Rail transit fault prediction method and system

Abstract

The invention discloses a rail transit fault prediction method and system. The method comprises the following steps: S1, acquiring current running state data and a timestamp of a vehicle door system; s2, standardizing the operation state data, inputting the operation state data into a fault prediction neural network model, and predicting future operation state data and time; performing cyclic prediction, and outputting the running state data and the corresponding fault time when the future running state data is the running state data of the fault; wherein the fault prediction neural network model is constructed based on a long short-term memory network and is trained based on historical operation state data of the vehicle door system; s3, inputting the running state data predicted and output in the step S2 into a fault diagnosis model, and outputting to obtain one or more fault categories; wherein the fault diagnosis model is established based on a support vector machine and is trained based on historical operation state data of the vehicle door system. The fault type and the fault occurrence time of the vehicle door system can be accurately predicted.

Description

Technical Field

[0001] The present invention relates to the field of intelligent management technology for rail transit, and more specifically, to a method and system for predicting faults in rail transit. Background Art

[0002] The door system of rail transit vehicles is directly related to the personal safety of passengers and is an important subsystem of rail transit vehicles. Its reliability and safety have always been the focus of urban rail transit operation departments. Due to the short distance between rail transit stations, the doors are opened and closed very frequently, the door control electrical and pneumatic components are easily damaged, and there are many factors of passenger interference, which makes the door failures frequent. This not only affects the normal operation of the train, delays passengers' travel, but also endangers the personal safety of passengers. According to existing technical statistics, door system failures account for more than 30% of the total number of rail transit vehicle failures, ranking first among the subsystems of rail transit vehicles.

[0003] Due to the complex structure and strong electromechanical coupling of rail transit door systems, it is difficult to establish an accurate physical model to describe the mechanism of its state change. Combined with the development and application of existing sensor technology, data storage technology, communication technology and other technologies, it is possible to monitor the state of the door system in real time, and to transmit and store it in real time, making it possible to analyze the monitoring data based on the door system. However, since the changes in monitoring data caused by door system failures are diverse and related, it is necessary to establish a prediction method and system to capture the potential characteristic relationship between monitoring data and failures in order to achieve failure prediction of the door system. Summary of the invention

[0004] An object of the present invention is to solve at least the above problems and to provide at least the advantages which will be described hereinafter.

[0005] In order to achieve these purposes and other advantages according to the present invention, a rail transit fault prediction method is provided, comprising the following steps: S1, obtaining the current operating status data and timestamp of the door system; S2, standardize the operation status data, input it into a fault prediction neural network model, and predict the future operation status data and time; Cyclic prediction: when the future operating status data is the operating status data of a fault, the operating status data and the corresponding fault time are output; Wherein, the fault prediction neural network model is constructed based on a long short-term memory network and trained based on historical operating status data of the vehicle door system; S3, inputting the running status data predicted and outputted in step S2 into a fault diagnosis model, and outputting one or more fault categories; The fault diagnosis model is constructed based on a support vector machine and trained based on historical operating status data of the vehicle door system.

[0006] Preferably, the operating status data includes motor current, motor encoder position, relay switching time and frequency, door position signal, door lock status, sensor output signal of obstacle detection system, electromagnetic lock current, communication network status, indicator light signal and software CPU usage.

[0007] Preferably, the parameters of the fault prediction neural network model are set as follows: the number of hidden layers is 6, the number of neurons in each hidden layer is 30, and the number of recursive steps folded within a time step is 20.

[0008] Preferably, the fault diagnosis model includes 6 fault categories, namely, electrical fault, mechanical fault, obstacle detection fault, electromagnetic lock fault, communication network fault and software fault, and the normal state is 1 normal category, including 7 categories in total. For each category, the fault diagnosis model is provided with a support vector machine model; In step S3, specifically: the operating status data is diagnosed using all seven support vector machine models respectively, and each category of the support vector machine model outputs a score, and the output score is higher than the preset diagnosis threshold value of the fault type.

[0009] Preferably, it also includes: When a fault occurs in the door system, the acquired operating status data of the door system at the time of the fault is input into the fault diagnosis model after being standardized, and one or more fault categories are outputted.

