Train intelligent operation and maintenance system based on state monitoring

Abstract

The application provides a train intelligent operation and maintenance system based on state monitoring, and relates to the technical field of train intelligent operation and maintenance, comprising a basic evaluation module, a trend prediction module, a fusion correction module and an operation and maintenance decision module.The application realizes the integration and correlation analysis of all couplers and all bogies by constructing the space topology relationship of all couplers and all bogies, and combining the single distance influence weight and the longitudinal position importance weight to fuse and correct the prediction basic health degree, can fully fit the component interaction law in the actual operation process of the train, effectively improves the accuracy and rationality of the health state evaluation and trend prediction, and can generate an operation and maintenance priority sequence and a health risk prediction heat map, thereby providing an intuitive and reliable guidance basis for train operation and maintenance work.

Description

Technical Field

[0001] This invention belongs to the field of intelligent train operation and maintenance technology, and specifically relates to an intelligent train operation and maintenance system based on condition monitoring. Background Technology

[0002] Intelligent train operation and maintenance is a core technological means to ensure the safety of rail transit operations, improve operational efficiency, and reduce maintenance costs. Through real-time monitoring, health assessment, trend prediction, and operation and maintenance decision-making of the operating status of key train components, it achieves a shift from scheduled maintenance to predictive maintenance, which is of great significance to the efficient and reliable operation of rail transit. Couplers and bogies, as the most core and critical components of the train's running and coupling system, directly determine the safety and stability of train operation and are key monitoring targets for intelligent train operation and maintenance.

[0003] During actual train operation, the couplers and bogies do not operate in isolation; there are significant interactions and coupling effects between the components. Deterioration of one component can affect surrounding components through force transmission and vibration conduction, thereby altering the overall health of the train. Furthermore, the safety importance of components varies significantly across different locations on the train; components in critical areas such as the front and rear of the train have a far greater impact on operational safety than those in the middle of the carriages.

[0004] However, most existing train health monitoring and maintenance systems only analyze individual couplers or bogies independently, without considering the actual spatial relationships and mutual influences between components, or the differences in importance caused by the longitudinal position of the components in the train. They cannot truly reflect the actual state of the collaborative work of multiple train components, which makes the health assessment results of existing systems inaccurate and the trend prediction bias large. It is difficult to form maintenance decisions that fit the actual working conditions and cannot meet the high-precision, intelligent and collaborative maintenance needs of modern trains. Summary of the Invention

[0005] This invention provides a train intelligent operation and maintenance system based on condition monitoring to solve at least one of the technical problems mentioned above.

[0006] To address the aforementioned technical problems, this invention discloses a train intelligent operation and maintenance system based on condition monitoring, comprising: The basic assessment module is used to collect real-time operating status data of all couplers and bogies of the train, and calculate the basic health status of each coupler and each bogie at each monitoring moment of the current shift based on the operating status data of all couplers and bogies of the train. The trend prediction module is used to construct the health change curves of each coupler and each bogie in the current shift based on the historical basic health of each coupler and each bogie in the current shift, and to predict the predicted basic health of each coupler and each bogie in the next finite number of monitoring times in the current shift. The fusion correction module is used to construct the spatial topology relationship of all couplers and bogies of the train, calculate the weight of the single distance influence of all adjacent components on the target component and the weight of the longitudinal position importance of the target component, and based on the spatial topology relationship, the weight of the single distance influence and the weight of the longitudinal position importance, the fusion correction is performed on the predicted basic health of each coupler and each bogie for the current shift in the future finite number of monitoring times, so as to obtain the final predicted health of each coupler and each bogie for the current shift in the future finite number of monitoring times. The operation and maintenance decision module is used to calculate the difference between the ordinate of the final predicted health status of the last monitoring time in multiple future monitoring times of the current shift and the preset health status threshold line, based on the preset health status threshold line of each coupler and each bogie. Based on the ordinate difference, it generates an operation and maintenance priority sequence and outputs a heat map of the predicted health risks of train components.

[0007] Preferably, the basic evaluation module includes: The coupler status acquisition submodule is used to collect data on the coupler coupling status, coupler lock gap, coupler impact amplitude, coupler air tightness, coupler crush tube stress, and coupler centering angle of all couplers in the train, forming the operating status data of each coupler at each monitoring time of the current shift. The bogie status acquisition submodule is used to collect data on bogie axle temperature, bogie vibration effective value, bogie wheel diameter wear, bogie braking pressure, bogie air spring pressure, and bogie lateral displacement of all bogies in the train, forming the operating status data of each bogie at each monitoring moment of the current shift; The basic health determination submodule calculates the basic health of each coupler and each bogie at each monitoring moment of the current shift, based on the operating status data of all couplers and bogies of the train at each monitoring moment of the current shift.

[0008] Preferably, the basic health determination submodule includes: The map construction unit is used to construct real-time radar maps of each coupler and each bogie at each monitoring time of the current shift, based on the operating status data of all couplers and bogies of the train at each monitoring time of the current shift. The template storage unit is used to store standard radar maps of couplers and standard radar maps of bogies; the standard radar maps of couplers include health maps, sub-health maps, abnormal maps, and fault maps of couplers; the standard radar maps of bogies include health maps, sub-health maps, abnormal maps, and fault maps of bogies. The feature matching unit is used to compare the real-time radar map of each coupler at each monitoring moment of the current shift with the standard radar map of the coupler to obtain the overall matching degree of the coupler; and to compare the real-time radar map of each bogie with the standard radar map of the bogie to obtain the overall matching degree of the bogie. The health rating unit is used to take the health rating of the standard radar map of the coupler with the highest overall matching degree as the basic health rating of the coupler at each monitoring time of the current shift; and to take the health rating of the standard radar map of the bogie with the highest overall matching degree as the basic health rating of each bogie at each monitoring time of the current shift.

[0009] Preferably, the feature matching unit includes: The coupler comprehensive matching degree calculation subunit is used to calculate the similarity of the coupler radar map and the area of ​​the coupler radar map contour deviation based on the real-time radar map of each coupler at each monitoring time of the current shift and the standard radar map of the coupler. The coupler comprehensive matching degree is obtained by weighted fusion based on the similarity of the coupler radar map and the area of ​​the coupler radar map contour deviation. The bogie overall matching degree calculation subunit is used to calculate the similarity of the bogie radar map and the area of ​​the bogie radar map profile deviation based on the real-time radar map of each bogie and the standard radar map of the bogie, and to obtain the overall matching degree of the bogie based on the weighted fusion of the similarity of the bogie radar map and the area of ​​the bogie radar map profile deviation.

[0010] Preferably, the trend prediction module includes: The current shift health curve construction submodule is used to construct the health change curve of each coupler in the current shift based on the historical basic health of each coupler in the current shift; and to construct the health change curve of each bogie in the current shift based on the historical basic health of each bogie in the current shift. The previous shift health curve acquisition submodule is used to acquire the health change curve of each coupler and the health change curve of each bogie in the previous shift. The submodule for obtaining the predicted health recursive parameters is used to obtain the slope and acceleration of the curve at each monitoring moment in the health change curve of each coupler and the health change curve of each bogie in the previous shift, which are respectively used as the coupler health recursive parameters and bogie health recursive parameters at the corresponding monitoring moment in the current shift. The predictive health recursive execution submodule, based on the coupler health recursive parameters and bogie health recursive parameters at each monitoring time of the current shift, recursively obtains the predicted basic health of each coupler and each bogie for a finite number of future monitoring times of the current shift.

