Method and device for determining a rail vehicle maintenance strategy
By acquiring operational data and image information from rail vehicles, and using anomaly detection models and knowledge graphs for fusion analysis, automated maintenance strategies are generated. This addresses the shortcomings of manual inspection in existing technologies, enabling more accurate and timely maintenance and ensuring the long-term stability of the equipment.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CRRC NANJING PUZHEN CO LTD
- Filing Date
- 2026-05-08
- Publication Date
- 2026-06-12
AI Technical Summary
Existing rail vehicle maintenance methods rely on periodic manual inspections, which can lead to omissions, misjudgments, and difficulties in ensuring accuracy and timeliness. Furthermore, they may ignore the actual usage conditions of the equipment, potentially resulting in over-maintenance or failure to detect potential problems in a timely manner.
By acquiring operational data from rail vehicles, utilizing anomaly judgment rules and anomaly detection models, and combining them with a rail vehicle anomaly knowledge graph for fusion analysis, maintenance strategies are generated. This includes data processing, image recognition, and knowledge graph fusion, enabling automated maintenance.
It improves the timeliness and accuracy of maintenance, avoids omissions and misjudgments, ensures the long-term stability of equipment, and avoids over-maintenance or failure to detect problems in a timely manner.
Smart Images
Figure CN122198945A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of urban rail transit vehicle maintenance technology, and in particular to a method and apparatus for determining rail vehicle maintenance strategies. Background Technology
[0002] Ensuring the safety and reliability of rail vehicles is crucial in their daily operation. To guarantee their performance during operation, regular maintenance work, such as daily and monthly inspections, is typically required.
[0003] Existing rail vehicle maintenance methods mainly rely on periodic manual inspections. Manual inspections are often limited by the experience and physical condition of the inspectors, which may lead to omissions and misjudgments during the inspection process. The accuracy and timeliness are often difficult to guarantee, which in turn affects the operational safety of the equipment. Secondly, although periodic inspections ensure the stability of the equipment to a certain extent, they ignore the actual usage condition of the equipment, which may lead to over-maintenance or failure to detect potential equipment problems in a timely manner. Summary of the Invention
[0004] This invention provides a method for determining a maintenance strategy for rail vehicles, which improves the timeliness and accuracy of rail vehicle maintenance, avoids omissions, misjudgments, over-maintenance, or failure to detect potential equipment problems in a timely manner, and ensures the long-term stability of the equipment. The method includes: Obtain vehicle operation data for rail vehicles; Based on preset anomaly judgment rules, the vehicle operation data is processed to obtain the first anomaly information; the first anomaly information includes the abnormal component information and data anomaly type information corresponding to the abnormal operation data. Obtain the image information of the component corresponding to the abnormal component information; Image information is input into a pre-trained anomaly detection model to obtain second anomaly information; the second anomaly information includes component anomaly type information; wherein, the anomaly detection model is obtained by training a machine learning model using historical images of each component of the rail vehicle and historical vehicle anomaly information; Based on a pre-set knowledge graph of rail vehicle anomalies, the first and second anomaly information are fused and analyzed to obtain a comprehensive anomaly analysis result. Based on the comprehensive anomaly analysis results, the final rail vehicle maintenance strategy is generated.
[0005] Optionally, based on preset anomaly detection rules, the vehicle operation data is processed to obtain first anomaly information, including: Based on the preset anomaly judgment rules, the vehicle operation data is processed using a forward reasoning algorithm to obtain the first anomaly information.
[0006] Optionally, based on a pre-defined knowledge graph of rail vehicle anomalies, the first and second anomaly information are fused and analyzed to obtain a comprehensive anomaly analysis result, including: Based on a pre-set knowledge graph of rail vehicle anomalies, the correlation between the first and second anomaly information is analyzed to determine the related information, which includes the related state or conflict state. Based on the first anomaly information, the second anomaly information, and the related information, the comprehensive anomaly analysis results are determined.
[0007] Optionally, based on the comprehensive anomaly analysis results, a final rail vehicle maintenance strategy is generated, including: Based on the pre-defined correlation between abnormal information and maintenance strategies, the comprehensive abnormal analysis results are analyzed to determine the final rail vehicle maintenance strategy.
[0008] Optionally, the anomaly detection model is a convolutional neural network model.