[0010] The present invention also provides a rail transit fault prediction system, comprising: A data acquisition module, which is used to obtain the current operating status data and timestamp of the door system; A fault prediction module is used to standardize the operation status data, input it into a fault prediction neural network model, and predict the future operation status data and time; when the future operation status data is the operation status data of the fault, the operation status data and the corresponding fault time are output; wherein the fault prediction neural network model is constructed based on the long short-term memory network and trained based on the historical operation status data of the vehicle door system; A fault diagnosis module is used to input the predicted output operating status data into a fault diagnosis model, and output one or more fault categories; wherein the fault diagnosis model is constructed based on a support vector machine and trained based on the historical operating status data of the vehicle door system.

[0011] Preferably, when a fault occurs in the door system, the fault diagnosis module is further used to obtain the operating status data of the door system when the fault occurs, subject it to standardization processing, input it into the fault diagnosis model, and output one or more fault categories.

[0012] The present invention provides an electronic device, comprising: one or more processors; and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor executes the rail transit fault prediction method.

[0013] The present invention provides a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, the rail transit fault prediction method is implemented.

[0014] The present invention has at least the following beneficial effects: First, by obtaining the current operating status data and timestamp of the door system, the current operating condition of the door system can be accurately reflected. Then, through the prediction of the fault prediction neural network model, it can be known that the door system may fail at a certain point in the future, thereby assisting maintenance personnel to inspect the door system in advance and reduce the probability of actual failure of the door system.

[0015] Second, in order to further improve the maintenance efficiency, the predicted operating status data of the door system is used to diagnose the fault type through a fault diagnosis model, thereby assisting maintenance personnel to accurately inspect and repair numerous components in the door system and improve maintenance efficiency.

[0016] Third, based on the operating status data of the door system at the time of failure, the fault diagnosis model can be used to directly diagnose the fault category of the current fault, that is, the fault type, to guide maintenance personnel to arrange time for inspection and maintenance in advance.

[0017] Other advantages, objectives and features of the present invention will be embodied in part through the following description, and in part will be understood by those skilled in the art through study and practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 The figure is a schematic diagram of the process framework of one of the technical solutions of the present invention. DETAILED DESCRIPTION

[0019] The present invention is further described in detail below in conjunction with the accompanying drawings so that those skilled in the art can implement the invention with reference to the description.

[0020] It should be noted that the experimental methods described in the following embodiments are conventional methods unless otherwise specified, and the reagents and materials are commercially available unless otherwise specified; in the description of the present invention, the orientation or positional relationship indicated by the terms is based on the orientation or positional relationship shown in the accompanying drawings, which is only for the convenience of describing the present invention and simplifying the description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention.

[0021] like Figure 1 As shown, the rail transit fault prediction method includes the following steps: S1. Obtain the current operating status data and timestamp of the vehicle door system; the operating status data includes normal operating status data and fault operating status data. Specifically, based on the types of faults that often occur in the vehicle door system and their correlation with each other, as well as the limitations of data acquisition equipment under existing conditions, the operating status data is usually collected as motor current, motor encoder position, relay switching time and frequency, door position signal, door lock status, sensor output signal of the obstacle detection system, electromagnetic lock current, communication network status, indicator light signal, software CPU usage rate, etc.

[0022] S2, standardize the operation status data, input it into a fault prediction neural network model, and predict the future operation status data and time; Cyclic prediction: when the future operating status data is the operating status data of a fault, the operating status data and the corresponding fault time are output; Among them, the fault prediction neural network model is constructed based on the long short-term memory network and trained based on the historical operation status data of the door system; due to the short distance between rail transit stations, the doors are opened and closed very frequently, and the door controller, obstacle detection system, electromagnetic lock and other components are very easy to be damaged. Based on the above operation law of the door system, it is appropriate to use the long short-term memory network to capture the potential characteristic relationship between the operation status data and the fault. Specifically, the fault prediction neural network model has a state unit, which can store the historical operation status data, timestamp and whether it is a fault state of the door system, and update it. The number of stored data groups is set to 20 groups. The operation status data and timestamp of the door system based on the history can reflect the actual state of each component of the door system, such as wear and tear, remaining service life, etc. The fault prediction neural network model predicts the operation status data, timestamp and whether it is a fault state of the next group of door systems in the future based on the 20 groups of operation status data, timestamp and whether it is a fault state in the state unit. When it is a fault state, the corresponding operation status data, timestamp and fault state are output. When the prediction is a normal state, the data of the state unit is updated with the predicted running state data, timestamp, and whether it is a fault state, and the prediction is cyclically performed until the predicted running state data is a fault state, and the fault state, running state data, and timestamp are output. That is, the time of the future fault is obtained, so that maintenance preparations can be made in advance. Specifically, the method of judging whether the running state data is fault data or normal state can be classified by the support vector machine method.

