Motor train unit wheel set turning repair management system

Abstract

The invention provides a motor train unit wheel set turning repair management system, which relates to the field of vehicle operation and maintenance systems and comprises an equipment layer, a data layer, a model layer and an application layer which are connected in sequence, the equipment layer is used for collecting wheel set detection data and uploading the wheel set detection data to the data layer; the model layer calls the train operation mileage data, the train predicted turning repair time, the fault level of the train wheel set and the historical maintenance frequency of the train wheel set in the data layer to determine the turning repair emergency degree value of the train; and the application layer visually displays the train cards of the trains according to the turning repair emergency degree values of the trains from large to small. According to the invention, the application of the wheel set detection data is realized through the model layer, and the quantification of the train maintenance emergency degree is realized based on the wheel set detection data, so that the arrangement sequence of train cards is determined based on the turning repair emergency degree value in the application layer; and operation and maintenance administrative personnel can conveniently obtain a turning repair scheduling plan with higher resource utilization rate and maintenance efficiency based on the arrangement sequence of the train cards.

Description

Technical Field

[0001] This invention relates to the field of vehicle operation and maintenance system technology, and in particular to a high-speed train wheelset turning and maintenance management system. Background Technology

[0002] Currently, in the field of high-speed train maintenance, the method of manually turning wheels according to a predetermined plan faces a series of technical problems. The primary problem lies in the fact that the traditional turning method requires manual measurement of the turning machine before turning, and turning resources are relatively scarce.

[0003] Meanwhile, the lack of digital means and reasonable planning for the full-life management of wheelsets makes it impossible to accurately assess their condition and maintenance needs. Furthermore, the current wheel turning schedule and processes have room for optimization; frequent use of periodic wheel turning not only consumes significant maintenance resources but also results in serious waste of maintenance costs. Summary of the Invention

[0004] This invention provides a trainset wheelset turning and maintenance management system to address the shortcomings of existing technologies that involve significant waste of maintenance resources in periodic wheelset turning, thereby achieving a turning and maintenance management system with higher maintenance efficiency.

[0005] This invention provides a high-speed train wheelset turning and maintenance management system, comprising a device layer, a data layer, a model layer, and an application layer connected in sequence; The device layer is used to collect wheelset detection data and upload the wheelset detection data to the data layer; The model layer calls the train operating mileage data, predicted train maintenance time, train wheelset fault level, and historical maintenance count of the train wheelset in the data layer to determine the train maintenance urgency value. The application layer displays the train cards visually in descending order of the train's repair urgency value.

[0006] According to the present invention, a high-speed train wheelset turning and maintenance management system is provided, wherein the model layer is specifically used for: For each train, the train's operating mileage data, predicted repair time, wheelset fault level, and historical repair frequency of the wheelset are determined separately, and the weighted sum of these scores is used as the repair urgency value of the train in the current repair cycle. The scoring weights of the train's operating mileage data in the current repair cycle, the train's predicted repair time, the train wheelset's fault level, and the train wheelset's historical repair count are obtained based on the output of a pre-built neural network model.

[0007] According to the present invention, a train wheelset maintenance management system is provided, wherein the neural network model is a multilayer perceptron. The multilayer perceptron outputs corresponding scoring weights based on the input train operating mileage data, the train predicted maintenance time, the train wheelset fault level, and the historical maintenance count of the train wheelset. The multilayer perceptron is trained based on the train operating mileage data, the train predicted maintenance time, the train wheelset fault level, and the historical maintenance count of the train wheelset during the historical maintenance cycle.

[0008] According to the present invention, a high-speed train wheelset turning and maintenance management system is provided, wherein the model layer is specifically used for: Calculate the first difference between the train's operating mileage during the current maintenance cycle and the minimum operating mileage corresponding to the train's operating line; Calculate the second difference between the maximum and minimum operating mileage of the train's operating line; The ratio of the first difference to the second difference is used as the score for the train's operating mileage data.

[0009] According to the present invention, a high-speed train wheelset turning and maintenance management system is provided, wherein the model layer is specifically used for: The predicted maintenance time score of the train is determined based on the predicted maintenance time of the train: ,in, T To predict the repair date, T 0 represents the preset time threshold. k The steepness coefficient.

[0010] According to the present invention, a high-speed train wheelset turning and maintenance management system is provided, wherein the model layer is specifically used for: The fault level of the train is determined based on its fault severity. ,in, F The fault level.