[0009] Optionally, it also includes: Acquire historical data of rail vehicles; each piece of historical data includes: historical vehicle operation data, historical anomaly data, and historical maintenance strategy data; Based on the historical data, multiple anomaly judgment rules for rail vehicles, a rail vehicle anomaly knowledge graph, and the correlation between anomaly information and maintenance strategies are determined.
[0010] This invention also provides a device for determining a rail vehicle maintenance strategy, used to improve the timeliness and accuracy of rail vehicle maintenance, avoid omissions, misjudgments, over-maintenance, or failure to detect potential equipment problems in a timely manner, and ensure the long-term stability of the equipment. The device includes: The data acquisition module is used to acquire vehicle operation data of rail vehicles; The rule matching module is used to process vehicle operation data based on preset anomaly judgment rules to obtain first anomaly information; the first anomaly information includes abnormal component information and data anomaly type information corresponding to the abnormal operation data. The image acquisition module is used to acquire image information of the component corresponding to the abnormal component information; The image processing module is used to input image information into a pre-trained anomaly detection model to obtain second anomaly information; the second anomaly information includes component anomaly type information; wherein, the anomaly detection model is trained on a machine learning model using historical images of each component of the rail vehicle and historical vehicle anomaly information; The fusion analysis module is used to perform fusion analysis on the first and second anomaly information based on the preset rail vehicle anomaly knowledge graph to obtain a comprehensive anomaly analysis result. The strategy determination module is used to generate the final rail vehicle maintenance strategy based on the comprehensive anomaly analysis results.
[0011] This invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the above-described method for determining the rail vehicle maintenance strategy.
[0012] This invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for determining rail vehicle maintenance strategies.
[0013] This invention also provides a computer program product, which includes a computer program that, when executed by a processor, implements the above-described method for determining the maintenance strategy of rail vehicles.
[0014] In this embodiment of the invention, vehicle operation data of a rail vehicle is acquired; based on preset anomaly judgment rules, the vehicle operation data is processed to obtain first anomaly information; the first anomaly information includes abnormal component information and data anomaly type information corresponding to the abnormal operation data; image information of the component corresponding to the abnormal component information is acquired; the image information is input into a pre-trained anomaly detection model to obtain second anomaly information; the second anomaly information includes component anomaly type information; wherein, the anomaly detection model is trained using historical images of each component of the rail vehicle and historical vehicle anomaly information; based on a preset rail vehicle anomaly knowledge graph, the first and second anomaly information are fused and analyzed to obtain a comprehensive anomaly analysis result; based on the comprehensive anomaly analysis result, a final rail vehicle maintenance strategy is generated. Therefore, this embodiment of the invention utilizes the combination of expert rule judgment and visual recognition technology to comprehensively analyze the working status of each device in the rail vehicle from both real-time data and image levels, avoiding interference from human factors on the maintenance results, improving the timeliness and efficiency of maintenance, enhancing the accuracy of maintenance strategies, avoiding over-maintenance, or failure to detect potential equipment problems in a timely manner, and ensuring the long-term stability of the equipment. Attached Figure Description
[0015] 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 some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0016] In the attached diagram: Figure 1 This is a flowchart illustrating a method for determining a rail vehicle maintenance strategy according to an embodiment of the present invention. Figure 2 The flowchart of the method for fusing and analyzing first and second anomaly information based on a preset knowledge graph of rail vehicle anomalies in this embodiment of the invention to obtain a comprehensive anomaly analysis result is provided. Figure 3 This is a flowchart illustrating the implementation of the method for determining the maintenance strategy for rail vehicles provided in this embodiment of the invention. Figure 4 This is a schematic diagram of a device for determining a rail vehicle maintenance strategy provided in an embodiment of the present invention; Figure 5 This is a schematic diagram of a computer device provided in an embodiment of the present invention. Detailed Implementation
[0017] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings. Here, the illustrative embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.
[0018] In the description of this specification, the terms "comprising," "including," "having," and "containing" are open-ended terms, meaning that they include but are not limited to. The terms "an embodiment," "a specific embodiment," "some embodiments," and "for example," etc., refer to specific features, structures, or characteristics described in connection with that embodiment or example that are included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, or characteristics described can be combined in any suitable manner in one or more embodiments or examples. The order of steps involved in the various embodiments is used to illustrate the implementation of this application, and the order of steps is not limited and can be adjusted appropriately as needed.
