Track line maintenance prediction system based on big data analysis

Abstract

The invention relates to the technical field of track maintenance prediction, in particular to a track line maintenance prediction system based on big data analysis, which is provided with a data acquisition module, a data analysis module, a feature recognition module, a feature judgment module and a maintenance regulation and control module. Screening a feature monitoring track section, performing morphological analysis processing on a ballast bed section model of the feature monitoring track section through a feature recognition module, performing section state analysis on a ballast bed morphological representation image, and judging an abnormal tendency category of the feature monitoring track section through a feature judgment module; and a track line maintenance regulation and control mode is determined through the maintenance regulation and control module. According to the method, the track section with the abnormal tendency is rapidly screened, the track line maintenance method is adaptively adjusted according to the actual track bed condition of the track section, and the efficiency and reliability of a track line maintenance prediction system are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of track maintenance prediction, and particularly relates to a track line maintenance prediction system based on big data analysis. BACKGROUND

[0002] As the core backbone of the modern transportation system, rail transit has the advantages of large capacity, high speed, high safety, energy saving and environmental protection, and has become the key infrastructure supporting the interconnection between cities and the development of regional economy. As the core bearing part of rail transit, the running state of the track line directly determines the driving safety, passenger comfort and transportation efficiency. As an important part of the track line, the track bed bears the core functions of supporting sleepers, dispersing train loads, absorbing vibration energy, and maintaining the stability of track geometry and position. The structural integrity and performance stability of the track bed are crucial to the long-term reliable operation of the track line. With the continuous expansion of the rail transit network, the extension of the operation period, and the continuous improvement of train speed and capacity, the track bed is easily affected by multiple factors such as long-term load, environmental erosion, ballast wear and pollution, and is prone to problems such as ballast heterogenization (e.g. ballast mixing, soil intrusion, particle wear, and cross-section deformation such as width reduction and height reduction). This further leads to increased track vibration and decreased stability. If not timely discovered and targetedly maintained, it may cause safety hazards such as track deformation and track bed collapse, and in severe cases, even lead to line shutdown, causing huge economic losses and social impact. Traditional track line maintenance mainly relies on periodic manual inspection, geometric parameter detection by track inspection vehicles, and post-fault repair response mode. Manual inspection is highly subjective and inefficient, and it is difficult to discover the internal state of the track bed. For the quantitative detection of the track bed cross-section state, although advanced methods such as laser scanning are available, the data is mainly used for post-recording and local evaluation, and it is difficult to provide predictive solutions for track maintenance in a timely manner, affecting the efficiency and reliability of track line maintenance prediction. Therefore, it is an urgent technical problem to improve the efficiency and reliability of the track line maintenance prediction system.

[0003] For example, Chinese patent application publication No. CN116468419A discloses a rail transit big data operation and maintenance decision analysis method, device and storage medium. The method comprises the following steps: S1, a perception layer contains a rail transit associated basic database; S2, a data acquisition layer obtains fault alarm data and analog quantity data from the perception layer and transmits them to a data management center; S3, the data management center has a health evaluation model inside, which performs real-time health diagnosis; S4, when the health index of equipment i reaches the health index maintenance threshold, the data management center sends equipment maintenance disposal suggestion work orders to a PMS system. The application stores the fault alarm data associated with the rail transit equipment according to the upper and lower hierarchical architecture, constructs a health evaluation model inside the data management center, and quantitatively calculates the health indexes of each device under the station, line, and network. Then, the warning time is accurately predicted to guide the condition-based maintenance of the rail transit equipment.

[0004] The prior art also has the following problems: The prior art does not consider that the ballast is prone to heterogenization under the influence of multiple factors such as long-term load, environmental erosion, ballast wear and pollution. Different influencing factors have different effects on the ballast. The prior art cannot quickly screen rail sections with abnormal tendencies, and cannot adaptively adjust the rail line maintenance method according to the actual ballast condition of the rail section, affecting the efficiency and reliability of the rail line maintenance prediction system. SUMMARY

[0005] Therefore, the present application provides a rail line maintenance prediction system based on big data analysis to overcome the problems that the prior art cannot quickly screen rail sections with abnormal tendencies, and cannot adaptively adjust the rail line maintenance method according to the actual ballast condition of the rail section, affecting the efficiency and reliability of the rail line maintenance prediction system.

[0006] To achieve the above-mentioned purpose, the present application provides a rail line maintenance prediction system based on big data analysis, comprising: A data acquisition module, which comprises a vibration response unit for acquiring vibration parameters of a to-be-monitored rail line, and a point cloud acquisition unit for acquiring point cloud data of the to-be-monitored rail line through a monitoring device of a rail inspection vehicle, and constructing a plurality of ballast section models based on the point cloud data; A data analysis module connected to the data acquisition module, for dividing the to-be-monitored rail line into a plurality of monitoring rail sections, constructing a plurality of vibration response curves based on a plurality of vibration parameters of each monitoring rail section, determining a vibration response approaching parameter according to the comparison between a plurality of vibration response curves of the monitoring rail section within a preset monitoring period, and screening a characteristic monitoring rail section; a feature identification module connected with the data acquisition module and the data analysis module, configured to perform morphological analysis on the ballast section model of the feature monitoring track section to obtain a ballast morphology representation image, and perform section state analysis on the ballast morphology representation image to obtain a ballast heterogeneity representation parameter and a ballast stability trend representation parameter; a feature determination module connected with the feature identification module, configured to determine an abnormal tendency category of the feature monitoring track section based on the section state analysis corresponding to the feature monitoring track section; a maintenance regulation module connected with the data analysis module and the feature determination module, configured to determine a track line maintenance regulation mode according to the abnormal tendency category, the track line maintenance regulation mode including increasing a ballast cleaning target depth of the feature monitoring track section, increasing a ballast cleaning frequency of the feature monitoring track section, and determining a section shape according to the ballast section model to supplement ballast and increase a tamping strength of the feature monitoring track section.

