Railway vehicle operation safety diagnosis system and diagnosis method

Abstract

To provide a railway vehicle operation safety diagnosis system and diagnosis method that can detect abnormal conditions in railway vehicles and tracks in real time, classify the occurrence of abnormal conditions in the railway vehicles or the tracks, and when the occurrence of an abnormal condition is diagnosed, perform emergency measures such as decelerating or stopping the train that includes the relevant vehicle or the train that travels on the relevant track, and perform inspections and maintenance of the part where the abnormal condition was diagnosed.SOLUTION: Vibrations generated in the traveling train are measured and analyzed through one or more vibration sensors installed in each vehicle of the train, and operation information including traveling position and traveling speed of the vehicle is added, then transmitted to a central control system. By analyzing integrated operation data through pre-learned artificial intelligence, locations where abnormalities occur in the vehicle or the track are predicted, and artificial intelligence is re-learned through big data analysis.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a diagnostic system for diagnosing abnormalities affecting the safety of running, such as the occurrence of derailment of a railroad vehicle, in real time, and more particularly, to a system and method for diagnosing the safety of running a railroad vehicle, which measures and analyzes vibrations generated in a running vehicle through one or more vibration sensors installed in each train car, adds operation information including the running position and running speed of the vehicle, transmits the information to a central control system, predicts the location of an abnormality in the vehicle or track by analyzing the integrated operation data through pre-trained artificial intelligence, and re-trains the artificial intelligence through big data analysis. [Background technology]

[0002] A railway is a track facility for quickly transporting large quantities of passengers and goods using railway vehicles. A pair of parallel tracks on the ground guide the direction of travel of wheels that are arranged opposite each other along both sides of the railway vehicle, allowing the railway vehicle to travel in a straight line or on curves.

[0003] The wheels, which are connected to the power source of the railroad vehicle, roll on the tracks due to frictional force generated when the wheels come into contact with the tracks, and the railroad vehicle travels along the tracks. To prevent the wheels from coming off the tracks while the railroad vehicle is traveling, the inner wheel treads of both sides of the wheels that come into contact with the tracks protrude outward from the wheel radius to form flanges.

[0004] When a railroad vehicle is in operation, the wheels move along the upper surface of the track to allow the vehicle to travel, and the lateral movement direction relative to the direction of travel is controlled by flanges. This means that the railroad vehicle is constantly at risk of derailment, which can cause a safety accident. In order to prevent derailment accidents caused by abnormalities in the railroad vehicle, the running stability of the railroad vehicle is ensured through a diagnostic system that monitors the condition of the railroad vehicle in real time.

[0005] Vehicle diagnostic systems for diagnosing abnormal conditions while a railcar is in operation usually follow the Condition Based Maintenance (CBM) system, in which sensors and diagnostic devices are installed at key locations that affect the safety of the railcar's operation, and detection signals that are measured through the sensors and diagnostic devices while the railcar is in operation are transmitted to the engineer or the vehicle control system. When an abnormal condition is detected in the vehicle, emergency measures such as decelerating or emergency stop are taken.

[0006] As an example of such a railway vehicle diagnostic system, Korean Patent Publication No. 10-2235625 (registered on March 29, 2021) proposes a railway vehicle abnormal condition diagnosis system in which, when the wheels of a railway vehicle pass through a diagnosis section, the track is vibrated through a vibration unit to generate a sine wave with a certain period on the track, the vibration generated in the railway vehicle by the vibration is measured through a measurement unit, and the abnormal condition of the railway vehicle is diagnosed based on the measured vibration.

[0007] Due to the nature of trains consisting of many coupled cars, a derailment can result in numerous personal injuries, and so it is necessary to continuously monitor abnormal conditions while the train is in operation. However, such railcar diagnostic systems have a problem in that they cannot diagnose abnormal conditions of the railcar in sections other than the diagnostic section.

[0008] In addition, even if a railcar is operating normally, a derailment can occur if a problem occurs on the tracks. However, existing railcar diagnostic systems that check for abnormal conditions in railcars by analyzing vibration patterns measured using vibration sensors have the problem of not being able to detect abnormalities on the tracks. [Prior art documents] [Patent documents]

[0009] (Patent Document 1) Republic of Korea Registered Patent Publication No. 10-2235625 (registered on March 29, 2021) Summary of the Invention [Problem to be solved by the invention]

[0010] The present invention aims to provide a railway vehicle operation safety diagnosis system and method that detects abnormal conditions of railway vehicles and tracks in real time, classifies and detects the occurrence of abnormal conditions of railway vehicles or tracks, and when an abnormal condition is diagnosed, performs emergency measures such as decelerating or emergency stopping a train including the corresponding vehicle or a train operating on the corresponding track, and performs inspection and maintenance of parts where an abnormal condition has been diagnosed. [Means for solving the problem]