[0019] Research has found that existing methods of regular manual inspection for rail vehicle maintenance may lead to omissions and misjudgments during the inspection process, making it difficult to guarantee accuracy and timeliness. Furthermore, these methods may overlook the actual usage conditions of the equipment, resulting in over-maintenance or failure to detect potential equipment problems in a timely manner.
[0020] In response to the above research, embodiments of the present invention provide a scheme for determining the maintenance strategy of rail vehicles and an automated maintenance method to improve the timeliness and accuracy of rail vehicle maintenance, avoid omissions, misjudgments, over-maintenance, or failure to detect potential equipment problems in a timely manner, and ensure the long-term stability of the equipment.
[0021] like Figure 1 The diagram shows a flowchart of a method for determining a rail vehicle maintenance strategy according to an embodiment of the present invention. The method may include: Step 101: Obtain the vehicle operation data of the rail vehicle; Step 102: Based on preset anomaly judgment rules, process the vehicle operation data to obtain first anomaly information; the first anomaly information includes abnormal component information and data anomaly type information corresponding to the abnormal operation data. Step 103: Obtain the image information of the component corresponding to the abnormal component information; Step 104: Input the image information into the pre-trained anomaly detection model to obtain the second anomaly information; the second anomaly information includes component anomaly type information; wherein, the anomaly detection model is obtained by training a machine learning model using historical images of each component of the rail vehicle and historical vehicle anomaly information; Step 105: Based on the preset knowledge graph of rail vehicle anomalies, perform fusion analysis on the first anomaly information and the second anomaly information to obtain the comprehensive anomaly analysis result; Step 106: Based on the comprehensive anomaly analysis results, generate the final rail vehicle maintenance strategy.
[0022] In this embodiment of the invention, combining vehicle operation data and component image information of the rail vehicle, the vehicle operation data is first processed based on preset anomaly judgment rules to determine the first anomaly information. Then, the image information of the component corresponding to the anomaly component information is processed through an anomaly detection model to determine the second anomaly information. Thus, based on a preset rail vehicle anomaly knowledge graph, the first and second anomaly information are fused and analyzed to determine the final rail vehicle maintenance strategy. Compared with existing manual inspection maintenance methods, this embodiment of the invention can improve the maintenance efficiency of rail vehicles and conduct comprehensive anomaly analysis of component anomalies from both vehicle operation data and component image levels. This improves the accuracy and timeliness of fault diagnosis and maintenance strategies, avoids omissions and misjudgments, and prevents over-maintenance or failure to detect potential equipment problems in a timely manner, ensuring the long-term stability of the equipment.
[0023] Before implementing the method for determining the maintenance strategy of rail vehicles in this embodiment of the invention, it is necessary to set up anomaly judgment rules, a rail vehicle anomaly knowledge graph, and the relationship between components, anomaly types, and maintenance strategies. This can be achieved in the following way: Acquire historical data of rail vehicles; each piece of historical data includes: historical vehicle operation data, historical anomaly data, and historical maintenance strategy data; Based on historical data, we determined multiple anomaly judgment rules for rail vehicles, a rail vehicle anomaly knowledge graph, and the correlation between components, anomaly types, and maintenance strategies.
[0024] In practice, multiple historical data points of rail vehicles can be collected. Each historical data point can include: historical vehicle operation data, historical anomaly data, and historical maintenance strategy data. The historical maintenance strategy data can include maintenance method data and / or component replacement data, etc.
[0025] Historical data can be represented as follows: Where D is the historical data set, For the first historical data, This represents the total number of historical data entries.
[0026] Then, historical data can be preprocessed to remove noise and outliers, ensuring data accuracy and integrity. This preprocessing can employ smoothing algorithms, which can be expressed as follows: in, For the cleaned data, This refers to the window size.
[0027] For the preprocessed historical data, multiple anomaly detection rules are established based on expert experience. These anomaly detection rules can be expressed as follows: Where R is the set of anomaly detection rules. For the first Anomaly detection rules This represents the total number of exception handling rules. Each exception handling rule can include exception handling conditions.
[0028] For example, the anomaly judgment rule can be: if the traction motor bearing temperature is >90 degrees Celsius, then the "bearing temperature too high" anomaly is triggered.
[0029] Based on the preprocessed historical data and combined with expert experience, a knowledge graph of anomalies in rail vehicles is constructed.
