An artificial intelligence-based intelligent maintenance assistance system for rail transit vehicles

Abstract

The present invention discloses an artificial intelligence-based intelligent maintenance auxiliary system for rail transit vehicles, comprising a maintenance mode acquisition module, a first acquisition module, a second acquisition module, a data processing module, and an information sending module; the maintenance mode acquisition module is used to collect maintenance modes, which include on-track maintenance and derailment maintenance; the first acquisition module is used to collect on-track maintenance-related information during on-track maintenance; the second acquisition module is used to collect derailment maintenance-related information during derailment maintenance; the data processing module is used to process on-track maintenance-related information and derailment maintenance-related information to generate on-track maintenance auxiliary information and derailment maintenance auxiliary information; the information sending module is used to send the on-track maintenance auxiliary information and derailment maintenance auxiliary information to a preset receiving terminal. The present invention can better assist rail transit vehicles in performing maintenance operations.

Description

Technical Field

[0001] The present invention relates to the field of auxiliary systems, and in particular to an artificial intelligence-based intelligent maintenance auxiliary system for rail transit vehicles. Background Art

[0002] With the acceleration of urbanization, rail transit plays an increasingly critical role in people's daily travel, and the safety and reliability of its operation have attracted much attention.

[0003] On the one hand, the continuous expansion of rail transit networks and the increasing complexity of routes have led to a sharp increase in vehicle operation frequency, which has accelerated the wear and aging of vehicle components and increased the probability of failure. Traditional maintenance methods rely primarily on manual experience, with maintenance personnel relying on visual observation and simple tool inspections to determine faults and repair needs. This approach is not only inefficient but also difficult to accurately locate the root cause of complex faults. This can easily lead to missed hidden dangers, resulting in incomplete repairs and affecting the vehicle's return to service and operational safety.

[0004] On the other hand, rail transit vehicle maintenance scenarios are diverse, including on-track maintenance and derailment maintenance. On-track maintenance needs to be completed quickly and safely within a limited time window without affecting the normal operation of the line, which places extremely high demands on the establishment of protective measures and the deployment of maintenance personnel. Derailment maintenance involves high-risk operations such as the lifting of large equipment. Once a cable breaks, lifting imbalances, or personnel safety accidents occur, it will not only cause serious economic losses, but may also delay the resumption of line operations, causing great inconvenience to citizens.

[0005] Therefore, it is necessary to use a rail transit vehicle maintenance assistance system for maintenance assistance. To this end, an artificial intelligence-based intelligent maintenance assistance system for rail transit vehicles is proposed. Summary of the Invention

[0006] The technical problem to be solved by the present invention is: how to solve the problem that the existing maintenance assistance system has low maintenance efficiency but certain safety hazards, and provides an artificial intelligence-based intelligent maintenance assistance system for rail transit vehicles.

[0007] The present invention solves the above technical problems through the following technical solutions. The present invention includes a maintenance mode acquisition module, a first acquisition module, a second acquisition module, a data processing module and an information sending module;

[0008] The maintenance mode collection module is used to collect maintenance modes, which include on-track maintenance and off-track maintenance;

[0009] The first acquisition module is used to collect information related to on-orbit maintenance during on-orbit maintenance;

[0010] The second collection module is used to collect information related to derailment maintenance during derailment maintenance;

[0011] The data processing module is used to process the on-orbit maintenance related information and the derailment maintenance related information to generate on-orbit maintenance auxiliary information and derailment maintenance auxiliary information;

[0012] The information sending module is used to send on-track maintenance auxiliary information and derailment maintenance auxiliary information to a preset receiving terminal.

[0013] Furthermore, the specific process of the first acquisition module collecting information related to on-orbit maintenance is as follows:

[0014] When conducting on-orbit maintenance, maintenance personnel need to set up protective signs. Position sensors and angle acquisition devices are installed on the protective signs to collect the position and angle of the protective signs.

[0015] At the same time, maintenance personnel need to wear location sensing equipment to collect their real-time location;

[0016] At the same time, information on maintenance reasons and the number of real-time maintenance personnel are collected;

[0017] That is, track maintenance related information includes the location and placement angle of protective signs, the real-time location of maintenance personnel, maintenance reason information and the real-time number of maintenance personnel;

[0018] The data processing module processes the location and placement angle of the protective signs, the real-time location of the maintenance personnel, the maintenance reason information, and the real-time number of maintenance personnel to generate on-orbit maintenance auxiliary information;

[0019] On-orbit maintenance auxiliary information includes protective sign prompt information and personnel prompt information.

[0020] Furthermore, the process of obtaining the placement angle is as follows:

[0021] Two tilt sensors are set on the side of the protective sign to collect the tilt information of the protective sign in real time and mark it as the first tilt angle and the second tilt angle;

[0022] A collection frequency is preset, and the first inclination angle and the second inclination angle are collected multiple times according to the preset collection frequency to obtain multiple first inclination angles and multiple second inclination angles, which constitute placement angle information.

[0023] Furthermore, the process of obtaining the protection mark prompt information is as follows:

[0024] Extracting the collected placement angles, and extracting a plurality of first inclination angles and a plurality of second inclination angles from the placement angles;

[0025] The multiple first inclination angles are processed, and whether there is a first inclination angle among the multiple first inclination angles that differs from the preset standard inclination angle by more than a first preset amplitude (for example, there is a first inclination angle that differs from the preset standard inclination angle by more than 20%) is first detected. When there is a first inclination angle that differs from the preset standard inclination angle by more than a preset amplitude, a protective mark prompt message is generated, and the average of the multiple first inclination angles is calculated. When the average of the multiple first inclination angles differs from the preset standard inclination angle by more than a second preset amplitude, a protective mark prompt message is also generated. In this case, the specific content of the protective mark prompt message is that the protective mark may overturn and the angle needs to be adjusted;

[0026] The multiple second inclination angles are processed in the same manner as the multiple first inclination angles to determine whether to generate a protective sign prompt message. In this case, the protective sign prompt message specifically indicates that the protective sign may overturn and needs to adjust its angle.