[0011] According to the present invention, a high-speed train wheelset turning and maintenance management system is provided, wherein the model layer is specifically used for: The historical maintenance frequency score (H) of a train is determined based on its historical maintenance frequency. scoer =min(H,5)×(1+H / 10), where H is the number of times the train has undergone historical maintenance.

[0012] According to the present invention, a train set wheelset turning and maintenance management system is provided, wherein the equipment layer includes a trackside wheelset online detection system, a trackside running quality detection system, a wheel turning machine without wheel drop, and a vehicle analysis system.

[0013] A method for managing the turning and maintenance of wheelsets on high-speed trains includes: Wheelset detection data is collected at the device layer and uploaded to the data layer. The urgency level of train repair is determined by calling train operating mileage data, predicted train repair time, train wheelset fault level, and historical repair count of train wheelsets from the data layer in the model layer. The application layer visualizes the train cards according to the train's repair urgency value from highest to lowest.

[0014] This invention provides a wheelset turning and maintenance management system for high-speed trains. It collects wheelset inspection data at the equipment layer, stores and manages this data at the data layer, and applies the data at the model layer. Based on the wheelset inspection data, it quantifies the urgency of train maintenance. At the application layer, it determines the order of train cards based on the urgency value, enabling maintenance personnel to obtain a more efficient and resource-efficient turning schedule based on the order of the train cards, thus ensuring the scientific and efficient nature of the turning schedule. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0016] Figure 1 This is an architecture diagram of a high-speed train wheelset turning and maintenance management system provided by the present invention; Figure 2 This is a schematic diagram of the arrangement of train cards in a high-speed train wheelset turning and maintenance management system provided by the present invention; Figure 3 This is a schematic diagram of the contents of a train card in a high-speed train wheelset turning and maintenance management system provided by the present invention; Figure 4 This is a flowchart illustrating a method for managing the turning and repair of wheelsets in high-speed trains, provided by the present invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0018] The following is combined with Figures 1 to 3 This invention introduces a wheelset turning and maintenance management system for high-speed trains, such as... Figure 1As shown, it includes a device layer, a data layer, a model layer, and an application layer connected in sequence; The device layer is used to collect wheelset detection data and upload the wheelset detection data to the data layer; Specifically, such as Figure 1 As shown, the equipment layer in this embodiment specifically includes a trackside wheelset online detection system, a trackside running quality detection system, and a wheel lathe that does not fall off the track, all located beside the track, as well as a vehicle analysis system for monitoring and analyzing train data.

[0019] The above system collects wheelset inspection data for each wheel of the train.

[0020] Optionally, the wheelset inspection data includes the wheelset dimensions, profile curves, tread images, flaw detection maps, and out-of-roundness data, and may also include the wheelset's running mileage and wheel diameter difference.

[0021] The health status of wheelsets is assessed by using wheelset inspection data, so that wheelsets can be repaired based on their health status, thereby ensuring the safe operation of the train set.

[0022] The equipment layer transmits the collected wheelset inspection data to the data layer, which then cleans, stores, and manages the received wheelset inspection data.

[0023] Optionally, the data layer includes a data management and analysis system for recording and managing the dimensional data of train wheelsets in each maintenance cycle, and for analyzing the historical wear trends of each wheelset based on the stored wheelset inspection data.

[0024] The model layer calls the train operating mileage data, predicted train maintenance time, train wheelset fault level, and historical maintenance count of the train wheelset in the data layer to determine the train maintenance urgency value. Optionally, such as Figure 1 As shown, the model layer includes a wheel turning profile suggestion model for generating wheel turning profile suggestions, and a full-line wheel operation statistics model for statistical analysis of wheel operation.

[0025] Based on this, the model layer in this embodiment is also used to call the wheelset detection data stored and managed in the data layer, and to determine the train operating mileage data, the train predicted maintenance time, the mileage wheelset fault level, and the historical maintenance number of the train wheelsets based on the wheelset detection data, thereby determining the maintenance urgency value of each train managed by the system in the current maintenance cycle based on the above-mentioned data.

[0026] Among them, the train operating mileage data represents the operating mileage of each wheelset of the current train. The larger the train operating mileage data, the higher the urgency of the train's wheelsets needing repair.

[0027] Train predicted maintenance time represents the time until the train's next maintenance. It is predicted based on wheelset inspection data. The shorter the predicted maintenance time, the higher the urgency of the train's wheelsets needing maintenance.

[0028] The wheelset fault level standard for trains is based on the wear condition of the train wheelsets, with fault levels predefined. Optionally, a higher fault level indicates more severe wheelset wear, and consequently, a higher wheelset fault level indicates a greater urgency for wheelset repair.

