, or sandstorms. The approximate formula for wind pressure calculation is P. w =v 2 / 1.6, plus an additional 20% negative pressure.

[0037] Fitting the wind speed and wind pressure curves, such as Figure 1c As shown, the influence of wind speed on the deflection of the pantograph carbon sliding plate is analyzed by adding wind pressure inside and outside the curve. To monitor the dynamic deflection of the pantograph carbon sliding plate during operation, detection points are added at different positions on the lead and intermediate vehicles to analyze the influence of the pantograph carbon sliding plate installation position on the pull-out value. The dynamic deflection of the pantograph carbon sliding plate is obtained through dynamic analysis calculation.

[0038] The numerical simulation data above was divided into 16 sets of lateral deviation data for carbon sliding plates under typical working conditions. Each set of data contained 6 input features and 3 output targets. The input features included: vehicle speed (km / h), track type (straight / curved), superelevation status (none / present), lateral wind pressure status (none / inner / outer), curve radius, and installation position. The output targets were the positive lateral deviation value, the negative lateral deviation value, and the deviation range. The training sample set consisted of the input feature vectors and simulation output data from multiple sets of working conditions.

[0039] Based on the multi-rigid-body dynamics simulation model, representative typical working conditions are simulated to obtain training samples. This can realistically reflect the coupled vibration between carriages of the train and the dynamic response differences when the pantograph is installed in different positions, ensuring the accuracy and reliability of the training samples.

[0040] Optionally, before performing dynamic simulations on various preset working conditions based on the multi-rigid-body dynamics simulation model, the method further includes: In the multi-rigid-body dynamics simulation model, a track center marker and a carbon slide center marker are established so that the center of the carbon slide coincides with the center of the track under static conditions. Using the track center marker as the measurement reference, the lateral displacement of the carbon skateboard center marker during the simulation is directly taken as the lateral deviation value relative to the track center.

[0041] In this optional embodiment, before performing dynamic simulations on various preset working conditions based on the multi-rigid-body dynamics simulation model, the method further includes: establishing a track center marker and a carbon slide center marker in the multi-rigid-body dynamics simulation model, such that the center of the carbon slide coincides with the center of the track under static conditions. Using the track center marker as a measurement reference, the lateral displacement of the carbon slide center marker during the simulation is directly taken as the lateral deviation value relative to the track center.

[0042] Specifically, the center position of the pantograph carbon slide plate is established, and a reference coordinate system is used to establish a relative reference point between the vehicle body and the track. At the same time, a relative reference point between the center position of the pantograph carbon slide plate and the vehicle body is established, so that the center of the carbon slide plate coincides with the center of the track under static conditions, ensuring that the positional deviation of the carbon slide plate center is a relative displacement with respect to the center of the track during vehicle movement.

[0043] By setting the track center marker and the carbon skateboard center marker in the model and making them statically coincident, the lateral displacement of the carbon skateboard center marker can be directly output as the lateral deviation value relative to the track center during the simulation. No additional coordinate transformation or post-processing calculation is required after the simulation, thus avoiding the cumulative error that may be introduced during the coordinate transformation process.

[0044] S130. Obtain the static lateral deviation value. The static lateral deviation value is a fixed deviation determined based on the wheel-rail clearance and construction deviation.

[0045] In this embodiment of the invention, the static lateral deviation value provided by the user can be directly obtained; this value is fixed. If the user does not provide it, the static lateral deviation value can be determined based on the wheel-rail clearance and construction deviation. Specifically, based on the wheel-rail clearance deviation, the center deviation of the pantograph carbon sliding plate is determined, and then this deviation is summed with the construction deviation to obtain the static lateral deviation value.

[0046] S140. Based on the predicted values ​​of dynamic and static lateral deviations, determine the predicted value of total lateral deviation.

[0047] The total lateral deviation prediction value is the final predicted range of lateral deviation between the pantograph carbon slide center and the track center, obtained by superimposing the dynamic lateral deviation prediction value and the static lateral deviation value. Specifically, the total lateral deviation prediction value includes the total positive lateral deviation prediction value, the total negative lateral deviation prediction value, and the total deviation range.

[0048] In this embodiment of the invention, the static lateral deviation value is accumulated in the positive direction to the dynamic lateral positive deviation prediction value to obtain the total lateral positive deviation prediction value. Similarly, the static lateral deviation value is accumulated in the negative direction to the dynamic lateral negative deviation prediction value to obtain the total lateral negative deviation prediction value. Finally, based on the total lateral positive deviation prediction value and the total lateral negative deviation prediction value, the total deviation range is calculated. The total lateral positive deviation prediction value, the total lateral negative deviation prediction value, and the total deviation range are collectively used as the total lateral deviation prediction value.