[0007] Further, the data analysis module is configured to screen the feature monitoring track section, wherein, the data analysis module screens the monitoring track section as the feature monitoring track section based on a determination result that a vibration response trend parameter of the monitoring track section does not exceed a preset vibration response trend parameter threshold; the vibration response trend parameter is a mean value of coincidence degrees of a plurality of vibration response curves of the monitoring track section within a preset monitoring period; the vibration response curve is constructed with time as the horizontal axis and vibration parameters as the vertical axis.

[0008] Further, the feature identification module is configured to perform morphological analysis on the ballast section model to obtain a ballast morphology representation image, wherein, the feature identification module is configured to set a plurality of virtual segmentation lines perpendicular to the train running direction along the ballast section model in the train running direction, and segment the ballast section model into a plurality of ballast section sub-models according to the virtual segmentation lines; the feature identification module is configured to obtain a plurality of ballast morphology sub-images along the train running direction based on each ballast section sub-model, respectively, and fit the plurality of ballast morphology sub-images to obtain the ballast morphology representation image.

[0009] Further, the feature identification module is configured to perform section state analysis on the ballast morphology representation image of the feature monitoring track section to obtain a ballast heterogeneity representation parameter, wherein, the feature identification module obtains laser intensity parameters of a plurality of points on the ballast morphology representation image, and determines a laser intensity parameter variance as the ballast heterogeneity representation parameter.

[0010] Further, the feature identification module is configured to perform cross-section state analysis on the ballast bed morphology representation image of the feature monitoring track section to obtain a ballast bed stability tendency representation parameter, wherein, The feature identification module obtains a plurality of cross-section length values along the horizontal direction of the cross-section of the ballast bed in the ballast bed morphology representation image, and a plurality of cross-section height values along the direction perpendicular to the track; The ratio of the maximum cross-section length value to the minimum cross-section height value is determined as the ballast bed stability tendency representation parameter.

[0011] Further, the feature determination module is configured to determine the abnormal tendency category of the feature monitoring track section, wherein, The feature determination module determines that the abnormal tendency category of the feature monitoring track section is a strong explicit abnormal tendency category based on the determination result that the cross-section state analysis corresponding to the feature monitoring track section meets the strong explicit abnormal tendency condition; The feature determination module determines that the abnormal tendency category of the feature monitoring track section is a weak explicit abnormal tendency category based on the determination result that the cross-section state analysis corresponding to the feature monitoring track section does not meet the strong explicit abnormal tendency condition; The strong explicit abnormal tendency condition is that the ballast bed heterogeneity representation parameter exceeds a preset ballast bed heterogeneity representation parameter threshold, and the ballast bed stability tendency representation parameter exceeds a preset ballast bed stability tendency representation parameter interval.

[0012] Further, the maintenance control module is configured to determine the track line maintenance control mode according to the abnormal tendency category, wherein, If the abnormal tendency category is the strong explicit abnormal tendency category, the maintenance control module determines that the track line maintenance control mode is to determine the ballast cleaning target depth of the feature monitoring track section according to the ballast bed heterogeneity representation parameter and the vibration response tendency parameter; If the abnormal tendency category is the weak explicit abnormal tendency category, the maintenance control module determines that the track line maintenance control mode is to determine the ballast cleaning frequency of the feature monitoring track section according to the ballast bed heterogeneity representation parameter, or to determine the tamping intensity of the supplementary ballast according to the vibration response tendency parameter.

[0013] Further, the maintenance control module is configured to determine the growth amplitude of the ballast cleaning target depth of the feature monitoring track section, wherein, The growth amplitude of the ballast cleaning target depth is positively correlated with the ballast bed heterogeneity representation parameter and negatively correlated with the vibration response tendency parameter.

[0014] Further, the maintenance control module is configured to determine the track line maintenance control mode based on the feature monitoring track section being a weak explicit abnormal tendency category, wherein, If the ballast heterogeneity characterization parameter of the feature monitoring track section exceeds a preset ballast heterogeneity characterization parameter threshold, the maintenance control module determines a ballast cleaning frequency for the feature monitoring track section according to the ballast heterogeneity characterization parameter; If the ballast stability trend characterization parameter of the feature monitoring track section exceeds a preset ballast stability trend characterization parameter interval, the maintenance control module determines a ballast section shape according to the ballast section model to supplement ballast, and determines a tamping strength for supplementing ballast based on the vibration response trend parameter.

[0015] Further, the maintenance control module is used to determine an increase range of the ballast cleaning frequency and an increase range of the tamping strength, wherein, The increase range of the ballast cleaning frequency is positively correlated with the ballast heterogeneity characterization parameter; The increase range of the tamping strength is negatively correlated with the vibration response trend parameter.

[0016] Compared with the prior art, the beneficial effects of the present application are that the present application sets a data acquisition module, a data analysis module, a feature recognition module, a feature determination module, and a maintenance control module, determines a vibration response trend parameter according to the comparison between a plurality of vibration response curves of the monitoring track section in a preset monitoring period by the data analysis module to screen a feature monitoring track section, performs morphological analysis processing on a ballast section model of the feature monitoring track section by the feature recognition module to obtain a ballast morphological characterization image, performs section state analysis on the ballast morphological characterization image, determines an abnormal tendency category of the feature monitoring track section based on the corresponding section state analysis of the feature monitoring track section by the feature determination module, and determines a track line maintenance control mode according to the abnormal tendency category by the maintenance control module, thereby realizing rapid screening of track sections with abnormal tendencies, adaptively adjusting the track line maintenance method according to the actual ballast condition of the track section, and improving the efficiency and reliability of the track line maintenance prediction system.