[0011] Preferably, the current shift health curve construction submodule includes: The current shift coupler health curve construction unit is used to map the basic health of each coupler in the historical basic health of the current shift to the two-dimensional coordinate system of the current shift coupler health according to the time sequence, and connect the basic health of each coupler at each monitoring time in sequence to form the health change curve of each coupler in the current shift. The current shift bogie health curve construction unit is used to map the basic health of each bogie in the historical basic health of each bogie in the current shift to the two-dimensional coordinate system of bogie health in the current shift according to time sequence, and connect the basic health of each bogie at each monitoring time in sequence to form the health change curve of each bogie in the current shift.

[0012] Preferably, the fusion correction module includes: The spatial topology construction submodule is used to construct a spatial topology diagram with couplers and bogies as nodes, based on the actual installation positions of all couplers and bogies of the train. The predicted value labeling submodule is used to label the predicted basic health of each coupler and the predicted basic health of each bogie for the current shift at a finite number of future monitoring times obtained by the trend prediction module to the corresponding nodes in the spatial topology diagram, forming a spatial topology diagram corresponding to each monitoring time in the finite number of future monitoring times. The neighboring component filtering submodule is used to filter all couplers and bogies that are less than a preset distance threshold and have no other components between them and the target component in the spatial topology diagram corresponding to each monitoring time in a finite number of future monitoring times. The target components include couplers or bogies. The multi-source distance weight calculation submodule is used to calculate the single distance influence weight of each neighboring component on the target component at the corresponding future monitoring time, based on the spatial distance of each neighboring component to the target component and the predicted basic health of the neighboring component itself at the corresponding future monitoring time. The step-by-step fusion correction submodule is used to first perform a first round of fusion correction on the predicted basic health of the target component at each future monitoring time based on the single-item distance influence weight of all neighboring components on the target component at each future monitoring time, and then perform a second round of fusion correction based on the longitudinal position importance weight of the target component to obtain the final predicted health of each coupler and each bogie at each future monitoring time of the current shift.

[0013] Preferably, the step-by-step fusion correction submodule includes: The first round of fusion correction unit is used to weight and fuse the predicted basic health of the target component with the single distance influence weight of all neighboring components at each future monitoring time to obtain the first round of corrected health of the target component at each future monitoring time. The second-round fusion correction unit is used to weight and fuse the first-round corrected health status of the target component with the longitudinal position importance weight of the target component at each future monitoring time to obtain the final predicted health status of each coupler and each bogie at each future monitoring time of the current shift.

[0014] Preferably, the operation and maintenance decision module includes: The threshold acquisition submodule is used to acquire the preset health threshold line corresponding to each coupler and each bogie in the current shift; The difference calculation submodule is used to calculate the difference in the vertical axis between the final predicted health of each coupler and each bogie and the corresponding preset health threshold line at the last monitoring time in the finite number of monitoring times in the current shift. The priority generation submodule is used to generate the maintenance priority sequence of all couplers and all bogies in the current shift based on the sign and absolute value of the difference between the final predicted health of each coupler and each bogie and the corresponding preset health threshold line. The heatmap generation submodule is used to generate and output a heatmap of train component health risk prediction based on the final predicted health status of each coupler and bogie at a finite number of future monitoring times for the current shift.

[0015] Preferably, the priority generation submodule includes: The first-level priority determination unit is used to determine the components whose ordinate difference between the final predicted health level and the corresponding preset health level threshold line is less than the first preset threshold as first-level maintenance priority. The secondary priority determination unit is used to determine the components whose ordinate difference between the final predicted health level and the corresponding preset health level threshold line is between the first preset threshold and zero as secondary maintenance priority. The third-level priority determination unit is used to determine the components whose ordinate difference between the final predicted health level and the corresponding preset health level threshold line is between zero and the second preset threshold as third-level maintenance priority. The fourth-level priority determination unit is used to determine the components whose ordinate difference between the final predicted health level and the corresponding preset health level threshold line is greater than the second preset threshold as fourth-level maintenance priority.

[0016] Compared with the prior art, the present invention has the following beneficial effects: This invention constructs the spatial topological relationship of all couplers and bogies of a train, and combines the influence weight of individual distance and the importance weight of longitudinal position to fuse and correct the predicted basic health status. This achieves a holistic and correlated analysis of all couplers and bogies, which can fully conform to the interaction law of components in the actual operation of the train, effectively improve the accuracy and rationality of health status assessment and trend prediction, and generate maintenance priority sequence and health risk prediction heat map, providing intuitive and reliable guidance for train maintenance and repair work. Attached Figure Description

[0017] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a schematic diagram of a train intelligent operation and maintenance system based on condition monitoring according to the present invention. Detailed Implementation

[0018] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0019] Furthermore, in this invention, the use of terms such as "first" and "second" is for descriptive purposes only and does not specifically refer to any order or sequence, nor is it intended to limit the invention. They are merely used to distinguish components or operations described using the same technical terms and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined with "first" or "second" may explicitly or implicitly include at least one of those features. Additionally, the technical solutions and features of the various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. If a combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed by this invention.

[0020] The present invention provides the following embodiments: Example 1

[0021] This embodiment provides a train intelligent operation and maintenance system based on condition monitoring, such as... Figure 1 As shown, it includes: The basic assessment module is used to collect real-time operating status data of all couplers and bogies of the train, and calculate the basic health status of each coupler and each bogie at each monitoring moment of the current shift based on the operating status data of all couplers and bogies of the train. The trend prediction module is used to construct the health change curves of each coupler and each bogie in the current shift based on the historical basic health of each coupler and each bogie in the current shift, and to predict the predicted basic health of each coupler and each bogie in the next finite number of monitoring times in the current shift. The fusion correction module is used to construct the spatial topology relationship of all couplers and bogies of the train, calculate the weight of the single distance influence of all adjacent components on the target component and the weight of the longitudinal position importance of the target component, and based on the spatial topology relationship, the weight of the single distance influence and the weight of the longitudinal position importance, the fusion correction is performed on the predicted basic health of each coupler and each bogie for the current shift in the future finite number of monitoring times, so as to obtain the final predicted health of each coupler and each bogie for the current shift in the future finite number of monitoring times. The operation and maintenance decision module is used to calculate the difference between the ordinate of the final predicted health status of the last monitoring time in multiple future monitoring times of the current shift and the preset health status threshold line, based on the preset health status threshold line of each coupler and each bogie. Based on the ordinate difference, it generates an operation and maintenance priority sequence and outputs a heat map of the predicted health risks of train components.