[0049] The technical solution of this invention involves obtaining a set of input parameters for the train's operating conditions to be predicted, preprocessing the input parameter set to obtain an input feature vector, using the K-nearest neighbor regression method to predict lateral deviation based on the input feature vector and a pre-constructed training sample set, obtaining a dynamic lateral deviation prediction value, and obtaining a static lateral deviation value. The static lateral deviation value is a fixed deviation determined based on wheel-rail clearance and construction deviation. Based on the dynamic lateral deviation prediction value and the static lateral deviation value, a total lateral deviation prediction value is determined. This method can predict the lateral deviation of the pantograph's carbon sliding plate under different train operating conditions, providing a reliable reference for pantograph selection, installation location optimization, and vehicle clearance design.

[0050] Example 2 Figure 2aThis is a flowchart of a pantograph carbon slider lateral deviation prediction method provided in Embodiment 2 of the present invention. This embodiment further refines the above embodiment, providing specific steps for predicting lateral deviation using the K-nearest neighbor regression method based on the input feature vector and a pre-constructed training sample set to obtain dynamic lateral deviation prediction values, and specific steps for determining the total lateral deviation prediction value based on the dynamic lateral deviation prediction value and the static lateral deviation value. Figure 2a As shown, the method includes: S210. Obtain the input parameter set of the train's operating conditions to be predicted, and preprocess the input parameter set to obtain the input feature vector.

[0051] Optionally, the input parameter set includes vehicle speed, track type, curve radius, superelevation status, lateral wind pressure status, and pantograph mounting position on the vehicle body.

[0052] Optionally, the input parameter set is preprocessed to obtain the input feature vector, including: The input parameters that belong to the category type are converted into numerical form according to a preset ordered encoding rule to obtain the feature values ​​corresponding to the category type parameters; the category type parameters include line type, superelevation status and lateral wind pressure status. The input parameters that are numerical are normalized to obtain the characteristic values ​​corresponding to the numerical parameters; the numerical parameters include vehicle speed, curve radius and pantograph mounting position on the vehicle body. The feature values ​​corresponding to each parameter in the input parameter set are combined in a preset order to form the input feature vector.

[0053] In the two optional embodiments described above, the specific content of the input parameter set and the specific method for preprocessing the input parameter set to obtain the input feature vector are provided as follows: Since the K-assertion algorithm is based on distance metrics, it is necessary to unify all input parameters into numerical types and eliminate the influence of dimensions. For each parameter type in the input parameter set, preprocessing is performed separately: Parameters belonging to the categorical type in the input parameter set are converted into numerical forms according to a preset ordered encoding rule to obtain the feature values ​​corresponding to the categorical parameters. Specifically, the categorical parameters include line type, superelevation status, and lateral wind pressure status. Among them, in the line type, "straight line" corresponds to a value of 0, and "curve" corresponds to a value of 1; in the superelevation status, "none" corresponds to a value of 0, and "present" corresponds to a value of 1; in the lateral wind pressure status, "none" corresponds to a value of 0, "inner wind pressure" corresponds to a value of 1, and "outer wind pressure" corresponds to a value of 2.

[0054] Simultaneously, the numerical parameters in the input parameter set are normalized to obtain the corresponding feature values. Specifically, the numerical parameters include vehicle speed, curve radius, and the pantograph's mounting position on the vehicle body. The Min-Max normalization method is used, and the calculation formula is as follows: , in, These are the normalized eigenvalues, and x is the initial value of the numerical parameter. , These are the maximum and minimum values ​​of the current numerical parameter in the training set, respectively. When the range is 0, it is set to 1 to avoid division by zero.

[0055] Finally, the feature values ​​corresponding to each parameter in the input parameter set are combined in a preset order to form the input feature vector. By performing category encoding and numerical normalization preprocessing on the input parameter set, all input features are unified into a dimensionless numerical format, which can correctly participate in the K-nearest neighbor regression calculation based on Euclidean distance, avoiding the distance metric distortion problem caused by differences in feature dimensions and improving the accuracy of prediction results.

[0056] S220. Calculate the spatial distance between the input feature vector and the feature vector of each sample in the training sample set, and based on the spatial distance, select a set number of neighboring sample feature vectors that are close to the input feature vector; the set number is predetermined by cross-validation.

[0057] Each training sample includes a sample feature vector, as well as corresponding positive and negative lateral deviation values. The training sample set is obtained through simulation under multiple preset working conditions using a multi-rigid-body dynamics simulation model of wheel-rail and vehicle body.

[0058] In this embodiment of the invention, the method for predicting the lateral deviation of the pantograph carbon sliding plate is as follows: Figure 2b As shown, after preprocessing the input parameters to obtain the input feature vector, the spatial distance between the input feature vector and the feature vector of each sample in the training sample set is calculated. The specific formula is as follows: , in, It is the input feature vector With the feature vector of the i-th training sample The Euclidean distance between them, where m is the dimension of the feature vector. , These are the values ​​of the input feature vector and the i-th training sample feature vector on the j-th feature, respectively.

