CN114633774A - Rail transit fault detection system based on artificial intelligence - Google Patents
Rail transit fault detection system based on artificial intelligence Download PDFInfo
- Publication number
- CN114633774A CN114633774A CN202210332961.7A CN202210332961A CN114633774A CN 114633774 A CN114633774 A CN 114633774A CN 202210332961 A CN202210332961 A CN 202210332961A CN 114633774 A CN114633774 A CN 114633774A
- Authority
- CN
- China
- Prior art keywords
- track
- unit
- detection
- data
- artificial intelligence
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 167
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 20
- 238000012545 processing Methods 0.000 claims abstract description 69
- 238000006073 displacement reaction Methods 0.000 claims abstract description 20
- 238000012544 monitoring process Methods 0.000 claims abstract description 11
- 238000013528 artificial neural network Methods 0.000 claims description 50
- 238000013135 deep learning Methods 0.000 claims description 38
- 238000005452 bending Methods 0.000 claims description 27
- 230000002159 abnormal effect Effects 0.000 claims description 20
- 230000006870 function Effects 0.000 claims description 7
- 230000009471 action Effects 0.000 claims description 5
- 230000002068 genetic effect Effects 0.000 claims description 4
- 238000005516 engineering process Methods 0.000 claims description 2
- 238000004148 unit process Methods 0.000 claims 1
- 238000000034 method Methods 0.000 abstract description 19
- 230000008569 process Effects 0.000 abstract description 17
- 230000005856 abnormality Effects 0.000 description 4
- 238000012423 maintenance Methods 0.000 description 4
- 238000011084 recovery Methods 0.000 description 4
- 238000012360 testing method Methods 0.000 description 3
- 238000013499 data model Methods 0.000 description 2
- 238000005265 energy consumption Methods 0.000 description 2
- 238000009434 installation Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B61—RAILWAYS
- B61K—AUXILIARY EQUIPMENT SPECIALLY ADAPTED FOR RAILWAYS, NOT OTHERWISE PROVIDED FOR
- B61K9/00—Railway vehicle profile gauges; Detecting or indicating overheating of components; Apparatus on locomotives or cars to indicate bad track sections; General design of track recording vehicles
- B61K9/08—Measuring installations for surveying permanent way
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B61—RAILWAYS
- B61L—GUIDING RAILWAY TRAFFIC; ENSURING THE SAFETY OF RAILWAY TRAFFIC
- B61L23/00—Control, warning or like safety means along the route or between vehicles or trains
- B61L23/04—Control, warning or like safety means along the route or between vehicles or trains for monitoring the mechanical state of the route
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B61—RAILWAYS
- B61L—GUIDING RAILWAY TRAFFIC; ENSURING THE SAFETY OF RAILWAY TRAFFIC
- B61L23/00—Control, warning or like safety means along the route or between vehicles or trains
- B61L23/04—Control, warning or like safety means along the route or between vehicles or trains for monitoring the mechanical state of the route
- B61L23/042—Track changes detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B21/00—Alarms responsive to a single specified undesired or abnormal condition and not otherwise provided for
- G08B21/18—Status alarms
- G08B21/24—Reminder alarms, e.g. anti-loss alarms
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B7/00—Radio transmission systems, i.e. using radiation field
- H04B7/14—Relay systems
- H04B7/15—Active relay systems
- H04B7/185—Space-based or airborne stations; Stations for satellite systems
- H04B7/18502—Airborne stations
- H04B7/18506—Communications with or from aircraft, i.e. aeronautical mobile service
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/06—Protocols specially adapted for file transfer, e.g. file transfer protocol [FTP]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/12—Protocols specially adapted for proprietary or special-purpose networking environments, e.g. medical networks, sensor networks, networks in vehicles or remote metering networks
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04Q—SELECTING
- H04Q9/00—Arrangements in telecontrol or telemetry systems for selectively calling a substation from a main station, in which substation desired apparatus is selected for applying a control signal thereto or for obtaining measured values therefrom
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04Q—SELECTING
- H04Q2209/00—Arrangements in telecontrol or telemetry systems
- H04Q2209/30—Arrangements in telecontrol or telemetry systems using a wired architecture
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04Q—SELECTING
- H04Q2209/00—Arrangements in telecontrol or telemetry systems
- H04Q2209/40—Arrangements in telecontrol or telemetry systems using a wireless architecture
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Physics & Mathematics (AREA)
- Signal Processing (AREA)
- Mechanical Engineering (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Theoretical Computer Science (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- Computational Linguistics (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Business, Economics & Management (AREA)
- Emergency Management (AREA)
- Medical Informatics (AREA)
- Astronomy & Astrophysics (AREA)
- Aviation & Aerospace Engineering (AREA)
- Train Traffic Observation, Control, And Security (AREA)
Abstract
The invention belongs to the technical field of safety monitoring, and particularly relates to a rail transit fault detection system based on artificial intelligence, which comprises a cloud big data center and a detection unit; the cloud big data center comprises an analysis unit, a comparison unit, a processing unit and a database, wherein the analysis unit is used for analyzing and classifying the detection data, the comparison unit is used for processing and comparing the detection data to give a detection result, and the processing unit is used for processing according to the detection result; the detection unit comprises displacement sensors, the displacement sensors are arranged on the track along a fixed interval, and the displacement sensors are used for monitoring the position offset of the spikes on the track; according to the invention, the data acquired by the detection unit is processed and compared through artificial intelligence, the states of the track and the train are obtained in time, and workers are informed to process through the processing unit, so that the safety and stability of train running are improved.
Description
Technical Field
The invention belongs to the technical field of safety monitoring, and particularly relates to a rail transit fault detection system based on artificial intelligence.