[0027] Extract the location of the protective sign, then collect the location of the rail transit vehicle that needs maintenance, collect its center point and mark it as the reference point;

[0028] The position of the protective sign is marked as A1, and the reference point is marked as A2. The distance between the protective sign position A1 and the reference point A2 is monitored in real time to obtain the evaluation distance. When the evaluation distance is less than the preset standard placement distance, a protective sign prompt message is generated. The specific content of the protective sign prompt message is that there is an abnormality in the placement of the protective sign and the position needs to be readjusted to ensure the warning effect.

[0029] The assessment distance is collected at a preset frequency. When the assessment distance collected at the preset frequency changes continuously and the amplitude of the change is greater than the preset value, a protective sign prompt message is generated. At this time, the specific content of the protective sign prompt message is that there is an abnormal displacement of the protective sign.

[0030] Furthermore, the specific process of obtaining the personnel prompt information is as follows:

[0031] Extract the real-time location of maintenance personnel, maintenance reason information and the real-time number of maintenance personnel from track maintenance related information;

[0032] Import the maintenance reason information into the preset database, retrieve the standard maintenance personnel number from the preset database, and calculate that when the difference between the real-time maintenance personnel number and the standard maintenance personnel number is less than 0 or greater than the preset value a, generate personnel prompt information;

[0033] Then collect the distance information between the maintenance personnel's real-time position and the reference point A2 to obtain the personnel distance information. When the personnel distance information is greater than the preset value for more than the preset time, a personnel prompt message is generated;

[0034] At the same time, the distance between all personnel performing maintenance tasks is collected. When the distance between all personnel performing maintenance tasks is greater than the preset value, a personnel prompt message is generated;

[0035] Collect maintenance task information assigned to maintenance personnel, extract the task execution time point from the maintenance task information, and when there is a preset time left before the task execution time point, extract the personnel distance information of all maintenance personnel performing the maintenance task. When there is personnel distance information greater than the preset value, generate personnel prompt information.

[0036] Furthermore, when performing derailment maintenance, the second collection module collects derailment maintenance related information in the following specific process:

[0037] Before lifting and derailing a rail transit vehicle, inspect the lifting cable and obtain the lifting cable information;

[0038] During the hoisting process, real-time hoisting image collection is performed;

[0039] The lifting cable information and real-time lifting images constitute the derailment repair related information.

[0040] Furthermore, the process of obtaining the hoisting cable information is as follows:

[0041] The steel cable is inspected using a magnetic particle inspection method, i.e., the steel cable is magnetized and then magnetic powder is sprayed on its surface. After the spraying is completed, the number of regions with an area greater than a preset value for adsorbing magnetic powder and the number of regions with a length greater than a preset length are collected. When the area of ​​a region with an area greater than a preset value for adsorbing magnetic powder and a length greater than a preset length are both recorded, only one region is recorded. The total number of regions with an area greater than a preset value for adsorbing magnetic powder and the total number of regions with a length greater than a preset length are calculated to obtain the abnormal region and the number of abnormal regions, i.e., the first abnormal information;

[0042] At the same time, the ultrasonic detection method is used to detect the steel cable, and the abnormal area of ​​the steel cable and the number of abnormal areas of the steel cable are obtained, that is, the second abnormal information;

[0043] The process of acquiring the real-time hoisting image is as follows: a safety area is set up on the hoisting path, and an image acquisition device is set up in the safety area. Image information in the safety area during the hoisting of the rail transit vehicle is acquired in real time, that is, the real-time hoisting image is acquired.

[0044] Furthermore, the derailment repair auxiliary information includes hoisting cable auxiliary information and hoisting process auxiliary information;

[0045] The process of obtaining the auxiliary information of the hoisting cable is as follows:

[0046] Extracting the acquired hoisting cable information, and extracting the first abnormality information and the second abnormality information therefrom;

[0047] When either the number of abnormal areas in the first abnormal information or the number of abnormal areas in the second abnormal information is greater than a preset value m1, auxiliary information for hoisting steel cables is generated;

[0048] When the number of abnormal areas of the first abnormal information and the number of abnormal areas of the second abnormal information are both less than the preset value m2 but greater than 0, the abnormal areas of the first abnormal information and the abnormal areas of the second abnormal information are extracted. When the abnormal areas of the first abnormal information and the abnormal areas of the second abnormal information are different areas, the lifting steel cable auxiliary information is generated. At this time, the content of the lifting steel cable auxiliary information is that there is an abnormality in the lifting steel cable.

[0049] The process of obtaining auxiliary information during the lifting process is as follows:

[0050] The real-time hoisting image is extracted and a human body recognition model is imported into the real-time hoisting image. When a human body model is recognized in the safe area, auxiliary information of the hoisting process is generated. At this time, the auxiliary information of the hoisting process indicates that there are abnormal people in the safe area and they need to be expelled;

[0051] Then, the real-time hoisting image is processed and analyzed during the hoisting process to obtain the hoisting speed and release speed;

[0052] When either the lifting speed or the releasing speed is greater than a preset value, auxiliary information of the lifting process is generated. At this time, the content of the auxiliary information of the lifting process is that the lifting speed / releasing speed is too fast and needs to be adjusted.