[0059] Furthermore, based on spatial distance, a predetermined number of neighboring sample feature vectors are selected, for example, K neighboring sample feature vectors. This predetermined number is determined in advance using cross-validation. Each training sample includes a sample feature vector, as well as corresponding positive and negative lateral bias values.

[0060] S230. Determine the weight of each neighboring sample feature vector based on the spatial distance between each neighboring sample feature vector and the input feature vector.

[0061] Among them, the closer the spatial distance between the neighboring sample feature vector and the input feature vector, the greater the corresponding weight.

[0062] In this embodiment of the invention, the weight of each neighboring sample feature vector is determined based on the spatial distance between each neighboring sample feature vector and the input feature vector. The principle is that the smaller the spatial distance between a neighboring sample feature vector and the input feature vector, the greater its corresponding weight. Specifically, the reciprocal of each spatial distance between the neighboring sample feature vector and the input feature vector is calculated as the initial weight value for that neighboring sample. Then, the initial weight values ​​of all neighboring samples are summed to obtain a total weight, and the initial weight value of each neighboring sample is divided by this total weight to obtain the normalized weight value for that neighboring sample.

[0063] Alternatively, based on the spatial distance between the feature vectors of each neighboring sample and the input feature vector, an initial weight value for each neighboring sample can be calculated using a Gaussian kernel function. The output value of the Gaussian kernel function decays exponentially with increasing distance. Then, the initial weight values ​​of all neighboring samples are normalized to obtain the normalized weight value for each neighboring sample.

[0064] Optionally, the weight of each neighboring sample feature vector is determined based on the spatial distance between each neighboring sample feature vector and the input feature vector, including: The inverse of the spatial distance between each neighboring sample feature vector and the input feature vector is used as the initial weight of that neighboring sample feature vector; The initial weights of the predetermined number of neighboring sample feature vectors are summed to obtain the initial weight sum. The ratio of the initial weight of each neighboring sample feature vector to the initial weight sum is calculated and used as the weight of each neighboring sample feature vector.

[0065] In this optional embodiment, a specific method is provided for determining the weight of each neighboring sample feature vector based on the spatial distance between each neighboring sample feature vector and the input feature vector: the reciprocal of the spatial distance between each neighboring sample feature vector and the input feature vector is used as the initial weight of that neighboring sample feature vector. The initial weights of a set number of neighboring sample feature vectors are summed to obtain a total initial weight. The ratio of the initial weight of each neighboring sample feature vector to the total initial weight is calculated as the weight of each neighboring sample feature vector. The specific calculation formula is as follows: , in, It is the weight of the feature vector of the i-th neighboring sample. It is the distance in space between the feature vector of the i-th neighboring sample and the input feature vector. It is the spatial distance between the feature vector of the j-th neighboring sample and the infeed feature vector.

[0066] S240. Based on the lateral positive deviation value, lateral negative deviation value and weight corresponding to the feature vector of each neighboring sample, predict the dynamic lateral positive deviation prediction value, dynamic lateral negative deviation prediction value and dynamic deviation range under the train's operating conditions to be predicted, and use them as the dynamic lateral deviation prediction value.

[0067] In this embodiment of the invention, based on the lateral positive deviation value, lateral negative deviation value, and weight corresponding to each neighboring sample feature vector, the dynamic lateral positive deviation prediction value, dynamic lateral negative deviation prediction value, and dynamic deviation range under the train's operating condition are predicted, and these are used as the dynamic lateral deviation prediction value. Specifically, the lateral positive deviation value of each neighboring sample feature vector is multiplied by its corresponding weight and then summed to obtain the dynamic lateral positive deviation prediction value under the current operating condition. Similarly, the lateral negative deviation value of each neighboring sample feature vector is multiplied by its corresponding weight and then summed to obtain the dynamic lateral negative deviation prediction value under the current operating condition. The specific calculation formula is as follows: , in, It is the predicted value of dynamic lateral positive deviation. The lateral positive deviation value of the feature vector of the i-th neighboring sample. It is the weight of the feature vector of the i-th neighboring sample. Dynamic lateral negative bias prediction value, The lateral negative bias value of the feature vector of the i-th neighboring sample.

[0068] Furthermore, based on the predicted values ​​of dynamic lateral positive deviation and dynamic lateral negative deviation, the dynamic deviation range is calculated using the following formula: , in, It is the dynamic deviation range. It is the predicted value of dynamic lateral positive deviation. This is the predicted value of the dynamic lateral negative bias. Ultimately, the predicted values ​​of the dynamic lateral positive bias, dynamic lateral negative bias, and the dynamic bias range are used as the predicted value of the dynamic lateral bias. Dynamic lateral deviation prediction based on K-nearest neighbor regression can quickly estimate any new working condition using simulation data of small sample typical working conditions, avoiding the need for high-cost and long-cycle dynamic simulation calculations for each prediction, thus improving the iterative efficiency of engineering design.