Background
With the continuous development of economy, the operation forms of rail transit are more and more abundant, rail transit also becomes the preferred choice of people's trip gradually, and because rail transit has the characteristics of great bearing capacity, the monitoring of rail transit operation safety is also more and more important.
Chinese patent CN201510230665.6 discloses a drive board fault detection device in rail transit, which includes a computer, an output end of the computer is connected with an input end of a PLC industrial control board, an output end of the PLC industrial control board is connected with an input end of a second relay and an input end of a third relay respectively, an output end of the second relay is connected with an input end of a second switching power supply, and an output end of the second switching power supply is connected with an input end of an interface board; the output end of the third relay is connected with the input end of the pulse trigger main board; the output end of the pulse trigger mainboard is connected with the input end of the interface board; the output end of the interface board is connected with the input end of the Hall voltage and current sensor, the output end of the Hall voltage and current sensor is connected with the input end of the USB data acquisition card, and the input end of the USB data acquisition card is connected with the input end of the computer, so that the problems that the conventional passenger car drive board is time-consuming and labor-consuming to troubleshoot and low in working efficiency due to manual judgment in fault maintenance are solved;
however, the above patent does not consider that the existing rail transit operation safety monitoring mainly focuses on safety monitoring of rail train driving states, rail deformation and the like, and obviously, the existing rail transit operation safety monitoring still has certain disadvantages.
In view of this, the present invention provides a rail transit fault detection system based on artificial intelligence to solve the above technical problems.
Disclosure of Invention
In order to make up for the defects of the prior art and solve the problems that intelligent monitoring and analysis cannot be carried out on rail faults and the running safety and stability of a train are ensured, the invention provides a rail transit fault detection system based on artificial intelligence.
The technical scheme adopted by the invention for solving the technical problem is as follows: the invention relates to a rail transit fault detection system based on artificial intelligence, which comprises a cloud big data center and a detection unit;
the cloud big data center comprises an analysis unit, a comparison unit, a processing unit and a database, wherein the analysis unit is used for receiving the data uploaded by the detection unit, analyzing and classifying the data, the comparison unit is a deep learning neural network, the comparison unit analyzes the analysis unit, processes and compares the classified data to give a detection result, the detection result is synchronously sent to the processing unit, and the processing unit performs processing action according to the detection result;
the detection unit is arranged on a track and/or a vehicle and exchanges data with the cloud big data center through a wired network and/or a 5G network;
the detection unit comprises a displacement sensor, the displacement sensor is arranged on the track along a fixed distance, and the displacement sensor is used for monitoring the position offset of the spike on the track.
Preferably, the cloud big data center further includes a decision unit, the deep learning neural network in the comparison unit includes but is not limited to any one of a bp (back) neural network, an rbf (radial basis function) neural network, and a genetic algorithm neural network, the comparison unit has odd number of parallel deep learning neural networks with the same structure, data obtained by analysis and classification by the analysis unit is simultaneously transmitted to the plurality of deep learning neural networks in the comparison unit for processing and comparison, results obtained by the plurality of deep learning neural networks in the comparison unit are output to the decision unit, and the decision unit synthesizes results obtained by the deep learning neural networks by using most of the results complying with minority principles to give a final detection result.
Preferably, the detecting unit further comprises a detection unmanned aerial vehicle, the detection result shows that there is an abnormality, then the processing unit sends a detection command to the detection unmanned aerial vehicle, the detection unmanned aerial vehicle patrols and takes a picture of the track according to the planned path, and sends the shot picture and video to the cloud big data center, the picture and video uploaded by the detection unmanned aerial vehicle received by the cloud big data center are firstly analyzed and classified by the analyzing unit, and then processed by the deep learning neural network in the comparison unit to obtain the processing priority that the track has the abnormality, and the processing unit gives an alarm according to the processing priority.
Preferably, be provided with the positioning mark on the track, it shoots to detect unmanned aerial vehicle there is the positioning mark department, the positioning mark is located the spike department on the track, it shoots photo or video with positioning mark and spike simultaneously to detect unmanned aerial vehicle.
Preferably, the detection unmanned aerial vehicle is at fixed height when taking a picture or a video, the detection unmanned aerial vehicle takes two rails of the track simultaneously and enters a picture when taking the picture, the positioning mark also has a position corresponding to the track inner side and the spike, and the detection unmanned aerial vehicle takes the positioning mark on the track inner side together and enters the picture with the two rails.
Preferably, the detection unit further comprises a first vibration sensor, the first vibration sensor is fixedly installed on the track along a fixed interval, the interval between any two vibration sensors is larger than the length of a train running on the track, and the position data of the first vibration sensor is uploaded to the cloud big data center synchronously when the first vibration sensor uploads the data.
Preferably, a second vibration sensor is mounted on a gearbox of the train running on the track.
Preferably, two vibration sensor have the multiunit altogether, vibration sensor two uses the gear box center of train as the initial point along the axial direction equidistance distribution of X axle, Y axle, Z axle respectively, vibration sensor two is the same in the quantity in each axial direction.
Preferably, the detection unit further comprises a grating sensor, the grating sensor is fixedly installed on the track along a fixed interval, the grating sensor is used for detecting bending deformation data of the track in the vertical direction, and the grating sensor uploads the detected bending deformation data to the cloud big data center.
Preferably, the grating sensor and the displacement sensor are integrated into a MEMS sensor having a function of detecting displacement and bending deformation by MEMS technology.
The invention has the following beneficial effects:
1. according to the rail transit fault detection system based on artificial intelligence, the detection unit is arranged to collect different data of a rail and a train during operation, then the collected detection data are sent to the analysis unit through the network for preliminary processing and classification, invalid or redundant data in the data are removed, the data are conveniently processed and compared by the deep learning neural network in the subsequent comparison unit, the detection result is finally obtained, finally the processing unit is used for processing different detection results, and a worker is timely informed to overhaul the rail or the train under the condition that the detection result is abnormal, so that potential safety hazards are avoided, and the safety and the stability of the train during operation are improved.