[0011] The railway vehicle operation safety diagnosis system according to the present invention includes a vibration diagnosis module that measures vibrations generated in a vehicle during operation through one or more vibration sensors installed in each vehicle of the train, acquires a vibration signal, and analyzes the vibration signal to form vibration data; a vehicle computer that is installed in each vehicle and receives the vibration data formed by the vibration diagnosis module of the corresponding vehicle and generates operation data by adding operation information including the vehicle's running position and running speed; a train comprehensive information device that aggregates the operation data received from the vehicle computers of the vehicles constituting the train, generates integrated operation data, and transmits it to a central control system; a server that stores the integrated operation data transmitted to the central control system and constructs big data; and a big data analysis module that analyzes the integrated operation data through a pre-trained artificial intelligence, predicts an abnormality occurrence location of a vehicle or a track, transmits it to the central control system, and re-learns the artificial intelligence from data obtained by comparing the predicted result with the actual measurement result obtained by an abnormality occurrence inspection; and the operation data and the integrated operation data are obtained by dividing the entire section of the track on which the train operates into certain sections, and generating vibration data measured in the vehicle for each section of the track, so that the big data analysis module predicts an abnormality occurrence location of a specific vehicle or track.

[0012] According to an embodiment of the present invention, the vibration data is generated through the vibration diagnosis module through an analysis method including any one of the following representative values ​​that can represent the dynamic state of the vehicle: RMS, vibration level, ride comfort coefficient, and ride comfort level.

[0013] According to an embodiment of the present invention, the vibration sensors of the vibration diagnostic module are mounted on the main vehicle parts including the wheel sets or bogies.

[0014] According to an embodiment of the present invention, the vibration diagnosis module determines whether the measured vibration data falls within a preset allowable standard range of vibration values ​​for the corresponding track section and transmits the determination result together with the vibration data to the vehicle computer, and the set allowable standard range of vibration values ​​is continuously updated by feeding back the artificial intelligence re-learning results through big data analysis by the big data analysis module.

[0015] According to an embodiment of the present invention, the allowable standard range of vibration values ​​is set by classification including the type of train or vehicle, the model of train or vehicle, or the number of vehicles in a train set, and the integrated operation data transmitted from the train information device to the central control system includes type-specific data including the type of train or vehicle, the model of train or vehicle, or the number of vehicles in a train set.

[0016] According to an embodiment of the present invention, the big data analysis module diagnoses a specific track section as a dangerous section when abnormal vibration data is measured from multiple trains operating on the specific track section, or diagnoses a specific vehicle as having an abnormality when abnormal vibration data is measured from all tracks, thereby constructing an abnormality diagnosis map of the operating track and operating vehicles and predicting the location of the abnormality on the track or vehicle.

[0017] According to an embodiment of the present invention, re-learning of artificial intelligence through a big data analysis module is performed in the following order: a data selection process of integrated operational data stored on a server, a pre-processing process of the selected data, a conversion process of the pre-processed data, a data mining process, and a pattern analysis process.

[0018] According to an embodiment of the present invention, the pattern analysis process comprises any one of the following methods selected from linear regression, artificial neural network, k-nearest neighbor method, or unsupervised learning.

[0019] The railway vehicle operation safety diagnosis method according to the present invention includes a vibration data forming step of measuring vibrations generated in a moving vehicle through a vibration diagnosis module that receives a measurement signal from one or more vibration sensors installed in each vehicle of the train, acquiring a vibration signal, and analyzing the vibration signal to form vibration data; an operation data generating step of receiving the vibration data formed by the vibration diagnosis module of the corresponding vehicle through a vehicle computer installed in each vehicle and generating operation data by adding operation information including the running position and running speed of the vehicle; a step of aggregating the operation data received from the vehicle computers of the vehicles constituting the train through a train comprehensive information device to generate integrated operation data, and then generating the integrated operation data to be transmitted to a central control system; and a step of storing the integrated operation data transmitted to the central control system on a server and analyzing the big data. The integrated operation data transmitted is analyzed through the pre-trained artificial intelligence included in the AI, and when analyzing the integrated operation data, the entire section of the track on which the train is operating is divided into certain sections and vibration data measured by the vehicle is divided and generated by each section of the track, thereby diagnosing the corresponding track section as a dangerous section when abnormal vibration data is measured from multiple trains operating in the specific track section, or diagnosing that an abnormality has occurred in the corresponding vehicle when abnormal vibration data is measured from a specific vehicle operating in all track sections, and providing the predicted track or vehicle abnormality occurrence result to the central control system; an abnormality occurrence response step in which the central control system takes response measures according to the abnormality occurrence situation when a track or vehicle abnormality is predicted; and a re-learning step in which the artificial intelligence is re-learned through a comparison between the predicted result and the actual measurement result by abnormality occurrence inspection.