[0022] In this embodiment, based on the spatial topology of all couplers and bogies of the train, all couplers and bogies that are less than a preset distance threshold from the target component and have no other components between them are identified as neighboring components of the target component; the target component includes couplers and bogies, and each target component may correspond to one or more neighboring components.

[0023] In this embodiment, the maintenance priority sequence is used to guide maintenance personnel to carry out inspection and maintenance work in order of priority, and the health risk prediction heat map is used to intuitively display the future risk distribution of all vehicle components and quickly locate high-risk components.

[0024] This embodiment constructs the spatial topological relationship of all couplers and bogies of the train, and combines the influence weight of individual distance and the importance weight of longitudinal position to fuse and correct the predicted basic health status. This achieves a holistic and correlated analysis of all couplers and bogies, which can fully conform to the interaction law of components in the actual operation of the train, effectively improve the accuracy and rationality of health status assessment and trend prediction, and generate maintenance priority sequence and health risk prediction heat map, providing intuitive and reliable guidance for train maintenance and repair work. Example 2

[0025] Based on Example 1, the basic evaluation module includes: The coupler status acquisition submodule is used to collect data on the coupler coupling status, coupler lock gap, coupler impact amplitude, coupler air tightness, coupler crush tube stress, and coupler centering angle of all couplers in the train, forming the operating status data of each coupler at each monitoring time of the current shift. The bogie status acquisition submodule is used to collect data on bogie axle temperature, bogie vibration effective value, bogie wheel diameter wear, bogie braking pressure, bogie air spring pressure, and bogie lateral displacement of all bogies in the train, forming the operating status data of each bogie at each monitoring moment of the current shift; The basic health determination submodule calculates the basic health of each coupler and each bogie at each monitoring moment of the current shift, based on the operating status data of all couplers and bogies of the train at each monitoring moment of the current shift.

[0026] In this embodiment, the coupling status of the car coupler is detected by a position sensor located at the position of the car coupler lock; The coupler-lock gap is detected by a displacement sensor installed on the coupler meshing surface; The impact amplitude of the coupler is detected by an acceleration sensor installed on the coupler body; The airtightness of the coupler is detected by a pressure sensor installed at the air duct joint of the coupler. The stress in the coupler crush tube is obtained by measuring strain gauges placed on the surface of the crush tube. The coupler alignment angle is detected by an angle sensor installed at the coupler alignment device. The bogie axle temperature is detected by a temperature sensor located at the axle box. The effective value of bogie vibration is obtained by vibration sensors installed at the frame location; Bogie wheel diameter wear is detected by displacement sensors; The bogie braking pressure is detected by a pressure sensor installed in the braking unit; The bogie air spring pressure is detected by a pressure sensor located at the air spring interface; Lateral displacement of the bogie is detected by displacement sensors installed between the frame and the wheelset.

[0027] The beneficial effects of the above technical solution are as follows: This embodiment collects six operational status data of the coupler and six operational status data of the bogie through the coupler status acquisition submodule and the bogie status acquisition submodule, respectively. It can comprehensively and completely reflect the real operational status of each coupler and each bogie at each monitoring moment of the current shift. At the same time, it clarifies the sensor installation location and detection method corresponding to each status data, ensuring that the data source is stable and reliable, and providing real and standardized data support for subsequent basic health calculation, radar map construction and trend prediction. Example 3

[0028] Based on Example 2, the basic health determination submodule includes: The map construction unit is used to construct real-time radar maps of each coupler and each bogie at each monitoring time of the current shift, based on the operating status data of all couplers and bogies of the train at each monitoring time of the current shift. The template storage unit is used to store standard radar maps of couplers and standard radar maps of bogies; the standard radar maps of couplers include health maps, sub-health maps, abnormal maps, and fault maps of couplers; the standard radar maps of bogies include health maps, sub-health maps, abnormal maps, and fault maps of bogies. The feature matching unit is used to compare the real-time radar map of each coupler at each monitoring moment of the current shift with the standard radar map of the coupler to obtain the overall matching degree of the coupler; and to compare the real-time radar map of each bogie with the standard radar map of the bogie to obtain the overall matching degree of the bogie. The health rating unit is used to take the health rating of the standard radar map of the coupler with the highest overall matching degree as the basic health rating of the coupler at each monitoring time of the current shift; and to take the health rating of the standard radar map of the bogie with the highest overall matching degree as the basic health rating of each bogie at each monitoring time of the current shift.

[0029] In this embodiment, the operating status data corresponding to the i-th coupler at the j-th monitoring time of the current shift are represented as: a six-dimensional vector. ; in: : The coupling status of the i-th coupler at the j-th monitoring time (normalized value, dimensionless, interval) ) : The coupler-lock gap at the j-th monitoring time (normalized value, dimensionless, interval) ); : Impact amplitude of the i-th coupler at the j-th monitoring time (normalized value, dimensionless, interval) ); Air tightness of the coupler at the j-th monitoring time (normalized value, dimensionless, interval) ); : Stress of the i-th coupler crush tube at the j-th monitoring time (normalized value, dimensionless, interval) ); : The centering angle of the coupler at the j-th monitoring time (normalized value, dimensionless, interval) ); The method for plotting the radar image of the i-th coupler at the j-th monitoring time is as follows: a six-dimensional vector The six components are mapped to six axes (the included angles of each axis) based on the coupler coupling status, coupler-lock gap, coupler impact amplitude, coupler air tightness, coupler crush tube stress, and coupler centering angle. axial length range On the radar image, six coordinate points are obtained; the six points are connected by straight lines in a preset dimensional order to form a closed irregular hexagon, which is the real-time radar image of the i-th coupler at the j-th monitoring time.

[0030] In this embodiment, the operating status data corresponding to the i-th bogie at the j-th monitoring time of the current shift are represented as: a six-dimensional vector. ; in: : The bogie axle temperature at the j-th monitoring time (normalized value, dimensionless, interval) ); : The effective value of vibration of the i-th bogie at the j-th monitoring time (normalized value, dimensionless, interval) ); : Wear of the i-th bogie wheel diameter at the j-th monitoring time (normalized value, dimensionless, interval) ); Braking pressure of bogie at monitoring time j (normalized value, dimensionless, interval) ); : Air spring pressure of bogie at monitoring time j (normalized value, dimensionless, interval) ); : Lateral displacement of the i-th bogie at the j-th monitoring time (normalized value, dimensionless, interval) ); The method for plotting the radar image of the i-th bogie at the j-th monitoring time is as follows: a six-dimensional vector The six components are mapped to six axes (including the bogie axle temperature, bogie vibration RMS value, bogie wheel diameter wear, bogie brake pressure, bogie air spring pressure, and bogie lateral displacement) as the axes (angle of each axis). axial length range On the radar image, six coordinate points are obtained; the six points are connected by straight lines in a preset dimensional order to form a closed irregular hexagon, which is the real-time radar image of the i-th bogie at the j-th monitoring time.