[0029] The historical maintenance count of a train is the total number of maintenance counts of the train wheelsets recorded in the data layer during the historical maintenance cycle. It can be understood that the higher the historical maintenance count, the higher the usage level of the wheelsets, the greater the possibility of hidden damage to the wheelsets, and therefore the higher the urgency of the train's wheelsets needing to be repaired.

[0030] For the reasons mentioned above, optionally, a wheelset maintenance level model can be pre-built in the model layer to determine the urgency value of the train based on the train's operating mileage data, the train's predicted repair time, the fault level of the train wheelset, and the historical number of repairs of the train wheelset, so as to characterize the urgency of the train wheelset requiring repair.

[0031] The application layer displays the train cards visually in descending order of the train's repair urgency value.

[0032] like Figure 1 As shown, the application layer generates wheel repair suggestions and guidance, integrates and displays wheel data, outputs wheel repair schedule suggestions, and performs multi-dimensional statistical analysis and display of wheelset data based on the output of the model layer or by calling the data stored in the data layer.

[0033] Among them, the maintenance schedule suggestion is generated based on the maintenance urgency value of the train output by the model layer and visualized in the form of train cards.

[0034] The page layout of multiple train cards in the wheelset turning and maintenance management system of this invention is as follows: Figure 2 As shown, the train card includes the following contents: Figure 3 As shown. Specifically, the train card includes at least the train number, the warning level of the train wheelset, and the recommended maintenance method.

[0035] Optionally, the alarm levels are divided into 1 to 6, with the urgency level increasing progressively, determined based on wheelset detection data.

[0036] Optionally, the recommended maintenance methods include immediate maintenance and recommended maintenance. Immediate maintenance indicates that the wheelset is worn to a high degree and needs to be repaired as soon as possible, while recommended maintenance indicates that the wheelset is worn but can still operate normally for a period of time.

[0037] In the visualization scheme of this embodiment, such as Figure 2 As shown, different colored circular icons are configured for each alarm level. The redder the color, the higher the alarm level, and the greener the color, the lower the alarm level.

[0038] When different degrees of wear appear in the train's wheelsets, each alarm level is displayed, and the number of wheelsets at that alarm level is marked. Therefore, the train card is generally larger than the card corresponding to trains with no wear or low wear. Figure 2 The first three train cards in the first row are shown in comparison with the rightmost card.

[0039] Therefore, by observing the size of the train card and the number of wheelsets with high alarm levels on the train card, simple wheelset repair prompts can be given to the staff, helping them determine the urgency of wheelset repair for each train.

[0040] It is understandable that multiple trains on the same operating line, within the same maintenance cycle, will have roughly similar overall mileage and wheelset wear due to the shared track. This means that multiple train cards with similar urgency levels may appear in a short period, requiring refining work on the trains corresponding to those cards.

[0041] However, due to limitations in maintenance resources, such as the train's line being under high operating pressure during holidays, tight maintenance time, limited wheel turning machine data at the maintenance site, or limitations in maintenance personnel, it may be impossible to carry out maintenance on multiple trains with similar alarm levels and requiring maintenance during the same maintenance period, such as on the same day.

[0042] Therefore, this embodiment introduces a repair urgency value to quantify the repair urgency of trains. The train cards are arranged in descending order of repair urgency value to construct a train card display scheme. This allows maintenance and administrative personnel who are not front-line maintenance technicians to determine a train repair scheduling scheme with higher accuracy, resource utilization, and maintenance efficiency based on the display order of train cards and actual repair resources, even without relevant direct technical experience.

[0043] This invention collects wheelset inspection data at the equipment layer, stores and manages the data at the data layer, applies the data at the model layer, and quantifies the urgency of train maintenance based on the data. Then, at the application layer, the order of train cards is determined based on the urgency value, allowing maintenance personnel to obtain a more efficient and resource-efficient maintenance schedule based on the order of the train cards, thus ensuring the scientific and efficient nature of the maintenance plan.

[0044] In the EMU wheelset turning and maintenance management system of the present invention, the model layer is specifically used for: For each train, the train's operating mileage data, predicted repair time, wheelset fault level, and historical repair frequency of the wheelset are determined separately, and the weighted sum of these scores is used as the repair urgency value of the train in the current repair cycle. The scoring weights of the train's operating mileage data in the current repair cycle, the train's predicted repair time, the train wheelset's fault level, and the train wheelset's historical repair count are obtained based on the output of a pre-built neural network model.

[0045] For each train managed by the maintenance management system, the model layer calls the train's operating mileage data in the current maintenance cycle stored in the data layer to obtain the train operating mileage data in units of tens of thousands of kilometers.