[0020] According to an embodiment of the present invention, in the vibration data formation step, the vibration diagnosis module determines whether the measured vibration data falls within a preset allowable standard range of vibration values ​​for the corresponding track section, and transmits the determination result together with the vibration data to the vehicle computer. In the re-learning step, the artificial intelligence re-learning result through big data analysis by the big data analysis module is fed back to the vibration diagnosis module, and the set allowable standard range of vibration values ​​is continuously updated.

[0021] According to an embodiment of the present invention, in the re-learning step, the re-learning of artificial intelligence through a big data analysis module is performed in the following order: a data selection process of integrated operation data stored in a server, a pre-processing process of the selected data, a conversion process of the pre-processed data, a data mining process, and a pattern analysis process, and the pattern analysis process is performed through any one of linear regression, artificial neural network, k-nearest neighbor method, or unsupervised learning. Effect of the Invention

[0022] According to the present invention, the entire section of the track on which the train runs is divided into certain sections according to the topographical conditions and the necessary conditions for safe operation management and maintenance, and section units are generated for each track. Vibration data is individually generated for each set section unit of the track from the vibration signal acquired through the vibration sensor. This makes it possible to distinguish and detect abnormal conditions in the vehicle or track when diagnosing the occurrence of an abnormality through vibration data analysis, and makes it possible to identify the position of the track where the abnormality occurred, thereby enabling quick response measures to be taken for the abnormality.

[0023] According to the present invention, it is possible to improve the accuracy of predicting abnormalities in vehicles or tracks through a railway vehicle operation safety diagnosis system through big data analysis of operation data generated from vibration signals acquired from each vehicle of a train.

[0024] According to the present invention, artificial intelligence that predicts the occurrence of vehicle and track abnormalities is re-learned using data acquired through big data analysis of operational data, and the allowable standard range of vibration values ​​acquired through the artificial intelligence re-learning is fed back and periodically updated, thereby further improving the accuracy of railway vehicle operation safety diagnosis. [Brief description of the drawings]

[0025] [Figure 1] 1 is a diagram showing an overall structure of a railway vehicle operation safety diagnosis system according to the present invention. [Diagram 2] 4 is a diagram illustrating a data compilation process performed in the process of generating integrated operation data from vibration signals measured from vibration sensors mounted on a vehicle in the railway vehicle operation safety diagnosis system according to the present invention; [Diagram 3] 1 is a diagram showing a track section unit formed by dividing a track in order to classify an abnormality occurring on a track or vehicle in the railway vehicle operation safety diagnosis system according to the present invention and to identify the track on which an abnormal state occurs when an abnormal state of the track is detected; [Figure 4] 1 is a diagram showing an example of identifying a section location of a track where an abnormality is predicted to occur by comparing vibration data measured by a vibration sensor according to the present invention with an allowable reference range of vibration values; [Diagram 5] 1 is a diagram showing an example of identifying a track on which an abnormality occurs based on vibration data measured from a plurality of trains in the railway vehicle operation safety diagnosis system according to the present invention; [Figure 6] 1 is a diagram illustrating an example of identifying a vehicle in which an abnormality has occurred based on vibration data measured from a specific vehicle of a train in operation in the railway vehicle operation safety diagnosis system according to the present invention; [Figure 7] 1 is a diagram showing a sequence of a railway vehicle operation safety diagnosis method according to the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0026] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings.

[0027] The following detailed description will focus on the parts necessary for understanding the operation and function of the present invention.

[0028] In describing the embodiments of the present invention, technical details that are widely known in the technical field to which the present invention pertains and are not directly related to the present invention will not be described.

[0029] This is for the purpose of more clearly communicating the gist of the present invention without obscuring it by omitting unnecessary explanations.

[0030] In addition, in describing components of the present invention, components with the same name may be given different reference numbers depending on the drawings, and the same reference number may be given to components in different drawings.

[0031] However, even in such cases, this does not mean that the corresponding components have different functions depending on the embodiment, or that they have the same function in different embodiments, and the function of each component should be determined based on the description of each component in the corresponding embodiment.

[0032] Furthermore, unless otherwise defined in this specification, technical terms used in this specification should be interpreted in the sense that they are commonly understood by a person having ordinary knowledge in the technical field to which this invention belongs, and should not be interpreted in an overly comprehensive or narrow sense.

[0033] Furthermore, as used herein, singular expressions include plural expressions unless otherwise indicated in the context.

[0034] In this application, terms such as "comprise" or "include" should not be construed as including all of the various components or steps described in the specification, but should be construed as meaning that some of the components or steps may not be included, or that additional components or steps may be further included.

[0035]

[0036] In this specification, the train 10 and the vehicle 11 are formed by combining a bogie, which is a running device consisting of wheels and axles, an axle box assembly, a suspension system, a traction system, a braking system and a coupling, with a car body on which passengers board and driving equipment are mounted, and the train 10 refers to a single railroad vehicle formed by connecting a number of vehicles 11 in a row through couplings.