[0031] In this embodiment, the standard radar map of the coupler is a six-dimensional normalized vector and its radar map corresponding to the operating status of each coupler with a known health level in history. Each map in the library is accompanied by a health level given by expert evaluation or maintenance records, with a value range of [value range missing]. , For the best coupler health, The health status is the worst, and the coupler is completely ineffective; Coupler Health Map: Indicates the health status A subset of historical maps; Car coupler sub-health map: refers to the level of health in A subset of historical maps; Coupler anomaly graph: refers to the health status at... A subset of historical maps; Coupler Fault Diagram: Indicates Health Status A subset of historical maps; Acquisition method: Collect and normalize the six-dimensional operational data of each coupler at a specific monitoring moment throughout history. Simultaneously, experts or regular inspection records will calibrate continuous health values ​​to form... Entries, among which, These represent the six-dimensional normalized vector corresponding to the m-th coupler standard radar map and the health status of the m-th coupler standard radar map, respectively. All entries are summarized to form the coupler standard radar map library.

[0032] The standard radar map of a bogie is a six-dimensional normalized vector and its radar map corresponding to the operating status of each bogie with a known health level throughout history. Each map in the library is accompanied by a health level given by experts or maintenance records, with a value range of [value range missing]. , For the best bogie health, The bogie is in the worst condition and has completely failed. Bogie Health Chart: Indicates health status A subset of historical maps; Bogie Sub-health Map: Indicates the level of health of the bogie A subset of historical maps; Bogie anomaly graph: refers to the health status at... A subset of historical maps; Bogie Fault Diagram: Indicating Health Status A subset of historical maps; Acquisition method: Collect and normalize the six-dimensional operational data of each bogie at a specific monitoring moment throughout history. Simultaneously, experts or routine inspection records are used to calibrate continuous health values, forming... Entries, among which, These represent the six-dimensional normalized vector corresponding to the m-th bogie standard radar map and the health status of the m-th bogie standard radar map, respectively. All of them are summarized to form the bogie standard radar map library.

[0033] The beneficial effects of the above technical solution are as follows: This embodiment, through the cooperation of the map construction unit, template storage unit, feature matching unit and health rating unit, transforms multi-dimensional operating status data into real-time radar maps and compares them with standard radar maps. This can objectively and stably determine the basic health of each coupler and each bogie at each monitoring moment of the current shift, avoiding errors caused by relying on a single indicator or human experience for judgment, and enabling the basic health to more realistically reflect the actual performance status of the components. Example 4

[0034] Based on Example 3, the feature matching unit includes: The coupler comprehensive matching degree calculation subunit is used to calculate the similarity of the coupler radar map and the area of ​​the coupler radar map contour deviation based on the real-time radar map of each coupler at each monitoring time of the current shift and the standard radar map of the coupler. The coupler comprehensive matching degree is obtained by weighted fusion based on the similarity of the coupler radar map and the area of ​​the coupler radar map contour deviation. The bogie overall matching degree calculation subunit is used to calculate the similarity of the bogie radar map and the area of ​​the bogie radar map profile deviation based on the real-time radar map of each bogie and the standard radar map of the bogie, and to obtain the overall matching degree of the bogie based on the weighted fusion of the similarity of the bogie radar map and the area of ​​the bogie radar map profile deviation.

[0035] In this embodiment, the radar image of the i-th coupler at the j-th monitoring time of the current shift is compared with the m-th standard radar image of the coupler, and the similarity of the coupler radar images is calculated using the following formula: In the formula: The similarity between the i-th coupler and the m-th coupler standard map at the j-th monitoring time; For the j-th cycle, the i-th coupler... Dimensional normalized value; For the m-th standard drawing of the coupler Dimensional normalized value; The radar image of the i-th coupler at the j-th monitoring time of the current shift is compared with the m-th standard radar image of the coupler. The area of ​​the profile deviation of the coupler radar image is calculated using the following formula: In the formula: Let be the area of ​​the contour deviation between the i-th coupler and the m-th coupler standard map at the j-th monitoring time, which is the absolute value of the difference between the areas enclosed by the contours of the two radar maps. For the j-th cycle, the i-th coupler... Dimensional normalized value; For the m-th standard drawing of the coupler Dimensional normalized value; , ; The overall matching degree of the coupler is obtained by weighted fusion of the similarity between the radar image of the i-th coupler at the j-th monitoring time and the area of ​​the contour deviation of the radar image of the coupler. The calculation formula is as follows: In the formula: The comprehensive matching degree between the i-th coupler and the m-th coupler standard map at the j-th monitoring time is given. Similarity weight coefficient (default) ); This represents the area of ​​the maximum possible deviation in the coupler pattern (normalization constant).

[0036] In this embodiment, the radar image of the i-th bogie at the j-th monitoring time of the current shift is compared with the m-th standard radar image of the bogie, and the similarity of the bogie radar images is calculated using the following formula: In the formula: Let represent the similarity between the i-th bogie and the m-th bogie standard map at the j-th monitoring time. For the i-th bogie in the j-th cycle Dimensional normalized value; For the m-th bogie standard drawing, Dimensional normalized value; In this embodiment, the radar image of the i-th bogie at the j-th monitoring time of the current shift is compared with the m-th standard radar image of the bogie, and the area of ​​the profile deviation of the bogie radar image is calculated. The calculation formula is as follows: In the formula: Let be the area of ​​the contour deviation between the i-th bogie and the m-th standard map of the j-th bogie, that is, the absolute value of the difference between the areas enclosed by the contours of the two radar maps. For the i-th bogie in the j-th cycle Dimensional normalized value; For the m-th bogie standard drawing, Dimensional normalized value; , ; The overall bogie matching degree is obtained by weighted fusion of the similarity between the radar image of the i-th bogie at the j-th monitoring time and the area of ​​the contour deviation of the radar image of the bogie. The calculation formula is as follows: In the formula: Let be the overall matching degree between the i-th bogie and the m-th bogie standard map at the j-th monitoring time. Similarity weight coefficient (default) ); This represents the area of ​​the maximum possible deviation in the bogie pattern.

[0037] The beneficial effects of the above technical solution are as follows: This embodiment calculates the similarity of the radar spectrum of the coupler and the bogie and the area of ​​the radar spectrum contour deviation, and then weights and fuses the two to obtain the comprehensive matching degree. It can comprehensively reflect the difference between the real-time state and the standard state from both the dimensional value and the overall contour. All calculation processes use normalized dimensionless values ​​to avoid the influence of differences in the magnitude of different parameters, effectively improve the reliability and distinguishability of the comprehensive matching degree, and thus improve the accuracy of the basic health assessment. Example 5

[0038] Based on Example 1, the trend prediction module includes: The current shift health curve construction submodule is used to construct the health change curve of each coupler in the current shift based on the historical basic health of each coupler in the current shift; and to construct the health change curve of each bogie in the current shift based on the historical basic health of each bogie in the current shift. The previous shift health curve acquisition submodule is used to acquire the health change curve of each coupler and the health change curve of each bogie in the previous shift. The submodule for obtaining the predicted health recursive parameters is used to obtain the slope and acceleration of the curve at each monitoring moment in the health change curve of each coupler and the health change curve of each bogie in the previous shift, which are respectively used as the coupler health recursive parameters and bogie health recursive parameters at the corresponding monitoring moment in the current shift. The predictive health recursive execution submodule, based on the coupler health recursive parameters and bogie health recursive parameters at each monitoring time of the current shift, recursively obtains the predicted basic health of each coupler and each bogie for a finite number of future monitoring times of the current shift.