[0053] Compared with the existing technology, the present invention has the following advantages: the artificial intelligence-based intelligent maintenance assistance system for rail transit vehicles has more targeted maintenance methods, can clearly distinguish between on-track maintenance and derailment maintenance, and collects relevant information for different maintenance methods through the first acquisition module and the second acquisition module, making the maintenance assistance more targeted and able to meet the needs of different maintenance scenarios.

[0054] On-orbit maintenance has higher safety guarantees, and protective signs are accurately monitored. By setting position sensing equipment and angle acquisition equipment on the protective signs, the position and placement angle of the protective signs are collected in real time. Abnormal angles, positions and displacements of the protective signs can be detected in time, and corresponding prompt information can be generated to ensure the warning effect of the protective signs and provide safety guarantees for maintenance operations.

[0055] Maintenance personnel management is more efficient and effective. Maintenance personnel wear location sensors, and the system collects their real-time location. Combined with maintenance reason information and the real-time maintenance personnel headcount, this system enables multi-dimensional maintenance personnel management. For example, it can promptly identify abnormal personnel headcounts based on the standard maintenance personnel headcounts in a pre-set database. It can also monitor the distance between maintenance personnel and reference points, as well as the distance between personnel themselves, to prevent personnel from being in danger or affecting maintenance efficiency. Furthermore, it reminds personnel to be in position when the task execution time is approaching, ensuring the smooth progress of maintenance tasks.

[0056] The risk of derailment repair is controllable. During derailment repair, the lifting cables are inspected using magnetic particle detection and ultrasonic detection methods, which can comprehensively obtain information such as abnormal areas and the number of abnormal areas of the cables. By analyzing this information, safety hazards in the cables can be discovered in a timely manner. When the number of abnormal areas reaches a certain level or the abnormal areas of different detection methods are different, auxiliary information for the lifting cables is generated to indicate that there are abnormalities in the lifting cables, effectively reducing the risks during the lifting process. Image acquisition equipment is set up in the safe area of ​​the lifting path to collect image information in real time during the lifting of rail transit vehicles. Using the human body recognition model, abnormal personnel in the safe area can be discovered in a timely manner and auxiliary information can be generated to remind them to leave; the real-time lifting images are analyzed to obtain the lifting speed and release speed. When the speed exceeds the preset value, a prompt message is generated to remind you to adjust the speed to ensure the safety and smoothness of the lifting process, making the system more worthy of promotion and use. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 It is a system block diagram of the present invention. DETAILED DESCRIPTION

[0058] The following is a detailed description of an embodiment of the present invention. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process. However, the protection scope of the present invention is not limited to the following embodiment.

[0059] like Figure 1 As shown, this embodiment provides a technical solution: an artificial intelligence-based intelligent maintenance auxiliary system for rail transit vehicles, including a maintenance mode acquisition module, a first acquisition module, a second acquisition module, a data processing module and an information sending module;

[0060] The maintenance mode collection module is used to collect maintenance modes, which include on-track maintenance and off-track maintenance;

[0061] The first acquisition module is used to collect information related to on-orbit maintenance during on-orbit maintenance;

[0062] The second collection module is used to collect information related to derailment maintenance during derailment maintenance;

[0063] The data processing module is used to process the on-orbit maintenance related information and the derailment maintenance related information to generate on-orbit maintenance auxiliary information and derailment maintenance auxiliary information;

[0064] The information sending module is used to send on-track maintenance auxiliary information and derailment maintenance auxiliary information to a preset receiving terminal.

[0065] When the first acquisition module is used for on-orbit maintenance, the specific process of collecting on-orbit maintenance related information is as follows:

[0066] When conducting on-orbit maintenance, maintenance personnel need to set up protective signs. Position sensors and angle acquisition devices are installed on the protective signs to collect the position and angle of the protective signs.

[0067] At the same time, maintenance personnel need to wear location sensing equipment to collect their real-time location;

[0068] At the same time, information on maintenance reasons and the number of real-time maintenance personnel are collected;

[0069] That is, track maintenance related information includes the location and placement angle of protective signs, the real-time location of maintenance personnel, maintenance reason information and the real-time number of maintenance personnel;

[0070] The data processing module processes the location and placement angle of the protective signs, the real-time location of the maintenance personnel, the maintenance reason information, and the real-time number of maintenance personnel to generate on-orbit maintenance auxiliary information;

[0071] On-orbit maintenance auxiliary information includes protective sign prompt information and personnel prompt information.

[0072] The process of obtaining the placement angle is as follows:

[0073] Two tilt sensors are set on the side of the protective sign to collect the tilt information of the protective sign in real time and mark it as the first tilt angle and the second tilt angle;

[0074] A collection frequency is preset, and the first inclination angle and the second inclination angle are collected multiple times according to the preset collection frequency to obtain multiple first inclination angles and multiple second inclination angles, which constitute placement angle information;

[0075] This data collection process encompasses multiple aspects, including the location and angle of protective signs, the real-time location of maintenance personnel, maintenance reasons, and the number of personnel involved. This comprehensive information reflects the on-orbit maintenance site, from safety signs in the maintenance environment to the actual conditions of maintenance personnel. It provides a rich and comprehensive foundation for subsequent data processing and auxiliary information generation, facilitating precise control and analysis of the entire maintenance process.

[0076] Based on the characteristics and requirements of on-orbit maintenance scenarios, key information closely related to maintenance safety and efficiency was selected for collection. For example, the position and angle of protective signs are directly related to the warning effect and safety of the maintenance site, while the location and number of maintenance personnel are closely related to the execution and coordination of maintenance operations. This targeted collection can better meet the actual needs of on-orbit maintenance and enhance the practicality of the maintenance assistance system.