[0037] The railway vehicle operation safety diagnosis system according to the present invention is configured to generate operation data from vibration signals acquired from each vehicle 11 constituting a train 10 as shown in FIG. 1, transmit the operation data to a central control system 1, analyze the operation data through a big data analysis module 200, predict where an abnormality will occur in the vehicle 11 or the track, and analyze the operation data stored in a server 2 and convert it into big data, thereby improving the accuracy of predicting the occurrence of an abnormality in the vehicle 11 or the track through the railway vehicle operation safety diagnosis system.

[0038] As shown in FIG. 2, one or more vibration sensors 101 are installed in each car 11 of the train 10 to measure vibrations transmitted from the tracks to the car 11 through the wheels or vibrations generated in the car 11 to obtain vibration signals. The vibration signals obtained through the vibration sensors 101 are transmitted to a vibration diagnosis module 110 mounted on each car 11, and the measured vibration signals are analyzed through the vibration diagnosis module 110 to form vibration data.

[0039] It is preferable that the vibration sensors 111 are mounted on the wheelset, bogie, and car body, which are the main parts of each vehicle 11, but they may be mounted only at limited locations on the wheelset and bogie in order to reduce the increase in cost of building a railway vehicle operation safety diagnosis system due to the installation of a large number of vibration sensors 111, the overload of vibration signal analysis and vibration data transmission in the vibration diagnosis module 110, or the overload of the storage capacity of the big data formed by the operation data stored in the server 2.

[0040] Vibration data is formed through the vibration diagnosis module 110 through the analysis of representative values ​​that can represent the dynamic state of the vehicle 11. Representative values ​​obtained from the vibration sensor 111 include RMS (Root Mean Square), vibration level, ride comfort coefficient, ride comfort level, etc., and vibration data is formed through any one of the analysis methods selected from these.

[0041] RMS represents the effective output that is the standard for the average output of the vibration signal continuously output through the vibration sensor 111, and is calculated by taking the arithmetic mean of the squared vibration signal output value and then taking the square root of the arithmetic mean value, and the vibration level is expressed by converting the strength of the vibration signal measured through the vibration sensor 111 into decibels (dB) as an absolute value.

[0042] The ride comfort coefficient is calculated by passing a vibration acceleration signal due to the linear vibration in the front-rear direction, the linear vibration in the left-right direction, and the linear vibration in the up-down direction of the vehicle 11 measured by the vibration sensor 111 through a weighting filter, calculating an effective value at regular intervals to obtain a cumulative frequency distribution, calculating the vibration acceleration corresponding to 95% of the cumulative frequency distribution, and substituting the obtained value into the following ride comfort coefficient formula.

[0043] Ride comfort coefficient TIFF2025015512000002.tif12155

[0044] At this time, TIFF2025015512000003.tif1730 shows the vibration acceleration that corresponds to 95% of the cumulative frequency distribution of forward and backward vibrations. TIFF2025015512000004.tif1433 shows the vibration acceleration that corresponds to 95% of the cumulative frequency distribution of lateral vibration. TIFF2025015512000005.tif1332 represents the vibration acceleration that corresponds to 95% of the cumulative frequency distribution of vertical vibration.

[0045] The ride comfort level is expressed in decibels (dB) by applying a ride comfort filter corresponding to the human perception evaluation sensitivity to the effective value of the body vibration acceleration of the vehicle 11 measured through the vibration sensor 111 based on ISO2631, and is calculated by substituting it into the following ride comfort level formula.

[0046] Ride comfort level TIFF2025015512000006.tif14151

[0047] At this time, TIFF2025015512000007.tif1419 is the effective value of vibration acceleration with the ride feel filter applied, TIFF2025015512000008.tif1626 shows the reference value of vibration with the ride feel filter applied.

[0048] In addition, the vibration diagnosis module 110 compares the vibration data formed in the above manner with an allowable reference range of vibration values ​​preset in the vibration diagnosis module 110 to determine whether the data is within the allowable reference range or exceeds the allowable reference range.

[0049] Since railroad tracks are installed in various topographical environments, such as flat ground, tunnels, or bridges, the vibration data acquired through the vibration diagnosis module 110 may be formed differently for each section of the railroad tracks. If the vibration diagnosis module 110 is uniformly set to a preset allowable reference range for vibration values, it may be possible that vibration data acquired from a railroad track or vehicle 11 in a normal state may be mistaken for an abnormal state depending on the topographical environment in which the railroad tracks are installed, or conversely, vibration data acquired from a railroad track or vehicle 11 in an abnormal state may be mistaken for a normal state.

[0050] Therefore, as shown in FIG. 3, the entire section of the track on which the train 10 runs is divided into sections at certain locations according to the topographical conditions and the necessary conditions for safe operation management and maintenance, and section units are generated for each track, and accordingly, the vibration signal measured by the vibration sensor 111 and the vibration data acquired by the vibration diagnosis module 110 are also generated separately for each section unit of the track.

[0051] Also, the allowable reference range of vibration values ​​preset in the vibration diagnosis module 110 is set individually for each corresponding line section.