[0039] In this embodiment, a shift refers to a complete route cycle from the start of a train's operation out of the depot to its return to the depot. Within a shift, a fixed sequence of monitoring times is set at equal time intervals, with the sequence number starting from... to (For example, each) (Once every minute). Different shifts have highly similar operating routes, stop plans, and load patterns, even with the same serial number. The operating sections, speed conditions, and stress environments corresponding to the monitoring times are highly similar. Therefore, the degradation trend (slope, acceleration) of the health status change curve of the previous shift at that time can reliably serve as a recursive parameter for the degradation trend of the current shift at the same stage. In other words, the degradation trend of the current shift at the [missing information] stage... The monitoring time is the same as the previous shift. Although the monitoring moments are different in absolute time, they are in the same relative position in the mission profile, forming a one-to-one correspondence.

[0040] In this embodiment, the baseline health of each coupler at each monitoring moment in the historical baseline health of the previous shift is mapped sequentially to the two-dimensional coordinate system of the coupler health of the previous shift, and the baseline health of the coupler at each monitoring moment is connected sequentially to form the health change curve of each coupler in the previous shift; the previous shift's... The health status change curve of each coupler is represented as follows: ; The baseline health of each bogie at each monitoring moment in the historical baseline health data of the previous shift is mapped sequentially to the two-dimensional coordinate system of bogie health data of the previous shift, and then the baseline health data of each bogie at each monitoring moment are connected sequentially to form the health change curve of each bogie in the previous shift; the previous shift's... The health change curve of each bogie is represented as follows: .

[0041] In this embodiment, let the current monitoring time of the current shift be the j-th monitoring time, and the future finite number of monitoring times be the next 3 monitoring times, namely the (j+1)-th, (j+2)-th, and (j+3)-th monitoring times of the current shift, then: The recursive parameters for the coupler health status at the (j+1)th monitoring time of the current shift include the slope and acceleration of the coupler health status change curve at the jth monitoring time of the previous shift. The slope and acceleration of the coupler health status change curve at the jth monitoring time of the previous shift are expressed as follows: and ; The recursive parameters for the coupler health status at the (j+2)th monitoring time of the current shift include the slope and acceleration of the coupler health status change curve at the (j+1)th monitoring time of the previous shift. The slope and acceleration of the coupler health status change curve at the (j+1)th monitoring time of the previous shift are expressed as follows: and ; The recursive parameters for the coupler health status at the (j+3)th monitoring time of the current shift include the slope and acceleration of the coupler health status change curve at the (j+2)th monitoring time of the previous shift. The slope and acceleration of the coupler health status change curve at the (j+2)th monitoring time of the previous shift are expressed as follows: and ; The recursive parameters for bogie health at the (j+1)th monitoring time of the current shift include the slope and acceleration of the bogie health change curve at the jth monitoring time of the previous shift. The slope and acceleration of the bogie health change curve at the jth monitoring time of the previous shift are expressed as follows: and ; The recursive parameters for bogie health at the (j+2)th monitoring time of the current shift include the slope and acceleration of the bogie health change curve at the (j+1)th monitoring time of the previous shift. The slope and acceleration of the bogie health change curve at the (j+1)th monitoring time of the previous shift are expressed as follows: and ; The recursive parameters for bogie health at the (j+3)th monitoring time of the current shift include the slope and acceleration of the bogie health change curve at the (j+2)th monitoring time of the previous shift. The slope and acceleration of the bogie health change curve at the (j+2)th monitoring time of the previous shift are expressed as follows: and ; Based on the baseline health status of the i-th coupler at the j-th monitoring time of the current shift. The slope of the health status change curve of the coupler at the j-th monitoring time compared to the previous shift. and curvilinear acceleration Calculate the predicted baseline health of the i-th coupler at the (j+1)-th monitoring time of the current shift. The calculation formula is: ; The predicted basic health of the coupler is based on the (j+1)th monitoring time of the current shift for the i-th coupler. The slope of the health status change curve of the coupler at the (j+1)th monitoring time compared to the previous shift and curvilinear acceleration Calculate the predicted baseline health of the i-th coupler at the (j+2)-th monitoring time of the current shift. The calculation formula is: ; The predicted basic health of the coupler is based on the coupler of the i-th coupler at the (j+2)-th monitoring time of the current shift. The slope of the health status change curve of the coupler at the (j+2)th monitoring time compared to the previous shift. and curvilinear acceleration The formula for calculating the predicted baseline health of the i-th coupler at the (j+3)-th monitoring time of the current shift is: ; Based on the baseline health status of the i-th bogie at the j-th monitoring time of the current shift. The slope of the health change curve of the bogie at the j-th monitoring time compared to the previous shift. and curvilinear acceleration Calculate the predicted baseline health of the i-th bogie at the (j+1)-th monitoring time of the current shift. The calculation formula is as follows: ; Based on the predicted baseline health of the i-th bogie at the (j+1)-th monitoring time of the current shift. The slope of the bogie health change curve at the (j+1)th monitoring time point compared to the previous shift. and curvilinear acceleration The formula for calculating the predicted basic health of the i-th bogie at the (j+2)-th monitoring time of the current shift is: ; Based on the predicted baseline health of the i-th bogie at the (j+2)-th monitoring time of the current shift. The slope of the bogie health change curve at the (j+2)th monitoring time point compared to the previous shift. and curvilinear acceleration The formula for calculating the predicted baseline health of the i-th bogie at the (j+3)-th monitoring time of the current shift is as follows: ; in, This represents the time interval between adjacent monitoring moments.

[0042] The beneficial effects of the above technical solution are as follows: This embodiment uses the slope and acceleration of the health status change curve of the previous shift as the recursive parameters of the current shift, and recursively obtains the predicted basic health status of the current shift for a limited number of future monitoring times. It can make full use of the similarity of the operating conditions of different train shifts, so that the trend prediction is more in line with the actual health status degradation law of the components. At the same time, by combining the slope and acceleration to jointly characterize the change trend, it can more accurately reflect the dynamic change characteristics of health status and reduce the deviation between the prediction results and the actual state. Example 6

[0043] Based on Example 5, the current shift health curve construction submodule includes: The current shift coupler health curve construction unit is used to map the basic health of each coupler in the historical basic health of the current shift to the two-dimensional coordinate system of the current shift coupler health according to the time sequence, and connect the basic health of each coupler at each monitoring time in sequence to form the health change curve of each coupler in the current shift. The current shift bogie health curve construction unit is used to map the basic health of each bogie in the historical basic health of each bogie in the current shift to the two-dimensional coordinate system of bogie health in the current shift according to time sequence, and connect the basic health of each bogie at each monitoring time in sequence to form the health change curve of each bogie in the current shift.