[0046] Optionally, the model layer calls the wheelset inspection data of the train in the current maintenance cycle stored in the data layer. After preprocessing the called data, different curves are selected to fit the mileage and dimensional parameters to calculate the wear rate per 10,000 kilometers for wheel diameter, flange thickness, and flange height. Then, based on the dimensional parameters reaching the limit standard, the mileage to the next turning limit is predicted.

[0047] Based on this, the distance to the next re-maintenance deadline for each wheelset of the train is calculated. Combined with the train's operating route and timetable, the predicted re-maintenance time for each wheelset is determined on a daily basis. The predicted re-maintenance time for the wheelset with the longest time is then taken as the predicted re-maintenance time for the entire train.

[0048] Optionally, the model vehicle can access the wheelset inspection data of the train in the current maintenance cycle stored in the data layer to determine the fault level of the train wheelsets.

[0049] In this embodiment, the fault levels are divided into slight wear (F=1), moderate wear (F=2), and severe wear or cracks (F=3).

[0050] Optionally, the model vehicle calls the wheelset inspection data of the train in the historical maintenance cycle stored in the data layer to determine the number of historical maintenance operations of the train, which is used as the number of historical maintenance operations for the train in the current cycle, that is, in this calculation.

[0051] Optionally, the maintenance cycle is determined based on the available time of actual maintenance resources; in this embodiment, it is on a daily scale.

[0052] Based on this, the score corresponding to each of the above data categories is determined by the train's operating mileage data in the current cycle, the train's predicted maintenance time, the train wheelset's fault level, and the train wheelset's historical maintenance frequency.

[0053] It is understandable that the data scales of the four types of data mentioned above are different. Therefore, the score corresponding to each type of data can be determined by normalization and other methods in order to reduce the gap between data scales and facilitate the calculation of the urgency value of repair.

[0054] Furthermore, in order to achieve a more accurate calculation of the urgency level of the repair, a neural network model is pre-built. The weights corresponding to the scores of each type of data are obtained based on the neural network model, and then the weighted sum of the scores of each type of data is used as the urgency level of the train's repair.

[0055] In the EMU wheelset maintenance management system of the present invention, the neural network model is a multilayer perceptron. The multilayer perceptron outputs corresponding scoring weights based on the input train operating mileage data, train predicted maintenance time, train wheelset fault level, and train wheelset historical maintenance count. The multilayer perceptron is trained based on the train operating mileage data, train predicted maintenance time, train wheelset fault level, and train wheelset historical maintenance count during the historical maintenance cycle.

[0056] Specifically, in this embodiment, a neural network model for weight prediction is constructed and trained using a multilayer perceptron (MLP).

[0057] In other feasible approaches, a neural network model for weight prediction can be built based on other models, such as self-learning models.

[0058] The multilayer perceptron includes an input layer, a hidden layer, and an output layer. Finally, the output layer outputs the weights of the operating mileage (D1), the predicted overhaul time (D2), the wheelset failure level (D3), and the historical maintenance count (D4).

[0059] The input layer has four nodes, which correspond to train operating mileage data, predicted train maintenance time, train wheelset fault level, and historical maintenance count of train wheelsets, respectively.

[0060] The hidden layers use the ReLU activation function, and the design includes 2-3 hidden layers. The number of nodes in each layer is determined based on the amount and complexity of the data.

[0061] The output layer has four nodes, which output the weights of operating mileage (D1), predicted overhaul time (D2), wheelset failure level (D3), and historical maintenance frequency (D4), respectively.

[0062] Based on this, train operating mileage data, predicted train maintenance time, train wheelset fault levels, and historical maintenance frequency of train wheelsets are collected in advance to construct a dataset for MLP training.

[0063] Simultaneously, a loss function is constructed based on the mean squared error (MSE) for training the MLP: ; In the formula, Predicted Score i The predicted repair urgency value is calculated using a weighted score sum method. The Target Score is a pre-set target value based on experience, and N is the total number of iterations. i Indicates the i The next iteration.

[0064] On the constructed dataset, the model is trained iteratively multiple times, and the loss is repeatedly calculated until the loss converges or the maximum number of iterations is reached. Finally, a trained neural network model is obtained, which is used to predict the weights of each class of data.

[0065] In the EMU wheelset turning and maintenance management system of the present invention, the model layer is specifically used for: Calculate the first difference between the train's operating mileage during the current maintenance cycle and the minimum operating mileage corresponding to the train's operating line; Calculate the second difference between the maximum and minimum operating mileage of the train's operating line; The ratio of the first difference to the second difference is used as the score for the train's operating mileage data.