[0017] Especially, the application determines the vibration response approaching parameter according to the comparison between a plurality of vibration response curves of the monitored track section by the data analysis module to screen the characteristic monitoring track section, and it can be understood that the mass, stiffness and damping distribution of the track section with relatively stable structure state are determined, so that the vibration response curve generated under the same or similar train excitation input will present a highly repeatable form feature in the time domain or frequency domain, if the track bed appears compaction, subsidence or loose disease, the structure dynamic characteristics will change, resulting in different responses under the same excitation, the vibration response approaching parameter is small, the optimization configuration and efficient use of detection resources are realized, the characteristic monitoring track section with abnormal dynamic behavior is quickly identified, and the huge resource consumption caused by indiscriminate general survey of the whole line is avoided, and then, the track section with abnormal tendency is quickly screened, and the efficiency and reliability of the track line maintenance prediction system are improved.

[0018] Especially, the application performs morphological analysis processing on the track bed cross-section model of the characteristic monitoring track section by the feature recognition module to obtain a track bed morphological representation image, and it can be understood that the traditional manual inspection or photography can only obtain discrete point non-standard cross-section information, and through automatic sequence slicing and imaging of the point cloud model, a track full-line continuous, standard and unified two-dimensional morphological representation image can be generated, so that the key dimensions of the track bed shoulder width, slope and ballast surface profile can be measured and monitored, providing a complete data base for evaluating the macro stability of the track bed, laying a standardized foundation for subsequent intelligent feature extraction, realizing the balance between information compression and key feature retention, converting massive three-dimensional point cloud data into two-dimensional representation images, reducing the burden of data storage and processing, and at the same time, through the carefully designed slicing and projection direction, the key morphological features that can best reflect the bearing and stability state of the track bed are retained, and redundant spatial details are discarded, and then, the track bed morphological representation image is obtained, and the efficiency and reliability of the track line maintenance prediction system are improved.

[0019] Especially, the application can understand that for the compound abnormal tendency section that has both internal serious fouling and overall dynamic instability, radical treatment measures need to be taken, and the growth range of the screening target depth is jointly driven by the two parameters, which ensures that the intervention depth can not only remove the identified deep fouling source, but also match the severity of the overall instability of the track. The track bed heterogeneity characterization parameter quantifies the material homogeneity inside the track bed, and the smaller the vibration response approaching parameter means that the dynamic response of the track under train load is relatively discrete and unpredictable, which is often caused by surface fouling, and it also implies that the underlying foundation support may have been unstable or have deep defects, and more thorough and deeper screening intervention is needed to fundamentally rebuild the stable foundation. Further, the track line maintenance method is adaptively adjusted according to the actual track bed condition of the track section, and the efficiency and reliability of the track line maintenance prediction system are improved.

[0020] Especially, the application can understand that for the compound abnormal tendency section that has both internal serious fouling and overall dynamic instability, radical treatment measures need to be taken, and the growth range of the screening target depth is jointly driven by the two parameters, which ensures that the intervention depth can not only remove the identified deep fouling source, but also match the severity of the overall instability of the track. The track bed heterogeneity characterization parameter quantifies the material homogeneity inside the track bed, and the smaller the vibration response approaching parameter means that the dynamic response of the track under train load is relatively discrete and unpredictable, which is often caused by surface fouling, and it also implies that the underlying foundation support may have been unstable or have deep defects, and more thorough and deeper screening intervention is needed to fundamentally rebuild the stable foundation. Further, the track line maintenance method is adaptively adjusted according to the actual track bed condition of the track section, and the efficiency and reliability of the track line maintenance prediction system are improved.

[0021] In particular, the application, by maintaining the regulation module under the weakly dominant abnormal tendency category, if the track bed stability tendency characteristic parameter of the feature monitoring track section exceeds the preset track bed stability tendency characteristic parameter interval, determines the section shape according to the track bed section model to supplement the ballast, and determines the tamping strength of the supplemented ballast based on the vibration response approaching parameter. It can be understood that the track bed stability tendency parameter exceeds the normal interval, indicating that the macro-geometric profile of the track bed has been deformed. By filling sufficient ballast material, the physical shape of the track bed is restored to a standard form that can provide sufficient support area and slope stability. Tamping can compact and embed the loose ballast into the existing track bed system, and the strength directly determines the compactness and integrity of the repaired foundation. The lower the vibration response approaching parameter, the more discrete and unstable the dynamic response of the section under train load, and stronger tamping operation is required to quickly rebuild a stable foundation. Therefore, according to the clear quantitative data, the deviation and uncertainty of artificial experience judgment are reduced, and the limited maintenance resources are more reasonably allocated to the key link that can improve the overall line stability, optimizing the allocation of maintenance resources, and further, the track line maintenance method is adaptively adjusted according to the actual track bed condition of the track section, improving the efficiency and reliability of the track line maintenance prediction system. BRIEF DESCRIPTION OF DRAWINGS

[0022] Figure 1 The functional block diagram of the track line maintenance prediction system based on big data analysis of the embodiment of the application; Figure 2 The logic flowchart of the data analysis module of the embodiment of the application for screening feature monitoring track sections; Figure 3 The logic flowchart of the feature determination module of the embodiment of the application for determining the abnormal tendency category; Figure 4 The logic flowchart of the maintenance regulation module of the embodiment of the application for determining the track line maintenance regulation mode. DETAILED DESCRIPTION

[0023] In order to make the objects and advantages of the application clearer, the application will be further described below with reference to the embodiments. It should be understood that the specific embodiments described herein are only used to explain the application and do not limit the application.

[0024] The preferred embodiments of the application will be described below with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are only used to explain the technical principles of the application and do not limit the protection scope of the application.