[0044] In this embodiment, the two-dimensional coordinate system for the health of the coupler in the current shift is a coordinate system with the monitoring time sequence number as the horizontal axis and the basic health of the coupler as the vertical axis.

[0045] The current shift's bogie health status two-dimensional coordinate system is a coordinate system with the monitoring time sequence number as the horizontal axis and the bogie foundation health status as the vertical axis.

[0046] The beneficial effects of the above technical solution are as follows: This embodiment constructs a two-dimensional health coordinate system with the monitoring time sequence number as the horizontal axis and the basic health level as the vertical axis, transforming discrete basic health level data into a continuous health level change curve. This can fully present the health status change process of each coupler and each bogie within the current shift, providing an accurate and intuitive data basis for the extraction of slope and acceleration. The unified construction rules also enable a unified comparison benchmark for the change trends between different components. Example 7

[0047] Based on Example 1, the fusion correction module includes: The spatial topology construction submodule is used to construct a spatial topology diagram with couplers and bogies as nodes, based on the actual installation positions of all couplers and bogies of the train. The predicted value labeling submodule is used to label the predicted basic health of each coupler and the predicted basic health of each bogie for the current shift at a finite number of future monitoring times obtained by the trend prediction module to the corresponding nodes in the spatial topology diagram, forming a spatial topology diagram corresponding to each monitoring time in the finite number of future monitoring times. The neighboring component filtering submodule is used to filter all couplers and bogies that are less than a preset distance threshold and have no other components between them and the target component in the spatial topology diagram corresponding to each monitoring time in a finite number of future monitoring times. The target components include couplers or bogies. The multi-source distance weight calculation submodule is used to calculate the single distance influence weight of each neighboring component on the target component at the corresponding future monitoring time, based on the spatial distance of each neighboring component to the target component and the predicted basic health of the neighboring component itself at the corresponding future monitoring time. The step-by-step fusion correction submodule is used to first perform a first round of fusion correction on the predicted basic health of the target component at each future monitoring time based on the single-item distance influence weight of all neighboring components on the target component at each future monitoring time, and then perform a second round of fusion correction based on the longitudinal position importance weight of the target component to obtain the final predicted health of each coupler and each bogie at each future monitoring time of the current shift.

[0048] In this embodiment, the spatial topology diagram uses all the couplers and bogies of the train as nodes, and the node positions are consistent with the actual installation positions of the components, in order to reflect the spatial positional relationships between the components.

[0049] In this embodiment, the predicted value labeling submodule labels the predicted basic health of each coupler and each bogie at a finite number of future monitoring times for the current shift on the corresponding nodes of the spatial topology graph, so that each future monitoring time corresponds to an independent spatial topology graph with health labeling.

[0050] In this embodiment, the multi-source distance weight calculation submodule calculates for each future monitoring time separately. For each neighboring component of the target component, the single-item distance influence weight of the neighboring component on the target component is obtained based on the spatial distance between the neighboring component and the target component and the predicted basic health of the neighboring component at the corresponding monitoring time. Let the current monitoring time of the current shift be the j-th monitoring time, and take the i-th coupler at the 3rd future monitoring time, i.e. the j+3rd monitoring time, as the target component. Let its adjacent components include the i-th bogie and the i+1-th bogie. Weight of the single-distance influence of the i-th bogie on the target component, i.e., the i-th coupler, at the (j+3)-th monitoring time. The calculation formula is: Among them, the The first coupler and the first The longitudinal spatial distance between the bogies is , No. One bogie in The predicted baseline health level at any given time is The distance attenuation parameter is The unit is meters; Weight of the single-distance influence of the bogie on the target component, i.e., the i-th coupler, at the (j+3)-th monitoring time. The calculation formula is: Among them, the The first coupler and the first The longitudinal spatial distance between the bogies is , No. One bogie in The predicted baseline health level at any given time is .

[0051] The beneficial effects of the above technical solution are as follows: This embodiment constructs a spatial topology diagram with couplers and bogies as nodes, and marks the predicted basic health of a limited number of future monitoring times on the corresponding nodes. At the same time, it selects neighboring components that meet the distance requirements and have no component gaps. Combining the spatial distance with the predicted basic health of the neighboring components themselves, it calculates the single-item distance influence weight. This can truly reflect the spatial coupling influence relationship between components in the health analysis process, making the subsequent fusion correction more in line with the actual operation mechanism of the train. Example 8

[0052] Based on Example 7, the step-by-step fusion correction submodule includes: The first round of fusion correction unit is used to weight and fuse the predicted basic health of the target component with the single distance influence weight of all neighboring components at each future monitoring time to obtain the first round of corrected health of the target component at each future monitoring time. The second-round fusion correction unit is used to weight and fuse the first-round corrected health status of the target component with the longitudinal position importance weight of the target component at each future monitoring time to obtain the final predicted health status of each coupler and each bogie at each future monitoring time of the current shift.

[0053] In this embodiment, the first one is still used. The monitoring time of the first monitoring moment The first coupler is the target component, and its adjacent component is the second coupler. The bogie, the first For each bogie, the original predicted baseline health of the target component is... The first round of fusion correction uses a distance-weighted average, calculated using the following formula: In the formula For the first A car coupler in The first round of health adjustment at any time; in the denominator The base weights corresponding to the target component itself; The second round of fusion correction introduces the first The importance weight of the longitudinal position of each coupler (Target components in the front or rear section of the vehicle) Target components in the central section The first round of revised health score is weighted and fused with the longitudinal position importance weights to obtain the final predicted health score. : That is, when the health of neighboring components decreases or the weight of positional importance increases, the final predicted health will decrease monotonically.

[0054] The beneficial effects of the above technical solution are as follows: This embodiment first performs a first round of fusion correction based on the single-item distance influence weight of all adjacent components, and then performs a second round of fusion correction based on the longitudinal position importance weight. This can sequentially reflect the spatial coupling influence between components and the importance difference of the longitudinal installation position, so that the final predicted health reflects the three aspects of its own prediction trend, the influence of adjacent components and the position importance. Compared with the single calculation method, it is more in line with the actual operation and maintenance risk judgment logic. Example 9

[0055] Based on Example 1, the operation and maintenance decision module includes: The threshold acquisition submodule is used to acquire the preset health threshold line corresponding to each coupler and each bogie in the current shift; The difference calculation submodule is used to calculate the difference in the vertical axis between the final predicted health of each coupler and each bogie and the corresponding preset health threshold line at the last monitoring time in the finite number of monitoring times in the current shift. The priority generation submodule is used to generate the maintenance priority sequence of all couplers and all bogies in the current shift based on the sign and absolute value of the difference between the final predicted health of each coupler and each bogie and the corresponding preset health threshold line. The heatmap generation submodule is used to generate and output a heatmap of train component health risk prediction based on the final predicted health status of each coupler and bogie at a finite number of future monitoring times for the current shift.

