Railway vehicle anomaly detection method, device and equipment and storage medium

Abstract

The invention discloses a rail vehicle anomaly detection method, device and equipment and a storage medium, and relates to the field of rail transit vehicles, and the method comprises the steps: determining a target performance parameter of a target system in a target rail vehicle, constructing an anomaly detection model by using the target performance parameter and a target parameter having a preset influence relationship with the target performance parameter; performing prediction based on the anomaly detection model and the vehicle state parameter of the target railway vehicle to obtain a prediction value corresponding to the target performance parameter; determining a dynamic threshold band function about the target performance parameter by using the predicted value, and performing anomaly detection based on the dynamic threshold band function and the current vehicle state parameter; the dynamic threshold band function is a function representing that the data of the target performance parameter is in a preset normal interval. Therefore, the accuracy of anomaly detection can be improved.

Description

Technical Field

[0001] This invention relates to the field of rail transit vehicles, and in particular to a method, apparatus, equipment, and storage medium for detecting anomalies in rail vehicles. Background Technology

[0002] Prognostic and Health Management (PHM) technology for rail transit vehicles is crucial to the level of vehicle intelligence and operation and maintenance costs, and technologies such as anomaly detection of the target system are essential. However, existing technologies have many shortcomings, which seriously restrict the development of the industry.

[0003] First, in the PHM (Prognostics and Management) technology of rail vehicle target systems, existing anomaly detection and fault diagnosis methods have limited ability to detect and uncover early-stage anomalies and fault characteristics. When early data anomalies and fault characteristics appear in the target system, they are difficult to detect in a timely manner, resulting in the inability to achieve early warning of anomalies and faults. This makes it difficult to grasp the development trend of system anomalies and faults, and the maintenance and management of spare parts cannot be effectively planned, hindering the transformation of rail vehicles from condition-based maintenance to predictive maintenance. Second, after discovering anomaly characteristics, existing anomaly detection technologies can only provide alarms or results, and cannot perform correlation and cause analysis on other system parameters related to the anomaly. Therefore, when system anomalies occur, it is impossible to provide emergency response suggestions based on anomaly correlation analysis, which greatly reduces the efficiency of handling system anomalies and early faults.

[0004] Therefore, improving the accuracy of anomaly detection when performing anomaly detection on the subsystems of rail vehicles is a pressing technical problem that needs to be solved. Summary of the Invention

[0005] In view of this, the purpose of this invention is to provide a method, apparatus, device, and storage medium for anomaly detection of rail vehicles, which can improve the accuracy of anomaly detection when performing anomaly detection on subsystems of rail vehicles. The specific solution is as follows:

[0006] Firstly, this application provides an anomaly detection method for rail vehicles, comprising:

[0007] In the target rail vehicle, the target performance parameters of the target system are determined, and an anomaly detection model is constructed using the target performance parameters and target parameters that have a preset influence relationship with the target performance parameters.

[0008] Based on the anomaly detection model and the vehicle state parameters of the target rail vehicle, a prediction is made to obtain a predicted value corresponding to the target performance parameters.

[0009] The predicted value is used to determine a dynamic threshold band function for the target performance parameter, and anomaly detection is performed based on the dynamic threshold band function and the current vehicle state parameters; the dynamic threshold band function is a function that characterizes the data of the target performance parameter being within a preset normal range.

[0010] Optionally, the step of constructing an anomaly detection model using the target performance parameter and target parameters that have a preset influence relationship with the target performance parameter includes:

[0011] Historical data for determining the target performance parameters;

[0012] Determine the vehicle operating parameters, environmental parameters, and vehicle maintenance parameters that have a preset influence relationship with the target performance parameters, so as to determine the corresponding target parameters;

[0013] The historical data and the target parameters are trained using a preset machine learning algorithm to construct a corresponding anomaly detection model.

[0014] Optionally, the step of constructing an anomaly detection model using the target performance parameter and target parameters that have a preset influence relationship with the target performance parameter includes:

[0015] The physical laws governing the target performance parameters are determined in the target rail vehicle, and a mechanistic model of the target performance parameters is constructed based on the physical laws.

[0016] An anomaly detection model is constructed based on the aforementioned mechanism model and the target parameters that have a preset influence relationship with the target performance parameters.

[0017] Optionally, the step of predicting based on the anomaly detection model and the vehicle state parameters of the target rail vehicle to obtain a predicted value corresponding to the target performance parameters includes:

[0018] The vehicle operating conditions, vehicle operating parameters, environmental parameters, and vehicle maintenance parameters of the target rail vehicle are determined as vehicle status parameters.