[0025] It should be noted that in the description of the present application, the terms indicating the direction or positional relationship of "upper", "lower", "inner", "outer" and the like are based on the direction or positional relationship shown in the drawings, which is only for the convenience of description, and does not indicate or imply that the device or element must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation on the present application.

[0026] In addition, it should be further pointed out that in the description of the present application, unless otherwise explicitly specified and limited, the terms "mounting", "connection" should be understood broadly, for example, it can be fixed connection, or detachable connection, or integral connection, it can be mechanical connection, or electrical connection, it can be directly connected, or indirectly connected through intermediate medium, it can be the communication inside two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0027] Please refer to Figure 1 The function block diagram of the track line maintenance prediction system based on big data analysis of the embodiment of the present application is shown, and the track line maintenance prediction system based on big data analysis of the present application comprises: The data acquisition module comprises a vibration response unit for acquiring vibration parameters of the track line to be monitored, and a point cloud acquisition unit for acquiring point cloud data of the track line to be monitored through a monitoring device of a track inspection vehicle, and constructing a plurality of ballast section models based on the point cloud data; Specifically, the structure of the vibration response unit and the point cloud acquisition unit is not limited in the embodiment of the present application, and preferably, it can be a sensor array comprising an acceleration sensor and a displacement sensor, which can be arranged on a special vehicle frame of a sleeper, a rail or a track inspection vehicle, to acquire vibration parameters, and a laser radar system, which is a core component of a modern track inspection vehicle, integrates high-precision inertial navigation and an odometer, synchronously records the spatial absolute position of each laser point, and constructs a ballast section model of the monitored track line through point cloud processing algorithms such as filtering, classification and modeling, which will not be repeated here.

[0028] The data analysis module is connected with the data acquisition module, and is used to divide the track line to be monitored into a plurality of monitoring track sections, construct a plurality of vibration response curves based on a plurality of vibration parameters of each monitoring track section, determine vibration response approaching parameters according to the comparison between a plurality of vibration response curves of the monitoring track section within a preset monitoring period, and screen characteristic monitoring track sections. Specifically, the structure of the data analysis module is not limited in the embodiment of the present application, and preferably, it can be a microprocessor, which is used to screen characteristic monitoring track sections, and will not be repeated here.

[0029] Specifically, the monitoring track sections can be uniformly divided by a person skilled in the art according to the accuracy requirement of track line maintenance prediction, the higher the accuracy requirement is, the more the number of the divided monitoring track sections is, and 40 monitoring track sections can be divided as an example.

[0030] Specifically, the preset monitoring period can be uniformly divided by a person skilled in the art according to the accuracy requirement of track line maintenance prediction, the higher the accuracy requirement is, the longer the preset monitoring period is, and the value range can be [1, 5] with the interval unit being day, and preferably, the preset monitoring period can be 2 days.

[0031] The feature recognition module is connected with the data acquisition module and the data analysis module respectively, and is used for performing morphological analysis processing on the ballast section model of the feature monitoring track section to obtain a ballast morphology representation image, and performing section state analysis on the ballast morphology representation image to obtain a ballast heterogeneity representation parameter and a ballast stability trend representation parameter. Specifically, the structure of the feature recognition module is not limited in the embodiment of the application, and preferably, the feature recognition module can be a microprocessor, which is used for performing section state analysis on the ballast morphology representation image, and details are not repeated here.

[0032] The feature determination module is connected with the feature recognition module, and is used for determining an abnormal tendency category of the feature monitoring track section based on the section state analysis corresponding to the feature monitoring track section. Specifically, the structure of the feature determination module is not limited in the embodiment of the application, and preferably, the feature determination module can be a processor used in a computer, which is used for determining the abnormal tendency category, and details are not repeated here.

[0033] The maintenance control module is connected with the data analysis module and the feature determination module respectively, and is used for determining a track line maintenance control mode according to the abnormal tendency category, the track line maintenance control mode including increasing a ballast cleaning target depth of the feature monitoring track section, increasing a ballast cleaning frequency of the feature monitoring track section, and determining a section shape according to the ballast section model to supplement ballast, and increasing a tamping intensity of the feature monitoring track section.

[0034] Specifically, the structure of the maintenance control module is not limited in the embodiment of the application, and preferably, the maintenance control module can be a microprocessor, which is used for determining the track line maintenance control mode, and details are not repeated here. Please refer to Figure 2 The figure is a logic flow chart of the data analysis module of the embodiment of the application for screening the feature monitoring track section, the data analysis module is used for screening the feature monitoring track section, wherein, The data analysis module screens the monitoring track section as the feature monitoring track section based on the determination result that the vibration response trend parameter of the monitoring track section does not exceed the preset vibration response trend parameter threshold. based on the determination result that the vibration response approaching parameter of the monitored track section exceeds the preset vibration response approaching parameter threshold, the monitored track section is not screened; The vibration response approaching parameter is a mean value of coincidence degrees of a plurality of vibration response curves of the monitored track section in a preset monitoring period. The vibration response curve is constructed with time as the horizontal axis and vibration parameter as the vertical axis.

[0035] Specifically, the preset vibration response approaching parameter threshold is a product of a vibration response approaching parameter reference value and a vibration factor, the vibration response approaching parameter reference value is a mean value of vibration response approaching parameters under the same working condition in historical data, and the vibration factor can be set by a person skilled in the art according to the accuracy requirement of track line maintenance prediction, the higher the accuracy requirement, the smaller the set value, and the value range can be [1.05, 1.2], preferably 1.1.

[0036] The way of constructing each vibration response curve is not limited, for example, the vibration response curve can be fitted by using matlab related fitting software, which will not be repeated here.