[0052] In addition, as the type of vehicle 11 changes, such as the type of vehicle 11 such as a locomotive, passenger car, or freight car, the model of train 10 or vehicle 11 consisting of a specific model, the number of vehicles 11 in a train, the connection method of vehicles 11, or changes in the vibration pattern of vehicle 11 due to the application of diesel or electric power, the vibration signal measured through vibration sensor 111 and the vibration data acquired by vibration diagnosis module 110 change. Therefore, the allowable standard range of vibration values ​​preset in vibration diagnosis module 110 is individually set by classifying it according to various types including the type of vehicle 11 such as a locomotive, passenger car, or freight car, the model of train 10 or vehicle 11, the number of vehicles 11 in a train, the connection method of vehicles 11, or changes in the vibration pattern of vehicle 11 due to the application of diesel or electric power.

[0053] The vibration data formed by the vibration diagnosis module 110 is transmitted to the vehicle computer 100 installed in each vehicle 11, and each vehicle computer 100 receives the vibration data formed by the vibration diagnosis module 110 of the corresponding vehicle 11, and then generates operation data by adding operation information including the running position and running speed of the vehicle 11 obtained from the GPS data transmitted through the Train Control and Monitoring System (TCMS, 12).

[0054] The operation information added to the operation data makes it possible to identify the location of the track section where the vibration data included in the operation data was obtained, thereby enabling emergency measures to be taken quickly to carry out inspection and maintenance work on the relevant track when an abnormality in the track is determined.

[0055] The operation data generated by each vehicle computer 100 installed in each vehicle 11 that makes up the train 10 is transmitted to the train information device 12, which consolidates the received operation data to generate integrated operation data, and then transmits it to the central control system 1.

[0056] The integrated operation data transmitted from the train comprehensive information device 12 of each train 10 is transmitted to and stored in a server 2 connected to the central control system 1 via a network, and big data for a railway vehicle operation safety diagnosis system is constructed from the integrated operation data accumulated and stored in the server 2.

[0057] The big data analysis module 200 is connected to the central control system 1 and the server 2 via a network, receives integrated operation data transmitted from the train comprehensive information device 12 of each train 10 in real time, and analyzes the integrated operation data through pre-trained artificial intelligence.

[0058] If the big data analysis module 200 analyzes the integrated operation data and determines that the vibration data measurement value exceeds the allowable standard range of vibration values ​​and an abnormality is predicted in the vehicle 11 or the track, as shown in Figure 4, it identifies the vehicle 11 of the train 10 or the section of the track where the abnormality is predicted to occur and transmits the prediction result to the central control system 1.

[0059] The central control system 1 that receives the prediction result performs emergency measures such as slowing down or making an emergency stop on the train 10 connected to the vehicle 11 in which an abnormality is predicted to occur, depending on the severity of the symptoms of the prediction result, or performs emergency measures such as slowing down or making an emergency stop on the train 10 approaching the section of track in which an abnormality is predicted to occur.

[0060] When slowing down the train, personnel for inspection and maintenance work are deployed to the track section or vehicle 11 where the abnormality is predicted to occur after the end of railway operating hours, and when emergency stop measures are taken, personnel for inspection and maintenance work are immediately deployed to the relevant track section or vehicle 11.

[0061] In addition, the big data analysis module 200 compares the predicted abnormality occurrence results for the vehicles 11 and the tracks with the actual measurement results of the abnormality occurrence inspection as a result of the implementation of emergency measures by the central control system 1 to obtain big data on the accuracy of the abnormality occurrence prediction analysis, and the artificial intelligence is re-trained from the obtained data.

[0062] The acceptable standard range of vibration values ​​acquired as a result of artificial intelligence re-learning through big data analysis by the big data analysis module 200 is fed back into the acceptable standard range of vibration values ​​set in the vibration diagnosis module 110 and the big data analysis module 200, thereby updating the previously set acceptable standard range of vibration values ​​of the vibration diagnosis module 110 and the big data analysis module 200.

[0063] At regular intervals set by the railway vehicle operation safety diagnosis system administrator, artificial intelligence re-learning is performed through big data analysis of integrated operation data that is periodically accumulated and stored on the server 2, and the allowable standard range of vibration values ​​is continuously updated, thereby improving the accuracy of the big data analysis module 200's prediction results for vehicle 11 and track abnormality occurrence.

[0064] The re-learning of artificial intelligence through the big data analysis module 200 is carried out in the following order: a process of selecting data from the integrated operational data stored in the server 2, a pre-processing process of pre-processing the selected data, a conversion process of converting the pre-processed data, a data mining process, and a pattern analysis process.

[0065] In the pattern analysis process, the pattern analysis is performed using any one of linear regression, artificial neural network, k-nearest neighbor, or unsupervised learning.

[0066] Linear regression is an analytical technique that models a linear correlation (regression equation) of already known data values, such as vibration data measured and acquired through the vibration sensor 111 and integrated operation data, using techniques such as the least squares method, and then estimates unknown data and parameters from the modeled linear correlation.