[0077] Multi-sensor real-time monitoring: Two tilt sensors are installed on the sides of the guard sign to collect the primary and secondary tilt angles in real time. This multi-sensor setup captures the guard sign's tilt information from different angles, avoiding the limitations of a single sensor. This improves the accuracy and reliability of the collected data and provides a more comprehensive and accurate picture of the guard sign's placement angle.

[0078] A preset collection frequency is used to repeatedly collect inclination information, generating multiple first and second inclination angles. This multiple collection process effectively reduces the impact of random factors on the results. By comprehensively analyzing large amounts of data, the system can more accurately identify the changing trends in the placement angles of protective signs, providing more reliable data support for subsequent assessments of any anomalies. This further enhances the accuracy and stability of the system's monitoring of protective sign status.

[0079] Auxiliary information generated based on accurately collected data can help maintenance personnel promptly understand potential risks and problems at the maintenance site, such as abnormal conditions of protective signs and the rationality of personnel arrangements, so as to make more scientific and reasonable decisions, ensure the safe and efficient progress of maintenance work, improve the intelligence level and management efficiency of the entire rail transit vehicle on-track maintenance, and thus better assist rail transit vehicles.

[0080] The process of obtaining the protection sign prompt information is as follows:

[0081] Extracting the collected placement angles, and extracting a plurality of first inclination angles and a plurality of second inclination angles from the placement angles;

[0082] The multiple first inclination angles are processed, and whether there is a first inclination angle among the multiple first inclination angles that differs from the preset standard inclination angle by more than a first preset amplitude (for example, there is a first inclination angle that differs from the preset standard inclination angle by more than 20%) is first detected. When there is a first inclination angle that differs from the preset standard inclination angle by more than a preset amplitude, a protective mark prompt message is generated, and the average of the multiple first inclination angles is calculated. When the average of the multiple first inclination angles differs from the preset standard inclination angle by more than a second preset amplitude, a protective mark prompt message is also generated. In this case, the specific content of the protective mark prompt message is that the protective mark may overturn and the angle needs to be adjusted;

[0083] The multiple second inclination angles are processed in the same manner as the multiple first inclination angles to determine whether to generate a protective sign prompt message. In this case, the protective sign prompt message specifically indicates that the protective sign may overturn and needs to adjust its angle.

[0084] Extract the location of the protective sign, then collect the location of the rail transit vehicle that needs maintenance, collect its center point and mark it as the reference point;

[0085] The position of the protective sign is marked as A1, and the reference point is marked as A2. The distance between the protective sign position A1 and the reference point A2 is monitored in real time to obtain the evaluation distance. When the evaluation distance is less than the preset standard placement distance, a protective sign prompt message is generated. The specific content of the protective sign prompt message is that there is an abnormality in the placement of the protective sign and the position needs to be readjusted to ensure the warning effect.

[0086] The assessment distance is collected at a preset frequency. When the assessment distance collected at the preset frequency changes continuously and the change amplitude is greater than a preset value, a protective sign prompt message is generated. The specific content of the protective sign prompt message at this time is that the protective sign has abnormal displacement;

[0087] The first preset amplitude: assuming it is set to 20%, that is, when the absolute value of the difference between a certain first inclination angle and the standard inclination angle of 5° exceeds 5°×20%=1°, the protection mark prompt information generation mechanism will be triggered, reminding maintenance personnel that there may be a sudden change in the inclination angle in a single direction due to external force collision, local sinking of soft foundation, etc., which requires attention and adjustment.

[0088] The second preset amplitude: for example, it is set to 10%. When the calculated average of multiple first inclination angles differs from the standard inclination angle of 5° by more than 5°×10%=0.5°, it means that the protective sign may show an overall slow tilting trend due to the cumulative impact of wind, small vibrations, etc., and the angle also needs to be adjusted in time to ensure the warning effect.

[0089] Standard placement distance: Based on the type and size of the maintenance vehicle, the space required for maintenance operations, and the visual range and warning distance of the protective sign itself, the standard placement distance is set at 10 meters from the center of the maintenance vehicle. This ensures that the protective sign is not too close to the vehicle to hinder the movement of maintenance personnel and tools, but is not too far away to cause passing vehicles to ignore the warning and fail to make evasive preparations in advance.

[0090] Preset frequency: The frequency of collecting assessment distances can be set to every 5 minutes, taking into account the duration of maintenance operations, the frequency of factors that may affect the position of protective signs, etc. If the system continuously monitors at this frequency during a 30-minute maintenance operation, once it finds that the assessment distance collected several times has gradually decreased, such as from the initial 10 meters to 8 meters or 7 meters, it will promptly prompt that there is an abnormal displacement of the protective sign, which may be caused by being hit by passing large equipment, slight vibration of the track, etc., and the maintenance personnel need to fix it immediately;

[0091] Through the above process, we can ensure that the angle of the protective signs is compliant and the effectiveness of the warning is guaranteed:

[0092] By processing multiple first and second inclination angles collected by the two inclination sensors installed on the protective sign, the system not only detects deviations from the preset standard inclination angle but also calculates the average and compares it with the standard. This multi-angle and multi-dimensional analysis method can comprehensively and accurately determine whether the protective sign is at risk of overturning. If an abnormal inclination angle is detected, a prompt message is generated to prompt maintenance personnel to adjust the angle, ensuring that the protective sign is always in the optimal warning state, reliably serving as a warning to passing vehicles and pedestrians, and avoiding safety accidents caused by the sign's tilt, which may cause the warning to fail.