[0037] Specifically, the way of determining the coincidence degree is not limited, for example, the cosine similarity method can be used to determine the coincidence degree, the vibration response curve segment is vectorized to calculate the cosine similarity, and the obtained cosine similarity is determined as the coincidence degree, of course, other methods can also be used, which will not be repeated here.

[0038] Specifically, the embodiment of the present application determines the vibration response approaching parameter according to the comparison between the vibration response curves of the monitored track section, so as to screen the characteristic monitoring track section. It can be understood that for the track section with relatively stable structure state, the mass, stiffness and damping distribution are determined, so that under the same or similar train excitation input, the vibration response curve generated will present a highly repeatable form feature in the time domain or frequency domain. The vibration response approaching parameter is the average coincidence degree between the vibration response curves collected multiple times in a preset period. If the track section structure state does not deteriorate, the multiple response curves should be relatively similar, and the vibration response approaching parameter is larger. If the track bed appears to be hardened, sunken or loose, the structure dynamic characteristics will change, resulting in different responses to the same excitation, and the vibration response approaching parameter is smaller. When the vibration response approaching parameter of a certain track section is relatively small, it indicates that the inherent dynamic fingerprint has been unstable, and the internal structure state is likely to have changed abnormally. The optimization of detection resources and efficient use are realized, the characteristic monitoring track section with abnormal dynamic behavior is quickly identified, and the subsequent cost higher and more detailed track bed section detection is only carried out on these suspicious sections, avoiding the huge resource consumption caused by indiscriminate general survey of the whole line. At the same time, the change of the vibration response is often the early warning signal of the internal state of the track structure, such as the local softening of the track bed supporting stiffness and the appearance of cavities. By monitoring the stability of the response curve, the early and weak abnormal signs can be captured, the maintenance intervention time is greatly advanced, the change from post-repair to pre-prevention is realized, and then the track section with abnormal tendency is quickly screened, and the efficiency and reliability of the track line maintenance prediction system are improved.

[0039] Specifically, the feature recognition module is used to perform morphological analysis processing on the track bed section model to obtain a track bed morphology representation image, wherein, The feature recognition module is used to set a plurality of virtual segmentation lines perpendicular to the train running direction on the track bed section model along the train running direction, and divide the track bed section model into a plurality of track bed section sub-models according to the virtual segmentation lines. The feature recognition module is used to obtain a plurality of track bed morphology sub-images along the train running direction based on each track bed section sub-model respectively, and fit the plurality of track bed morphology sub-images to obtain the track bed morphology representation image.

[0040] Specifically, the embodiment of the present application obtains the track bed morphology representation image by the feature recognition module for morphology analysis processing of the track bed cross-section model of the feature monitoring track section. It can be understood that the traditional manual inspection or photographing can only obtain discrete point non-standard cross-section information. Through the automatic sequence slicing and imaging of the point cloud model, a track full-line continuous, standard and unified two-dimensional morphology representation image can be generated, so that the key dimensions of the track bed shoulder width, slope and ballast surface profile can be measured and monitored, a complete data base for evaluating the macro stability of the track bed is provided, a standardized foundation for subsequent intelligent feature extraction is laid, the balance between information compression and key feature retention is achieved, the massive three-dimensional point cloud data is converted into a two-dimensional representation image, the burden of data storage and processing is reduced, and at the same time, through the careful design of the slicing and projection direction, the key morphology features that can best reflect the load bearing and stability state of the track bed are retained, and the redundant spatial details are discarded. Further, the track bed morphology representation image is obtained, and the efficiency and reliability of the track line maintenance prediction system are improved.

[0041] Specifically, it can be understood that a series of virtual division lines with equal intervals are arranged along the direction of train travel, and the division lines are perpendicular to the track direction, so that the complete track bed cross-section is divided into a plurality of continuous track bed cross-section sub-models, each sub-model records the track bed information of the specific cross-section position, and the profile line or surface texture is extracted along the direction of train travel, i.e. the direction perpendicular to the slicing plane, to form a two-dimensional track bed morphology sub-image reflecting the local morphology of the cross-section. All sub-images are spliced and fitted according to their spatial order, and finally a complete track bed morphology representation image is generated, which is expanded along the track direction. The image is essentially a sequence mapping of the three-dimensional surface morphology of the track bed on the two-dimensional plane, which not only retains the key geometric features of the track bed cross-section shape, but also reflects the continuous change of the features along the track longitudinal direction through the image sequence. Therefore, the three-dimensional morphology problem which is difficult to calculate is converted into a standardized two-dimensional problem which can be extracted and analyzed by mature image processing technology, and further, the track bed morphology representation image is obtained, and the efficiency and reliability of the track line maintenance prediction system are improved.

[0042] Specifically, the feature recognition module is used to analyze the cross-section state of the track bed morphology representation image of the feature monitoring track section to obtain track bed heterogeneity representation parameters. The feature recognition module obtains the laser intensity parameters of a plurality of point positions on the track bed morphology representation image, and determines the laser intensity parameter variance as the track bed heterogeneity representation parameter.

[0043] Specifically, the sampling point positions of the laser intensity parameters can be uniformly distributed to comprehensively reflect the track bed heterogeneity.

[0044] Specifically, the feature recognition module is used to perform section state analysis on the track bed morphology representation image of the feature monitoring track section to obtain a track bed stability trend representation parameter, wherein, The feature recognition module obtains a plurality of section length values along the horizontal direction of the cross section of the track bed section in the track bed morphology representation image, and a plurality of section height values along the direction perpendicular to the track; The ratio of the maximum section length value to the minimum section height value is determined as the track bed stability trend representation parameter.