[0056] In this embodiment, each coupler and each bogie is configured with an independent preset health threshold line based on its own component type and installation location. The preset health threshold lines for different components are independent of each other and are not used interchangeably. The difference on the ordinate is calculated by subtracting the corresponding preset health threshold line from the final predicted health level. A negative difference indicates that the final predicted health level is lower than the threshold line, and the component has operational risks; a positive difference indicates that the final predicted health level is higher than the threshold line, and the component is relatively reliable.

[0057] In this embodiment, the maintenance priority sequence is sorted according to the risk level of the difference in the vertical axis. The more the final predicted health level is lower than the preset health level threshold, the higher the maintenance priority. Components whose final predicted health level is higher than the preset health level threshold but close to the threshold have the next highest maintenance priority. Components whose status is far above the preset health level threshold have the lowest maintenance priority.

[0058] In this embodiment, the health risk prediction heatmap uses color grading to represent the risk level. The lower the final predicted health level, the closer the color is to red, indicating a higher health risk; the higher the final predicted health level, the closer the color is to green, indicating a lower health risk. This can intuitively present the future health risk distribution of key components of the entire vehicle.

[0059] The beneficial effects of the above technical solution are as follows: This embodiment can objectively quantify the future risk level of each coupler and each bogie by calculating the difference between the final predicted health level and the preset health level threshold line of the last time in the finite number of monitoring times in the current shift, and generate a maintenance priority sequence accordingly. At the same time, combined with the health risk prediction heat map, it can intuitively display the future risk distribution of all vehicle components, which makes it easier for maintenance personnel to quickly identify high-risk components and reasonably arrange the maintenance sequence. Example 10

[0060] Based on Example 9, the priority generation submodule includes: The first-level priority determination unit is used to determine the components whose ordinate difference between the final predicted health level and the corresponding preset health level threshold line is less than the first preset threshold as first-level maintenance priority. The secondary priority determination unit is used to determine the components whose ordinate difference between the final predicted health level and the corresponding preset health level threshold line is between the first preset threshold and zero as secondary maintenance priority. The third-level priority determination unit is used to determine the components whose ordinate difference between the final predicted health level and the corresponding preset health level threshold line is between zero and the second preset threshold as third-level maintenance priority. The fourth-level priority determination unit is used to determine the components whose ordinate difference between the final predicted health level and the corresponding preset health level threshold line is greater than the second preset threshold as fourth-level maintenance priority.

[0061] In this embodiment, the first preset threshold is a negative value, used to define whether the component has undergone substantial deterioration and entered a high-risk state. When the final predicted health level is lower than the preset health level threshold line, and the difference in the vertical axis is less than the first preset threshold, it indicates that the health status of the component has significantly deviated from the safe range, the degree of deterioration is severe, and immediate repair or replacement measures are required.

[0062] The specific value of the first preset threshold can be determined based on the health degradation records of all couplers and bogies of the same line and model within a preset historical period. The quantile distribution of the difference between the final predicted health of each component and the corresponding preset health threshold line is statistically analyzed, and the quantile point in the lower quantile range is selected as the value of the first preset threshold; the reference quantile range is 5% to 10%. When actual maintenance safety requirements increase or the scope of high-risk component capture needs to be expanded, a quantile point closer to 10% can be selected; when maintenance resources are limited or it is necessary to focus on extremely degraded components, a quantile point closer to 5% can be selected.

[0063] The second preset threshold is a positive value used to define whether a component is in a sufficiently safe and normal state that requires no additional attention. When the final predicted health level is higher than the preset health level threshold and the difference in the vertical axis is greater than the second preset threshold, it indicates that the component's health status is far above the warning level, the deterioration trend is gradual, and planned maintenance can be performed according to the regular cycle.

[0064] The specific value of the second preset threshold can be determined based on the health degradation records of all couplers and bogies of the same line and model within a preset historical period. The quantile distribution of the difference between the final predicted health of each component and the corresponding preset health threshold line is statistically analyzed, and the quantile point in the higher quantile range is selected as the value of the second preset threshold; the reference quantile range is 80% to 95%. When maintenance resources are sufficient and it is desired to reduce the tracking scope of low-risk components, a quantile point closer to 95% can be selected; when maintenance resources are tight and more components need to be included in planned attention, a quantile point closer to 80% can be selected.

[0065] The beneficial effects of the above technical solution are as follows: This embodiment divides the operation and maintenance priorities by using a multi-level priority determination unit. It can clearly and objectively classify all couplers and all bogies according to the difference between the final predicted health level and the preset health level threshold line. This allows components with different risk levels to have clear handling methods, which can not only ensure that high-risk components are handled first, but also achieve reasonable allocation of operation and maintenance resources, thereby improving the efficiency and standardization of the overall operation and maintenance work.

[0066] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A train intelligent operation and maintenance system based on condition monitoring, characterized in that, include: The basic assessment module is used to collect real-time operating status data of all couplers and bogies of the train, and calculate the basic health status of each coupler and each bogie at each monitoring moment of the current shift based on the operating status data of all couplers and bogies of the train. The trend prediction module is used to construct the health change curves of each coupler and each bogie in the current shift based on the historical basic health of each coupler and each bogie in the current shift, and to predict the predicted basic health of each coupler and each bogie in the next finite number of monitoring times in the current shift. The fusion correction module is used to construct the spatial topology relationship of all couplers and bogies of the train, calculate the weight of the single distance influence of all adjacent components on the target component and the weight of the longitudinal position importance of the target component, and based on the spatial topology relationship, the weight of the single distance influence and the weight of the longitudinal position importance, the fusion correction is performed on the predicted basic health of each coupler and each bogie for the current shift in the future finite number of monitoring times, so as to obtain the final predicted health of each coupler and each bogie for the current shift in the future finite number of monitoring times. The operation and maintenance decision module is used to calculate the difference between the ordinate of the final predicted health status of the last monitoring time in multiple future monitoring times of the current shift and the preset health status threshold line, based on the preset health status threshold line of each coupler and each bogie. Based on the ordinate difference, it generates an operation and maintenance priority sequence and outputs a heat map of the predicted health risks of train components.

2. The train intelligent operation and maintenance system based on condition monitoring according to claim 1, characterized in that, The basic assessment module includes: The coupler status acquisition submodule is used to collect data on the coupler coupling status, coupler lock gap, coupler impact amplitude, coupler air tightness, coupler crush tube stress, and coupler centering angle of all couplers in the train, forming the operating status data of each coupler at each monitoring time of the current shift. The bogie status acquisition submodule is used to collect data on bogie axle temperature, bogie vibration effective value, bogie wheel diameter wear, bogie braking pressure, bogie air spring pressure, and bogie lateral displacement of all bogies in the train, forming the operating status data of each bogie at each monitoring moment of the current shift; The basic health determination submodule calculates the basic health of each coupler and each bogie at each monitoring moment of the current shift, based on the operating status data of all couplers and bogies of the train at each monitoring moment of the current shift.