[0093] Two judgment criteria are employed: a single inclination angle deviation exceeding a first preset magnitude, and a mean angle deviation exceeding a second preset magnitude from the preset standard inclination angle. This increases the rigor and reliability of the judgment. Different judgment criteria can address a variety of possible situations. Whether it's a sudden change in individual inclination angle caused by an external impact at a moment's notice, or a gradual deviation from the standard due to environmental factors over a long period of time, these can be accurately captured, effectively reducing the risk of misjudgment and maximizing the accuracy of the safety sign's angle.

[0094] Accurately control the location of protective signs to enhance on-site safety;

[0095] The protective sign's location is marked as A1, and the center point of the maintenance vehicle is set as the reference point A2. The distance between the two is monitored in real time and compared with the preset standard placement distance. When the assessed distance is less than the standard value, a prompt message is immediately generated to remind maintenance personnel to readjust the position. This process is like placing a precise "navigator" for the protective sign, ensuring that it is always within the specified safe and effective warning range. This prevents improper placement, such as being too close to the maintenance vehicle, affecting the maintenance operation space, or too far away to effectively warn approaching vehicles, thereby improving the safety of the entire maintenance site.

[0096] Dynamic Displacement Monitoring and Control: Distances are collected and assessed at a preset frequency, and the trend of distance changes is observed to determine whether the protective markings have shifted abnormally. If the distance is found to be constantly changing, it means that the protective markings may have shifted due to external forces, ground vibration, or other factors during maintenance. A prompt message is generated, allowing maintenance personnel to immediately detect and take measures to fix the markings. This dynamic monitoring mechanism proactively prevents potential safety hazards, nipping risks in the bud and creating a continuously stable and safe environment for on-orbit maintenance operations.

[0097] Comprehensively improve the intelligent level of maintenance assistance to facilitate efficient maintenance;

[0098] The entire process of acquiring protective sign information relies on a wealth of real-time data collected by sensors, from multi-angle inclination data to dynamic position data. Through rigorous algorithms and analysis compared with pre-set standards, highly targeted prompts are generated for maintenance personnel. Maintenance personnel no longer need to rely on subjective experience to judge the status of protective signs. Instead, they can make scientific decisions about adjusting protective signs based on the clear prompts provided by the system, significantly improving the accuracy and efficiency of maintenance decisions.

[0099] The system automatically and continuously monitors all parameters of protective signs and issues prompt notifications if any anomalies are detected, overcoming the long time intervals and omissions that can occur with manual inspections. Whether it's late at night or early in the morning, when workers are prone to fatigue, or during busy maintenance situations with limited staffing, the system ensures that protective signs are always under intelligent monitoring, providing a solid safety barrier for on-track maintenance work and comprehensively enhancing the intelligent management of rail transit vehicle maintenance.

[0100] The specific process of obtaining the personnel prompt information is as follows:

[0101] Extract the real-time location of maintenance personnel, maintenance reason information and the real-time number of maintenance personnel from track maintenance related information;

[0102] Import the maintenance reason information into the preset database, retrieve the standard maintenance personnel number from the preset database, and calculate that when the difference between the real-time maintenance personnel number and the standard maintenance personnel number is less than 0 or greater than the preset value a, generate personnel prompt information;

[0103] Then collect the distance information between the maintenance personnel's real-time position and the reference point A2 to obtain the personnel distance information. When the personnel distance information is greater than the preset value for more than the preset time, a personnel prompt message is generated;

[0104] At the same time, the distance between all personnel performing maintenance tasks is collected. When the distance between all personnel performing maintenance tasks is greater than the preset value, a personnel prompt message is generated;

[0105] Collect maintenance task information assigned to maintenance personnel, extract the task execution time from the maintenance task information, and when there is a preset time left before the task execution time, extract the personnel distance information of all maintenance personnel performing the maintenance task. If the personnel distance information is greater than the preset value, a personnel prompt information is generated;

[0106] Personnel quantity control based on maintenance reasons, by importing maintenance reasons into a preset database to retrieve the standard number of maintenance personnel, and then comparing it with the real-time number of maintenance personnel, can accurately determine whether the current manpower allocation is reasonable. For example, when dealing with complex circuit fault repairs, the database provides a standard configuration of 8 professional maintenance personnel based on past similar repair cases and professional standards. If the real-time number of personnel on site is 6 and the difference is less than 0, the system generates a prompt message to prompt timely dispatch of personnel to avoid maintenance delays due to insufficient manpower; on the contrary, if there are 12 personnel on site, which exceeds the reasonable number, it may cause congestion on site and confusion in division of labor. The system prompts to optimize personnel arrangements to ensure an efficient and smooth maintenance process.

[0107] Real-time location monitoring optimizes personnel layout and collects distance information between the real-time location of maintenance personnel and reference point A2. When the distance between personnel is too large and exceeds the preset time, such as in a key component replacement and maintenance task, the preset distance is 15 meters from the reference point and the preset time is 10 minutes. If a maintenance personnel exceeds this range for 10 minutes, the system prompts the personnel to stay away from the key operation area, so that management personnel can adjust the personnel's position in time, ensure that maintenance personnel concentrate on the effective operation range, avoid poor collaboration due to personnel dispersion, and improve overall maintenance efficiency.

[0108] Strengthen teamwork and reduce safety risks;

[0109] Personnel distance monitoring to prevent collision risks: The distance between all personnel performing maintenance tasks is collected. Once it is found that the distance between personnel is greater than the preset value, for example, when performing precision equipment maintenance in a narrow rail car, the preset personnel distance is 2 meters. If the distance between two people reaches 3 meters, it may cause communication difficulties, tool transfer difficulties, and collisions when people move around. At this time, a prompt message is generated to remind maintenance personnel to maintain a reasonable distance, ensure close and safe team collaboration, and reduce the probability of accidents caused by negligence.