[0019] The historical or current data of the vehicle state parameters are input into the anomaly detection model so that the anomaly detection model outputs a predicted value corresponding to the target performance parameter.

[0020] Optionally, determining the dynamic threshold band function with respect to the target performance parameter using the predicted value includes:

[0021] Determine the standard deviation of the predicted value, and determine the product of the standard deviation and a preset multiple;

[0022] Add the standard deviation and the product value together to obtain the sum;

[0023] The sum of the predicted value and the summed value is determined as the upper limit, and the difference between the predicted value and the summed value is determined as the lower limit.

[0024] A dynamic threshold band is determined based on the predicted value, the upper limit value, and the lower limit value, and a dynamic threshold band function is determined using the dynamic threshold band.

[0025] Optionally, the anomaly detection based on the dynamic threshold band function and the current vehicle state parameters includes:

[0026] The current vehicle state parameters of the target rail vehicle are input into the dynamic threshold band function to obtain the corresponding target dynamic threshold band;

[0027] The current data of the target performance parameter is compared with the target dynamic threshold band, and the corresponding comparison results are obtained.

[0028] If the comparison results show that the current data is within the range represented by the target dynamic threshold band at the same time, then the current state of the target performance parameter is determined to be normal.

[0029] If the comparison results show that the current data is not within the range represented by the target dynamic threshold band at the same time, then the current state of the target performance parameter is determined to be abnormal.

[0030] Optionally, after determining that the current state of the target performance parameter is abnormal, the method further includes:

[0031] The moment when the current state is abnormal is defined as the first moment, and the moment within the target time period is defined as the second moment; the target time period is the time period earlier than the first moment.

[0032] Record the target performance parameter data between the first time point and the second time point so that the obtained recorded data can be used for data analysis and data visualization.

[0033] Secondly, this application provides an anomaly detection device for rail vehicles, comprising:

[0034] The model building module is used to determine the target performance parameters of the target system in the target rail vehicle, and to build an anomaly detection model using the target performance parameters and target parameters that have a preset influence relationship with the target performance parameters.

[0035] The prediction value acquisition module is used to make predictions based on the anomaly detection model and the vehicle state parameters of the target rail vehicle to obtain the predicted values ​​corresponding to the target performance parameters.

[0036] An anomaly detection module is used to determine a dynamic threshold band function for the target performance parameter using the predicted value, and to perform anomaly detection based on the dynamic threshold band function and the current vehicle state parameters; the dynamic threshold band function is a function that characterizes the data of the target performance parameter being within a preset normal range.

[0037] Thirdly, this application provides an electronic device, comprising:

[0038] Memory, used to store computer programs;

[0039] A processor is used to execute the computer program to implement the aforementioned method for detecting anomalies in rail vehicles.

[0040] Fourthly, this application provides a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the aforementioned method for detecting anomalies in rail vehicles.

[0041] In this application, within the target rail vehicle, target performance parameters of the target system are determined, and an anomaly detection model is constructed using the target performance parameters and target parameters that have a preset influence relationship with the target performance parameters. Based on the anomaly detection model and the vehicle state parameters of the target rail vehicle, predictions are made to obtain predicted values ​​corresponding to the target performance parameters. The predicted values ​​are used to determine a dynamic threshold band function for the target performance parameters, and anomaly detection is performed based on the dynamic threshold band function and the current vehicle state parameters. The dynamic threshold band function is a function characterizing that the data of the target performance parameters is within a preset normal range. As can be seen from the above, in the target rail vehicle, this application first determines the target performance parameters of the target system, then constructs an anomaly detection model based on the target performance parameters and target parameters that have a preset influence relationship with the target performance parameters. Next, predictions are made using the anomaly detection model and the vehicle state parameters of the target rail vehicle to obtain predicted values ​​corresponding to the target performance parameters. Then, this predicted value is used to determine a dynamic threshold band function for the target performance parameters. Finally, anomaly detection is performed based on the dynamic threshold band function and the current vehicle state parameters. In this way, this application can improve the accuracy of anomaly detection when performing anomaly detection on the subsystems of rail vehicles, thereby effectively ensuring the safety and stability of rail vehicle operation. Attached Figure Description

[0042] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0043] Figure 1 This is a flowchart of an anomaly detection method for rail vehicles disclosed in this application;

[0044] Figure 2 This is a schematic diagram of a system anomaly detection based on real-time dynamic threshold disclosed in this application;

[0045] Figure 3 This is a correlation analysis diagram of bearing temperature with train running speed, motor traction force, and motor temperature disclosed in this application;

[0046] Figure 4 This application discloses a correlation analysis diagram of bearing temperature with ambient temperature, altitude, and line mileage markers.