[0030] The knowledge graph of anomalies in rail vehicles can include the relationships between various types of anomalies in rail vehicles. For example, excessively high bearing temperature (obtained through analysis of vehicle operation data) is often accompanied by excessive vibration of the bearing's characteristic frequency, while fatigue wear (obtained through image analysis) manifests as excessive vibration of the bearing's characteristic frequency.
[0031] Based on the preprocessed historical data and combined with expert experience, a correlation between anomaly information and maintenance strategies is constructed.
[0032] For example, the correlation between anomaly information and maintenance strategies can include the correlation between components, anomaly types, and maintenance strategies. If the component is a bearing and the anomaly type is bearing overheating / damage, then the maintenance strategy could be immediate repair. Another example is the correlation between anomaly information and maintenance strategies. If the first anomaly is abnormal, the second anomaly is normal, and the correlation information is a conflicting state (possibly sensor drift or internal fault (image not visible)), then the maintenance strategy could be to wait and observe, increase the image detection frequency of the component, or trigger more precise detection (such as ultrasonic testing).
[0033] The following is about Figure 1 The method for determining the maintenance strategy for rail vehicles is described in detail.
[0034] In step 101 above, the vehicle operation data of the rail vehicle can be obtained.
[0035] In practice, vehicle operation data can come from onboard sensors, such as motor temperature, or from vehicle environment and passenger service systems, such as air conditioning system data, ventilation system data, etc.; or from operating performance data, energy consumption data, etc.
[0036] Vehicle operation data can be represented as: in, For vehicle operation data set, For the first Vehicle operation data.
[0037] In step 102 above, the vehicle operation data can be processed based on preset anomaly judgment rules to obtain the first anomaly information.
[0038] In practice, the vehicle operation data can be inferred based on the anomaly judgment rules to obtain the abnormal component information and data anomaly type information corresponding to the abnormal operation data.
[0039] In one embodiment, step 102 may specifically include: processing vehicle operation data using a forward reasoning algorithm based on preset anomaly judgment rules to obtain first anomaly information.
[0040] In practice, the reasoning process can be implemented using a forward reasoning algorithm. The specific process of the forward reasoning algorithm includes: taking vehicle operation data as initial facts, traversing all anomaly judgment rules, finding all anomaly judgment rules that can be satisfied by the data in the initial facts, and then obtaining the abnormal component information and data anomaly type information corresponding to the rules satisfied by the data in the initial facts.
[0041] In this way, processing vehicle operation data through forward reasoning algorithms can improve processing efficiency and accuracy, while avoiding the opacity of deep learning black box models and providing a clear, reliable, and auditable diagnostic logic chain.
[0042] In practice, the first anomaly information may also include the first severity information and the first confidence level information.
[0043] The severity level can be determined based on the deviation of abnormal operating data from normal operating data, or based on the type of data anomaly.
[0044] The first confidence level can be calculated using the Bayesian algorithm. The Bayesian algorithm can be expressed by the following formula: in, In the data The following malfunction occurred The probability of (data anomaly type F), For a given fault Data The probability, Let be the prior probability of the fault. This represents the total probability of the data.
[0045] The first severity information and the first confidence information are then fused and analyzed in subsequent steps to improve the accuracy of anomaly detection.
[0046] In step 103 above, image information of the component corresponding to the abnormal component information can be obtained.
[0047] In practice, an image acquisition system installed on the rail vehicle can be used to collect image information of the components corresponding to the abnormal parts. The image acquisition system can include an adjustable-angle camera and supplementary lighting to ensure clear images are captured under different lighting conditions.
[0048] Then, the image information can be preprocessed, including image enhancement, denoising, and grayscale conversion. Image preprocessing can be achieved using the following formula: in, The image after preprocessing. For the original image, The weighting function is Gaussian. For the size of the filter window, The selection can be based on image quality and noise level, typically 100%. or size.
[0049] in, denoted as the standard deviation of the Gaussian distribution.
[0050] In step 104 above, the image information is input into the pre-trained anomaly detection model to obtain the second anomaly information.
[0051] In practice, anomaly detection models can be used to identify abnormal information in image information, and the second abnormal information includes component anomaly type information.
[0052] In practice, the second anomaly information may also include the second severity information and the second confidence information output by the anomaly detection model.
[0053] The anomaly detection model is trained by using historical images of various components of the rail vehicle and historical vehicle anomaly information.