[0110] Personnel management and control when task execution is approaching: Extract the task execution time point from the maintenance task information, and check the personnel distance information when there is a preset time left before the task is executed. For example, 5 minutes before the critical operation of track connection, the preset distance between key personnel is 3 meters. If it is found that the distance between some personnel is too far at this time, prompt information will be provided to remind personnel to quickly take their positions and work closely together to ensure that all personnel work together at the critical moment of the task, avoid affecting the maintenance quality and progress due to the failure of personnel to arrive in time, and ensure that the maintenance task is completed on time and safely.

[0111] When performing derailment maintenance, the specific process of the second collection module collecting derailment maintenance related information is as follows:

[0112] Before lifting and derailing a rail transit vehicle, inspect the lifting cable and obtain the lifting cable information;

[0113] During the hoisting process, real-time hoisting image collection is performed;

[0114] The lifting cable information and real-time lifting images constitute the derailment repair related information.

[0115] The process of obtaining the hoisting cable information is as follows:

[0116] The steel cable is inspected using a magnetic particle inspection method, i.e., the steel cable is magnetized and then magnetic powder is sprayed on its surface. After the spraying is completed, the number of regions with an area greater than a preset value for adsorbing magnetic powder and the number of regions with a length greater than a preset length are collected. When the area of ​​a region with an area greater than a preset value for adsorbing magnetic powder and a length greater than a preset length are both recorded, only one region is recorded. The total number of regions with an area greater than a preset value for adsorbing magnetic powder and the total number of regions with a length greater than a preset length are calculated to obtain the abnormal region and the number of abnormal regions, i.e., the first abnormal information;

[0117] At the same time, the ultrasonic detection method is used to detect the steel cable, and the abnormal area of ​​the steel cable and the number of abnormal areas of the steel cable are obtained, that is, the second abnormal information;

[0118] The real-time hoisting image acquisition process is as follows: a safety zone is set up on the hoisting path, and an image acquisition device is set up in the safety zone to collect image information in the safety zone during the hoisting of the rail transit vehicle in real time, that is, to obtain the real-time hoisting image;

[0119] Multiple methods are used to inspect the hoisting cables to ensure safe lifting. Using the magnetic particle inspection method, after the cables are magnetized and sprayed with magnetic powder, detailed information is collected about areas where the magnetic powder is adsorbed and the length is greater than the preset length. For example, in a subway vehicle derailment repair scenario, the preset magnetic powder adsorption area is 5 square centimeters and the length is 3 centimeters. When an area with an adsorption area of ​​8 square centimeters and a length of 4 centimeters in a certain section of the cable is detected, it is accurately marked as an abnormal area. By rigorously calculating the number of such abnormal areas, i.e., the first abnormal information, subtle defects on the cable surface caused by wear, rust, etc. can be accurately detected, providing a key basis for evaluating the cable's load-bearing capacity in advance and determining whether it is suitable for hoisting operations, effectively avoiding hoisting accidents caused by hidden dangers of the cable itself.

[0120] Ultrasonic testing is also used to detect cable anomalies from an internal structural perspective. This method can penetrate the cable and detect underlying issues that are invisible to the naked eye, such as broken wires and uneven stranding. This method provides the location and number of cable anomalies, representing secondary anomaly information. This method complements magnetic particle testing, providing a comprehensive internal and external inspection of the cable's condition, significantly improving the accuracy of cable safety performance assessments. This ensures that lifting is initiated only when the cable meets acceptable conditions, ensuring a foolproof lifting process for derailed vehicles.

[0121] Real-time image acquisition strengthens monitoring throughout the entire lifting process. Image acquisition equipment is installed in safe areas along the lifting route to capture real-time image information throughout the entire process of lifting rail transit vehicles. This is like providing "eyes" for the lifting operation, allowing managers to instantly see the actual situation at each stage of the vehicle's lifting, translation, and lowering. Regardless of day or night, and in complex weather conditions, they can clearly grasp the dynamics of the site. For example, during nighttime repairs, real-time images can accurately observe key details such as whether the lifting cable is vertical and whether the vehicle's posture in the air is stable. This allows for the timely detection and correction of potential problems such as deflection and shaking, ensuring that the lifting operation is carried out strictly according to standard procedures.

[0122] Based on real-time images, on the one hand, the human body recognition model can be used to quickly detect unauthorized personnel who intrude into the safe area. Once the human body model is recognized, auxiliary information for the lifting process is immediately generated to prompt the person to leave and avoid casualties. On the other hand, the lifting speed and release speed are obtained through image analysis and processing. When the speed exceeds the preset value, the operator is promptly reminded to adjust to prevent risks such as uneven force on the steel cable and vehicle collision with surrounding facilities due to excessive speed, thereby ensuring the safety and orderly progress of the lifting operation in all aspects.

[0123] The derailment maintenance auxiliary information includes hoisting cable auxiliary information and hoisting process auxiliary information;

[0124] The process of obtaining the auxiliary information of the hoisting cable is as follows:

[0125] Extracting the acquired hoisting cable information, and extracting the first abnormality information and the second abnormality information therefrom;

[0126] When either the number of abnormal areas in the first abnormal information or the number of abnormal areas in the second abnormal information is greater than a preset value m1, auxiliary information for hoisting steel cables is generated;

[0127] When the number of abnormal areas of the first abnormal information and the number of abnormal areas of the second abnormal information are both less than the preset value m2 but greater than 0, the abnormal areas of the first abnormal information and the abnormal areas of the second abnormal information are extracted. When the abnormal areas of the first abnormal information and the abnormal areas of the second abnormal information are different areas, the lifting steel cable auxiliary information is generated. At this time, the content of the lifting steel cable auxiliary information is that there is an abnormality in the lifting steel cable.