[0047] Figure 5 This is a schematic diagram of the structure of an anomaly detection device for a rail vehicle disclosed in this application;

[0048] Figure 6 This is a structural diagram of an electronic device disclosed in this application. Detailed Implementation

[0049] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0050] Currently, in PHM (Prognostics and Management) technology for rail vehicle target systems, existing anomaly detection and fault diagnosis methods have limited ability to detect and uncover early-stage anomalies and fault characteristics. When early data anomalies and fault characteristics appear in the target system, they are difficult to detect in a timely manner, resulting in the inability to provide early warnings of anomalies and faults. This makes it difficult to grasp the development trend of system anomalies and faults, and hinders the effective planning of spare parts maintenance, impeding the transformation of rail vehicles from condition-based maintenance to predictive maintenance. Secondly, existing anomaly detection technologies can only provide alarms or results after detecting anomalies, without being able to perform correlation and cause analysis on other system parameters related to the anomaly. Therefore, when system anomalies occur, emergency response suggestions cannot be provided based on anomaly correlation analysis, greatly reducing the efficiency of handling system anomalies and early faults. Therefore, this application provides an anomaly detection method, device, equipment, and storage medium for rail vehicles, which can improve the accuracy of anomaly detection when performing anomaly detection on subsystems of rail vehicles.

[0051] See Figure 1 As shown in the figure, an embodiment of the present invention discloses an anomaly detection method for rail vehicles, including:

[0052] Step S11: In the target rail vehicle, determine the target performance parameters of the target system, and construct an anomaly detection model using the target performance parameters and target parameters that have a preset influence relationship with the target performance parameters.

[0053] In this embodiment, firstly, the target performance parameters of the target system in the target rail vehicle need to be clearly defined. The target system refers to the key subsystems within the target rail vehicle, that is, the most important systems among the various systems of the target rail vehicle. Key subsystems include the running gear system, traction system, ventilation system, braking system, pantograph system, etc., and different subsystems can correspond to different performance parameters. For example, the bogie bearing temperature in the running gear system and the traction motor current in the traction system can both be used as target performance parameters. When determining the target performance parameters, it is necessary to combine the functional characteristics of the key subsystems with the actual operation monitoring requirements, and select parameters that can effectively reflect the system's operating status from system design specifications, historical fault data, and operation and maintenance experience.

[0054] Next, an anomaly detection model needs to be constructed. In one specific implementation, historical data of the target performance parameters is acquired. Historical data often covers the operational data of the target rail vehicle under different operating conditions, environmental conditions, and maintenance statuses. The data acquisition cycle can be determined based on the parameter characteristics and detection accuracy requirements. For example, for slowly changing parameters such as temperature, a minute-level acquisition frequency can be used; for rapidly changing parameters such as current, a second-level or higher acquisition frequency is required. It is understood that the acquired data can be preprocessed, including noise reduction, missing value imputation, and normalization, to ensure that the data quality meets the requirements for model construction.

[0055] Furthermore, target parameters with a predetermined influence relationship with the target performance parameters are identified. These target parameters include vehicle operating parameters, environmental parameters, and vehicle maintenance parameters. Vehicle operating parameters, such as operating speed, traction, load, uphill / downhill conditions, and tunnel driving status, directly affect the workload and operating status of the subsystem. Environmental parameters encompass altitude, season, and weather conditions; environmental factors can significantly impact equipment performance, for example, bearing temperature may rise in high-temperature environments. Additionally, vehicle maintenance parameters include mileage, system time since last maintenance, and maintenance records; maintenance status affects the health level of the equipment, such as the potential performance degradation of components that have not been maintained for a long time.

[0056] After determining the target parameters, anomaly detection models can be constructed in two ways. The first approach is a data-driven model building method, which uses a pre-defined machine learning algorithm to train historical data and target parameters. Suitable machine learning algorithms include vector machines and convolutional neural networks. Taking convolutional neural networks as an example, preprocessed historical data of the target performance parameters are used as output variables, and the target parameters are used as input variables to construct a multi-layer neural network model. Through training on a large amount of data, the model learns the relationship between the target performance parameters, thereby enabling it to predict the normal range of values ​​for the target performance parameters based on the current target parameters.