[0067] An artificial neural network is a machine learning model that mimics the structure of human neurons through software, and embodies artificial intelligence by iteratively learning in a direction that is expected to minimize errors with pre-trained data and updating weights and thresholds.

[0068] The k-nearest neighbor algorithm is an algorithm that uses Euclidean distance calculation to classify data by referring to the labels of other data that are close to the data. It searches for K pieces of data adjacent to the data to be classified and classifies the data into the category where the corresponding data has the most labels. Since the algorithm is simple, the training stage is quick and it shows excellent performance in numerical-based data classification work.

[0069] Since the k-nearest neighbor algorithm is based on a distance-based model, in order to facilitate the analysis of the differences between each variable, normalization such as min-max normalization or z-score standardization needs to be applied to rescale the variable values ​​to a standard range, and the number of K is preferably set to an odd number for classification of the data.

[0070] Unsupervised learning is an analytical technique that uses machine learning algorithms to analyze unlabeled data sets and group data that have the same or similar characteristics into clusters, allowing the system to learn and improve autonomously without explicit programming by providing it with large amounts of data.

[0071] At this time, the allowable standard range of vibration values ​​is set according to various types, such as the vehicle type of train 10 or vehicle 11, such as KTX or SRT, or a detailed model classification thereof, the model type of train 10 or vehicle 11, such as a power-centralized KTX or a power-distributed KTX joint, and the number of vehicles 11, such as a 10-car KTX or a 20-car KTX. When predicting the occurrence of abnormality in vehicles 11 or on tracks by comparing the vibration data with the allowable standard range of vibration values ​​in the vibration diagnosis module 110, or by comparing the integrated operation data with the allowable standard range of vibration values ​​in the big data analysis module 200, the allowable standard range of vibration values ​​set according to type is applied to minimize the occurrence of errors in the prediction results of abnormality in vehicles 11 or on tracks.

[0072] In addition, unlike existing vehicle diagnostic systems that use various sensors to check the status of major devices of the vehicle in real time and apply a Condition Based Maintenance (CBM) method to collect and analyze fault history and maintenance operation data, the railway vehicle operation safety diagnostic system of the present invention can not only diagnose abnormalities in the vehicle 11 but also predict abnormalities in the track, and the process of distinguishing and predicting abnormalities in the vehicle 11 and the track from vibration data measured through the same vibration sensor 111 is performed as follows.

[0073] As shown in Figure 5, when vibration data measured from multiple trains 10 operating on a specific track section exceeds the allowable standard range of vibration values ​​for the track section and the type of vehicle 11 of the train 10, indicating that abnormal vibration data has occurred, the track section is diagnosed as a dangerous section, and in order to prevent errors from occurring when determining whether the track is a dangerous section, the presence or absence of a track abnormality is determined based on vibration data measured from at least three or more trains 10.

[0074] As shown in Figure 6, if vibration data measured from a specific vehicle 11 of a train 10 operating along all track sections is measured to have exceeded the allowable standard range of vibration values ​​for the track section and the type of vehicle 11 of the train 10 periodically or intermittently in multiple track sections, indicating that abnormal vibration data has occurred, and if abnormal vibration data of the same or similar pattern does not occur in the vehicles 11 of other trains 10, it is diagnosed that an abnormality has occurred in the vehicle 11 in which the abnormal vibration data was measured.

[0075] In order to classify and predict the occurrence of abnormalities in the vehicle 11 or track, the big data analysis module 200 constructs an abnormality diagnosis map of the operating track and the operating vehicle 11 through the above-mentioned judgment method, and classifies and diagnoses the occurrence of abnormalities in the track or vehicle 11 based on the abnormality diagnosis map, and accurately identifies the track section or vehicle 11 where the abnormal condition occurred.

[0076]

[0077] The railway vehicle operation safety diagnosis method using the railway vehicle operation safety diagnosis system configured as above is carried out in the steps shown in FIG.

[0078] In the vibration data formation step (S10), the vibration diagnosis module 110 receives measurement signals from the vibration sensors 111 installed in each car 11 of the train 10 to obtain vibration signals generated in the cars 11 during operation, and analyzes the vibration signals through the vibration diagnosis module 110 to form vibration data.

[0079] In the operation data generation step (S20), the vibration data formed in the vibration data formation step (S10) is received by a vehicle computer 100 mounted on the same vehicle 11 as the vehicle 11 that acquired the vibration data, and operation information including the running position, which is the position of the track section on which the vehicle 11 was running when the vibration data was acquired, and the running speed, which is the speed at which the vehicle 11 passes through the track section, is added to the vibration data to generate operation data.

[0080] In the step of generating integrated operation data (S30), the operation data received from the vehicle computers 100 of each vehicle 11 constituting the train 10 in the operation data generation step (S20) is compiled through the train comprehensive information device 12 to generate integrated operation data, and the train comprehensive information device 12 transmits the generated integrated operation data to the central control system 1.