[0023] S3, inputting the running status data predicted and outputted in step S2 into a fault diagnosis model, and outputting one or more fault categories; The fault diagnosis model is constructed based on a support vector machine and trained based on the historical operating status data of the door system. One-to-one diagnosis based on the common fault types of the door system can significantly improve the accuracy, precision and recall of fault diagnosis and improve maintenance efficiency.

[0024] In the above technical solution, by obtaining the current operating status data and timestamp of the door system, the current operating condition of the door system can be accurately reflected, and then through the prediction of the fault prediction neural network model, it can be known that the door system may fail at a certain time point in the future, thereby assisting maintenance personnel to inspect the door system in advance and reduce the probability of actual failure of the door system. In order to further improve the inspection efficiency, the predicted operating status data of the door system is also diagnosed through the fault diagnosis model, so that maintenance personnel can accurately inspect and repair many components in the door system and improve the inspection efficiency.

[0025] In another technical solution, the operating status data includes motor current, motor encoder position, relay switching time and frequency, door position signal, door lock status, sensor output signal of obstacle detection system, electromagnetic lock current, communication network status, indicator light signal and software CPU usage.

[0026] In the above technical solution, in order to more accurately reflect the actual operating conditions of the door system, a detailed investigation of the fault types was conducted. The types of faults in the rail transit door system are mainly analyzed as follows: Electrical failure: This type of failure mainly occurs in door controllers, power supplies, indicator lights, relays, etc. When this error occurs, it will affect the normal opening and closing of the door. For example, rail vehicles cannot receive control commands in time because the vehicle's door opening / closing control system cannot receive control commands in time due to abnormal interference caused by external signals, which in turn prolongs the door opening and closing time. At the same time, when the door is closed, the signal transmission is interrupted, and the current increases a lot instantly. It will also cause the motor control to be interrupted, ultimately affecting the safe and reliable operation of the vehicle; Mechanical failure: There are many reasons for this failure, among which the relevant parts of the door are deformed due to extrusion, or the dimensional error is relatively large during production, so the hardware conditions cannot meet the relevant requirements. At the same time, during the operation of the rail vehicle, the frequency of the door opening and closing is relatively high, so the hardware condition of the door changes, which is also the main reason for this type of failure. When the passenger flow is relatively large, the passengers will be crowded when getting on and off the car, and many passengers will be squeezed out of the door position. In the long run, the door will be deformed, which will cause common faults such as slow opening and closing of the door, which is not conducive to the normal operation of the rail vehicle; Obstacle detection failure: mainly includes the detection of obstacles when opening and the detection of obstacles when closing. In rail transit, how to detect obstacles in rail transit during the locking process of rail transit is a very important task. If the obstacle detection system fails, it will not be able to detect obstacles correctly, resulting in the door being unable to close or open normally; Electromagnetic lock failure: Electromagnetic lock is an important mechanical component in the process of door opening and closing. Analyzing and rectifying its defects can effectively improve the operational stability of the platform door system. Electromagnetic lock failure may affect the normal locking and unlocking of the door, causing the door to be unable to open and close normally; Communication network failure: Failure of the train communication network may also affect the door obstacle detection system, because the door control system relies on the train communication network to receive and send signals. If the communication network has frame errors, process data problems, frame interval problems, signal amplitude problems, signal rise time problems or interference, etc., the door obstacle detection system may not work properly; Software and system failure: Problems with the system software may cause the door to be closed but the lights to be off or flickering, affecting the normal opening and closing of the door.