[0066] Since train operating mileage data is usually large, in order to reduce the impact of large mileage on the calculation when calculating the urgency of repair, the train operating mileage data needs to be mapped to the range of [0,1].

[0067] Specifically, calculate the train's operating mileage during the current maintenance cycle. M Minimum operating mileage corresponding to the train's operating line M min The first difference between them, where, M min The shortest operating mileage for a train on the current operating line.

[0068] Calculate the maximum operating mileage corresponding to the train's operating line. M max With minimum operating mileage M min The second difference between them, where, M max This refers to the longest operating mileage of the train on the current operating line.

[0069] The ratio of the first difference to the second difference is then calculated and used as the score for the train operating mileage data. W : ; By using the above method, the score corresponding to the train's operating mileage data within the current repair cycle can be obtained.

[0070] In the EMU wheelset turning and maintenance management system of the present invention, the model layer is specifically used for: The predicted maintenance time score of the train is determined based on the predicted maintenance time of the train: ,in, T To predict the repair date, T 0 represents the preset time threshold. k The steepness coefficient.

[0071] Optionally, in this embodiment, T 0 is set to 3 days.

[0072] Using the above method, the predicted maintenance time score of the train is calculated based on the predicted maintenance time of the train. T score This makes in T and T The smaller the difference of 0, the higher the urgency of the predicted repair time, resulting in a sharp increase in the score for this item.

[0073] In the EMU wheelset turning and maintenance management system of the present invention, the model layer is specifically used for: The fault level of the train is determined based on its fault severity. ,in, F The fault level.

[0074] By employing the above method, the fault severity score increases non-linearly, thus better reflecting the escalating effect of fault severity.

[0075] In the EMU wheelset turning and maintenance management system of the present invention, the model layer is specifically used for: The historical maintenance frequency score (H) of a train is determined based on its historical maintenance frequency. scoer =min(H,5)×(1+H / 10), where H is the number of times the train has undergone historical maintenance.

[0076] The more times a train has undergone historical maintenance, the faster its score increases.

[0077] Based on this, the Severity Score for turning repairs in this embodiment is: Severity Score = D1 × W +D2× T scoer +D3× F scoer +D4×Hscoer ; This enables intelligent assessment and dynamic prioritization of wheel refinishing status. It deeply integrates multi-source data, including vehicle operating mileage, predicted refinishing time, fault type, and historical refinishing frequency, among other factors. Through nonlinear modeling and multi-factor interactive analysis, it accurately quantifies the urgency of wheel refinishing. Combined with real-time data feedback and an online learning mechanism, the system can adaptively optimize weight allocation, ensuring the scientific and efficient nature of the refinishing plan. Through intelligent scheduling and resource optimization, the system effectively avoids the accumulation of refinished wheels, ensuring normal vehicle operation and efficient line operation, and providing strong technical support for intelligent operation and maintenance of rail transit.

[0078] In one specific implementation, the original repair times for trains numbered 8011, 8013, and 2027 are shown in Table 1 below: Table 1

[0079] In other words, trains numbered 8011, 8013, and 2027 have similar wheelset wear levels because they run on the same lines. Therefore, their original repair times are also similar in date, making it difficult to determine the actual wheelset wear levels of these three trains in order to reasonably schedule the repairs.

[0080] Based on this, in the maintenance management system of the present invention, the order of the train cards of the three trains is determined by calculating the maintenance urgency value of the three trains, so that the staff can intuitively generate a more accurate maintenance schedule based on the order.

[0081] The calculated urgency values ​​for the lathe repair are shown in Table 2 below: Table 2

[0082] Meanwhile, the system updates data every 24 hours, recalculates the urgency value of each train's maintenance, fine-tunes the model using the new data, and dynamically adjusts the weight allocation. Based on the daily determined maintenance urgency value, the system updates the train card ranking daily.

[0083] In addition, it should be noted that the wheelset turning and repair management system in this embodiment is implemented using a web application. Users access the system through a browser. The system architecture includes a front-end interface implemented using HTML, CSS, and JavaScript technologies, which displays wheelset-related information and functional operations through a user-friendly interface. The back-end uses web frameworks such as Spring Boot to handle user operation requests, execute turning and repair scheduling, and calculate turning and repair profile suggestions, among other business logic. The system uses a MySQL database and file system to store wheelset dimensions, profile curves, tread images, flaw detection maps, out-of-roundness data, etc. Data transmission considers encryption and secure transmission, and authentication and access control ensure that users can only access data and functions they are authorized to, preventing unauthorized access and data leakage. The wheelset turning and repair management system achieves cross-platform, distributed access, allowing users to conveniently query and manage wheelset-related information anytime, anywhere, improving work efficiency and management level.