Drawings
The invention will be further explained with reference to the drawings.
FIG. 1 is a system flow diagram of the present invention.
Detailed Description
As shown in fig. 1, the rail transit fault detection system based on artificial intelligence of the present invention includes a cloud big data center and a detection unit;
the cloud big data center comprises an analysis unit, a comparison unit, a processing unit and a database, wherein the analysis unit is used for receiving the data uploaded by the detection unit, analyzing and classifying the data, the comparison unit is a deep learning neural network, the comparison unit analyzes the analysis unit, processes and compares the classified data, a detection result is given, the detection result is synchronously sent to the processing unit, and the processing unit performs processing action according to the detection result;
the detection unit is arranged on a track and/or a vehicle and exchanges data with the cloud big data center through a wired network and/or a 5G network;
the detection unit comprises displacement sensors, the displacement sensors are arranged on the track along a fixed interval, and the displacement sensors are used for monitoring the position offset of the spikes on the track;
in the working process, data collected by the detection unit is quickly uploaded to a cloud big data center through a wired network or a 5G network, then the received data is preliminarily processed through the analysis unit, so that the data is analyzed and classified, useless and redundant data in the received detection data are removed, the data is conveniently analyzed and compared by using a deep learning neural network in a comparison unit subsequently, an accurate detection result is quickly obtained, meanwhile, after the data preliminarily processed by the analysis unit is sent to the comparison unit, the data is analyzed and compared by using the deep learning neural network, a data model is detected by a component, so that the detection data is predicted according to the data recorded once, then the predicted data is compared with the detection data, and if the predicted data and the detection data are consistent, the detection result is judged to be normal, When the track is not abnormal and the track is not consistent, the detection result is judged to be abnormal, the track is abnormal, meanwhile, the deep learning neural network in the comparison unit can utilize the uploaded detection result to carry out self learning and iteration, the deep learning neural network is updated, the judgment efficiency and the judgment accuracy of the deep learning neural network are improved, then, after the comparison unit obtains the detection result, the detection result is sent to the processing unit, the processing unit works according to the detection result, under the condition that the detection result is normal, the processing unit only stores the received detection data, the detection result and the detection time data into the database without carrying out other actions, when the detection result is abnormal, the processing unit also sends an alarm signal on the premise that the received detection data, the detection result and the detection time data are stored into the database, reminding workers, and informing the workers to go to the position of the track for field maintenance; meanwhile, through a displacement sensor installed on the track, after the train passes through, the position offset of the spikes on the track is detected, then, the position offset data of the spikes are sent to a cloud big data center, the analysis unit and a comparison unit of the cloud big data center are used for analyzing and judging the position offset and a standard value, whether the positions of the spikes are still normal or not is obtained, when the positions of the spikes are abnormal, namely the spikes are loosened, a processing unit timely informs a worker of the treatment, the potential safety hazard in train running is avoided, meanwhile, the analysis and the prejudgment of the comparison unit are carried out, before the positions of the spikes are in problem, the processing unit informs the worker of the treatment in advance, and the potential safety hazard is avoided.
As a specific implementation manner of the present invention, the cloud big data center further includes a decision unit, the deep learning neural network in the comparison unit includes but is not limited to any one of a bp (back propagation) neural network, an rbf (radial basis function) neural network, and a genetic algorithm neural network, the comparison unit has odd number of parallel deep learning neural networks with the same structure, the data obtained by the analysis and classification by the analysis unit is simultaneously transmitted to the plurality of deep learning neural networks in the comparison unit for processing and comparison, the results obtained by the plurality of deep learning neural networks in the comparison unit are output to the decision unit, and the decision unit adopts most of the minority to synthesize the results obtained by each deep learning neural network, so as to give a final detection result;
in the working process, the number of the parallel deep learning neural networks arranged in the comparison unit is at least 3, so that the judgment accuracy of the comparison unit is improved by taking the output result of most deep learning neural networks as the final detection result under the condition that most of the parallel deep learning neural networks obey the minority principle, and the influence on the driving safety of a train caused by the wrong detection result given by the comparison unit when only one deep learning neural network exists in the comparison unit and the deep learning neural network is interfered is avoided; meanwhile, the deep learning neural network used in the comparison unit can be any one of a BP neural network, an RBF neural network and a genetic algorithm neural network, and can also be other open source deep learning neural networks, so that a more accurate comparison result and higher comparison efficiency can be obtained, and the performance of the comparison unit can be improved.
As a specific implementation manner of the present invention, the detection unit further includes a detection unmanned aerial vehicle, and if the detection result indicates that there is an abnormality, the processing unit sends a detection command to the detection unmanned aerial vehicle, the detection unmanned aerial vehicle performs patrol and photograph on the track according to the planned path, and sends the photographed photo and video to the cloud big data center, the photo and video uploaded by the detection unmanned aerial vehicle and received by the cloud big data center are analyzed and classified by the analysis unit, and then processed by the deep learning neural network in the comparison unit, so as to obtain a processing priority level of the abnormality of the track, and the processing unit makes an alarm prompt according to the processing priority level;
in the working process, when the detection result given by the comparison unit is abnormal, the processing unit sends a detection command to the unmanned detection machine, so that the unmanned detection machine can detect the track on the spot, more data are provided for subsequent detection and judgment, meanwhile, the track is divided into different sections, unmanned machine stations are arranged in each section, the unmanned detection machine is parked in the unmanned machine stations for protection and charging when not used, meanwhile, a planned fixed flight path is recorded in the unmanned detection machine in each section, and the flight path is consistent with the track path in the section where the unmanned detection machine is located, so that the unmanned detection machine can rapidly and accurately go to the track to carry out on-spot detection after receiving the detection command sent by the processing unit, and meanwhile, the unmanned detection machine shoots the track through a high-definition camera carried by the unmanned detection machine, the shot pictures and videos are sent to the cloud big data center, the shot pictures and videos are analyzed and processed in an analysis unit and a comparison unit of the cloud big data center, whether obvious abnormity exists in the spikes on the track or not and whether a detection result given by the comparison unit is correct or not are judged, if the spikes are abnormal, a processing priority is given by the comparison unit, and alarm prompts of different levels are given by the processing unit according to the processing priority, so that workers can process the spikes with problems in a proper sequence, and potential safety hazards caused by untimely processing are avoided.