[0057] The second approach is a model-building method based on mechanistic analysis. First, the physical laws governing the target performance parameters in the target rail vehicle are analyzed. For example, the traction motor current has a clear electromagnetic and dynamic relationship with motor speed, load torque, and power supply voltage. Based on these physical laws, a mechanistic model of the target performance parameters is constructed. This mechanistic model typically describes the relationships between parameters in the form of mathematical equations. Then, the mechanistic model is combined with the target parameters to construct an anomaly detection model.

[0058] When building an anomaly detection model, it can be validated and optimized. A portion of historical data not used in training can be selected as a test set. The prediction results of the anomaly detection model on the test set can be compared with the actual measured values ​​to evaluate the prediction accuracy and generalization ability of the model. If the performance of the anomaly detection model does not meet the requirements, the model structure, algorithm parameters, or supplementary data need to be adjusted and retrained until the anomaly detection model achieves the expected detection accuracy and reliability.

[0059] Step S12: Based on the anomaly detection model and the vehicle state parameters of the target rail vehicle, a prediction is made to obtain the predicted value corresponding to the target performance parameters.

[0060] In this embodiment, after determining the anomaly detection model, the vehicle state parameters of the target rail vehicle need to be input into the model to obtain predicted values ​​of the target performance parameters. The vehicle state parameters include vehicle operating conditions, vehicle operating parameters, environmental parameters, and vehicle maintenance parameters. Among these, vehicle operating conditions cover operating scenarios such as uphill / downhill, tunnels, and different track conditions, which directly affect the load characteristics and operating status of the target rail vehicle subsystem.

[0061] Then, when inputting historical or current data of vehicle status parameters into the anomaly detection model, an appropriate data input method can be selected based on the model's characteristics and the application scenario. For example, when inputting historical data, a sliding window approach can be used to select a time series of data before the target time as input. This method can fully utilize the temporal characteristics of the data and improve the accuracy of prediction.

[0062] For anomaly detection models based on mechanistic analysis, vehicle state parameters need to be substituted into the mathematical equations of the mechanistic model for calculation to obtain the corresponding predicted values. For example, for predicting traction motor current, based on the electromagnetic and dynamic equations of the motor, parameters such as the current motor speed, load torque, and power supply voltage are substituted into the equations to calculate the theoretical current value. Simultaneously, the influence of environmental parameters on the model parameters is considered, and the calculation results are corrected accordingly.

[0063] The predicted values ​​of target performance parameters obtained based on the anomaly detection model and vehicle state parameters provide a dynamic reference standard for subsequent anomaly detection, enabling more accurate identification of abnormal states in the rail vehicle target system.

[0064] Step S13: Determine the dynamic threshold band function for the target performance parameter using the predicted value, and perform anomaly detection based on the dynamic threshold band function and the current vehicle state parameters; the dynamic threshold band function is a function that characterizes the data of the target performance parameter being within a preset normal range.

[0065] In this embodiment, after obtaining the predicted values ​​of the target performance parameters, a dynamic threshold band function needs to be constructed to achieve effective detection of system anomalies. The construction of the dynamic threshold band function is based on the statistical characteristics of the predicted values. First, the standard deviation of the predicted values ​​is determined, which can be calculated through statistical analysis of historical predicted values. The selection of the preset multiple needs to be determined according to the sensitivity requirements of anomaly detection; 30 can be selected as the preset multiple. Then, the standard deviation is multiplied by the preset multiple to obtain the product value, which is used to determine the width of the dynamic threshold band.

[0066] Understandably, adding the product to the standard deviation yields a sum, which represents the allowable fluctuation range based on the predicted value. Adding the predicted value to the sum gives the upper limit, and subtracting the predicted value from the sum gives the lower limit. The upper and lower limits together constitute the boundary of the dynamic threshold band, which dynamically adjusts as the predicted value changes, adapting to the normal operating range of the system under different operating conditions and environments.

[0067] After determining the dynamic threshold band based on the predicted value, upper limit value, and lower limit value, a dynamic threshold band function is constructed. This function takes the current vehicle state parameters of the target rail vehicle as input and outputs the corresponding dynamic threshold band. In other words, during anomaly detection, the current vehicle state parameters of the target rail vehicle are input into the dynamic threshold band function to obtain the corresponding target dynamic threshold band. The current vehicle state parameters include vehicle operating conditions, operating parameters, environmental parameters, and maintenance parameters. For example, during vehicle operation, parameters such as current speed, traction force, ambient temperature, and altitude are collected in real time and input into the dynamic threshold band function.