[0045] Please refer to Figure 3 The logic flow chart of the feature determination module of the embodiment of the present application is shown in the figure, which is used to determine the abnormal tendency category of the feature monitoring track section, wherein, The feature determination module determines the abnormal tendency category of the feature monitoring track section as a strong explicit abnormal tendency category based on the determination result that the section state analysis corresponding to the feature monitoring track section meets the strong explicit abnormal tendency condition; Based on the determination result that the section state analysis corresponding to the feature monitoring track section does not meet the strong explicit abnormal tendency condition, the feature determination module determines the abnormal tendency category of the feature monitoring track section as a weak explicit abnormal tendency category; The strong explicit abnormal tendency condition is that the track bed heterogeneity representation parameter exceeds the preset track bed heterogeneity representation parameter threshold, and the track bed stability trend representation parameter exceeds the preset track bed stability trend representation parameter interval.

[0046] Specifically, the preset track bed heterogeneity representation parameter threshold is the product of the track bed heterogeneity representation parameter reference value and the heterogeneity factor, the upper limit of the preset track bed stability trend representation parameter interval is the product of the track bed stability trend representation parameter reference value and the first stability factor, and the lower limit is the product of the track bed stability trend representation parameter reference value and the second stability factor. The track bed heterogeneity representation parameter reference value and the track bed stability trend representation parameter reference value are respectively the average value of the track bed heterogeneity representation parameter and the average value of the track bed stability trend representation parameter under the same working condition in the historical data. The heterogeneity factor, the first stability factor and the second stability factor can be set by the person skilled in the art according to the accuracy requirement of track line maintenance prediction. The higher the accuracy requirement, the smaller the heterogeneity factor. The value range can be [1.05, 1.15], preferably 1.1. The smaller the first stability factor, the value range can be [1.1, 1.2], preferably 1.15. The larger the second stability factor, the value range can be [0.8, 0.95], preferably 0.85.

[0047] Please refer to Figure 4As shown, it is a logic flow diagram for determining track line maintenance control mode by the maintenance control module of the embodiment of the present application, the maintenance control module is used to determine track line maintenance control mode according to the abnormal tendency category, wherein, If the abnormal tendency category is the strong explicit abnormal tendency category, the maintenance control module determines that the track line maintenance control mode is to determine the screening target depth of the characteristic monitoring track segment according to the ballast bed heterogeneity characterization parameter and the vibration response approaching parameter; If the abnormal tendency category is the weak explicit abnormal tendency category, the maintenance control module determines that the track line maintenance control mode is to determine the screening frequency of the characteristic monitoring track segment according to the ballast bed heterogeneity characterization parameter, or to determine the tamping strength of the supplemented ballast according to the vibration response approaching parameter.

[0048] Specifically, under the strong explicit abnormal tendency category, the maintenance control module determines the screening target depth of the characteristic monitoring track segment according to the ballast bed heterogeneity characterization parameter and the vibration response approaching parameter, it can be understood that for the compound abnormal tendency section that has both internal serious contamination and overall dynamic instability, radical treatment measures need to be taken, and the growth range of the screening target depth is jointly driven by the two parameters, which ensures that the intervention depth can not only remove the identified deep contamination source, but also match the severity of the overall instability of the track, the ballast bed heterogeneity characterization parameter quantifies the material homogeneity inside the ballast bed, the larger the ballast bed heterogeneity characterization parameter, the more significant the difference in the reflection characteristics of the ballast surface, which usually means that the invasion of internal contamination such as coal powder and soil is relatively serious, the gradation is deteriorated or there is uneven hardening, the disease may spread deeper in the depth direction, and the screening depth must be sufficient to completely remove the deteriorated layer, and the smaller the vibration response approaching parameter means that the dynamic response of the track under the train load is relatively discrete and unpredictable, which is often caused by surface contamination, and more implies that the lower part of the foundation support may have been unstable or have deep defects, and more thorough and deeper screening intervention is needed to fundamentally rebuild the stable foundation, and then, the track line maintenance method is adaptively adjusted according to the actual ballast bed condition of the track segment, and the efficiency and reliability of the track line maintenance prediction system are improved.

[0049] Specifically, the maintenance control module is used to determine the growth range of the screening target depth of the characteristic monitoring track segment, wherein, The growth range of the screening target depth is positively correlated with the ballast bed heterogeneity characterization parameter and negatively correlated with the vibration response approaching parameter.

[0050] Specifically, the growth range of the screening target depth is the heterogeneous weight factor x the ballast heterogeneity characterization parameter / the ballast heterogeneity characterization parameter reference value + the vibration weight factor x the vibration response trend parameter reference value / the vibration response trend parameter, the heterogeneous weight factor and the vibration weight factor can be the size of the influence factor selected by the person skilled in the art according to the influence degree of the ballast heterogeneity characterization parameter and the vibration response trend parameter in the historical data on the calculation result, the heterogeneous weight factor + the vibration weight factor = 1, preferably, the heterogeneous weight factor can be 0.7, and the vibration weight factor can be 0.3.

[0051] Specifically, the maintenance control module is used to determine the track line maintenance control mode based on the feature monitoring track section being a weakly dominant abnormality tendency category, wherein, If the ballast heterogeneity characterization parameter of the feature monitoring track section exceeds the preset ballast heterogeneity characterization parameter threshold value, the maintenance control module determines the screening frequency of the feature monitoring track section according to the ballast heterogeneity characterization parameter; If the ballast stability trend characterization parameter of the feature monitoring track section exceeds the preset ballast stability trend characterization parameter interval, the maintenance control module determines the cross-section shape according to the ballast cross-section model to supplement the ballast, and determines the tamping strength of the supplemented ballast based on the vibration response trend parameter.