As a specific embodiment of the invention, the track is provided with a positioning mark, the detection unmanned aerial vehicle shoots at the position with the positioning mark, the positioning mark is located at the spike on the track, and the detection unmanned aerial vehicle shoots the positioning mark and the spike into a photo or a video at the same time;
in the course of the work, through setting up the positioning mark on the track, make analysis unit and contrast unit when the photo and the video of analysis shooting, can regard as the reference with the positioning mark, thereby fix and known positioning mark through the specification carries out the analysis to the spike position in the photo, confirm the relative orbital accurate position of spike, avoid judging the position of spike and appear the deviation, influence the exactness of the testing result that contrast unit gave, and simultaneously, through the positioning mark that sets up, can conveniently detect unmanned aerial vehicle through the image detection function of taking certainly, fix a position self flight track, avoid detecting unmanned aerial vehicle and deviate from the good track of planning in flight process, influence the normal work that detects unmanned aerial vehicle.
As a specific embodiment of the invention, the detection unmanned aerial vehicle is at a fixed height when shooting a photo or a video, the detection unmanned aerial vehicle shoots two rails of a track into one photo at the same time when shooting, the positioning mark also has a position corresponding to a spike inside the track, and the detection unmanned aerial vehicle shoots the positioning mark inside the track into the photo with the two rails together;
in the course of the work, shoot at fixed height through making the detection unmanned aerial vehicle, thereby two orbital rails can appear simultaneously in the photo that makes unmanned aerial vehicle shoot and the video, afterwards, the rethread sets up at the track inboard and is detected the location mark that unmanned aerial vehicle shot in photo and the video, make analytical element and contrast unit in the big data center in the high in the clouds come to orbital gauge according to photo and the video that detection unmanned aerial vehicle shot and obtain, the rail is to the isoparametric judgement, confirm that the track is still in normal condition, avoid appearing the potential safety hazard, influence the driving safety of train.
As a specific implementation manner of the invention, the detection unit further comprises a first vibration sensor, the first vibration sensor is fixedly installed on the track along a fixed distance, the distance between any two vibration sensors is greater than the length of a train running on the track, and the first vibration sensor synchronously uploads position data of the first vibration sensor to a cloud big data center when uploading data;
in the working process, the track is divided into different sections, the length of each section is greater than that of the train, and when the train passes by, the train can cause the vibration of the track, therefore, through the data collected and uploaded by the vibration sensor, the analysis unit and the comparison unit of the cloud big data center can obtain the vibration data of the track in different sections, and then, the comparison unit compares the vibration data of the tracks in different sections with each other to obtain a deviation value, and under the condition that the deviation value is in a normal interval, the comparison unit can judge the track state is normal without additional detection and maintenance, and when the deviation value exceeds the normal range, the contrast unit can judge that the track state is unusual to in time inform the staff through the processing unit and go to check, maintain before, avoid the track more serious problem to appear, influence the safe operation of train.
As a specific embodiment of the invention, a second vibration sensor is arranged on a gear box of the train running on the track;
in the working process, the second vibration sensor is installed on the gear box of the train, so that vibration data of the train in operation is collected, meanwhile, position data of the train in operation can be synchronously uploaded to the cloud big data center along with the data collected by the second vibration sensor, the comparison unit can compare and analyze the vibration data of the train in different sections of the track with the vibration data of the track in the section, whether the vibration data and the vibration data are consistent or deviated within a normal range is judged, if the vibration data and the deviation are inconsistent or the deviation exceeds the normal range, the comparison unit judges that the train and/or the track are abnormal, and a processing unit informs workers to check and maintain the track, so that the track is normal, the train is safe to pass, and no potential safety hazard exists.
As a specific embodiment of the present invention, the second vibration sensors have a plurality of groups, the second vibration sensors are equidistantly distributed along the axial directions of the X axis, the Y axis and the Z axis respectively with the center of the gear box of the train as an origin, and the second vibration sensors have the same number in each axial direction;
in the course of the work, through set up the vibration sensor two that the axial direction of multiunit edge X axle, Y axle, Z axle distributes on the gear box of train, the running data of collection train operation in-process gear box that can be more accurate, through gathering more data, promote the contrast unit analysis, the degree of accuracy that reachs the testing result after judging, simultaneously, through set up a plurality of sensors in each axial direction, can avoid vibration sensor two to appear damaging the back, arouse the unsafe condition of testing data.
As a specific implementation manner of the present invention, the detection unit further includes a grating sensor, the grating sensor is fixedly installed on the track along a fixed interval, the grating sensor is configured to detect bending deformation data of the track in a vertical direction, and the grating sensor uploads the detected bending deformation data to the cloud big data center;
in the working process, when a train passes through, the train generates pressure on the track to cause the track to generate bending deformation, the bending deformation data of the track is detected by the grating sensor, then after the train passes through the track, the grating sensor detects the bending deformation data of the track again, then the bending deformation data obtained by the two detection processes are compared and analyzed by the comparison unit to judge whether the bending deformation data of the track is in a normal range and whether the recovery capability of the track after bending deformation meets the standard, when the track can not be completely recovered after bending deformation or the bending deformation data still exists after recovery and is larger than the standard requirement, a processing unit informs a worker to replace and maintain the track in the position in advance to avoid potential safety hazards, and meanwhile, when an unexpected condition such as earthquake occurs, the track is bent and deformed, after the grating sensor detects data, the detected data are sent to the cloud big data center, analysis is carried out through the analysis unit and the comparison unit, after the track is judged to be bent and damaged, a processing unit informs workers to carry out processing before the track is judged, and therefore the workers do not need to check the damaged place along the track, the labor intensity of the workers is reduced, and the repairing speed after the track is damaged is accelerated.