[0084] The wheelset turning and maintenance management system includes the following functions: Data visualization: The system provides various charts and images to display statistical information related to wheelsets, including distribution maps of wheelset fleet vehicle usage status, distribution maps of wheel diameter comprehensive wear, distribution maps of wheel wear rate, historical wear trend maps, distribution maps of wheel radial runout, and distribution maps of wheel flange and wheel diameter turning amount, etc. Through intuitive chart displays, it helps users quickly understand data distribution and trend changes.

[0085] Data filtering allows users to filter data according to their needs, selecting specific time ranges, vehicle numbers, repair status, and other conditions. The interface provides calculation functions for various statistical indicators, including average, standard deviation, maximum, minimum, etc., allowing for the analysis and comparison of different statistical indicators. Data comparison and analysis: The interface supports the comparison and analysis of multiple sets of data. Users can select different sets of data such as different time periods, different vehicle numbers, and different wheels to analyze the correlation and differences between the data. Report generation and export: Users can generate and export reports based on statistical analysis results. Anomaly detection and early warning: The system automatically detects anomalies in the data and provides early warning functions for wheelset size anomalies, wear rate anomalies, and wheel size limit anomalies.

[0086] The system uses a MySQL database and file system to store and manage wheelset-related data. The following is the data content that can be used for EMU vehicle maintenance and its application directions: Wheel fault data and detailed inspection data, including flange thickness, flange height, flaw detection, scratches, wheel diameter difference, radial runout, etc., can be used to guide planning and scheduling, perform data fusion and display, wear rate calculation, radial runout development rate calculation, and multi-dimensional statistical analysis.

[0087] Trackside wheelset dimensional inspection system: Used to collect dimensional data and profile curves of trackside wheelsets, as well as potential fault data. This data helps assess whether the wheelset dimensions meet standards and whether there is deformation or other faults.

[0088] Wheelset tread inspection system: Used to acquire image data of the wheelset tread and identify potential tread faults. This image data allows for the detection of wear, surface defects, and other issues on the wheelset tread, enabling timely fault identification and resolution.

[0089] Wheelset flaw detection system: Used to collect flaw detection data of wheelsets, as well as potential faults. This data can help detect internal defects, cracks, and other problems in the wheelsets.

[0090] Operational quality inspection system: Used to collect out-of-roundness data during vehicle operation. By detecting the out-of-roundness of the wheels, it is possible to assess whether the wheels are out of round, or if there are bearing problems, etc.

[0091] Mileage data can be used for wear rate calculation and multi-dimensional statistical analysis.

[0092] Data from the non-dismounting wheel lathe, including measurement data before and after lathe repair and a list of recommended deformations for lathe repair, needs to be obtained in real time. This data can be used for comprehensive optimization suggestions and multi-dimensional statistical analysis for selecting deformation profiles.

[0093] The present invention also provides a method for managing the wheelset turning of high-speed trains. The wheelset turning management method provided by the present invention can be referred to in correspondence with the wheelset turning management system described above. Therefore, the descriptions and definitions in the various embodiments of the aforementioned bridge maintenance digital twin construction method can be used to understand the various execution modules in the embodiments of the present invention.

[0094] like Figure 4 As shown, the present invention provides a method for managing the turning and repair of wheelsets on high-speed trains, comprising: Step 401: Collect wheelset detection data through the device layer and upload the wheelset detection data to the data layer; Specifically, such as Figure 1 As shown, the equipment layer in this embodiment specifically includes a trackside wheelset online detection system, a trackside running quality detection system, and a wheel lathe that does not fall off the track, all located beside the track, as well as a vehicle analysis system for monitoring and analyzing train data.

[0095] The above system collects wheelset inspection data for each wheel of the train.

[0096] Optionally, the wheelset inspection data includes the wheelset dimensions, profile curves, tread images, flaw detection maps, and out-of-roundness data, and may also include the wheelset's running mileage and wheel diameter difference.

[0097] The health status of wheelsets is assessed by using wheelset inspection data, so that wheelsets can be repaired based on their health status, thereby ensuring the safe operation of the train set.

[0098] The equipment layer transmits the collected wheelset inspection data to the data layer, which then cleans, stores, and manages the received wheelset inspection data.