[0069] S250. Obtain the static lateral deviation value. The static lateral deviation value is a fixed deviation determined based on the wheel-rail clearance and construction deviation.

[0070] Optionally, obtain the static lateral deviation value, including: Determine the wheel-rail clearance based on the preset maximum track gauge and minimum wheelset width; The static lateral deviation value is determined based on the preset construction deviation and wheel-rail clearance.

[0071] In this optional embodiment, a specific method for obtaining the static lateral deviation value is provided: firstly, based on the preset maximum track gauge L... max With minimum wheel pair width d min Determine the wheel-rail clearance δ max =L max -d min The resulting deviation of the pantograph's carbon sliding plate center is Δd1=δ max / 2.

[0072] Furthermore, based on the preset construction deviation Δa and wheel-rail clearance, the static lateral deviation value is determined. Specifically, the impact of the construction deviation Δa on the pantograph-catenary position is calculated as follows: the change in pantograph-catenary position caused by the construction deviation is Δd2 = Δa.

[0073] The final calculated static lateral deviation value is e sta =Δd1+Δd2.

[0074] S260. The dynamic lateral positive deviation prediction value in the dynamic lateral deviation prediction value is superimposed with the static lateral deviation value in the positive direction to obtain the total lateral positive deviation prediction value.

[0075] In this embodiment of the invention, the predicted value of dynamic positive lateral deviation in the dynamic lateral deviation prediction value is superimposed with the static lateral deviation value in the positive direction to obtain the total predicted value of positive lateral deviation. The specific calculation formula is as follows: , in, This is the predicted value of the total lateral positive deviation. It is the predicted value of dynamic lateral positive deviation. It is the static lateral deviation value.

[0076] S270. The dynamic lateral negative deviation prediction value in the dynamic lateral deviation prediction value is superimposed with the static lateral deviation value in the negative direction to obtain the total lateral negative deviation prediction value.

[0077] In this embodiment of the invention, the predicted value of dynamic lateral negative deviation in the dynamic lateral deviation prediction value is superimposed with the static lateral deviation value in the negative direction to obtain the total lateral negative deviation prediction value. The specific calculation formula is as follows: , in, This is the predicted value of the total lateral negative bias. Dynamic lateral negative bias prediction value, It is the static lateral deviation value.

[0078] S280. Based on the total positive lateral deviation prediction value and the total negative lateral deviation prediction value, calculate the total deviation range, and use the total positive lateral deviation prediction value, the total negative lateral deviation prediction value, and the total deviation range as the total lateral deviation prediction value.

[0079] In this embodiment of the invention, the total deviation range is calculated based on the predicted total positive lateral deviation and the predicted total negative lateral deviation. The specific calculation formula is as follows: , in, It is the total deviation range. This is the predicted value of the total lateral positive deviation. This is the predicted value of the total lateral negative bias. It is the dynamic deviation range. It is the static lateral deviation value.

[0080] Finally, the total lateral positive deviation prediction, the total lateral negative deviation prediction, and the total deviation range are used as the total lateral deviation prediction.

[0081] Optionally, embodiments of the present invention further include: The parameter input area and the static deviation input area are rendered in the interactive interface; the parameter input area contains a custom input box corresponding to each input parameter in the input parameter set, and the static deviation input area contains a custom input box for the static lateral deviation value. In response to the user's click on the prediction button, the system performs lateral bias prediction using the K-nearest neighbor regression method based on the user-provided input parameter set and the pre-built training sample set, to obtain the dynamic lateral bias prediction value and the total lateral bias prediction value. The prediction result display area is rendered in the interactive interface to display the dynamic lateral deviation prediction value and the total lateral deviation prediction value.

[0082] In this optional embodiment, it further includes: in such a way as Figure 2c The interactive interface shown renders a parameter input area and a static deviation input area. The parameter input area contains a custom input box corresponding to each input parameter in the input parameter set, and the static deviation input area contains a custom input box for the static lateral deviation value. Users can predict the lateral deviation of the pantograph carbon skid plate under corresponding driving conditions by inputting the parameter values ​​of each input parameter in the interactive interface.

[0083] In response to the user's click on the prediction button, the system performs lateral bias prediction using the K-nearest neighbor regression method based on the user-provided input parameter set and a pre-built training sample set, obtaining the dynamic lateral bias prediction value and the total lateral bias prediction value.