[0068] Furthermore, the current data of the target performance parameter is compared with the target dynamic threshold band to obtain the corresponding comparison results. If the comparison results show that the current data is within the range represented by the target dynamic threshold band at the same time, the current state of the target performance parameter is determined to be normal. If the comparison results show that the current data is not within the range represented by the target dynamic threshold band at the same time, the current state of the target performance parameter is determined to be abnormal.

[0069] When the current state of a target performance parameter is determined to be abnormal, this abnormal moment is designated as the first moment, and moments within a target time period are designated as the second moment. The target time period is the time period earlier than the first moment. The selection of the target time period needs to be determined based on the parameter characteristics and anomaly detection requirements; the target time period can be 1200 seconds.

[0070] Next, the target performance parameters are recorded between the first and second time points to facilitate data analysis and visualization. Analysis of this recorded data provides insights into the trends in system parameter changes before the anomaly occurs, offering a basis for tracing the cause of the anomaly and diagnosing the fault.

[0071] Various methods can be used when conducting data analysis. For example, analyzing the rate of change of the target performance parameter before the anomaly occurs may indicate a deterioration in system performance if the rate of change increases abnormally. The correlation between the target performance parameter and other relevant parameters can also be analyzed; if the correlation between certain parameters changes before the anomaly occurs, it may indicate a potential fault.

[0072] Data visualization can intuitively show the changes in system parameters before and after an anomaly occurs. For example, it can plot the change curve of a target performance parameter over time and mark the upper and lower limits of the dynamic threshold band to clearly observe whether the parameter exceeds the normal range. It can also plot scatter plots of the target performance parameter and other relevant parameters to analyze the changes in the relationship between the parameters.

[0073] Therefore, anomaly detection using a dynamic threshold band function can effectively overcome the limitations of traditional fixed threshold detection methods, enabling sensitive detection of early anomalies in rail vehicle target systems. Furthermore, recording, analyzing, and visualizing data before and after anomaly occurrence helps to gain a deeper understanding of the causes and development trends of anomalies, providing strong support for system fault prediction and diagnosis.

[0074] As can be seen from the above, this application first determines the target performance parameters of the target system within the target rail vehicle, and then constructs an anomaly detection model based on the target performance parameters and other target parameters that have a predetermined influence relationship with them. Next, predictions are made using this anomaly detection model and the vehicle state parameters of the target rail vehicle, thereby obtaining predicted values ​​corresponding to the target performance parameters. Then, these predicted values ​​are used to determine a dynamic threshold band function for the target performance parameters. Finally, anomaly detection is performed based on the dynamic threshold band function and the current vehicle state parameters. In this way, this application can improve the accuracy of anomaly detection when performing anomaly detection on the subsystems of the rail vehicle, thereby effectively ensuring the safety and stability of rail vehicle operation.

[0075] The following is combined Figure 2 , Figure 3 as well as Figure 4 The schematic diagram shown illustrates the technical solution of the embodiments of this application in detail.

[0076] Specifically, taking the bearing temperature (i.e., the target performance parameter) in the running gear system of a rail vehicle as an example, historical data on bogie bearing temperature, along with corresponding vehicle operating parameters, environmental parameters, and vehicle maintenance status (i.e., target parameters), are collected. Machine learning or deep learning methods, such as support vector machines or convolutional neural networks, are used to train a predictive model between the input and output parameters, thereby constructing a "performance baseline" predictive model (i.e., anomaly detection model) for the key performance parameter of bogie bearing temperature. Besides data-driven "performance baseline" predictive model construction, it is also possible to construct a "performance baseline" predictive model based on a mechanistic model. By establishing a mechanistic model of the heat generation, heat conduction, and heat dissipation processes of bogie bearing temperature, a bearing temperature rise mechanism model can be constructed using 3D thermal analysis modeling software such as Ansys, or system software thermal analysis modeling such as Amesim, thereby constructing a "performance baseline" predictive model for the key performance parameter of bogie bearing temperature. Based on the "performance baseline" prediction model constructed above, the predicted value of the vehicle bearing temperature can be obtained by inputting parameters such as the vehicle's current or historical operating parameters, environmental parameters, and the vehicle's own maintenance status into the model.