[0054] In one embodiment, the anomaly detection model is a convolutional neural network model.
[0055] In practical implementation, the formula for the convolutional neural network model can be: in, The state information output by the model. For the model's weight matrix, For the input image information, For bias, For the activation function, in this embodiment of the invention, the ReLU activation function can be selected: During model training, gradient descent can be performed using the Adam optimizer. The learning rate is set to 0.001, and the batch size is set to 32. The sample set includes historical images and historical state information of various components from different rail vehicles. The sample set can contain up to 10,000 images, including labeled images of normal states and abnormal states such as cracks, wear, and deformation. The labels include component type, abnormality type, and severity. The sample set is divided into training, validation, and test sets according to a preset ratio, with the training set accounting for 70%, the validation set for 15%, and the test set for 15%. The convolutional neural network model is trained using the training set, validated using the validation set, and tested using the test set.
[0056] In this way, by using the trained anomaly detection model to identify the abnormal state of the image information of the component to which each vehicle's abnormal information belongs, the accuracy of abnormal state identification can be improved.
[0057] In step 105 above, based on the preset knowledge graph of rail vehicle anomalies, the first anomaly information and the second anomaly information are fused and analyzed to obtain the comprehensive anomaly analysis result.
[0058] In practice, the knowledge graph of rail vehicle anomalies includes the relationships between various types of anomalies in rail vehicles. Based on the knowledge graph of rail vehicle anomalies, the first and second anomaly information can be fused and analyzed to obtain a comprehensive anomaly analysis result.
[0059] In one embodiment, step 105 above, as Figure 2 As shown, it can specifically include: Step 201: Based on the preset knowledge graph of rail vehicle anomalies, analyze the correlation between the first anomaly information and the second anomaly information to determine the related information, which includes the related state or conflict state. Step 202: Determine the comprehensive anomaly analysis result based on the first anomaly information, the second anomaly information, and the related information.
[0060] In practice, the association state can be considered as the data anomaly type and the component anomaly type, which are determined to be two different observed data anomaly types describing the same anomaly pattern. For example, the data anomaly type of the first anomaly information is "bearing temperature too high", and the component anomaly type of the second anomaly information is "bearing fatigue wear". Based on the knowledge graph, it can be known that bearing temperature too high is often accompanied by bearing characteristic frequency vibration exceeding the standard, while fatigue wear is manifested as bearing characteristic frequency vibration exceeding the standard. Therefore, the association relationship between the first anomaly information and the second anomaly information is an association state.
[0061] A conflicting state can be considered as a conclusion that is completely opposite to the conclusion that corresponds to the data anomaly type and the component anomaly type. For example, the data anomaly type is "vibration sensor shows that the bearing characteristic frequency amplitude is seriously out of standard", but the component anomaly type is "bearing raceway is smooth and there are no visible defects".
[0062] In practice, the first abnormal information, the second abnormal information, and the related information can be integrated into a comprehensive abnormality analysis result. For example, in the comprehensive abnormality analysis result, if the abnormal component is a bearing, the abnormality type is "bearing temperature is too high (data abnormality type) and bearing fatigue wear (component abnormality type)", and the two are related. Alternatively, if the abnormal component is a bearing, the abnormality types are "bearing characteristic frequency amplitude is seriously out of standard" (data abnormality type) and "bearing raceway is smooth and there are no visible defects" (component abnormality type), and the two are conflicting.
[0063] It should be noted that the confidence and severity of the anomaly type in the comprehensive anomaly analysis results can also be determined based on the first confidence level and first severity of the first anomaly information and the second confidence level and second severity of the second anomaly information, combined with the association status.
[0064] In step 106 above, the final rail vehicle maintenance strategy can be generated based on the comprehensive anomaly analysis results.
[0065] In practice, maintenance strategies for abnormal components are determined based on the abnormal components, abnormal types, and associated statuses identified in the comprehensive anomaly analysis results.
[0066] In one embodiment, step 106 may specifically include: analyzing the comprehensive anomaly analysis results based on the preset correlation between anomaly information and maintenance strategy, and determining the final rail vehicle maintenance strategy.
[0067] In practice, based on the associated status, first anomaly information and second anomaly information in the comprehensive anomaly analysis results, the maintenance strategy corresponding to the comprehensive anomaly analysis results is determined from the preset association relationship between anomaly information and maintenance strategy as the final rail vehicle maintenance strategy.