As a specific embodiment of the present invention, the grating sensor and the displacement sensor are integrated into an MEMS sensor having functions of detecting displacement and bending deformation by an MEMS technique;
in the working process, the MEMS sensor is used, the sensor installation workload is reduced, the engineering efficiency is improved, meanwhile, the working energy consumption of the system is reduced through the MEMS sensor, and the working stability is improved.
The specific working process is as follows:
in the working process, data collected by a detection unit is quickly uploaded to a cloud big data center through a wired network or a 5G network, then the received data is subjected to primary processing through an analysis unit, so that the data is analyzed and classified, useless and redundant data in the received detection data are removed, meanwhile, after the data subjected to primary processing by the analysis unit is sent to a comparison unit, the data is analyzed and compared through a deep learning neural network, a data model is detected by a component, so that the detection data is predicted according to the data recorded once, then the predicted data is compared with the detection data, if the data is consistent with the data, the detection result is normal, the track is not abnormal, if the data is inconsistent with the data, the detection result is judged to be abnormal, the track is abnormal, then after the comparison unit obtains the detection result, sending the detection result to a processing unit, wherein the processing unit works according to the detection result, only the received detection data, the detection result and the detection time data are stored in a database by the processing unit under the condition that the detection result is normal, and other actions are not carried out; meanwhile, after a train passes through the displacement sensor arranged on the track, the position offset of the spike on the track is detected, then the position offset data of the spike is sent to the cloud big data center, the position offset and the standard value are analyzed and judged through an analysis unit and a comparison unit of the cloud big data center, whether the position of the spike is still normal or not is obtained, and when the position of the spike is abnormal, namely the spike is loosened, a processing unit timely informs a worker of the fact that the spike is processed;
furthermore, the number of the parallel deep learning neural networks arranged in the comparison unit is at least 3, so that the result output by most deep learning neural networks is used as the final detection result under the condition that most deep learning neural networks are subjected to a minority principle;
when the detection result given by the comparison unit is abnormal, the processing unit sends a detection command to the unmanned detection machine, so that the unmanned detection machine can detect the track on the spot, more data are provided for subsequent detection and judgment, meanwhile, the track is divided into different sections, unmanned machine stations are arranged in each section, the unmanned detection machine is parked in the unmanned machine stations for protection and charging when not used, meanwhile, planned fixed flight paths are recorded in the unmanned detection machine in each section, and the flight paths are consistent with the track paths in the section where the unmanned detection machine is located, so that the unmanned detection machine can quickly and accurately go to the track for on-spot detection after receiving the detection command sent by the processing unit, meanwhile, the unmanned detection machine shoots the track through a high-definition camera carried by the unmanned detection machine, and sends the shot pictures and videos to a cloud large data center, analyzing and processing the shot pictures and videos in an analysis unit and a comparison unit of the cloud big data center, judging whether the spikes on the track are obviously abnormal and whether the detection result given by the comparison unit is correct, if the spikes are abnormal, giving processing priority by the comparison unit, and giving alarm reminders of different levels by the processing unit according to the processing priority;
furthermore, the positioning marks arranged on the track enable the analysis unit and the comparison unit to use the positioning marks as references when analyzing the shot pictures and videos, so that the positions of the spikes in the pictures are analyzed through the positioning marks with fixed and known specifications, and the accurate positions of the spikes relative to the track are determined;
furthermore, the detection unmanned aerial vehicle shoots at a fixed height, so that two rails of the track can appear in the photo and the video shot by the unmanned aerial vehicle, and then the analysis unit and the comparison unit in the cloud big data center can judge the track distance, the track direction and other parameters of the track according to the photo and the video shot by the detection unmanned aerial vehicle through the positioning marks which are arranged on the inner side of the track and shot by the detection unmanned aerial vehicle, so as to determine that the track is still in a normal state;
the track is divided into different sections, the length of each section is larger than that of a train, and when the train passes through, the train can cause the vibration of the track, so that data collected and uploaded by a vibration sensor is used, the analysis unit and the comparison unit of the cloud big data center obtain vibration data of the track in different sections, then the comparison unit compares the vibration data of the track in different sections mutually to obtain a deviation value, the comparison unit can judge that the track state is normal under the condition that the deviation value is in a normal interval without performing additional detection and maintenance, and when the deviation value exceeds a normal range, the comparison unit can judge that the track state is abnormal, so that a worker is timely informed of checking and maintaining before through the processing unit;
furthermore, a second vibration sensor is arranged on a gear box of the train, so that vibration data of the train in operation is collected, meanwhile, position data of the train in operation is synchronously uploaded to a cloud big data center along with the data collected by the second vibration sensor, so that a comparison unit can compare and analyze the vibration data of different sections of the train on the track with the vibration data of the track in the section, judge whether the two are consistent or have deviation within a normal range, if the two are inconsistent or have deviation beyond the normal range, the comparison unit judges that the train and/or the track are abnormal, and inform a worker to check and maintain the train and/or the track through a processing unit;
furthermore, a plurality of groups of vibration sensors II distributed along the axial directions of the X axis, the Y axis and the Z axis are arranged on the gearbox of the train, so that the running data of the gearbox in the running process of the train can be more accurately collected, and the accuracy of the detection result obtained after the analysis and the judgment of the comparison unit is improved by collecting more data;
when a train passes through, the train generates pressure on the track to cause the track to generate bending deformation, the bending deformation data of the track is detected through the grating sensor, then after the train passes through the track, the grating sensor detects the bending deformation data of the track again, then the bending deformation data obtained by the two detection processes are contrasted and analyzed through the contrast unit to judge whether the bending deformation data of the track is in a normal range and whether the recovery capability of the track after the bending deformation meets the standard, when the track can not be completely recovered after the bending deformation or the bending deformation data still meets the requirement that the bending deformation data is larger than the standard after the recovery, a worker is informed to replace and maintain the track in the position in advance through the processing unit, meanwhile, when an accident occurs, such as an earthquake, the track generates the bending deformation, and after the grating sensor detects the data, the detection data are sent to the cloud big data center, the analysis unit and the comparison unit are used for analyzing, and after the rail is judged to be bent and damaged, the processing unit is used for informing a worker to carry out processing;
furthermore, through using the MEMS sensor, the sensor installation workload is reduced, the engineering efficiency is improved, and meanwhile, the working energy consumption of the system is reduced through the MEMS sensor, and the working stability is improved.