[0128] The process of obtaining auxiliary information during the lifting process is as follows:

[0129] The real-time hoisting image is extracted and a human body recognition model is imported into the real-time hoisting image. When a human body model is recognized in the safe area, auxiliary information of the hoisting process is generated. At this time, the auxiliary information of the hoisting process indicates that there are abnormal people in the safe area and they need to be expelled;

[0130] Then, the real-time hoisting image is processed and analyzed during the hoisting process to obtain the hoisting speed and release speed;

[0131] When either the lifting speed or the release speed exceeds the preset value, the lifting process auxiliary information is generated. In this case, the lifting process auxiliary information indicates that the lifting speed / release speed is too fast and needs to be adjusted.

[0132] The above process quantitatively assesses and promptly warns of major hidden dangers. By setting a clear preset value, m1, to compare the number of abnormal areas in the first and second abnormal information, the system immediately generates auxiliary information for the hoisting cable. For example, at a light rail derailment repair site, m1 is preset to three locations. If magnetic particle testing reveals four areas where the adsorption area and length of magnetic powder exceed the standard, or if ultrasonic testing detects five internal abnormal areas, this indicates a serious safety risk to the cable and the possibility of breakage during the lifting process. This timely warning allows the maintenance team to decisively suspend operations and replace the cable, eliminating the accident at the source and ensuring a stable lifting foundation.

[0133] Carefully identify potential risks to nip them in the bud. For situations where the number of abnormal areas is less than the preset value m2 (assuming m2 is 2) but greater than 0, further analyze the location of the abnormal areas. If different detection methods find different abnormal areas, it means that although the steel cable does not have large-scale serious defects, there have been scattered minor problems, such as local slight wear and tear, and minor twisting of individual steel wires. These minor problems may also cause major failures if accumulated or when unevenly stressed. The generated auxiliary information for hoisting steel cables can guide maintenance personnel to carry out targeted maintenance, reinforcement or enhanced monitoring of the steel cables in advance, nip risks in the bud, and ensure the smooth progress of subsequent hoisting operations.

[0134] Real-time monitoring of the lifting process to ensure dynamic safety on site;

[0135] Utilizing a human recognition model imported from real-time hoisting images, the system instantly generates auxiliary information about the hoisting process as soon as a human model is identified within a safe zone, alerting and removing any unauthorized personnel. In busy rail transit maintenance stations, where other workers and uninformed passersby frequently pass through, the consequences of someone mistakenly entering the hoisting safety zone and coming to rest beneath a lifted load could be disastrous. This intelligent recognition and instant warning mechanism eliminates potential safety hazards immediately, ensuring the safety of personnel at the hoisting site.

[0136] Precise and efficient speed control: Real-time image analysis measures the lifting and release speeds, comparing them to preset values. If overspeed is detected, auxiliary information is immediately generated requesting adjustments. For example, when lifting a heavy subway car, the preset lifting speed is 0.5 meters per second. If an operator error or equipment failure causes the speed to soar to 1 meter per second, the excessive speed will subject the steel cables to significant impact loads and may cause the car to sway significantly in mid-air, potentially colliding with surrounding facilities. Precise speed control effectively avoids such hazards and ensures a smooth and orderly lifting process.

[0137] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of the technical features being referred to. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of the present invention, "plurality" means at least two, such as two, three, etc., unless otherwise specifically defined.

[0138] In the description of this specification, the reference terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" mean that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine different embodiments or examples described in this specification and features of different embodiments or examples without contradiction.

[0139] Although the embodiments of the present invention have been shown and described above, it will be understood that the above embodiments are illustrative and are not to be construed as limitations on the present invention. A person skilled in the art may change, modify, replace and modify the above embodiments within the scope of the present invention.