[0099] Optionally, the data layer includes a data management and analysis system for recording and managing the dimensional data of train wheelsets in each maintenance cycle, and for analyzing the historical wear trends of each wheelset based on the stored wheelset inspection data.

[0100] Step 402: The train repair urgency value is determined by calling the train operating mileage data, train predicted repair time, train wheelset fault level and train wheelset historical repair count from the data layer in the model layer. Optionally, such as Figure 1 As shown, the model layer includes a wheel turning profile suggestion model for generating wheel turning profile suggestions, and a full-line wheel operation statistics model for statistical analysis of wheel operation.

[0101] Based on this, the model layer in this embodiment is also used to call the wheelset detection data stored and managed in the data layer, and to determine the train operating mileage data, the train predicted maintenance time, the mileage wheelset fault level, and the historical maintenance number of the train wheelsets based on the wheelset detection data, thereby determining the maintenance urgency value of each train managed by the system in the current maintenance cycle based on the above-mentioned data.

[0102] Among them, the train operating mileage data represents the operating mileage of each wheelset of the current train. The larger the train operating mileage data, the higher the urgency of the train's wheelsets needing repair.

[0103] Train predicted maintenance time represents the time until the train's next maintenance. It is predicted based on wheelset inspection data. The shorter the predicted maintenance time, the higher the urgency of the train's wheelsets needing maintenance.

[0104] The wheelset fault level standard for trains is based on the wear condition of the train wheelsets, with fault levels predefined. Optionally, a higher fault level indicates more severe wheelset wear, and consequently, a higher wheelset fault level indicates a greater urgency for wheelset repair.

[0105] The historical maintenance count of a train is the total number of maintenance counts of the train wheelsets recorded in the data layer during the historical maintenance cycle. It can be understood that the higher the historical maintenance count, the higher the usage level of the wheelsets, the greater the possibility of hidden damage to the wheelsets, and therefore the higher the urgency of the train's wheelsets needing to be repaired.

[0106] For the reasons mentioned above, optionally, a wheelset maintenance level model can be pre-built in the model layer to determine the urgency value of the train based on the train's operating mileage data, the train's predicted repair time, the fault level of the train wheelset, and the historical number of repairs of the train wheelset, so as to characterize the urgency of the train wheelset requiring repair.

[0107] Step 403: The application layer visualizes the train cards according to the order of the train's repair urgency value from largest to smallest.

[0108] like Figure 1 As shown, the application layer generates wheel repair suggestions and guidance, integrates and displays wheel data, outputs wheel repair schedule suggestions, and performs multi-dimensional statistical analysis and display of wheelset data based on the output of the model layer or by calling the data stored in the data layer.

[0109] Among them, the maintenance schedule suggestion is generated based on the maintenance urgency value of the train output by the model layer and visualized in the form of train cards.

[0110] The page layout of multiple train cards in the wheelset turning and maintenance management system of this invention is as follows: Figure 2 As shown, the train card includes the following contents: Figure 3 As shown. Specifically, the train card includes at least the train number, the warning level of the train wheelset, and the recommended maintenance method.

[0111] Optionally, the alarm levels are divided into 1 to 6, with the urgency level increasing progressively, determined based on wheelset detection data.

[0112] Optionally, the recommended maintenance methods include immediate maintenance and recommended maintenance. Immediate maintenance indicates that the wheelset is worn to a high degree and needs to be repaired as soon as possible, while recommended maintenance indicates that the wheelset is worn but can still operate normally for a period of time.

[0113] In the visualization scheme of this embodiment, such as Figure 2 As shown, different colored circular icons are configured for each alarm level. The redder the color, the higher the alarm level, and the greener the color, the lower the alarm level.

[0114] When different degrees of wear appear in the train's wheelsets, each alarm level is displayed, and the number of wheelsets at that alarm level is marked. Therefore, the train card is generally larger than the card corresponding to trains with no wear or low wear. Figure 2 The first three train cards in the first row are shown in comparison with the rightmost card.

[0115] Therefore, by observing the size of the train card and the number of wheelsets with high alarm levels on the train card, simple wheelset repair prompts can be given to the staff, helping them determine the urgency of wheelset repair for each train.

[0116] It is understandable that multiple trains on the same operating line, within the same maintenance cycle, will have roughly similar overall mileage and wheelset wear due to the shared track. This means that multiple train cards with similar urgency levels may appear in a short period, requiring refining work on the trains corresponding to those cards.