[0052] Specifically, the maintenance control module in the weakly dominant abnormality tendency category determines the screening frequency of the feature monitoring track section according to the ballast heterogeneity characterization parameter if the ballast heterogeneity characterization parameter of the feature monitoring track section exceeds the preset ballast heterogeneity characterization parameter threshold value, it can be understood that the ballast heterogeneity characterization parameter exceeds the standard, indicating that the internal material homogeneity of the ballast begins to deteriorate, and relatively significant dirt aggregation or initial consolidation phenomenon occurs, since the ballast stability trend characterization parameter is still in the normal interval, it means that this material deterioration has not caused the collapse or serious deformation of the macro-geometric profile of the ballast, and the abnormality is still in the local, superficial layer or early development stage, at this time, large-scale and deep thorough screening may not be economically optimal, the larger the ballast heterogeneity characterization parameter, the more heterogeneous the material, the more obvious the dirt trend, and the faster the potential speed of further deterioration of its state, therefore, more frequent cleaning intervention is needed to curb the deterioration process and prevent it from upgrading from a simple dirt problem to a geometric deformation composite abnormality affecting structural stability, by adjusting the variable of the maintenance period in the time dimension, the cost-optimal control of early single disease is realized, the line state is stable, and the life cycle maintenance cost is minimized, and then, the track line maintenance method is adaptively adjusted according to the actual ballast condition of the track section, and the efficiency and reliability of the track line maintenance prediction system are improved.

[0053] Specifically, the maintenance control module is used to determine the growth range of the screening frequency and the growth range of the tamping strength, wherein, The growth amplitude of the screening frequency is positively correlated with the ballast heterogeneity characterization parameter; The growth amplitude of the tamping strength is negatively correlated with the vibration response approaching parameter.

[0054] Specifically, the growth amplitude of the screening frequency is a frequency factor multiplied by the ballast heterogeneity characterization parameter / the ballast heterogeneity characterization parameter reference value, and the growth amplitude of the tamping strength is a strength factor multiplied by the vibration response approaching parameter reference value / the vibration response approaching parameter, the frequency factor and the strength factor are calculated by a person skilled in the art according to the mean value of several experimental data, the value range of the frequency factor can be [0.1, 0.3], and the value range of the strength factor can be [0.2, 0.4], so as to avoid excessive or insufficient adjustment, preferably, the frequency factor can be 0.2, and the strength factor can be 0.3.

[0055] Specifically, according to the ballast section model, the section shape is determined to supplement the ballast under the weakly dominant abnormality tendency category of the maintenance control module, if the ballast stable tendency characterization parameter of the feature monitoring track section exceeds the preset ballast stable tendency characterization parameter interval, the tamping strength of the supplemented ballast is determined based on the vibration response approaching parameter, it can be understood that the ballast stable tendency parameter exceeds the normal interval, which represents that the macroscopic geometric profile of the ballast has collapsed, rheological or shoulder width deficiency deformation, and the designed mechanical bearing section is damaged, the section shape is determined according to the ballast section model to supplement the ballast, which restores the physical shape of the ballast to a standard form capable of providing sufficient support area and slope stability by filling sufficient ballast material, and only supplementing the bulk ballast cannot immediately restore the dynamic performance of the track, the newly supplemented ballast is in a loose state and cannot form an integrated bearing structure with the existing ballast, tamping can compact and embed the loose ballast into the existing ballast system, and the strength directly determines the compactness and integrity of the repaired foundation, the lower the vibration response approaching parameter, the more discrete and unstable the dynamic response of the section under the train load, and the more likely the underlying foundation has the fundamental defect of loose integrity and insufficient support stiffness, for such a section that has deformed and has unstable dynamics, stronger tamping operation is required to rearrange and tightly engage the ballast particles with higher energy, so as to quickly rebuild a stable foundation and fundamentally improve the dynamic response characteristics, therefore, the decision is made according to the clear quantitative data, the deviation and uncertainty of artificial experience judgment are reduced, the limited maintenance resources are more reasonably allocated to the key link that can most improve the overall line stability, the maintenance resource allocation is optimized, and then, the track line maintenance method is adaptively adjusted according to the actual ballast condition of the track section, and the efficiency and reliability of the track line maintenance prediction system are improved.

[0056] So far, the technical solutions of the present application have been described in combination with the preferred embodiments shown in the drawings, but it is easy for those skilled in the art to understand that the protection scope of the present application is obviously not limited to these specific embodiments. Those skilled in the art can make equivalent changes or replacements to the related technical features without departing from the principles of the present application, and the technical solutions after the changes or replacements will all fall within the protection scope of the present application.

[0057] The above only describes the preferred embodiments of the present application and is not intended to limit the present application; the present application can have various changes and variations for those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A track maintenance prediction system based on big data analysis, characterized in that, include: The data acquisition module includes a vibration response unit for acquiring vibration parameters of the track line to be monitored, and a point cloud acquisition unit for acquiring several point cloud data of the track line to be monitored through the monitoring device of the track inspection vehicle, and constructing several track bed cross-section models based on the point cloud data. The data analysis module, which is connected to the data acquisition module, is used to divide the track line to be monitored into several monitoring track segments, construct several vibration response curves based on several vibration parameters of each monitoring track segment, and determine vibration response approach parameters based on the comparison between several vibration response curves of the monitoring track segments within a preset monitoring period, so as to screen characteristic monitoring track segments. The feature recognition module is connected to the data acquisition module and the data analysis module respectively. It is used to perform morphological analysis processing on the track bed cross-section model of the track bed of the feature monitoring track section to obtain the track bed morphological characterization image, and to perform cross-sectional state analysis on the track bed morphological characterization image to obtain track bed heterogeneity characterization parameters and track bed stability tendency characterization parameters. The feature determination module, which is connected to the feature recognition module, is used to determine the abnormal tendency category of the feature monitoring track segment based on the cross-sectional state analysis corresponding to the feature monitoring track segment. The maintenance and control module is connected to the data analysis module and the feature determination module, respectively, and is used to determine the track line maintenance and control method according to the abnormal tendency category. The track line maintenance and control method includes increasing the target depth of the cleaning and screening of the track section under feature monitoring, increasing the cleaning and screening frequency of the track section under feature monitoring, determining the cross-sectional shape according to the track bed cross-section model to supplement ballast, and increasing the tamping intensity of the track section under feature monitoring.