3. The train intelligent operation and maintenance system based on condition monitoring according to claim 2, characterized in that, The basic health determination submodule includes: The map construction unit is used to construct real-time radar maps of each coupler and each bogie at each monitoring time of the current shift, based on the operating status data of all couplers and bogies of the train at each monitoring time of the current shift. The template storage unit is used to store standard radar maps of couplers and standard radar maps of bogies; the standard radar maps of couplers include health maps, sub-health maps, abnormal maps, and fault maps of couplers; the standard radar maps of bogies include health maps, sub-health maps, abnormal maps, and fault maps of bogies. The feature matching unit is used to compare the real-time radar map of each coupler at each monitoring moment of the current shift with the standard radar map of the coupler to obtain the overall matching degree of the coupler; and to compare the real-time radar map of each bogie with the standard radar map of the bogie to obtain the overall matching degree of the bogie. The health rating unit is used to take the health rating of the standard radar map of the coupler with the highest overall matching degree as the basic health rating of the coupler at each monitoring time of the current shift; and to take the health rating of the standard radar map of the bogie with the highest overall matching degree as the basic health rating of each bogie at each monitoring time of the current shift.

4. The train intelligent operation and maintenance system based on condition monitoring according to claim 3, characterized in that, Feature matching units include: The coupler comprehensive matching degree calculation subunit is used to calculate the similarity of the coupler radar map and the area of ​​the coupler radar map contour deviation based on the real-time radar map of each coupler at each monitoring time of the current shift and the standard radar map of the coupler. The coupler comprehensive matching degree is obtained by weighted fusion based on the similarity of the coupler radar map and the area of ​​the coupler radar map contour deviation. The bogie overall matching degree calculation subunit is used to calculate the similarity of the bogie radar map and the area of ​​the bogie radar map profile deviation based on the real-time radar map of each bogie and the standard radar map of the bogie, and to obtain the overall matching degree of the bogie based on the weighted fusion of the similarity of the bogie radar map and the area of ​​the bogie radar map profile deviation.

5. The train intelligent operation and maintenance system based on condition monitoring according to claim 1, characterized in that, The trend prediction module includes: The current shift health curve construction submodule is used to construct the health change curve of each coupler in the current shift based on the historical basic health of each coupler in the current shift; and to construct the health change curve of each bogie in the current shift based on the historical basic health of each bogie in the current shift. The previous shift health curve acquisition submodule is used to acquire the health change curve of each coupler and the health change curve of each bogie in the previous shift. The submodule for obtaining the predicted health recursive parameters is used to obtain the slope and acceleration of the curve at each monitoring moment in the health change curve of each coupler and the health change curve of each bogie in the previous shift, which are respectively used as the coupler health recursive parameters and bogie health recursive parameters at the corresponding monitoring moment in the current shift. The predictive health recursive execution submodule, based on the coupler health recursive parameters and bogie health recursive parameters at each monitoring time of the current shift, recursively obtains the predicted basic health of each coupler and each bogie for a finite number of future monitoring times of the current shift.

6. A train intelligent operation and maintenance system based on condition monitoring according to claim 5, characterized in that, The current shift's health curve construction submodule includes: The current shift coupler health curve construction unit is used to map the basic health of each coupler in the historical basic health of the current shift to the two-dimensional coordinate system of the current shift coupler health according to the time sequence, and connect the basic health of each coupler at each monitoring time in sequence to form the health change curve of each coupler in the current shift. The current shift bogie health curve construction unit is used to map the basic health of each bogie in the historical basic health of each bogie in the current shift to the two-dimensional coordinate system of bogie health in the current shift according to time sequence, and connect the basic health of each bogie at each monitoring time in sequence to form the health change curve of each bogie in the current shift.

7. The train intelligent operation and maintenance system based on condition monitoring according to claim 1, characterized in that, The fusion correction module includes: The spatial topology construction submodule is used to construct a spatial topology diagram with couplers and bogies as nodes, based on the actual installation positions of all couplers and bogies of the train. The predicted value labeling submodule is used to label the predicted basic health of each coupler and the predicted basic health of each bogie for the current shift at a finite number of future monitoring times obtained by the trend prediction module to the corresponding nodes in the spatial topology diagram, forming a spatial topology diagram corresponding to each monitoring time in the finite number of future monitoring times. The neighboring component filtering submodule is used to filter all couplers and bogies that are less than a preset distance threshold and have no other components between them and the target component in the spatial topology diagram corresponding to each monitoring time in a finite number of future monitoring times. The target components include couplers or bogies. The multi-source distance weight calculation submodule is used to calculate the single distance influence weight of each neighboring component on the target component at the corresponding future monitoring time, based on the spatial distance of each neighboring component to the target component and the predicted basic health of the neighboring component itself at the corresponding future monitoring time. The step-by-step fusion correction submodule is used to first perform a first round of fusion correction on the predicted basic health of the target component at each future monitoring time based on the single-item distance influence weight of all neighboring components on the target component at each future monitoring time, and then perform a second round of fusion correction based on the longitudinal position importance weight of the target component to obtain the final predicted health of each coupler and each bogie at each future monitoring time of the current shift.

8. A train intelligent operation and maintenance system based on condition monitoring according to claim 7, characterized in that, The step-by-step fusion correction submodule includes: The first round of fusion correction unit is used to weight and fuse the predicted basic health of the target component with the single distance influence weight of all neighboring components at each future monitoring time to obtain the first round of corrected health of the target component at each future monitoring time. The second-round fusion correction unit is used to weight and fuse the first-round corrected health status of the target component with the longitudinal position importance weight of the target component at each future monitoring time to obtain the final predicted health status of each coupler and each bogie at each future monitoring time of the current shift.

9. A train intelligent operation and maintenance system based on condition monitoring according to claim 1, characterized in that, The operation and maintenance decision module includes: The threshold acquisition submodule is used to acquire the preset health threshold line corresponding to each coupler and each bogie in the current shift; The difference calculation submodule is used to calculate the difference in the vertical axis between the final predicted health of each coupler and each bogie and the corresponding preset health threshold line at the last monitoring time in the finite number of monitoring times in the current shift. The priority generation submodule is used to generate the maintenance priority sequence of all couplers and all bogies in the current shift based on the sign and absolute value of the difference between the final predicted health of each coupler and each bogie and the corresponding preset health threshold line. The heatmap generation submodule is used to generate and output a heatmap of train component health risk prediction based on the final predicted health status of each coupler and bogie at a finite number of future monitoring times for the current shift.

10. A train intelligent operation and maintenance system based on condition monitoring according to claim 9, characterized in that, The priority generation submodule includes: The first-level priority determination unit is used to determine the components whose ordinate difference between the final predicted health level and the corresponding preset health level threshold line is less than the first preset threshold as first-level maintenance priority. The secondary priority determination unit is used to determine the components whose ordinate difference between the final predicted health level and the corresponding preset health level threshold line is between the first preset threshold and zero as secondary maintenance priority. The third-level priority determination unit is used to determine the components whose ordinate difference between the final predicted health level and the corresponding preset health level threshold line is between zero and the second preset threshold as third-level maintenance priority. The fourth-level priority determination unit is used to determine the components whose ordinate difference between the final predicted health level and the corresponding preset health level threshold line is greater than the second preset threshold as fourth-level maintenance priority.

Images