[0077] Next, based on the constructed bogie bearing temperature "performance baseline" prediction model, the predicted values ​​of the vehicle system performance parameters in the time series were obtained by inputting vehicle operating parameters, environmental parameters, and the vehicle's own maintenance status. The predicted values ​​of the vehicle system performance parameters in the time series are then added to the predicted values. The standard deviation of 30 times is used as the upper / lower error band of the prediction function curve. Based on the "performance baseline" prediction curve and its upper / lower error band, the "performance baseline" prediction value band (i.e., the dynamic threshold band) is formed. This "performance baseline" prediction value band is used as the real-time dynamic threshold band of the bogie bearing temperature performance parameter, thus forming the system real-time dynamic threshold band function based on the system "performance baseline" prediction model.

[0078] When detecting anomalies in the bearing temperature parameters of a vehicle bogie system, the system first acquires current operating parameters, environmental parameters, and the vehicle's own maintenance status. These parameters are then input into the system's real-time dynamic threshold function to obtain the real-time dynamic threshold range for the current bogie bearing temperature. The system compares the bogie bearing temperature performance parameters with this real-time dynamic threshold range. If the bearing temperature performance parameter value falls within the range, the bearing temperature is considered normal. If it falls outside the range, the bearing temperature performance parameter is considered abnormal. The moment of this abnormality is recorded, thus enabling system anomaly detection.

[0079] In addition, the bandwidth of the real-time dynamic threshold band, that is, the difference between the upper and lower limits of the dynamic threshold, can be appropriately adjusted according to the sensitivity requirements of early anomaly detection and fault diagnosis, thereby realizing early anomaly detection and fault diagnosis with autonomous sensitivity settings for the target.

[0080] After detecting and recording the abnormal temperature of the bogie bearings, the system extracts 3000 data points or 1200 seconds backward from the point of abnormality on the timeline. The extracted data points are saved and output as files with extensions such as .csv and .excel. The saved data includes current vehicle operating parameters, environmental parameters, vehicle maintenance status, and bogie bearing temperature performance parameters. Furthermore, the system performs data correlation analysis and data visualization on the saved bogie bearing temperature performance parameters and vehicle operating parameters, environmental parameters, and vehicle maintenance status, for example, by plotting data such as... Figure 3 The chart shown illustrates the correlation analysis between bearing temperature and operating speed, traction force, wheel speed, motor temperature, motor current, and traction fan speed, along with other data. Figure 4 The diagram showing the correlation between bearing temperature and environmental information such as ambient temperature, altitude, longitude, latitude, and route mileage markers analyzes the changing trends of other parameters when abnormal system parameters occur, thereby finding the correlation and causes of system anomalies from a data perspective.

[0081] Accordingly, see Figure 5 As shown in the figure, this application embodiment provides an anomaly detection device for rail vehicles, including:

[0082] The model building module 11 is used to determine the target performance parameters of the target system in the target rail vehicle, and to build an anomaly detection model using the target performance parameters and target parameters that have a preset influence relationship with the target performance parameters.

[0083] The prediction value acquisition module 12 is used to make predictions based on the anomaly detection model and the vehicle state parameters of the target rail vehicle to obtain the predicted value corresponding to the target performance parameters.

[0084] The anomaly detection module 13 is used to determine a dynamic threshold band function for the target performance parameter using the predicted value, and to perform anomaly detection based on the dynamic threshold band function and the current vehicle state parameters; the dynamic threshold band function is a function that characterizes the data of the target performance parameter being within a preset normal range.

[0085] As can be seen from the above, this application first determines the target performance parameters of the target system within the target rail vehicle, and then constructs an anomaly detection model based on the target performance parameters and other target parameters that have a predetermined influence relationship with them. Next, predictions are made using this anomaly detection model and the vehicle state parameters of the target rail vehicle, thereby obtaining predicted values ​​corresponding to the target performance parameters. Then, these predicted values ​​are used to determine a dynamic threshold band function for the target performance parameters. Finally, anomaly detection is performed based on the dynamic threshold band function and the current vehicle state parameters. In this way, this application can improve the accuracy of anomaly detection when performing anomaly detection on the subsystems of the rail vehicle, thereby effectively ensuring the safety and stability of rail vehicle operation.

[0086] In some specific embodiments, the model building module 11 specifically includes:

[0087] The data determination unit is used to determine the historical data of the target performance parameters;

[0088] The first parameter determination unit is used to determine vehicle operating parameters, environmental parameters, and vehicle maintenance parameters that have a preset influence relationship with the target performance parameters, so as to determine the corresponding target parameters;

[0089] The first model building unit is used to train the historical data and the target parameters using a preset machine learning algorithm to build a corresponding anomaly detection model.