[0081] In the abnormality occurrence prediction step (S40), the integrated operation data transmitted from the train comprehensive information device 12 to the central control system 1 is stored on the server 2, and the transmitted integrated operation data is analyzed in real time through the pre-trained artificial intelligence included in the big data analysis module 200 to diagnose the occurrence of an abnormality in the track or vehicle 11, and then the abnormality occurrence prediction result for the track or vehicle 11 is provided to the central control system 1.

[0082] In the abnormality response step (S50), if it is predicted that an abnormality has occurred on the track or vehicle 11 based on the abnormality prediction result predicted in the abnormality prediction step (S40), the central control system 1 performs response measures including decelerating or emergency stopping the train 10 running on the track or the train 10 including the relevant vehicle 11, and deploying personnel for inspection and maintenance work, depending on the type of abnormality situation and the severity of the abnormality symptoms, and when the inspection of the track or vehicle 11 in response to the abnormality is completed, the actual measurement results generated by the inspection are stored on the server 2.

[0083] In the re-learning step (S60), the artificial intelligence is re-learned through a comparison between the prediction results and the actual measurement results generated as a result of the inspection of the track or vehicle 11. After the re-learning is completed, the optimized allowable standard range of vibration values ​​is transmitted to the vibration diagnosis module 110 and the big data analysis module 200 to update the existing allowable standard range of vibration values.

[0084] The re-learning step (S60) is carried out in the following order: a data selection process (S61) for selecting integrated driving data stored in the server 2, a pre-processing process (S62) for pre-processing the selected data, a conversion process (S63) for the pre-processed data, a data mining process (S64), and a pattern analysis process (S65).

[0085] The optimization of the allowable standard range of vibration values ​​by performing the re-learning step (S60) and the updating of the existing allowable standard range of vibration values ​​in the vibration diagnosis module 110 and the big data analysis module 200 are performed automatically at intervals set by the administrator of the railway vehicle operation safety diagnosis system. Through continuous updates of the allowable standard range of vibration values, the accuracy of prediction results of abnormality occurrence in vehicles 11 and tracks is improved, thereby reducing the occurrence rate of safety accidents such as train 11 derailments.

[0086]

[0087] Although the embodiments of the present invention have been described with reference to the above, it will be understood by those skilled in the art to which the present invention pertains that the present invention may be embodied in other specific forms without changing the technical spirit or essential characteristics thereof.

[0088] Therefore, the above-described embodiments should be understood to be illustrative and not limiting in all respects, and the scope of the present invention described in the above detailed description is indicated by the claims below. All modifications and variations derived from the meaning and scope of the claims and their equivalent concepts should be construed as being included in the scope of the present invention. [Explanation of symbols]

[0089] 1: Central Control System 2: Server 10:Train 11: Vehicle 12:Train information system 100: Vehicle computer 110: Vibration diagnostic module 101: Vibration sensor 200: Big Data Analysis Module

Claims

1. a vibration diagnosis module (110) for measuring vibrations generated in the moving vehicles (11) through one or more vibration sensors (101) installed in each vehicle (11) of the train (10) to obtain vibration signals, and analyzing the vibration signals to form vibration data; A vehicle computer (100) that is installed in each vehicle (11) and receives vibration data formed by a vibration diagnosis module (110) of the vehicle (11) and generates operation data by adding operation information including the traveling position and traveling speed of the vehicle (11); a train control and monitoring system (TCMS, 12) that consolidates operation data received from the vehicle computers (100) of the vehicles (11) that make up the train (10) to generate integrated operation data, and then transmits the integrated operation data to the central control system 1; A server (2) for storing the integrated operation data transmitted to the central control system 1 and constructing big data; A big data analysis module (200) that predicts abnormality occurrence locations of the vehicle (11) or the track by analyzing the integrated operation data through a pre-trained artificial intelligence, transmits the prediction results to the central control system (1), and re-trains the artificial intelligence from data obtained by comparing the prediction results with the actual measurement results from the abnormality occurrence inspection; Although it is composed of The operation data and integrated operation data are generated by dividing the entire section of the track on which the train (10) operates into sections at certain locations, and the vibration data measured in the vehicle (11) is generated for each section of the track, thereby allowing the big data analysis module (200) to predict where an abnormality will occur in a specific vehicle (11) or track. This is a railway vehicle operation safety diagnosis system.

2. The formation of vibration data through the vibration diagnostic module (110) includes:

2. The railway vehicle operation safety diagnosis system according to claim 1, wherein the diagnosis is performed through an analysis method including any one of a root mean square (RMS), a vibration level, a ride comfort coefficient, and a ride comfort level, which are representative values ​​that can represent the dynamic state of the vehicle (11).