[0117] However, due to limitations in maintenance resources, such as the train's line being under high operating pressure during holidays, tight maintenance time, limited wheel turning machine data at the maintenance site, or limitations in maintenance personnel, it may be impossible to carry out maintenance on multiple trains with similar alarm levels and requiring maintenance during the same maintenance period, such as on the same day.

[0118] Therefore, this embodiment introduces a repair urgency value to quantify the repair urgency of trains. The train cards are arranged in descending order of repair urgency value to construct a train card display scheme. This allows maintenance and administrative personnel who are not front-line maintenance technicians to determine a train repair scheduling scheme with higher accuracy, resource utilization, and maintenance efficiency based on the display order of train cards and actual repair resources, even without relevant direct technical experience.

[0119] This invention collects wheelset inspection data at the equipment layer, stores and manages the data at the data layer, applies the data at the model layer, and quantifies the urgency of train maintenance based on the data. Then, at the application layer, the order of train cards is determined based on the urgency value, allowing maintenance personnel to obtain a more efficient and resource-efficient maintenance schedule based on the order of the train cards, thus ensuring the scientific and efficient nature of the maintenance plan.

[0120] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A wheel set turning and repair management system for EMUs, characterized in that: It includes the device layer, data layer, model layer and application layer connected in sequence; The device layer is used to collect wheelset detection data and upload the wheelset detection data to the data layer; The model layer calls the train operation mileage data, the train predicted turning and repair time, the fault level of the train wheelset and the historical maintenance times of the train wheelset in the data layer to determine the turning and repair urgency value of the train; The application layer visualizes the train cards of the trains in descending order of the repair urgency values ​​of the trains.

2. The EMU wheelset turning and repair management system according to claim 1 is characterized in that: The model layer is specifically used for: For each train, the train operating mileage data, the predicted repair time, the fault level of the train wheelset and the historical repair times of the train wheelset in the current repair cycle are determined respectively, and the weighted sum of the scores is used as the repair urgency value of the train in the current repair cycle; Among them, the scoring weights of the train operating mileage data of the train in the current repair cycle, the train predicted repair time, the fault level of the train wheelset and the historical maintenance times of the train wheelset are obtained based on the output of the pre-built neural network model.

3. The EMU wheelset turning and repair management system according to claim 2 is characterized in that: The neural network model is a multi-layer perceptron, which outputs corresponding scoring weights based on the input train operating mileage data, train predicted turning and repair time, train wheelset fault level and historical maintenance times of the train wheelset, and the multi-layer perceptron is trained based on the train operating mileage data, train predicted turning and repair time, train wheelset fault level and historical maintenance times of the train wheelset in the historical repair cycle.

4. The EMU wheelset turning and repair management system according to claim 2 is characterized in that: The model layer is specifically used for: Calculating a first difference between the operating mileage of the train in the current repair cycle and the minimum operating mileage corresponding to the train running line to which the train belongs; Calculating a second difference between the maximum operating mileage and the minimum operating mileage corresponding to the train running line to which the train belongs; The ratio of the first difference to the second difference is used as the score of the train operating mileage data of the train.

5. The EMU wheelset turning and repair management system according to claim 2 is characterized in that: The model layer is specifically used for: Determine the predicted turning time score of the train based on the predicted turning time of the train: ,in, T To predict the repair date, T 0 is the preset time threshold, k is the steepness coefficient.

6. The EMU wheelset turning and repair management system according to claim 2, characterized in that: The model layer is specifically used for: Determine the fault grade score of the train according to the fault grade of the train: ,in, F The fault level.

7. The EMU wheelset turning and repair management system according to claim 2, characterized in that: The model layer is specifically used for: Determine the historical maintenance score H of the train based on the historical maintenance times of the train scoer =min(H,5)×(1+H / 10), where H is the historical maintenance times of the train.

8. A wheel set turning and repair management system for EMUs according to any one of claims 1 to 7, characterized in that: The equipment layer includes a trackside wheelset online detection system, a trackside running quality detection system, a wheel-less lathe and a vehicle analysis system.

9. A wheel set turning and repair management system for EMUs according to any one of claims 1 to 7, characterized in that: The wheelset detection data includes wheel diameter, wheel rim thickness, wheel diameter difference, out-of-roundness, and mileage of wheel deviation.

10. A method for managing the turning and repair of EMU wheels, characterized in that: include: Collecting wheelset detection data through the equipment layer and uploading the wheelset detection data to the data layer; The model layer calls the train operation mileage data, train predicted repair time, train wheel set fault level and train wheel set historical maintenance times in the data layer to determine the repair urgency value of the train; The train cards of the trains are visualized through the application layer in descending order of the urgency of the repair of the trains.

Images