The front, the back, the left, the right, the upper and the lower are all based on figure 1 in the attached drawings of the specification, according to the standard of the observation angle of a person, the side of the device facing an observer is defined as the front, the left side of the observer is defined as the left, and so on.
In the description of the present invention, it is to be understood that the terms "center", "longitudinal", "lateral", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. indicate orientations or positional relationships based on those shown in the drawings, and are merely intended to facilitate the description of the present invention and to simplify the description, but do not indicate or imply that the device or element referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore should not be construed as limiting the scope of the present invention.
The foregoing shows and describes the general principles, principal features and advantages of the invention. It will be understood by those skilled in the art that the present invention is not limited to the embodiments described above, which are described in the specification and illustrated only to illustrate the principle of the present invention, but that various changes and modifications may be made therein without departing from the spirit and scope of the present invention, which fall within the scope of the invention as claimed. The scope of the invention is defined by the appended claims and equivalents thereof.
Claims (10)
1. The utility model provides a track traffic fault detection system based on artificial intelligence which characterized in that: the system comprises a cloud big data center and a detection unit;
the cloud big data center comprises an analysis unit, a comparison unit, a processing unit and a database, wherein the analysis unit is used for receiving the data uploaded by the detection unit and analyzing and classifying the data, the comparison unit is a deep learning neural network, the comparison unit processes and compares the data analyzed and classified by the analysis unit to give a detection result, the detection result is synchronously sent to the processing unit, and the processing unit performs processing action according to the detection result;
the detection unit is arranged on a track and/or a vehicle and exchanges data with the cloud big data center through a wired network and/or a 5G network;
the detection unit comprises a displacement sensor, the displacement sensor is arranged on the track along a fixed distance, and the displacement sensor is used for monitoring the position offset of the spikes on the track.
2. The rail transit fault detection system based on artificial intelligence of claim 1, characterized in that: the cloud big data center further comprises a decision unit, the deep learning neural network in the comparison unit comprises any one of a BP (back propagation) neural network, a RBF (radial basis function) neural network and a genetic algorithm neural network, the comparison unit is provided with odd number of parallel deep learning neural networks with the same structure, data obtained by analysis and classification of the analysis unit are simultaneously transmitted to the multiple deep learning neural networks in the comparison unit for processing and comparison, results obtained by the multiple deep learning neural networks in the comparison unit are output to the decision unit, and the decision unit integrates the results obtained by the deep learning neural networks by adopting majority obeying minority principles to obtain a final detection result.
3. The rail transit fault detection system based on artificial intelligence of claim 1, characterized in that: the detection unit further comprises a detection unmanned aerial vehicle, the detection result shows that the detection unmanned aerial vehicle is abnormal, then the processing unit sends a detection command to the detection unmanned aerial vehicle, the detection unmanned aerial vehicle patrols and takes a picture of the track according to a planned path and sends the taken picture and the taken video to the cloud big data center, the picture and the video which are uploaded by the detection unmanned aerial vehicle and are received by the cloud big data center are analyzed and classified by the analysis unit, then processed by the deep learning neural network in the comparison unit to obtain the processing priority that the track is abnormal, and the processing unit gives an alarm according to the processing priority.
4. The rail transit fault detection system based on artificial intelligence of claim 3, wherein: be provided with the positioning mark on the track, detect unmanned aerial vehicle and shoot in there being positioning mark department, the positioning mark is located the spike department on the track, detect unmanned aerial vehicle and shoot photo or video with positioning mark and spike simultaneously.
5. The rail transit fault detection system based on artificial intelligence of claim 4, wherein: detect unmanned aerial vehicle when taking photo or video, be in fixed height, in taking into a photo with orbital two rails simultaneously when detecting unmanned aerial vehicle and shoot, the location mark also has the position corresponding with the railway inboard and spike, it enters into the photo that has two rails with the inboard location mark of track together to detect unmanned aerial vehicle.
6. The rail transit fault detection system based on artificial intelligence of claim 2, characterized in that: the detection unit further comprises a first vibration sensor, the first vibration sensor is fixedly installed on the track along a fixed interval, the interval between any two vibration sensors is larger than the length of a train running on the track, and the position data of the first vibration sensor is uploaded to a cloud big data center synchronously when the first vibration sensor uploads the data.
7. The rail transit fault detection system based on artificial intelligence of claim 6, characterized in that: and a second vibration sensor is mounted on a gear box of the train running on the track.
8. The rail transit fault detection system based on artificial intelligence of claim 7, wherein: the second vibration sensor has a plurality of groups, the second vibration sensor uses the center of the gear box of the train as an original point and is distributed along the axial directions of the X axis, the Y axis and the Z axis respectively at equal intervals, and the second vibration sensor has the same quantity in each axial direction.