[0027] These fault causes cover multiple aspects from hardware to software, from mechanical to electrical, and need to be comprehensively considered and investigated to determine the specific fault causes and take corresponding maintenance measures. Based on the above existing fault conditions, the operating status data (including normal operating status data and operating status data during fault) of the relevant components of the door system are collected to form a data set, and the fault type and normal category are marked. The data set is standardized and processed, and then converted into vector data adapted to the SVM model. Each vector data corresponds to a label (fault type and normal category), and the data set is divided into a training set and a test set. Specifically, the operating status data includes motor current, motor encoder position, relay switching time and frequency, door position signal, door lock state, sensor output signal of the obstacle detection system, electromagnetic lock current, communication network status, indicator light signal, software CPU usage, etc. Among them, normal operating status data is easy to collect, while the operating status data samples during fault are relatively few. Therefore, this application adopts the establishment of a separate door system and the installation of corresponding data acquisition modules in the door system to artificially create various types of faults to collect operating status data samples during faults. For example, during the door closing process, an obstacle is placed and signal interference is set up nearby to collect relevant data on the obstacle detection fault type.

[0028] In another technical solution, the parameters of the fault prediction neural network model are set as follows: the number of hidden layers is 6, the number of neurons in each hidden layer is 30, and the number of recursive steps folded within the time step is 20. There are many types of faults based on the door system, and when a fault occurs, it is reflected in the data collected by the data acquisition device. There is a correlation relationship. Therefore, more potential features need to be extracted to more accurately reflect the relationship between the collected data and the fault. The present invention mines the relationship between the collected data and the fault by setting the number of hidden layers to 6 and the number of neurons in each hidden layer to 30, and the number of superimposed recursive steps to 20 to find the relationship between the historical operating status data and the time of the fault occurrence.

[0029] Specifically, the model parameters are set. The parameters of the fault prediction neural network model of the present application are set as follows: the number of neurons in the hidden layer (HIDDEN-SIZE) is 30, the number of hidden layers (NYM-LAYERS) is 6, the number of recursive steps folded within the time step (TIMESTEPS) is 20, the number of training steps (TRAIN-STEPS) is 2000, the training batch size (BATCH-SIZE) is 128, and the learning rate (LEARNING-RATE) is 0.1.

[0030] The model evaluation standard is set. The fault prediction neural network model of this application uses the mean square error MSE to evaluate the prediction results. A smaller MSE represents a higher description accuracy of the fault prediction neural network model for the data. The calculation method of MSE is: in, n Represents the total number of sample data, Represents the forecast data, Represents real measurement data.

[0031] The trained fault prediction neural network model is used to predict the faults of the rail transit door system. The data to be predicted (the data to be predicted is the operating status data and timestamp of the relevant components of the door system) is input, standardized, and then converted into a time series format suitable for LSTM input. It is input into the fault prediction neural network model, and the time of the next fault and the corresponding operating status data are output.

[0032] In another technical solution, the fault diagnosis model includes 6 fault categories, namely, electrical fault, mechanical fault, obstacle detection fault, electromagnetic lock fault, communication network fault and software fault, and the normal state is 1 normal category, including 7 categories in total. For each category, the fault diagnosis model is provided with a support vector machine model; In step S3, specifically: the operating status data is diagnosed using all seven support vector machine models respectively, and each category of the support vector machine model outputs a score, and the output score is higher than the preset diagnosis threshold value of the fault type.

[0033] In the above technical solution, a fault diagnosis model is constructed and trained based on the common fault categories of the door system. Specifically: Construct a fault diagnosis model. The fault diagnosis model is based on the support vector machine (SVM). The specific settings are divided into electrical fault, mechanical fault, obstacle detection fault, electromagnetic lock fault, communication network fault, software fault, and normal state. There are a total of 6 fault categories and 1 normal category. Train an SVM model for each category. Specifically: for a data set of 7 categories, 7 SVM models need to be trained. For each category, mark the samples of this category as positive (+1), and mark the samples of the other 6 categories as negative (-1). Then use the training set data to train each SVM model.

[0034] Model evaluation. The fault diagnosis model of this application is evaluated by accuracy, precision and recall. For each SVM model, there are four important indicators, namely, true positive (TP), which represents the number of samples with faults that are correctly judged as faulty; false positive (FP), which represents the number of samples without faults that are mistakenly judged as faulty; false negative (FN), which represents the number of samples with faults that are mistakenly judged as not faulty; true negative (TN), which represents the number of samples without faults that are correctly judged as not faulty.

[0035] The accuracy calculation formula is: Precision refers to the proportion of samples that are actually faulty among samples that are classified as faulty. The formula for calculating precision is: Recall refers to the proportion of samples that are correctly judged as faulty among samples that actually have faults. The calculation formula for recall is: Based on the above evaluation method, 7 SVM models are evaluated, that is, the fault diagnosis model is evaluated.