3. 2. The railway vehicle operation safety diagnosis system according to claim 1, wherein the vibration sensor (111) of the vibration diagnosis module (110) is mounted on a main part of the vehicle (11) including a wheel set or a bogie.

4. 2. The railway vehicle operation safety diagnosis system of claim 1, wherein the vibration diagnosis module (110) determines whether the measured vibration data falls within a preset allowable standard range of vibration values ​​for a corresponding track section, and transmits the determination result together with the vibration data to the vehicle computer (100), and the set allowable standard range of vibration values ​​is continuously updated by feeding back the artificial intelligence re-learning result through big data analysis of the big data analysis module (200).

5. The allowable standard range of vibration values ​​is set by classification including the type of train (10) or vehicle (11), the model of the train (10) or vehicle (11), or the number of vehicles (11) in a train configuration; 5. The railway vehicle operation safety diagnosis system according to claim 4, wherein the integrated operation data transmitted from the train comprehensive information device (12) to the central control system 1 includes type-specific data including the type of train (10) or vehicle (11), the model of the train (10) or vehicle (11), or the number of vehicles (11) in a train configuration.

6. The big data analysis module (200) When abnormal vibration data is measured from multiple trains (10) operating on a specific track section, the track section is diagnosed as a dangerous section, By diagnosing that an abnormality has occurred in a specific vehicle (11) in operation on all track sections when abnormal vibration data is measured, 2. The railway vehicle operation safety diagnosis system according to claim 1, further comprising: constructing an abnormality diagnosis map for the operating track and the operating vehicle (11) to predict an abnormality occurrence location of the track or the vehicle (11).

7. Re-learning artificial intelligence through the big data analysis module (200) 2. The railway vehicle operation safety diagnosis system according to claim 1, characterized in that the steps of: a data selection process of the integrated operation data stored in the server (2), a pre-processing process of the selected data, a conversion process of the pre-processed data, a data mining process, and a pattern analysis process are performed in this order.

8. The pattern analysis process is as follows:

8. The railway vehicle operation safety diagnosis system according to claim 7, characterized in that the diagnosis method is any one selected from the group consisting of linear regression, artificial neural network, k-nearest neighbor, and unsupervised learning.

9. A vibration data forming step (S10) of measuring vibrations occurring in the running vehicles (11) through a vibration diagnosis module (110) receiving a measurement signal from one or more vibration sensors 111 installed in each vehicle (11) of the train (10), acquiring a vibration signal, and analyzing the vibration signal to form vibration data; A driving data generating step (S20) of receiving vibration data formed by the vibration diagnosis module (110) of the vehicle (11) through the vehicle computer (100) installed in each vehicle (11) and generating driving data by adding driving information including the driving position and driving speed of the vehicle (11); A step of generating integrated operation data by aggregating operation data received from the vehicle computers (100) of the vehicles (11) constituting the train (10) through the train total information device (TCMS, 12) and then generating the integrated operation data to be transmitted to the central control system 1 (S30); An abnormality prediction step (S40) of storing the integrated operation data transmitted to the central control system 1 on the server (2) and analyzing the integrated operation data through a pre-trained artificial intelligence included in the big data analysis module (200), and dividing the entire section of the track on which the train (10) runs into certain sections and dividing and generating vibration data measured in the vehicle (11) by section unit of the track when analyzing the integrated operation data, thereby diagnosing the corresponding track section as a dangerous section when abnormal vibration data is measured from a plurality of trains (10) running in the specific track section, or diagnosing that an abnormality has occurred in the corresponding vehicle (11) when abnormal vibration data is measured from a specific vehicle (11) running in all track sections, and providing a track or vehicle (11) abnormality occurrence prediction result to the central control system 1; An abnormality response step (S50) in which the central control system 1 performs a response measure according to the abnormality occurrence situation when an abnormality occurrence is predicted for the track or vehicle (11); a re-learning step (S60) of re-learning the artificial intelligence through a comparison between the prediction result and the actual measurement result by the abnormality occurrence inspection.

10. In the vibration data forming step (S10), the vibration diagnosis module (200) determines whether the measured vibration data falls within a predetermined allowable standard range of vibration values ​​for the corresponding track section, and transmits the determination result together with the vibration data to the vehicle computer (100).

10. The railway vehicle operation safety diagnosis method of claim 9, wherein in the re-learning step (S60), the artificial intelligence re-learning result through the big data analysis of the big data analysis module (200) is fed back to the vibration diagnosis module (110) to continuously update the set allowable standard range of the vibration value.

11. In the re-learning step (S60), the re-learning of the artificial intelligence through the big data analysis module (200) is performed in the following order: data selection process (S61) of the integrated driving data stored in the server (2), pre-processing process (S62) of the selected data, conversion process (S63) of the pre-processed data, data mining process (S64), and pattern analysis process (S65).

10. The method of claim 9, wherein the pattern analysis step (S65) is performed through any one of a linear regression, an artificial neural network, a k-nearest neighbor method, and an unsupervised learning method.

Images