[0068] For example, in the comprehensive anomaly analysis results, the abnormal component is the bearing, and the anomaly type is "bearing temperature too high (data anomaly type) and bearing fatigue wear (component anomaly type)". The two are related, and the maintenance strategy can be determined to be immediate repair based on the relationship between the component, the anomaly type and the maintenance strategy.
[0069] Alternatively, if the abnormal component is a bearing, and the abnormality types are "bearing characteristic frequency amplitude severely exceeds the standard" (data abnormality type) and "bearing raceway is smooth, with no visible defects" (component abnormality type), these two are in conflict. Based on the correlation between the associated information and the maintenance strategy, the maintenance strategy corresponding to the conflicting state can be determined as to observe, increase the image detection frequency of the component, trigger more precise detection, or conduct manual inspection, etc. Alternatively, based on the correlation between the component, the abnormality type, and the maintenance strategy, the maintenance strategy corresponding to the abnormal component being a bearing with the abnormality type "bearing characteristic frequency amplitude severely exceeds the standard" can be taken as the final maintenance strategy.
[0070] Alternatively, if the abnormal component is a gear, and the abnormality type is "abnormal current fluctuation (data abnormality type) and slight wear on the gear surface (component abnormality type)", the two are related. It is believed that the wear may lead to more serious failures in the future. Based on the correlation between the component, the abnormality type and the maintenance strategy, the maintenance strategy can be determined to be planned replacement.
[0071] The method for determining a rail vehicle maintenance strategy provided in this invention involves: acquiring rail vehicle operation data; processing the operation data based on preset anomaly judgment rules to obtain first anomaly information; the first anomaly information includes information on abnormal components corresponding to the abnormal operation data and information on the data anomaly type; acquiring image information of the components corresponding to the abnormal component information; inputting the image information into a pre-trained anomaly detection model to obtain second anomaly information; the second anomaly information includes component anomaly type information; wherein, the anomaly detection model is trained using historical images of each component of the rail vehicle and historical vehicle anomaly information; based on a preset rail vehicle anomaly knowledge graph, the first and second anomaly information are fused and analyzed to obtain a comprehensive anomaly analysis result; and based on the comprehensive anomaly analysis result, a final rail vehicle maintenance strategy is generated. Therefore, this invention utilizes a combination of expert rule configuration and visual recognition technology to comprehensively analyze the working status of each device in the rail vehicle from both data and image perspectives, avoiding interference from human factors on the maintenance results, enhancing the accuracy of the maintenance strategy, and improving the timeliness and efficiency of maintenance; simultaneously, through historical data analysis and the ability to predict potential faults in advance, it significantly improves fault early warning capabilities and effectively avoids losses caused by sudden faults.
[0072] The following example illustrates the method for determining the above-mentioned rail vehicle maintenance strategy.
[0073] Figure 3 A flowchart illustrating the implementation of a method for determining a rail vehicle maintenance strategy as provided in an embodiment of the invention. (See attached flowchart.) Figure 3As shown, at the start of the daily inspection, the terminal connects to the onboard host to control the execution of daily inspection items. Specifically, execution instructions can be issued based on the type of daily inspection item. Then, anomaly detection is performed through a combination of a rule engine and an anomaly detection model. The rule engine includes pre-configured multiple anomaly judgment rules for the rail vehicle, a rail vehicle anomaly knowledge graph, and the correlation between anomaly information and maintenance strategies.
[0074] The anomaly detection process of rule engines and anomaly detection models can specifically include: The rule engine processes the acquired vehicle operation data (i.e., vehicle operation data corresponding to the item type) based on preset anomaly judgment rules to obtain the first anomaly information; the anomaly detection model processes the image information of the component corresponding to the acquired anomaly component information to obtain the second anomaly information; the rule engine performs fusion analysis on the first and second anomaly information based on the preset rail vehicle anomaly knowledge graph to obtain the comprehensive anomaly analysis result, and generates the final rail vehicle maintenance strategy based on the comprehensive anomaly analysis result.
[0075] Finally, based on the comprehensive anomaly analysis results and the corresponding final rail vehicle maintenance strategy, a rail vehicle maintenance task list can be generated and sent to maintenance personnel to guide them in subsequent maintenance operations. The task information in the rail vehicle maintenance task list can also be input into a time-series database to ensure the integrity of maintenance task records and subsequent queries. The time-series database uses timestamps as the primary key and records the execution time, task type, maintenance components, required tools, and personnel information for each maintenance task. The table structure design of the time-series database is as follows: Task table: contains fields for task ID, task description, assignee, task status, and timestamp.