[0084] Finally, the prediction results display area is rendered in the interactive interface, showing the dynamic lateral deviation prediction value and the total lateral deviation prediction value. The dynamic lateral deviation prediction value includes the dynamic positive lateral deviation prediction value, the dynamic negative lateral deviation prediction value, the dynamic deviation range, and the static deviation input. The total lateral deviation prediction value includes the total positive lateral deviation prediction value, the total negative lateral deviation prediction value, and the total deviation range.

[0085] It provides a visual graphical user interface and clear operational boundaries. An interactive interface was developed based on numerical simulation analysis software, supporting input of speed, line type, superelevation status, wind pressure status, curve radius, installation location, and static deviation values. It automatically completes data encoding, normalization, K-nearest neighbor regression prediction, and deviation synthesis, displaying dynamic prediction results and total deviation results in real time.

[0086] Optionally, before responding to a user's click on the prediction button, the following may also be included: Perform validity checks on parameters entered by users in custom input boxes; If abnormal parameters are detected, an alert message will be issued.

[0087] In this optional embodiment, before responding to the user's click on the prediction button, the method further includes: performing a validity check on the parameters entered by the user in the custom input box. If an abnormal parameter is detected, a warning message is issued. For example, if a negative vehicle speed is detected, the output is as follows: Figure 2d The warning message "Speed ​​must be positive" is displayed as a pop-up alert. The input validity check function triggers warnings when parameters are abnormal, lowering the barrier to entry and facilitating engineers' quick evaluation of different solutions.

[0088] To verify the effectiveness of the model, leave-one-out cross-validation was used to evaluate the samples in the training set, with the root mean square error (RMSE) as the evaluation metric. The specific formula is as follows: , Where n is the total number of test samples, It is the true value of the i-th test sample (i.e., the lateral deviation value obtained through simulation by the dynamic model, including the lateral positive deviation value and the lateral negative deviation value). It is the predicted value of the i-th test sample (i.e., the dynamic lateral deviation prediction value output by the K-nearest neighbor regression model for this sample, including the dynamic lateral positive deviation prediction value and the dynamic lateral negative deviation prediction value).

[0089] The technical solution of this invention involves obtaining a set of input parameters for the train's operating conditions to be predicted, preprocessing the input parameter set to obtain an input feature vector, using the K-nearest neighbor regression method to predict lateral deviation based on the input feature vector and a pre-constructed training sample set, obtaining a dynamic lateral deviation prediction value, and obtaining a static lateral deviation value. The static lateral deviation value is a fixed deviation determined based on wheel-rail clearance and construction deviation. Based on the dynamic lateral deviation prediction value and the static lateral deviation value, a total lateral deviation prediction value is determined. This method can predict the lateral deviation of the pantograph's carbon sliding plate under different train operating conditions, providing a reliable reference for pantograph selection, installation location optimization, and vehicle clearance design.

[0090] Example 3 Figure 3 This is a schematic diagram of a pantograph carbon sliding plate lateral deviation prediction device provided in Embodiment 3 of the present invention. Figure 3 As shown, the device includes: The input parameter acquisition module 310 is used to acquire the input parameter set of the train driving condition to be predicted, and to preprocess the input parameter set to obtain the input feature vector. The dynamic deviation prediction module 320 is used to predict the lateral deviation based on the input feature vector and the pre-constructed training sample set, using the K-nearest neighbor regression method to obtain the dynamic lateral deviation prediction value; the training sample set is obtained by simulating multiple preset working conditions through a multi-rigid-body dynamics simulation model of wheel-rail and vehicle body. The static deviation acquisition module 330 is used to acquire the static lateral deviation value, which is a fixed deviation determined based on the wheel-rail clearance and construction deviation. The total deviation determination module 340 is used to determine the total lateral deviation prediction value based on the dynamic lateral deviation prediction value and the static lateral deviation value.

[0091] The technical solution of this invention involves obtaining a set of input parameters for the train's operating conditions to be predicted, preprocessing the input parameter set to obtain an input feature vector, using the K-nearest neighbor regression method to predict lateral deviation based on the input feature vector and a pre-constructed training sample set, obtaining a dynamic lateral deviation prediction value, and obtaining a static lateral deviation value. The static lateral deviation value is a fixed deviation determined based on wheel-rail clearance and construction deviation. Based on the dynamic lateral deviation prediction value and the static lateral deviation value, a total lateral deviation prediction value is determined. This method can predict the lateral deviation of the pantograph's carbon sliding plate under different train operating conditions, providing a reliable reference for pantograph selection, installation location optimization, and vehicle clearance design.