[0090] In some specific embodiments, the model building module 11 specifically includes:

[0091] The second model building unit is used to determine the physical laws governing the target performance parameters in the target rail vehicle, and to build a mechanistic model of the target performance parameters based on the physical laws.

[0092] The third model construction unit is used to construct an anomaly detection model based on the mechanism model and the target parameters that have a preset influence relationship with the target performance parameters.

[0093] In some specific embodiments, the predicted value acquisition module 12 specifically includes:

[0094] The second parameter determination unit is used to determine the vehicle operating conditions, vehicle operating parameters, environmental parameters and vehicle maintenance parameters of the target rail vehicle as vehicle status parameters.

[0095] The prediction value determination unit is used to input the historical or current data of the vehicle state parameters into the anomaly detection model so that the anomaly detection model outputs a prediction value corresponding to the target performance parameter.

[0096] In some specific embodiments, the anomaly detection module 13 specifically includes:

[0097] A product value determination unit is used to determine the standard deviation of the predicted value and the product value of the standard deviation and a preset multiple.

[0098] The product value addition unit is used to add the standard deviation and the product value to obtain the sum value;

[0099] A numerical determination unit is used to determine the sum of the predicted value and the summed value as an upper limit value, and to determine the difference between the predicted value and the summed value as a lower limit value;

[0100] The function determination unit is used to determine a dynamic threshold band based on the predicted value, the upper limit value, and the lower limit value, and to determine a dynamic threshold band function using the dynamic threshold band.

[0101] In some specific embodiments, the anomaly detection module 13 specifically includes:

[0102] The threshold band determination unit is used to input the current vehicle state parameters of the target rail vehicle into the dynamic threshold band function to obtain the corresponding target dynamic threshold band.

[0103] The data comparison unit is used to compare the current data of the target performance parameter with the target dynamic threshold band and obtain the corresponding comparison results;

[0104] The normal determination unit is used to determine that the current state of the target performance parameter is normal if the comparison result shows that the current data is within the range represented by the target dynamic threshold band at the same time.

[0105] An anomaly determination unit is used to determine that the current state of the target performance parameter is abnormal if the comparison results show that the current data is not within the range represented by the target dynamic threshold band at the same time.

[0106] In some specific embodiments, the anomaly determination unit further includes:

[0107] The time determination subunit is used to determine the time when the current state is abnormal as the first time and the time within the target time period as the second time; the target time period is the time period earlier than the first time.

[0108] The data recording subunit is used to record the target performance parameters between the first time point and the second time point, so as to perform data analysis and data visualization on the obtained recorded data.

[0109] Furthermore, embodiments of this application also disclose an electronic device, Figure 6 This is a structural diagram of an electronic device 20 according to an exemplary embodiment. The content of the diagram should not be construed as limiting the scope of this application. The electronic device 20 may specifically include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. The memory 22 stores a computer program, which is loaded and executed by the processor 21 to implement the relevant steps in the anomaly detection method for rail vehicles disclosed in any of the foregoing embodiments. Furthermore, the electronic device 20 in this embodiment may specifically be an electronic computer.

[0110] In this embodiment, the power supply 23 is used to provide operating voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and external devices, and the communication protocol it follows can be any communication protocol applicable to the technical solution of this application, and is not specifically limited here; the input / output interface 25 is used to acquire external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs, and is not specifically limited here.

[0111] In addition, the memory 22, as a carrier for resource storage, can be a read-only memory, random access memory, disk or optical disk, etc. The resources stored thereon can include operating system 221, computer program 222, etc., and the storage method can be temporary storage or permanent storage.

[0112] The operating system 221 is used to manage and control the various hardware devices on the electronic device 20 and the computer program 222, which may be Windows Server, Netware, Unix, Linux, etc. In addition to including a computer program capable of performing the anomaly detection method for a rail vehicle executed by the electronic device 20 as disclosed in any of the foregoing embodiments, the computer program 222 may further include computer programs capable of performing other specific tasks.

[0113] Furthermore, this application also discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the aforementioned anomaly detection method for rail vehicles. Specific steps of this method can be found in the corresponding content disclosed in the foregoing embodiments, and will not be repeated here.

[0114] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.