[0036] Fault diagnosis, after data preprocessing, the operating status data collected before the fault occurs is obtained in a data form that is compatible with the fault diagnosis model, that is, the data to be diagnosed. The data to be diagnosed is predicted using all 7 SVM models respectively, and each SVM model will give a score, indicating the confidence that the data to be diagnosed belongs to the corresponding category. Usually, the category corresponding to the SVM model with the highest score is selected as the predicted category of the data to be diagnosed. However, due to the correlation between the various fault types in the door system, a diagnostic threshold is set based on experience, and all fault types above the diagnostic threshold are output for maintenance personnel to check. The diagnostic threshold can also be adjusted according to the maintenance data later to improve the accuracy of the fault type output and improve maintenance efficiency.

[0037] In another technical solution, it also includes: When a fault occurs in the door system, the acquired operating status data of the door system at the time of the fault is input into the fault diagnosis model after being standardized, and one or more fault categories are outputted.

[0038] In the above technical solution, the operating status data of the door system when it fails is directly diagnosed through the fault diagnosis model, and the fault category of the current fault, that is, the fault type, can be directly output to guide maintenance personnel to arrange time for inspection and maintenance in advance.

[0039] In another technical solution, a support vector machine algorithm is used to classify the operating state data predicted by the fault prediction neural network model into fault state and normal state. When classified as a fault state, the output is fault data, the time is the fault time, and the operating state data is the operating state data of the fault; when classified as a normal state, the predicted operating state data is used to update the data group stored in the state unit of the fault prediction neural network model, and the next group of operating state data and timestamps are cyclically predicted until the operating state data of the fault is obtained, and the operating state data and timestamp are output. That is, the time of the future fault is obtained, so that maintenance preparations can be made in advance.

[0040] Based on the same inventive concept, the present invention provides a rail transit fault prediction system, comprising: A data acquisition module, which is used to obtain the current operating status data and timestamp of the door system; A fault prediction module is used to standardize the operation status data, input it into a fault prediction neural network model, and predict the future operation status data and time; when the future operation status data is the operation status data of the fault, the operation status data and the corresponding fault time are output; wherein the fault prediction neural network model is constructed based on the long short-term memory network and trained based on the historical operation status data of the vehicle door system; A fault diagnosis module is used to input the predicted output operating status data into a fault diagnosis model, and output one or more fault categories; wherein the fault diagnosis model is constructed based on a support vector machine and trained based on the historical operating status data of the vehicle door system.

[0041] In the above technical solution, by obtaining the current operating status data and timestamp of the door system, the current operating condition of the door system can be accurately reflected, and then through the prediction of the fault prediction neural network model, it can be known that the door system may fail at a certain time point in the future, thereby assisting maintenance personnel to inspect the door system in advance and reduce the probability of actual failure of the door system. In order to further improve the inspection efficiency, the predicted operating status data of the door system is also diagnosed through the fault diagnosis model, so that maintenance personnel can accurately inspect and repair many components in the door system and improve the inspection efficiency.

[0042] In another technical solution, when a door system fails, the fault diagnosis module is also used to obtain the operating status data of the door system when the fault occurs, and input it into the fault diagnosis model after standardization, and output one or more fault categories.

[0043] In the above technical solution, the operating status data of the door system when it fails is directly diagnosed through the fault diagnosis model, and the fault category of the current fault, that is, the fault type, can be directly output to guide maintenance personnel to arrange time for inspection and maintenance in advance.

[0044] The division of modules in the embodiments of the present application is schematic and is only a logical function division. There may be other division methods in actual implementation. In addition, each functional module in each embodiment of the present invention may be integrated into a processor, or may exist physically separately, or two or more modules may be integrated into one module. The above-mentioned integrated modules may be implemented in the form of hardware or in the form of software functional modules.

[0045] The present invention provides an electronic device, comprising: one or more processors; and a memory in communication with the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor executes the rail transit fault prediction method. The electronic device may be any terminal device including a mobile phone, a laptop computer, a desktop computer, a tablet computer, a PDA (Personal Digital Assistant), a POS (Point of Sales), a vehicle-mounted computer, etc.