[0092] Optional, the dynamic deviation prediction module 320 includes: The neighbor sample filtering unit is used to calculate the spatial distance between the input feature vector and each sample feature vector in the training sample set, and based on the spatial distance, to filter a set number of neighbor sample feature vectors that are close to the input feature vector; the set number is predetermined by cross-validation; each training sample includes a sample feature vector, as well as corresponding lateral positive bias value and lateral negative bias value; The weight calculation unit is used to determine the weight of each neighboring sample feature vector based on the spatial distance between each neighboring sample feature vector and the input feature vector; the smaller the spatial distance between the neighboring sample feature vector and the input feature vector, the greater the corresponding weight. The dynamic deviation prediction unit is used to predict the dynamic positive lateral deviation, dynamic negative lateral deviation, and dynamic deviation range of the train under the predicted operating conditions based on the lateral positive deviation value, lateral negative deviation value, and weight corresponding to the feature vector of each neighboring sample, and to use these as the dynamic lateral deviation prediction value.

[0093] Optional, the total deviation determination module 340 is specifically used for: The dynamic positive lateral deviation prediction value in the dynamic lateral deviation prediction value is superimposed with the static lateral deviation value in the positive direction to obtain the total positive lateral deviation prediction value. The dynamic lateral negative deviation prediction value in the dynamic lateral deviation prediction value is superimposed on the static lateral deviation value in the negative direction to obtain the total lateral negative deviation prediction value. Based on the total positive lateral deviation prediction value and the total negative lateral deviation prediction value, the total deviation range is calculated, and the total positive lateral deviation prediction value, the total negative lateral deviation prediction value, and the total deviation range are used as the total lateral deviation prediction value.

[0094] Optional, the static deviation acquisition module 330 is specifically used for: Determine the wheel-rail clearance based on the preset maximum track gauge and minimum wheelset width; The static lateral deviation value is determined based on the preset construction deviation and the wheel-rail clearance.

[0095] Optionally, the pantograph carbon slider lateral deviation prediction device also includes: The first interface rendering module is used to render the parameter input area and the static deviation input area in the interactive interface; the parameter input area contains a custom input box corresponding to each input parameter in the input parameter set, and the static deviation input area contains a custom input box for the static lateral deviation value; The deviation prediction module is used to respond to the user's click operation on the prediction button, and to perform lateral deviation prediction based on the user-provided input parameter set and the pre-built training sample set, using the K-nearest neighbor regression method to obtain the dynamic lateral deviation prediction value and the total lateral deviation prediction value. The second interface rendering module is used to render the prediction result display area in the interactive interface and display the dynamic lateral deviation prediction value and the total lateral deviation prediction value.

[0096] Optional, weight calculation unit, specifically used for: The inverse of the spatial distance between each neighboring sample feature vector and the input feature vector is used as the initial weight of that neighboring sample feature vector; The initial weights of the predetermined number of neighboring sample feature vectors are summed to obtain the initial weight sum. The ratio of the initial weight of each neighboring sample feature vector to the initial weight sum is calculated and used as the weight of each neighboring sample feature vector.

[0097] Optionally, the pantograph carbon slider lateral deviation prediction device also includes a sample construction module, specifically used for: A multi-rigid-body dynamics simulation model of wheel-rail and vehicle body is established; the multi-rigid-body dynamics simulation model includes multiple vehicle body sections, multiple frames, multiple wheelsets and multiple axle box swing arms, and has multi-degree-of-freedom motion capability; Based on the multi-rigid-body dynamics simulation model, dynamics simulations are performed on various preset working conditions to obtain the lateral positive and negative deviation values ​​of the pantograph carbon slide plate center relative to the track center. The sample feature vector of the input parameter set associated with each preset working condition, along with the corresponding lateral positive and negative deviation values, are stored together as training samples to obtain the training sample set.

[0098] Optionally, the set of input parameters includes vehicle speed, track type, curve radius, superelevation status, lateral wind pressure status, and the pantograph's mounting position on the vehicle body.

[0099] Optionally, the input parameter acquisition module 310 is specifically used for: The input parameters that belong to the category type are converted into numerical form according to a preset ordered encoding rule to obtain the feature values ​​corresponding to the category type parameters; the category type parameters include line type, superelevation status and lateral wind pressure status. The input parameters that are numerical are normalized to obtain the characteristic values ​​corresponding to the numerical parameters; the numerical parameters include vehicle speed, curve radius and pantograph mounting position on the vehicle body. The feature values ​​corresponding to each parameter in the input parameter set are combined in a preset order to form the input feature vector.

[0100] Optionally, the pantograph carbon slider lateral deviation prediction device also includes: The parameter validation module is used to check the validity of parameters entered by the user in the custom input box; The alert message initiation module is used to issue alert messages when abnormal parameters are detected.

[0101] Optionally, the sample construction module is further specifically used for: Before performing dynamic simulations on various preset working conditions based on the multi-rigid-body dynamic simulation model, a track center marker point and a carbon slide center marker point are established in the multi-rigid-body dynamic simulation model so that the center of the carbon slide coincides with the center of the track under static conditions. Using the track center marker as the measurement reference, the lateral displacement of the carbon skateboard center marker during the simulation is directly taken as the lateral deviation value relative to the track center.