9. The rail transit fault detection system based on artificial intelligence of claim 1, characterized in that: the detection unit further comprises a grating sensor, the grating sensor is fixedly installed on the track along a fixed distance, the grating sensor is used for detecting bending deformation data of the track in the vertical direction, and the grating sensor uploads the detected bending deformation data to the cloud big data center.
10. The rail transit fault detection system based on artificial intelligence of claim 9, wherein: the grating sensor and the displacement sensor are integrated into an MEMS sensor with the functions of detecting displacement and bending deformation through an MEMS technology.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210332961.7A CN114633774A (en) | 2022-03-30 | 2022-03-30 | Rail transit fault detection system based on artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210332961.7A CN114633774A (en) | 2022-03-30 | 2022-03-30 | Rail transit fault detection system based on artificial intelligence |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114633774A true CN114633774A (en) | 2022-06-17 |
Family
ID=81952487
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210332961.7A Pending CN114633774A (en) | 2022-03-30 | 2022-03-30 | Rail transit fault detection system based on artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114633774A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114937040A (en) * | 2022-07-22 | 2022-08-23 | 珠海优特电力科技股份有限公司 | Train inspection method, device and system for rail transit vehicle section and storage medium |
| CN115675566A (en) * | 2022-11-11 | 2023-02-03 | 洛阳双瑞橡塑科技有限公司 | Track monitoring system and monitoring method |
Citations (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH09269217A (en) * | 1996-03-29 | 1997-10-14 | Kawasaki Heavy Ind Ltd | Abnormality detection method for railroad track and abnormality detection device |
| CN103699698A (en) * | 2014-01-16 | 2014-04-02 | 北京泰乐德信息技术有限公司 | Method and system for track traffic failure recognition based on improved Bayesian algorithm |
| CN103914735A (en) * | 2014-04-17 | 2014-07-09 | 北京泰乐德信息技术有限公司 | Failure recognition method and system based on neural network self-learning |
| CN106379376A (en) * | 2016-09-28 | 2017-02-08 | 成都奥克特科技有限公司 | on-line rail state monitoring method based on vibration and positioning monitoring |
| CN106682687A (en) * | 2016-12-13 | 2017-05-17 | 广东工业大学 | Multi-example learning method using deep learning technology |
| CN107433962A (en) * | 2016-06-18 | 2017-12-05 | 刘春梅 | A kind of method and system for being used for track traffic failure monitoring and intelligent early-warning |
| CN107858883A (en) * | 2017-11-29 | 2018-03-30 | 北京交通大学 | A kind of rail system safe condition comprehensive monitoring and intelligent analysis method |
| CN108848571A (en) * | 2018-07-04 | 2018-11-20 | 大连声文科技发展有限公司 | A kind of rail traffic safety monitoring system and monitoring method based on MEMS sensor |
| CN109278794A (en) * | 2018-09-27 | 2019-01-29 | 东莞方凡智能科技有限公司 | Train rail abnormality detection system |
| CN109460863A (en) * | 2018-10-29 | 2019-03-12 | 中国石油大学(华东) | Equipment state prediction method based on deep learning |
| CN110341761A (en) * | 2019-07-11 | 2019-10-18 | 郑州铁路职业技术学院 | A kind of railway track real-time monitoring system |
| CN111071300A (en) * | 2020-02-12 | 2020-04-28 | 太原理工大学 | High-speed train rail transit fault safety monitoring and early warning system and signal processing method |
| CN111332340A (en) * | 2020-03-06 | 2020-06-26 | 东莞理工学院 | Method and system for storing and processing rail transit monitoring data |
| CN111516727A (en) * | 2020-05-12 | 2020-08-11 | 上海数深智能科技有限公司 | High-speed rail defect abnormity intelligent diagnosis and detection system and method based on double vibration measurement sensors |
| CN111605578A (en) * | 2020-05-21 | 2020-09-01 | 哈尔滨市科佳通用机电股份有限公司 | Railway track inspection method for carrying 3D equipment by using unmanned aerial vehicle |
| CN113008998A (en) * | 2021-03-02 | 2021-06-22 | 南京审计大学 | Hidden engineering internal defect judgment method based on PCNN |
| CN113371033A (en) * | 2021-06-15 | 2021-09-10 | 武汉瑞辉科技发展有限公司 | Rail transit operation safety real-time online monitoring and early warning management cloud platform based on cloud computing |
| CN113537284A (en) * | 2021-06-04 | 2021-10-22 | 中国人民解放军战略支援部队信息工程大学 | Deep learning implementation method and system based on mimicry mechanism |
| CN214728845U (en) * | 2021-03-04 | 2021-11-16 | 宝鸡中车时代工程机械有限公司 | Rail engineering vehicle operation safety monitoring system |
| US20220017129A1 (en) * | 2020-07-15 | 2022-01-20 | Zahid F. Mian | Onboard Railway Health Monitoring |
-
2022
- 2022-03-30 CN CN202210332961.7A patent/CN114633774A/en active Pending
Patent Citations (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH09269217A (en) * | 1996-03-29 | 1997-10-14 | Kawasaki Heavy Ind Ltd | Abnormality detection method for railroad track and abnormality detection device |
| CN103699698A (en) * | 2014-01-16 | 2014-04-02 | 北京泰乐德信息技术有限公司 | Method and system for track traffic failure recognition based on improved Bayesian algorithm |
| CN103914735A (en) * | 2014-04-17 | 2014-07-09 | 北京泰乐德信息技术有限公司 | Failure recognition method and system based on neural network self-learning |
| CN107433962A (en) * | 2016-06-18 | 2017-12-05 | 刘春梅 | A kind of method and system for being used for track traffic failure monitoring and intelligent early-warning |