[0046] The present invention provides a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, the rail transit fault prediction method is implemented.

[0047] Through the description of the above implementation mode, the technicians in the relevant field can clearly understand that the present invention can be implemented by means of software plus necessary general hardware, and of course, it can also be implemented by special hardware including special integrated circuits, special CPUs, special memories, special components, etc. In general, all functions completed by computer programs can be easily implemented by corresponding hardware, and the specific hardware structure used to implement the same function can also be various, such as analog circuits, digital circuits or special circuits. However, for the present invention, software program implementation is a better implementation mode in more cases. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a readable storage medium, such as a computer floppy disk, a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, RandomAccess Memory), a disk or an optical disk, etc., including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present invention.

[0048] Although the embodiments of the present invention have been disclosed as above, they are not limited to the applications listed in the specification and the implementation modes, and they can be fully applied to various fields suitable for the present invention. For those familiar with the art, additional modifications can be easily implemented. Therefore, without departing from the general concept defined by the claims and the scope of equivalents, the present invention is not limited to the specific details and the illustrations shown and described herein.

Claims

1. A rail transit fault prediction method, characterized in that: The following steps are involved: S1, obtaining the current operating status data and timestamp of the door system; S2, standardize the operation status data, input it into a fault prediction neural network model, and predict the future operation status data and time; Cyclic prediction: when the future operating status data is the operating status data of a fault, the operating status data and the corresponding fault time are output; Wherein, the fault prediction neural network model is constructed based on a long short-term memory network and trained based on historical operating status data of the vehicle door system; S3, inputting the running status data predicted and outputted in step S2 into a fault diagnosis model, and outputting one or more fault categories; The fault diagnosis model is constructed based on a support vector machine and trained based on historical operating status data of the vehicle door system.

2. The rail transit fault prediction method according to claim 1, characterized in that: The operating status data includes motor current, motor encoder position, relay switching time and frequency, door position signal, door lock status, sensor output signal of obstacle detection system, electromagnetic lock current, communication network status, indicator light signal and software CPU usage.

3. The rail transit fault prediction method according to claim 1, characterized in that: The parameters of the fault prediction neural network model are set as follows: the number of hidden layers is 6, the number of neurons in each hidden layer is 30, and the number of recursive steps folded within a time step is 20.

4. The rail transit fault prediction method according to claim 1, characterized in that: The fault diagnosis model includes 6 fault categories, namely, electrical fault, mechanical fault, obstacle detection fault, electromagnetic lock fault, communication network fault and software fault, and the normal state is 1 normal category, including 7 categories in total. For each category, the fault diagnosis model is provided with a support vector machine model; In step S3, specifically: the operating status data is diagnosed using all seven support vector machine models respectively, and each category of the support vector machine model outputs a score, and the output score is higher than the preset diagnosis threshold value of the fault type.

5. The rail transit fault prediction method according to claim 1, characterized in that: Also includes: When a fault occurs in the door system, the acquired operating status data of the door system at the time of the fault is input into the fault diagnosis model after being processed in a standardized manner, and one or more fault categories are outputted.

6. Rail transit fault prediction system, characterized in that: include: A data acquisition module, which is used to obtain the current operating status data and timestamp of the door system; A fault prediction module is used to standardize the operation status data, input it into a fault prediction neural network model, and predict the future operation status data and time; when the future operation status data is the operation status data of the fault, the operation status data and the corresponding fault time are output; wherein the fault prediction neural network model is constructed based on the long short-term memory network and trained based on the historical operation status data of the vehicle door system; A fault diagnosis module is used to input the predicted output operating status data into a fault diagnosis model, and output one or more fault categories; wherein the fault diagnosis model is constructed based on a support vector machine and trained based on the historical operating status data of the vehicle door system.

7. The rail transit fault prediction system according to claim 6, characterized in that: When a fault occurs in the door system, the fault diagnosis module is further used to obtain the operating status data of the door system when the fault occurs, and input the data into the fault diagnosis model after standardization, and output one or more fault categories.

8. An electronic device, characterized in that include: one or more processors; and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor executes the rail transit fault prediction method according to any one of claims 1 to 5.

9. A computer-readable storage medium, characterized in that A computer program is stored on the storage medium, and when the program is executed by the processor, the rail transit fault prediction method described in any one of claims 1 to 5 is implemented.

Images