[0102] The pantograph carbon strip lateral deviation prediction device provided in this embodiment of the invention can execute the pantograph carbon strip lateral deviation prediction method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method.

[0103] In the technical solution of this invention, the information collected is information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of related data all comply with the relevant laws, regulations and standards of relevant countries and regions, take necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation entry points for users to choose to authorize or refuse.

[0104] Example 4 According to embodiments of the present invention, the present invention also provides an electronic device, a readable storage medium, and a computer program product.

[0105]

[01] Figure 4A schematic diagram of an electronic device 10, which can be used to implement embodiments of the present invention, is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0106]

[02] As Figure 4 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0107]

[03] Multiple components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, optical disk, etc.; and a communication unit 19, such as a network card, modem, wireless transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0108]

[04] The processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The processor 11 performs the various methods and processes described above, such as the pantograph carbon skid plate lateral deviation prediction method.

[0109]

[05] In some embodiments, the pantograph carbon skid plate lateral deviation prediction method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the pantograph carbon skid plate lateral deviation prediction method described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to perform the pantograph carbon skid plate lateral deviation prediction method by any other suitable means (e.g., by means of firmware).

[0110]

[06] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include: implementation in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0111]

[07] The computer program for implementing the method of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer program causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The computer program may be executed entirely on the machine, partially on the machine, as a standalone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0112]

[08] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0113]

[09] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0114]

[10] The systems and technologies described herein can be implemented in computing systems that include back-end components (e.g., as data servers), or computing systems that include middleware components (e.g., application servers), or computing systems that include front-end components (e.g., user computers with a graphical user interface or web browser through which a user can interact with embodiments of the systems and technologies described herein), or any combination of such back-end, middleware, or front-end components. Components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0115]

[11] A computing system may include clients and servers. Clients and servers are generally far apart and usually interact through a communication network. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0116] This application also discloses a computer program product, which includes a computer program that, when executed by a processor, implements the pantograph carbon skid plate lateral deviation prediction method provided in any embodiment of this application. This program product shares the same inventive concept as the pantograph carbon skid plate lateral deviation prediction method disclosed in the embodiments of this application, and therefore will not be described in detail here.

[0117] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0118] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for predicting the lateral deviation of a pantograph carbon sliding plate, characterized in that, include: Obtain the input parameter set of the train's operating conditions to be predicted, and preprocess the input parameter set to obtain the input feature vector; Based on the input feature vector and the pre-constructed training sample set, the K-nearest neighbor regression method is used to predict the lateral deviation and obtain the dynamic lateral deviation prediction value; the training sample set is obtained by simulating multiple preset working conditions through a multi-rigid-body dynamics simulation model of wheel-rail and vehicle body. Obtain the static lateral deviation value, which is a fixed deviation determined based on wheel-rail clearance and construction deviation; The total lateral deviation prediction value is determined based on the dynamic lateral deviation prediction value and the static lateral deviation value.

2. The method according to claim 1, characterized in that, Based on the input feature vector and the pre-constructed training sample set, the K-nearest neighbor regression method is used to predict lateral bias, resulting in a dynamic lateral bias prediction value, including: Calculate the spatial distance between the input feature vector and each sample feature vector in the training sample set, and based on the spatial distance, select a set number of neighboring sample feature vectors that are close to the input feature vector; the set number is predetermined by cross-validation; each training sample includes a sample feature vector, as well as corresponding positive and negative lateral deviation values; The weight of each neighboring sample feature vector is determined based on the spatial distance between each neighboring sample feature vector and the input feature vector; the neighboring sample feature vector with a smaller spatial distance from the input feature vector has a larger weight. Based on the lateral positive deviation value, lateral negative deviation value and weight corresponding to the feature vector of each neighboring sample, the dynamic lateral positive deviation prediction value, dynamic lateral negative deviation prediction value and dynamic deviation range under the train's operating conditions are predicted, and used as the dynamic lateral deviation prediction value.

3. The method according to claim 2, characterized in that, Based on the dynamic lateral deviation prediction value and the static lateral deviation value, the total lateral deviation prediction value is determined, including: The dynamic positive lateral deviation prediction value in the dynamic lateral deviation prediction value is superimposed on the static lateral deviation value in the positive direction to obtain the total positive lateral deviation prediction value. The dynamic lateral negative deviation prediction value in the dynamic lateral deviation prediction value is superimposed on the static lateral deviation value in the negative direction to obtain the total lateral negative deviation prediction value. Based on the total positive lateral deviation prediction value and the total negative lateral deviation prediction value, the total deviation range is calculated, and the total positive lateral deviation prediction value, the total negative lateral deviation prediction value, and the total deviation range are used as the total lateral deviation prediction value.