2. The track maintenance prediction system based on big data analysis according to claim 1, characterized in that, The data analysis module is used to filter track segments with specific features for monitoring, wherein, The data analysis module selects the monitored track segment as a feature monitoring track segment based on the determination result that the vibration response approach parameter of the monitored track segment does not exceed the preset vibration response approach parameter threshold. The vibration response approximation parameter is the average overlap of several vibration response curves of the monitored track segment within a preset monitoring period. The vibration response curve is constructed with time as the horizontal axis and vibration parameters as the vertical axis.

3. The track maintenance prediction system based on big data analysis according to claim 2, characterized in that, The feature recognition module is used to perform morphological analysis processing on the track bed cross-section model to obtain a track bed morphological representation image, wherein... The feature recognition module is used to set several virtual dividing lines perpendicular to the direction of train travel on the track bed cross-section model, and divide the track bed cross-section model into several track bed cross-section sub-models according to the virtual dividing lines. The feature recognition module is used to acquire several track bed morphology sub-images along the train travel direction based on each of the track bed cross-section sub-models, and to fit the several track bed morphology sub-images to obtain the track bed morphology representation image.

4. The track maintenance prediction system based on big data analysis according to claim 3, characterized in that, The feature recognition module is used to perform cross-sectional state analysis on the track bed morphology characterization image of the track section under feature monitoring to obtain track bed heterogeneity characterization parameters, wherein... The feature recognition module acquires the laser intensity parameters at several points on the track bed morphology characterization image, and determines the variance of the laser intensity parameters as the heterogeneous characterization parameters of the track bed.

5. The track maintenance prediction system based on big data analysis according to claim 4, characterized in that, The feature recognition module is used to perform cross-sectional state analysis on the track bed morphology characterization image of the track section under feature monitoring to obtain track bed stability trend characterization parameters, wherein... The feature recognition module obtains several cross-sectional length values ​​along the horizontal direction of the cross-section of the track bed in the track bed morphology characterization image, and obtains several cross-sectional height values ​​along the direction perpendicular to the track. The ratio of the maximum cross-sectional length value to the minimum cross-sectional height value is determined as the stability tendency characterization parameter of the track bed.

6. The track maintenance prediction system based on big data analysis according to claim 5, characterized in that, The feature determination module is used to determine the abnormal tendency category of the feature monitoring track segment, wherein, The feature determination module determines the abnormality tendency category of the feature monitoring track segment as a strong dominant abnormality tendency category based on the determination result of the cross-sectional state analysis corresponding to the feature monitoring track segment that meets the conditions of strong dominant abnormality tendency. Based on the determination result that the cross-sectional state analysis corresponding to the feature monitoring track segment does not meet the conditions for a strong dominant anomaly tendency, the anomaly tendency category of the feature monitoring track segment is determined to be a weak dominant anomaly tendency category. The strong dominant anomaly tendency condition is that the heterogeneous characterization parameter of the track bed exceeds the preset threshold of the heterogeneous characterization parameter of the track bed, and the stability tendency characterization parameter of the track bed exceeds the preset range of the stability tendency characterization parameter of the track bed.

7. The track maintenance prediction system based on big data analysis according to claim 6, characterized in that, The maintenance and control module is used to determine the track maintenance and control method based on the abnormal tendency category, wherein... If the abnormal tendency category is the strongly manifest abnormal tendency category, the maintenance and control module determines the track line maintenance and control method as follows: the target depth for cleaning the track section under characteristic monitoring is determined based on the heterogeneous characterization parameters of the track bed and the vibration response approach parameters. If the abnormal tendency category is the weakly manifest abnormal tendency category, the maintenance and control module determines the track line maintenance and control method as follows: determine the screening frequency of the characteristic monitoring track section based on the heterogeneous characterization parameters of the track bed, or determine the tamping intensity of the supplementary ballast based on the vibration response approach parameters.

8. The track maintenance prediction system based on big data analysis according to claim 7, characterized in that, The maintenance and control module is used to determine the rate of increase in the target depth of the screening process on the feature monitoring track segment, wherein, The increase in the target depth of the cleaning and screening is positively correlated with the heterogeneous characterization parameters of the track bed and negatively correlated with the vibration response approach parameters.

9. The track maintenance prediction system based on big data analysis according to claim 8, characterized in that, The maintenance and control module is used to determine the track line maintenance and control method based on the characteristic monitoring track segment being classified as having a weak and obvious anomaly tendency. If the heterogeneous characterization parameters of the track bed in the feature monitoring track section exceed the preset threshold for heterogeneous characterization parameters of the track bed, the maintenance and control module determines the cleaning frequency of the feature monitoring track section based on the heterogeneous characterization parameters of the track bed. If the track bed stability tendency characterization parameter of the track section under feature monitoring exceeds the preset track bed stability tendency characterization parameter range, the maintenance and control module determines the cross-sectional shape according to the track bed cross-section model to supplement the ballast, and determines the tamping strength of the supplemented ballast based on the vibration response approach parameter.

10. The track maintenance prediction system based on big data analysis according to claim 9, characterized in that, The maintenance and control module is used to determine the increase rate of the screening frequency and the increase rate of the compaction intensity, wherein, The increase in the cleaning frequency is positively correlated with the heterogeneity characterization parameters of the track bed. The increase in tamping intensity is negatively correlated with the vibration response approach parameter.

Images