[0076] Task Log Table: Includes fields for Task ID, Maintenance Start Time, Maintenance End Time, and Maintenance Result. The data query interface provides query functions based on time range, Task ID, and Task Status, supporting quick retrieval of historical maintenance data for easy subsequent analysis and report generation.
[0077] This invention also provides a device for determining a rail vehicle maintenance strategy, as described in the following embodiments. Since the principle by which this device solves the problem is similar to the method for determining a rail vehicle maintenance strategy described above, the implementation of this method can refer to the implementation of the method for determining a rail vehicle maintenance strategy, and repeated details will not be elaborated further.
[0078] like Figure 4 The diagram shown is a schematic of a device for determining a rail vehicle maintenance strategy according to an embodiment of the present invention. The device may include: The data acquisition module 401 is used to acquire the vehicle operation data of the rail vehicle; The rule matching module 402 is used to process vehicle operation data based on preset anomaly judgment rules to obtain first anomaly information; the first anomaly information includes abnormal component information and data anomaly type information corresponding to the abnormal operation data. Image acquisition module 403 is used to acquire image information of the component corresponding to the abnormal component information; Image processing module 404 is used to input image information into a pre-trained anomaly detection model to obtain second anomaly information; the second anomaly information includes component anomaly type information; wherein, the anomaly detection model is obtained by training a machine learning model using historical images of each component of the rail vehicle and historical vehicle anomaly information; The fusion analysis module 405 is used to perform fusion analysis on the first anomaly information and the second anomaly information based on the preset rail vehicle anomaly knowledge graph to obtain a comprehensive anomaly analysis result. The strategy determination module 406 is used to generate the final rail vehicle maintenance strategy based on the comprehensive anomaly analysis results.
[0079] In one embodiment, the rule matching module 402 can be used to: process vehicle operation data using a forward reasoning algorithm based on preset anomaly judgment rules to obtain first anomaly information.
[0080] In one embodiment, the fusion analysis module 405 can be specifically used to: analyze the correlation between the first abnormal information and the second abnormal information based on a preset knowledge graph of rail vehicle anomalies, determine the associated information, which includes the associated state or conflict state; and determine the comprehensive anomaly analysis result based on the first abnormal information, the second abnormal information and the associated information.
[0081] In one embodiment, the strategy determination module 406 can be used to: analyze the comprehensive anomaly analysis results based on the preset correlation between anomaly information and maintenance strategy, and determine the final rail vehicle maintenance strategy.
[0082] In one embodiment, the anomaly detection model is a convolutional neural network model.
[0083] In one embodiment, a configuration module may be further included for acquiring historical data of the rail vehicle; each piece of historical data includes: historical vehicle operation data, historical anomaly data, and historical maintenance strategy data; based on the historical data, multiple anomaly judgment rules for the rail vehicle, a rail vehicle anomaly knowledge graph, and the correlation between anomaly information and maintenance strategies are determined.
[0084] This invention also provides a computer device. Figure 5This is a schematic diagram of a computer device in an embodiment of the present invention. The computer device 500 includes a memory 510, a processor 520, and a computer program 530 stored in the memory 510 and executable on the processor 520. When the processor 520 executes the computer program 530, it implements the above-mentioned method for determining the maintenance strategy of rail vehicles.
[0085] This invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for determining rail vehicle maintenance strategies.
[0086] This invention also provides a computer program product, which includes a computer program that, when executed by a processor, implements the above-mentioned method for determining the maintenance strategy of rail vehicles.
[0087] In this embodiment of the invention, vehicle operation data of a rail vehicle is acquired; based on preset anomaly judgment rules, the vehicle operation data is processed to obtain first anomaly information; the first anomaly information includes abnormal component information and data anomaly type information corresponding to the abnormal operation data; image information of the component corresponding to the abnormal component information is acquired; the image information is input into a pre-trained anomaly detection model to obtain second anomaly information; the second anomaly information includes component anomaly type information; wherein, the anomaly detection model is trained using historical images of each component of the rail vehicle and historical vehicle anomaly information; based on a preset rail vehicle anomaly knowledge graph, the first and second anomaly information are fused and analyzed to obtain a comprehensive anomaly analysis result; based on the comprehensive anomaly analysis result, a final rail vehicle maintenance strategy is generated. Therefore, this embodiment of the invention utilizes the combination of expert rule judgment and visual recognition technology to comprehensively analyze the working status of each device in the rail vehicle from both real-time data and image levels, avoiding interference from human factors on the maintenance results, improving the timeliness and efficiency of maintenance, enhancing the accuracy of maintenance strategies, avoiding over-maintenance, or failure to detect potential equipment problems in a timely manner, and ensuring the long-term stability of the equipment.