Claims

1. An artificial intelligence-based intelligent maintenance assistance system for rail transit vehicles, characterized in that: It includes a maintenance mode acquisition module, a first acquisition module, a second acquisition module, a data processing module and an information sending module; The maintenance mode collection module is used to collect maintenance modes, which include on-track maintenance and off-track maintenance; The first acquisition module is used to collect information related to on-orbit maintenance during on-orbit maintenance; The second collection module is used to collect information related to derailment maintenance during derailment maintenance; The data processing module is used to process the on-orbit maintenance related information and the derailment maintenance related information to generate on-orbit maintenance auxiliary information and derailment maintenance auxiliary information; The information sending module is used to send on-track maintenance auxiliary information and derailment maintenance auxiliary information to a preset receiving terminal; The specific process of the first acquisition module collecting on-orbit maintenance related information is as follows: When conducting on-orbit maintenance, maintenance personnel need to set up protective signs. Position sensors and angle acquisition devices are installed on the protective signs to collect the position and angle of the protective signs. At the same time, maintenance personnel need to wear location sensing equipment to collect their real-time location; At the same time, information on maintenance reasons and the number of real-time maintenance personnel are collected; That is, track maintenance related information includes the location and placement angle of protective signs, the real-time location of maintenance personnel, maintenance reason information and the real-time number of maintenance personnel; The data processing module processes the location and placement angle of the protective signs, the real-time location of the maintenance personnel, the maintenance reason information, and the real-time number of maintenance personnel to generate on-orbit maintenance auxiliary information; On-orbit maintenance auxiliary information includes protective sign prompt information and personnel prompt information; The process of obtaining the placement angle is as follows: Two tilt sensors are set on the side of the protective sign to collect the tilt information of the protective sign in real time and mark it as the first tilt angle and the second tilt angle; A collection frequency is preset, and the first inclination angle and the second inclination angle are collected multiple times according to the preset collection frequency to obtain multiple first inclination angles and multiple second inclination angles, which constitute placement angle information; When performing derailment maintenance, the specific process of the second collection module collecting derailment maintenance related information is as follows: Before lifting and derailing a rail transit vehicle, inspect the lifting cable and obtain the lifting cable information; During the hoisting process, real-time hoisting image collection is performed; The lifting cable information and real-time lifting images constitute the derailment repair related information; The process of obtaining the hoisting cable information is as follows: The steel cable is inspected using a magnetic particle inspection method, i.e., the steel cable is magnetized and then magnetic powder is sprayed on its surface. After the spraying is completed, the number of regions with an area greater than a preset value for adsorbing magnetic powder and the number of regions with a length greater than a preset length are collected. When the area of ​​a region with an area greater than a preset value for adsorbing magnetic powder and a length greater than a preset length are both recorded, only one region is recorded. The total number of regions with an area greater than a preset value for adsorbing magnetic powder and the total number of regions with a length greater than a preset length are calculated to obtain the abnormal region and the number of abnormal regions, i.e., the first abnormal information; At the same time, the ultrasonic detection method is used to detect the steel cable, and the abnormal area of ​​the steel cable and the number of abnormal areas of the steel cable are obtained, that is, the second abnormal information; The process of acquiring the real-time hoisting image is as follows: a safety area is set up on the hoisting path, and an image acquisition device is set up in the safety area. Image information in the safety area during the hoisting of the rail transit vehicle is acquired in real time, that is, the real-time hoisting image is acquired.

2. The artificial intelligence-based intelligent maintenance assistance system for rail transit vehicles according to claim 1, characterized in that: The process of obtaining the protection sign prompt information is as follows: Extracting the collected placement angles, and extracting a plurality of first inclination angles and a plurality of second inclination angles from the placement angles; Processing the plurality of first inclination angles, first detecting whether there is a first inclination angle among the plurality of first inclination angles that differs from a preset standard inclination angle by more than a first preset margin, generating a protective sign prompt message when there is a first inclination angle that differs from the preset standard inclination angle by more than a preset margin, calculating an average of the plurality of first inclination angles, and also generating a protective sign prompt message when the average of the plurality of first inclination angles differs from the preset standard inclination angle by more than a second preset margin; The plurality of second inclination angles are processed in the same manner as the plurality of first inclination angles to determine whether to generate a protective sign prompt message; Extract the location of the protective sign, then collect the location of the rail transit vehicle that needs maintenance, collect its center point and mark it as the reference point; The position of the protective sign is marked as A1, and the reference point is marked as A2. The distance information between the protective sign position A1 and the reference point A2 is monitored in real time to obtain the evaluation distance. When the evaluation distance information is less than the preset standard placement distance, a protective sign prompt message is generated; The assessment distance is collected at a preset frequency, and when the assessment distance collected at the preset frequency gradually decreases, a protective sign prompt message is generated.

3. The artificial intelligence-based intelligent maintenance assistance system for rail transit vehicles according to claim 2, characterized in that: The specific process of obtaining the personnel prompt information is as follows: Extract the real-time location of maintenance personnel, maintenance reason information and the real-time number of maintenance personnel from track maintenance related information; Import the maintenance reason information into the preset database, retrieve the standard maintenance personnel number from the preset database, and calculate that when the difference between the real-time maintenance personnel number and the standard maintenance personnel number is less than 0 or greater than the preset value a, generate personnel prompt information; Then collect the distance information between the maintenance personnel's real-time position and the reference point A2 to obtain the personnel distance information. When the personnel distance information is greater than the preset value for more than the preset time, a personnel prompt message is generated; At the same time, the distance between all personnel performing maintenance tasks is collected. When the distance between all personnel performing maintenance tasks is greater than the preset value, a personnel prompt message is generated; Collect maintenance task information assigned to maintenance personnel, extract the task execution time point from the maintenance task information, and when there is a preset time left before the task execution time point, extract the personnel distance information of all maintenance personnel performing the maintenance task. When there is personnel distance information greater than the preset value, generate personnel prompt information.

4. The rail transit vehicle intelligent maintenance assistance system based on artificial intelligence according to claim 3 is characterized in that: The derailment maintenance auxiliary information includes hoisting cable auxiliary information and hoisting process auxiliary information; The process of obtaining the auxiliary information of the hoisting cable is as follows: Extracting the acquired hoisting cable information, and extracting the first abnormality information and the second abnormality information therefrom; When either the number of abnormal areas in the first abnormal information or the number of abnormal areas in the second abnormal information is greater than a preset value m1, auxiliary information for hoisting steel cables is generated; When the number of abnormal areas of the first abnormal information and the number of abnormal areas of the second abnormal information are both less than the preset value m2 but greater than 0, the abnormal areas of the first abnormal information and the abnormal areas of the second abnormal information are extracted. When the abnormal areas of the first abnormal information and the abnormal areas of the second abnormal information are different areas, the auxiliary information of the hoisting cable is generated; The process of obtaining auxiliary information during the lifting process is as follows: Extract the acquired real-time hoisting image, import the human body recognition model into the real-time hoisting image, and generate auxiliary information of the hoisting process when the human body model is recognized in the safe area; Then, the real-time hoisting image is processed and analyzed during the hoisting process to obtain the hoisting speed and release speed; When either the lifting speed or the releasing speed is greater than a preset value, auxiliary information of the lifting process is generated.

Images