[0115] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0116] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0117] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0118] The technical solutions provided in this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for detecting anomalies in rail vehicles, characterized in that, include: In the target rail vehicle, the target performance parameters of the target system are determined, and an anomaly detection model is constructed using the target performance parameters and target parameters that have a preset influence relationship with the target performance parameters. Based on the anomaly detection model and the vehicle state parameters of the target rail vehicle, a prediction is made to obtain a predicted value corresponding to the target performance parameters. The predicted value is used to determine a dynamic threshold band function for the target performance parameter, and anomaly detection is performed based on the dynamic threshold band function and the current vehicle state parameters. The dynamic threshold band function is a function that characterizes whether the data of the target performance parameter is within a preset normal range.

2. The anomaly detection method for rail vehicles according to claim 1, characterized in that, The step of constructing an anomaly detection model using the target performance parameters and target parameters that have a preset influence relationship with the target performance parameters includes: Historical data for determining the target performance parameters; Determine the vehicle operating parameters, environmental parameters, and vehicle maintenance parameters that have a preset influence relationship with the target performance parameters, so as to determine the corresponding target parameters; The historical data and the target parameters are trained using a preset machine learning algorithm to construct a corresponding anomaly detection model.

3. The anomaly detection method for rail vehicles according to claim 1, characterized in that, The step of constructing an anomaly detection model using the target performance parameters and target parameters that have a preset influence relationship with the target performance parameters includes: The physical laws governing the target performance parameters are determined in the target rail vehicle, and a mechanistic model of the target performance parameters is constructed based on the physical laws. An anomaly detection model is constructed based on the aforementioned mechanism model and the target parameters that have a preset influence relationship with the target performance parameters.

4. The anomaly detection method for rail vehicles according to claim 1, characterized in that, The step of predicting based on the anomaly detection model and the vehicle state parameters of the target rail vehicle to obtain predicted values ​​corresponding to the target performance parameters includes: The vehicle operating conditions, vehicle operating parameters, environmental parameters, and vehicle maintenance parameters of the target rail vehicle are determined as vehicle status parameters. The historical or current data of the vehicle state parameters are input into the anomaly detection model so that the anomaly detection model outputs a predicted value corresponding to the target performance parameter.

5. The anomaly detection method for rail vehicles according to claim 1, characterized in that, The step of determining the dynamic threshold band function with respect to the target performance parameter using the predicted value includes: Determine the standard deviation of the predicted value, and determine the product of the standard deviation and a preset multiple; Add the standard deviation and the product value together to obtain the sum; The sum of the predicted value and the summed value is determined as the upper limit, and the difference between the predicted value and the summed value is determined as the lower limit. A dynamic threshold band is determined based on the predicted value, the upper limit value, and the lower limit value, and a dynamic threshold band function is determined using the dynamic threshold band.

6. The anomaly detection method for rail vehicles according to any one of claims 1 to 5, characterized in that, The anomaly detection based on the dynamic threshold band function and the current vehicle state parameters includes: The current vehicle state parameters of the target rail vehicle are input into the dynamic threshold band function to obtain the corresponding target dynamic threshold band; The current data of the target performance parameter is compared with the target dynamic threshold band, and the corresponding comparison results are obtained. If the comparison results show that the current data is within the range represented by the target dynamic threshold band at the same time, then the current state of the target performance parameter is determined to be normal. If the comparison results show that the current data is not within the range represented by the target dynamic threshold band at the same time, then the current state of the target performance parameter is determined to be abnormal.

7. The anomaly detection method for rail vehicles according to claim 6, characterized in that, After determining that the current state of the target performance parameter is abnormal, the method further includes: The moment when the current state is abnormal is defined as the first moment, and the moment within the target time period is defined as the second moment; the target time period is the time period earlier than the first moment. Record the target performance parameter data between the first time point and the second time point so that the obtained recorded data can be used for data analysis and data visualization.

8. An anomaly detection device for rail vehicles, characterized in that, include: The model building module is used to determine the target performance parameters of the target system in the target rail vehicle, and to build an anomaly detection model using the target performance parameters and target parameters that have a preset influence relationship with the target performance parameters. The prediction value acquisition module is used to make predictions based on the anomaly detection model and the vehicle state parameters of the target rail vehicle to obtain the predicted values ​​corresponding to the target performance parameters. An anomaly detection module is used to determine a dynamic threshold band function for the target performance parameter using the predicted value, and to perform anomaly detection based on the dynamic threshold band function and the current vehicle state parameters; The dynamic threshold band function is a function that characterizes whether the data of the target performance parameter is within a preset normal range.

9. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor for executing the computer program to implement the anomaly detection method for rail vehicles as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, Used to store a computer program; wherein, when the computer program is executed by a processor, it implements the anomaly detection method for rail vehicles as described in any one of claims 1 to 7.

Images