[0088] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0089] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.
[0090] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.
[0091] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.
[0092] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.
Claims
1. A method for determining a maintenance strategy for rail vehicles, characterized in that, include: Obtain vehicle operation data for rail vehicles; Based on preset anomaly detection rules, the vehicle operation data is processed to obtain the first anomaly information; The first abnormal information includes information about the abnormal component corresponding to the abnormal operating data and information about the data abnormality type; Obtain the image information of the component corresponding to the abnormal component information; Image information is input into a pre-trained anomaly detection model to obtain second anomaly information; The second anomaly information includes component anomaly type information; wherein, the anomaly detection model is obtained by training a machine learning model using historical images of each component of the rail vehicle and historical vehicle anomaly information; Based on a pre-set knowledge graph of rail vehicle anomalies, the first and second anomaly information are fused and analyzed to obtain a comprehensive anomaly analysis result. Based on the comprehensive anomaly analysis results, the final rail vehicle maintenance strategy is generated.
2. The method as described in claim 1, characterized in that, Based on preset anomaly detection rules, vehicle operation data is processed to obtain the first anomaly information, including: Based on the preset anomaly judgment rules, the vehicle operation data is processed using a forward reasoning algorithm to obtain the first anomaly information.
3. The method as described in claim 1, characterized in that, Based on a pre-defined knowledge graph of rail vehicle anomalies, the first and second anomaly information are fused and analyzed to obtain comprehensive anomaly analysis results, including: Based on a pre-set knowledge graph of rail vehicle anomalies, the correlation between the first and second anomaly information is analyzed to determine the related information, which includes the related state or conflict state. Based on the first anomaly information, the second anomaly information, and the related information, the comprehensive anomaly analysis results are determined.
4. The method as described in claim 1, characterized in that, Based on the comprehensive anomaly analysis results, the final rail vehicle maintenance strategy is generated, including: Based on the pre-defined correlation between abnormal information and maintenance strategies, the comprehensive abnormal analysis results are analyzed to determine the final rail vehicle maintenance strategy.
5. The method as described in claim 1, characterized in that, The anomaly detection model is a convolutional neural network model.
6. The method as described in claim 1 or 4, characterized in that, Also includes: Acquire historical data of rail vehicles; Each piece of historical data includes: historical vehicle operation data, historical anomaly data, and historical maintenance strategy data; Based on the historical data, multiple anomaly judgment rules for rail vehicles, a rail vehicle anomaly knowledge graph, and the correlation between anomaly information and maintenance strategies are determined.
7. A device for determining a maintenance strategy for rail vehicles, characterized in that, include: The data acquisition module is used to acquire vehicle operation data of rail vehicles; The rule matching module is used to process vehicle operation data based on preset anomaly judgment rules to obtain the first anomaly information; The first abnormal information includes information about the abnormal component corresponding to the abnormal operating data and information about the data abnormality type; The image acquisition module is used to acquire image information of the component corresponding to the abnormal component information; The image processing module is used to input image information into a pre-trained anomaly detection model to obtain second anomaly information; The second anomaly information includes component anomaly type information; wherein, the anomaly detection model is obtained by training a machine learning model using historical images of each component of the rail vehicle and historical vehicle anomaly information; The fusion analysis module is used to perform fusion analysis on the first and second anomaly information based on the preset rail vehicle anomaly knowledge graph to obtain a comprehensive anomaly analysis result. The strategy determination module is used to generate the final rail vehicle maintenance strategy based on the comprehensive anomaly analysis results.
8. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method for determining the rail vehicle maintenance strategy according to any one of claims 1 to 6.
9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method for determining the rail vehicle maintenance strategy according to any one of claims 1 to 6.
10. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the method for determining the rail vehicle maintenance strategy as described in any one of claims 1 to 6.