4. The method according to claim 1, characterized in that, Obtain the static lateral deviation value, including: Determine the wheel-rail clearance based on the preset maximum track gauge and minimum wheelset width; The static lateral deviation value is determined based on the preset construction deviation and the wheel-rail clearance.

5. The method according to claim 1, characterized in that, Also includes: The parameter input area and the static deviation input area are rendered in the interactive interface; the parameter input area contains a custom input box corresponding to each input parameter in the input parameter set, and the static deviation input area contains a custom input box for the static lateral deviation value. In response to the user's click on the prediction button, the system performs lateral bias prediction using the K-nearest neighbor regression method based on the user-provided input parameter set and the pre-built training sample set, to obtain the dynamic lateral bias prediction value and the total lateral bias prediction value. The prediction result display area is rendered in the interactive interface to display the dynamic lateral deviation prediction value and the total lateral deviation prediction value.

6. The method according to claim 2, characterized in that, The weight of each neighboring sample feature vector is determined based on the spatial distance between each neighboring sample feature vector and the input feature vector, including: The inverse of the spatial distance between each neighboring sample feature vector and the input feature vector is used as the initial weight of that neighboring sample feature vector; The initial weights of the predetermined number of neighboring sample feature vectors are summed to obtain the initial weight sum. The ratio of the initial weight of each neighboring sample feature vector to the initial weight sum is calculated and used as the weight of each neighboring sample feature vector.

7. The method according to claim 1, characterized in that, The training sample set is constructed in the following manner: A multi-rigid-body dynamics simulation model of wheel-rail and vehicle body is established; the multi-rigid-body dynamics simulation model includes multiple vehicle body sections, multiple frames, multiple wheelsets and multiple axle box swing arms, and has multi-degree-of-freedom motion capability; Based on the multi-rigid-body dynamics simulation model, dynamics simulations are performed on various preset working conditions to obtain the lateral positive and negative deviation values ​​of the pantograph carbon slide plate center relative to the track center. The sample feature vector of the input parameter set associated with each preset working condition, along with the corresponding lateral positive and negative deviation values, are stored together as training samples to obtain the training sample set.

8. The method according to claim 1, characterized in that, The set of input parameters includes vehicle speed, track type, curve radius, superelevation status, lateral wind pressure status, and the pantograph's mounting position on the vehicle body.

9. The method according to claim 8, characterized in that, The input parameter set is preprocessed to obtain the input feature vector, including: The input parameters that belong to the category type are converted into numerical form according to a preset ordered encoding rule to obtain the feature values ​​corresponding to the category type parameters; the category type parameters include line type, superelevation status and lateral wind pressure status. The input parameters that are numerical are normalized to obtain the characteristic values ​​corresponding to the numerical parameters; the numerical parameters include vehicle speed, curve radius and pantograph mounting position on the vehicle body. The feature values ​​corresponding to each parameter in the input parameter set are combined in a preset order to form the input feature vector.

10. The method according to claim 5, characterized in that, Before responding to a user's click on the prediction button, it also includes: Perform validity checks on the parameters entered by the user in the custom input box; If abnormal parameters are detected, an alert message will be issued.

11. The method according to claim 7, characterized in that, Before performing dynamic simulations on various preset working conditions based on the multi-rigid-body dynamics simulation model, the following steps are also included: In the multi-rigid-body dynamics simulation model, a track center marker and a carbon slide center marker are established so that the center of the carbon slide coincides with the center of the track under static conditions. Using the track center marker as the measurement reference, the lateral displacement of the carbon skateboard center marker during the simulation is directly taken as the lateral deviation value relative to the track center.

12. A device for predicting the lateral deviation of a pantograph carbon sliding plate, characterized in that, include: The input parameter acquisition module is used to acquire the set of input parameters for the train's operating conditions to be predicted, and to preprocess the set of input parameters to obtain the input feature vector. The dynamic deviation prediction module is used to predict the lateral deviation based on the input feature vector and the pre-constructed training sample set, using the K-nearest neighbor regression method to obtain the dynamic lateral deviation prediction value; the training sample set is obtained by simulating multiple preset working conditions through a multi-rigid-body dynamics simulation model of wheel-rail and vehicle body. The static deviation acquisition module is used to acquire the static lateral deviation value, which is a fixed deviation determined based on the wheel-rail clearance and construction deviation. The total deviation determination module is used to determine the total lateral deviation prediction value based on the dynamic lateral deviation prediction value and the static lateral deviation value.

13. An electronic device, characterized in that, The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the pantograph carbon skid plate lateral deviation prediction method according to any one of claims 1-11.

14. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that cause a processor to execute the pantograph carbon skid plate lateral deviation prediction method according to any one of claims 1-11.

15. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the pantograph carbon skid plate lateral deviation prediction method according to any one of claims 1-11.

Images