| CN106379376A (en) * | 2016-09-28 | 2017-02-08 | 成都奥克特科技有限公司 | on-line rail state monitoring method based on vibration and positioning monitoring |
| CN106682687A (en) * | 2016-12-13 | 2017-05-17 | 广东工业大学 | Multi-example learning method using deep learning technology |
| CN107858883A (en) * | 2017-11-29 | 2018-03-30 | 北京交通大学 | A kind of rail system safe condition comprehensive monitoring and intelligent analysis method |
| CN108848571A (en) * | 2018-07-04 | 2018-11-20 | 大连声文科技发展有限公司 | A kind of rail traffic safety monitoring system and monitoring method based on MEMS sensor |
| CN109278794A (en) * | 2018-09-27 | 2019-01-29 | 东莞方凡智能科技有限公司 | Train rail abnormality detection system |
| CN109460863A (en) * | 2018-10-29 | 2019-03-12 | 中国石油大学(华东) | Equipment state prediction method based on deep learning |
| CN110341761A (en) * | 2019-07-11 | 2019-10-18 | 郑州铁路职业技术学院 | A kind of railway track real-time monitoring system |
| CN111071300A (en) * | 2020-02-12 | 2020-04-28 | 太原理工大学 | High-speed train rail transit fault safety monitoring and early warning system and signal processing method |
| CN111332340A (en) * | 2020-03-06 | 2020-06-26 | 东莞理工学院 | Method and system for storing and processing rail transit monitoring data |
| CN111516727A (en) * | 2020-05-12 | 2020-08-11 | 上海数深智能科技有限公司 | High-speed rail defect abnormity intelligent diagnosis and detection system and method based on double vibration measurement sensors |
| CN111605578A (en) * | 2020-05-21 | 2020-09-01 | 哈尔滨市科佳通用机电股份有限公司 | Railway track inspection method for carrying 3D equipment by using unmanned aerial vehicle |
| US20220017129A1 (en) * | 2020-07-15 | 2022-01-20 | Zahid F. Mian | Onboard Railway Health Monitoring |
| CN113008998A (en) * | 2021-03-02 | 2021-06-22 | 南京审计大学 | Hidden engineering internal defect judgment method based on PCNN |
| CN214728845U (en) * | 2021-03-04 | 2021-11-16 | 宝鸡中车时代工程机械有限公司 | Rail engineering vehicle operation safety monitoring system |
| CN113537284A (en) * | 2021-06-04 | 2021-10-22 | 中国人民解放军战略支援部队信息工程大学 | Deep learning implementation method and system based on mimicry mechanism |
| CN113371033A (en) * | 2021-06-15 | 2021-09-10 | 武汉瑞辉科技发展有限公司 | Rail transit operation safety real-time online monitoring and early warning management cloud platform based on cloud computing |
Non-Patent Citations (1)
| Title |
|---|
| 谷雨: "基于支持向量机的入侵检测算法研究", vol. 978, 西安交通大学出版社, pages: 126 - 128 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114937040A (en) * | 2022-07-22 | 2022-08-23 | 珠海优特电力科技股份有限公司 | Train inspection method, device and system for rail transit vehicle section and storage medium |
| CN114937040B (en) * | 2022-07-22 | 2022-11-18 | 珠海优特电力科技股份有限公司 | Train inspection method, device and system for rail transit vehicle section and storage medium |
| CN115675566A (en) * | 2022-11-11 | 2023-02-03 | 洛阳双瑞橡塑科技有限公司 | Track monitoring system and monitoring method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111507308B (en) | Transformer substation safety monitoring system and method based on video identification technology | |
| CN114633774A (en) | Rail transit fault detection system based on artificial intelligence | |
| CN104599249A (en) | Cable rope bearing bridge deck vehicle load distribution real-time detection method | |
| CN103266559B (en) | The method of BP bridge security inspection car and face, acquisition bridge surface phase | |
| CN106647819A (en) | Mobile image acquisition device and high-speed train primary maintenance intelligent auxiliary system | |
| CN211087009U (en) | Contact net inspection device based on aircraft | |
| KR102017870B1 (en) | Real-time line defect detection system | |
| CN103236181A (en) | Traffic signal lamp state monitoring system and method based on machine vision | |
| CN112381778A (en) | Transformer substation safety control platform based on deep learning | |
| CN115527340A (en) | Intelligent construction site safety monitoring system and method based on unmanned aerial vehicle and surveillance camera | |
| CN111523386B (en) | High-speed railway platform door monitoring and protecting method and system based on machine vision | |
| CN110455214A (en) | A kind of abrasion of pantograph pan condition monitoring system and method | |
| CN116110012B (en) | Dangerous violation identification method and system for intelligent construction site | |
| CN115527364B (en) | Traffic accident tracing method and system based on radar data fusion | |
| CN115465293B (en) | Multi-sensor safety self-cognition and safety processing device and method | |
| CN110909020A (en) | Vehicle-mounted contact net dynamic detection system | |
| CN112562407A (en) | ODD state pre-judging method and device and autonomous passenger-riding parking system | |
| CN110942221A (en) | Transformer substation fault rapid repairing method based on Internet of things | |
| CN115600124A (en) | Subway tunnel inspection system and inspection method | |
| CN116740833A (en) | Line inspection and card punching method based on unmanned aerial vehicle | |
| CN112285111A (en) | Pantograph front carbon sliding plate defect detection method, device, system and medium | |
| CN116664343A (en) | System and method based on intelligent building site digital model | |
| CN116630872A (en) | Building site safety monitoring method based on three-dimensional modeling | |
| CN114333331B (en) | Method and system for identifying vehicle passing information and vehicle weight of multi-lane bridge | |
| CN111267079A (en) | Intelligent inspection robot inspection method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |