Brake circuit leakage early warning method and system

Abstract

This invention discloses a method and system for early warning of brake circuit leakage, relating to the technical field of braking systems for rail transit vehicles. The method includes: collecting pressure data from multiple brake cylinders in the braking system; wherein the multiple brake cylinder pressure data includes at least pressure data from a first brake cylinder, pressure data from a second brake cylinder, and pressure data from a target brake cylinder; and identifying continuous data segments of the brake circuit in a steady-state pressure-maintaining condition based on the target brake cylinder pressure data. This invention addresses the problems of low efficiency, poor real-time performance, and reliance on maintenance personnel's experience in current manual inspections, making it difficult to detect early or intermittent leaks. While electronic control detection methods based on fixed thresholds achieve a certain degree of automation, their sensitivity and accuracy are limited, they are easily affected by operating condition fluctuations and sensor noise, and they cannot effectively distinguish leak types or quantitatively assess the degree of leakage.

Description

Technical Field

[0001] This invention relates to the field of braking system technology for rail transit vehicles, and in particular to a method and system for early warning of leakage in braking circuits. Background Technology

[0002] In the rail transit sector, the reliability of the braking system is directly related to train operation safety. As a critical air passage in a pneumatic braking system, the sealing integrity of the braking circuit is paramount. Braking circuit leakage is a common failure mode, leading to abnormal braking pressure, insufficient braking force, or delayed response, and in severe cases, potentially causing safety accidents.

[0003] Currently, the detection of brake circuit leaks mainly relies on periodic manual inspections or simple pressure threshold judgments. However, manual inspections are inefficient, lack real-time performance, and depend on the experience of maintenance personnel, making it difficult to detect early or intermittent leaks. Although electronic control detection methods based on fixed thresholds have achieved a certain degree of automation, their sensitivity and accuracy are limited, they are easily affected by fluctuations in operating conditions and sensor noise, and they cannot effectively distinguish between leak types (such as internal and external leaks), making it even more difficult to quantify the degree of leakage.

[0004] In the prior art, such as Chinese patent document with publication number CN112078558A, a test method for diagnosing leakage in the brake cylinder pipeline of a high-speed train is disclosed. This method relies on a specific pressure holding test procedure and cannot perform real-time online monitoring during normal train operation.

[0005] Another Chinese patent document with publication number CN117400897A discloses a method, device and electronic equipment for monitoring leakage faults in the brake piping system of railway freight cars. The method judges leakage by monitoring the pressure relationship between the main pipe and the auxiliary air cylinder. However, its monitoring object and judgment logic are mainly aimed at the air cylinder system specific to railway freight cars. It does not involve in-depth analysis and intelligent identification of the internal pressure fluctuation characteristics of the brake control circuit, nor does it apply machine learning methods to achieve more accurate fault mode differentiation. Summary of the Invention

[0006] Purpose of the invention: To provide a method and system for early warning of leakage in a braking circuit, so as to at least solve one of the problems existing in the prior art.

[0007] Technical solution: A method for early warning of leakage in a braking circuit, comprising: Collect pressure data from multiple brake cylinders in the braking system; wherein the pressure data from multiple brake cylinders includes at least pressure data from a first brake cylinder, pressure data from a second brake cylinder, and pressure data from a target brake cylinder. Based on the target brake cylinder pressure data, identify continuous data segments in which the brake circuit is in a steady-state pressure-holding condition; The first brake cylinder pressure data and the second brake cylinder pressure data in the continuous data segment are input into the trained unsupervised machine learning model; Based on the output of the unsupervised machine learning model, anomaly indicators characterizing the leakage state of the braking circuit are generated. In response to the abnormal indicator exceeding a preset threshold, a leak is determined to have occurred, and based on the correlation characteristics between the first brake cylinder pressure data and the second brake cylinder pressure data, the type of leak, whether internal or external, is distinguished.

[0008] Preferably, based on the target brake cylinder pressure data, continuous data segments in which the brake circuit is in a steady-state pressure-holding condition are identified, including: Filter out data segments in which the target brake cylinder pressure data remains stable for a duration exceeding a first preset time. The pressure data of the first brake cylinder and the pressure data of the second brake cylinder in the selected data segment are filtered to remove interference data with fluctuation amplitude less than or equal to a preset fluctuation threshold.

[0009] Preferably, the unsupervised machine learning model includes a collaborative model of generative adversarial networks and autoencoders; The generation of abnormal indicators characterizing the leakage state of the braking circuit based on the output of the unsupervised machine learning model includes: The reconstructed data of the input pressure data is obtained through the autoencoder model, and the first reconstruction error is calculated; The probability score of input stress data being judged as normal data is obtained through the discriminator of the generative adversarial network; The anomaly index is calculated based on the first reconstruction error and the probability score.

[0010] Preferably, the generative adversarial network is a conditional generative adversarial network, whose conditional information includes the target brake cylinder pressure value and / or braking command level; And / or, The distinction between leakage types includes: Calculate the real-time pressure difference and the rate of change of pressure difference between the pressure data of the first brake cylinder and the pressure data of the second brake cylinder; When the differential pressure remains below the first threshold and the abnormal indicator remains excessively high, it is determined to be an internal leak. When the pressure data of the first brake cylinder and the pressure data of the second brake cylinder show the same deviation in direction and have similar fluctuation patterns, it is determined to be an external leakage.

[0011] Preferably, the unsupervised machine learning model also includes a time series analysis model based on RNN-Conceptor; The generation of anomaly indicators characterizing the leakage state of the braking circuit based on the output of the unsupervised machine learning model further includes: The RNN-Conceptor model is used to extract the dynamic features of the input pressure data over time and generate the corresponding Conceptor matrix. Calculate the similarity between the Conceptor matrix and the pre-stored normal operating condition baseline Conceptor matrix; The anomaly index is generated based on the similarity.

[0012] Preferably, the feature is that it further includes: Monitor the long-term drift trend of the Conceptor matrix relative to a historical benchmark; Based on the long-term drift trend, the performance degradation status of the braking circuit is assessed, and predictive maintenance prompts are generated.

[0013] Preferably, in response to the abnormal indicator exceeding a preset threshold, a leak is determined to have occurred, and after distinguishing between internal and external leaks based on the correlation characteristics between the first brake cylinder pressure data and the second brake cylinder pressure data, the method further includes: Once a leak is detected, the leak rate is calculated based on the changing trends of the first brake cylinder pressure data and the second brake cylinder pressure data within the continuous data segment. The leakage level is determined based on the numerical range of the leakage rate; wherein the leakage level includes at least minor leakage, moderate leakage, and severe leakage.

[0014] Preferably, determining the leakage level includes: The leakage rate is normalized with the total air pressure of the current braking system to obtain the relative leakage rate; The leakage level is determined based on the percentage range of the relative leakage rate.

[0015] As a preferred option, a multi-model ensemble decision-making step is also included: Based on the first unsupervised learning model, the second unsupervised learning model, and the diagnostic model based on physical rules, respectively, the first anomaly determination result, the second anomaly determination result, and the third anomaly determination result are generated; The first anomaly determination result, the second anomaly determination result, and the third anomaly determination result are weighted and voted or fused with confidence to obtain the final leak diagnosis result; The first unsupervised learning model and the second unsupervised learning model are different types of models. To achieve the above objectives, according to another aspect of this application, a brake circuit leakage early warning device is provided.

[0016] The brake circuit leakage early warning device according to this application includes: The data acquisition module is used to acquire pressure data of multiple brake cylinders in the braking system; wherein, the pressure data of multiple brake cylinders includes at least the pressure data of the first brake cylinder, the pressure data of the second brake cylinder, and the pressure data of the target brake cylinder; The operating condition identification module is used to identify continuous data segments in which the braking circuit is in a steady-state pressure-holding condition based on the target brake cylinder pressure data. An edge computing module is used to input the first brake cylinder pressure data and the second brake cylinder pressure data in the continuous data segment into a trained unsupervised machine learning model; An anomaly indicator generation module is used to generate anomaly indicators characterizing the leakage state of the braking circuit based on the output of the unsupervised machine learning model. The leak type determination and differentiation module is used to determine that a leak has occurred in response to the abnormal index exceeding a preset threshold, and to differentiate between internal and external leaks based on the correlation characteristics between the first brake cylinder pressure data and the second brake cylinder pressure data.

[0017] Beneficial Effects: In this embodiment, an unsupervised machine learning model is used to collect pressure data from multiple brake cylinders in the braking system. This data includes at least first brake cylinder pressure data, second brake cylinder pressure data, and target brake cylinder pressure data. Based on the target brake cylinder pressure data, continuous data segments indicating a steady-state pressure-holding condition in the braking circuit are identified. The first and second brake cylinder pressure data from these continuous data segments are input into a trained unsupervised machine learning model. Based on the output of the unsupervised machine learning model, an abnormal index characterizing the leakage state of the braking circuit is generated. In response to the abnormal index exceeding a preset threshold, a leakage is determined, and the leakage is assessed based on the first and second brake cylinder pressure data. By identifying the correlation characteristics between the leaks and distinguishing between internal and external leaks, the system achieves high-frequency real-time monitoring of brake circuit leaks. This enables intelligent monitoring and diagnosis of brake circuit leaks, overcoming the shortcomings of traditional detection methods in terms of real-time performance and accuracy. Furthermore, it addresses the current limitations of relying on periodic manual inspections or simple pressure threshold judgments for brake circuit leak detection. Manual inspections are inefficient, lack real-time performance, and depend on the experience of maintenance personnel, making it difficult to detect early or intermittent leaks. While electronic control detection methods based on fixed thresholds have achieved a certain degree of automation, their sensitivity and accuracy are limited, they are easily affected by fluctuations in operating conditions and sensor noise, and they cannot effectively distinguish between leak types (such as internal and external leaks), let alone quantify the degree of leakage. Attached Figure Description

[0018] Figure 1 This is a schematic flowchart of a braking circuit leakage early warning method according to an embodiment of this application; Figure 2 This is a schematic flowchart of another braking circuit leakage early warning method according to an embodiment of this application; Figure 3 This is a comparison diagram of brake cylinder pressure fluctuations in normal and abnormal brake circuits according to the brake circuit leakage early warning method of the embodiment of this application. Figure 4 This is a comparison diagram of the external leakage fluctuation characteristics of the braking circuit leakage early warning method according to the embodiments of this application; Figure 5 This is a comparison diagram of the internal leakage fluctuation characteristics of the braking circuit leakage early warning method according to the embodiments of this application; Figure 6 This is a schematic diagram of the structure of a brake circuit leakage early warning device according to an embodiment of this application; and Figure 7 This is a schematic diagram of the electronic device of the braking circuit leakage early warning method according to an embodiment of this application. Detailed Implementation

[0019] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0020] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0021] like Figures 1-2 As shown in the embodiment of the present invention, a method for early warning of leakage in a braking circuit is provided, the method comprising the following steps S101 to S105: Step S101: Collect multi-channel brake cylinder pressure data from the braking system; wherein the multi-channel brake cylinder pressure data includes at least first brake cylinder pressure data, second brake cylinder pressure data, and target brake cylinder pressure data; Specifically, the pressure data of the first brake cylinder and the pressure data of the second brake cylinder are derived from the first brake cylinder and the second brake cylinder, which are usually located at different actuators of the same braking circuit, or at symmetrical or related positions such as the left and right wheels or the front and rear axles. The target brake cylinder pressure data comes from the target brake cylinder, which can be the master brake cylinder or a reference brake cylinder used to characterize the current braking state. Its pressure changes can reflect whether the braking system is in the pressurization, depressurization or pressure holding stage.

[0022] By collecting pressure data from multiple brake cylinders in the braking system using sensors, a multi-source, redundant, and cross-verifiable pressure data foundation is provided, thus providing the necessary conditions for subsequent identification of leakage modes and determination of leakage locations.

[0023] Step S102: Based on the target brake cylinder pressure data, identify continuous data segments in which the brake circuit is in a steady-state pressure-holding condition; According to an embodiment of the present invention, preferably, identifying continuous data segments in which the braking circuit is in a steady-state pressure-holding condition based on the target brake cylinder pressure data includes: Filter out data segments in which the target brake cylinder pressure data remains stable for a duration exceeding a first preset time. The pressure data of the first brake cylinder and the pressure data of the second brake cylinder in the selected data segment are filtered to remove interference data with fluctuation amplitude less than or equal to a preset fluctuation threshold.

[0024] Specifically, data points with constant target brake cylinder pressure for more than 6,000 hours are extracted from the obtained data. The data is then filtered to extract valid data segments and remove irrelevant fluctuation data.

[0025] It can effectively eliminate interference from normal control behaviors such as pressurization and depressurization in leak detection, thereby focusing on the operating conditions that are most likely to expose leak characteristics; at the same time, it can also reduce false alarm rate and improve diagnostic accuracy.

[0026] like Figures 3-5 As shown, pressure fluctuations in the first brake cylinder (BC1) and the second brake cylinder (BC2) under constant target brake cylinder (BC) pressure are analyzed to obtain pressure change curves. Data with fluctuations not exceeding the normal range (typically not exceeding 8 kPa) are filtered out due to potential sensor drift. In the figure, the blue solid line represents the emergency pre-control target pressure; the orange dotted line represents the AC pressure; the orange-yellow dashed line represents the target BC; the purple dashed line represents the BC1 pressure; and the green solid line represents the BC2 pressure.

[0027] Step S103: Input the first brake cylinder pressure data and the second brake cylinder pressure data in the continuous data segment into the trained unsupervised machine learning model; It enables the input of precise stress data into unsupervised machine learning models, thereby ensuring accurate output results.

[0028] Step S104: Based on the output of the unsupervised machine learning model, generate anomaly indicators characterizing the leakage state of the braking circuit; According to an embodiment of the present invention, preferably, the unsupervised machine learning model includes a collaborative model of generative adversarial networks and autoencoders; The generation of abnormal indicators characterizing the leakage state of the braking circuit based on the output of the unsupervised machine learning model includes: The reconstructed data of the input pressure data is obtained through the autoencoder model, and the first reconstruction error is calculated; The probability score of input stress data being judged as normal data is obtained through the discriminator of the generative adversarial network; The anomaly index is calculated based on the first reconstruction error and the probability score.

[0029] By utilizing generative adversarial networks (GANs) and auto-encoder models, the normal pressure distribution is reconstructed, and abnormal fluctuations in the pressure curve are identified through the reconstruction error.

[0030] According to an embodiment of the present invention, preferably, the unsupervised machine learning model further includes a time series analysis model based on RNN-Conceptor; The generation of anomaly indicators characterizing the leakage state of the braking circuit based on the output of the unsupervised machine learning model further includes: The RNN-Conceptor model is used to extract the dynamic features of the input pressure data over time and generate the corresponding Conceptor matrix. Calculate the similarity between the Conceptor matrix and the pre-stored normal operating condition baseline Conceptor matrix; The anomaly index is generated based on the similarity.

[0031] By using the unsupervised RNN-Conceptor method to extract dynamic features from time series data, and by comparing the similarity of time series Conceptor matrices, abnormal fluctuation patterns and leakage states are identified.

[0032] According to an embodiment of the present invention, preferably, it further includes: Monitor the long-term drift trend of the Conceptor matrix relative to a historical benchmark; Based on the long-term drift trend, the performance degradation status of the braking circuit is assessed, and predictive maintenance prompts are generated.

[0033] Step S105: In response to the abnormal index exceeding the preset threshold, a leak is determined to have occurred, and based on the correlation characteristics between the first brake cylinder pressure data and the second brake cylinder pressure data, the type of leak, whether internal or external, is distinguished.

[0034] According to an embodiment of the present invention, preferably, the generative adversarial network is a conditional generative adversarial network, and its condition information includes the target brake cylinder pressure value and / or braking command level; And / or, The distinction between leakage types includes: Calculate the real-time pressure difference and the rate of change of pressure difference between the pressure data of the first brake cylinder and the pressure data of the second brake cylinder; When the differential pressure remains below the first threshold and the abnormal indicator remains excessively high, it is determined to be an internal leak. When the pressure data of the first brake cylinder and the pressure data of the second brake cylinder show the same deviation in direction and have similar fluctuation patterns, it is determined to be an external leakage.

[0035] According to an embodiment of the present invention, preferably, after determining that a leak has occurred in response to the abnormal indicator exceeding a preset threshold, and distinguishing between internal and external leaks based on the correlation characteristics between the first brake cylinder pressure data and the second brake cylinder pressure data, the method further includes: Once a leak is detected, the leak rate is calculated based on the changing trends of the first brake cylinder pressure data and the second brake cylinder pressure data within the continuous data segment. The leakage level is determined based on the numerical range of the leakage rate; wherein the leakage level includes at least minor leakage, moderate leakage, and severe leakage.

[0036] According to an embodiment of the present invention, preferably, determining the leakage level includes: The leakage rate is normalized with the total air pressure of the current braking system to obtain the relative leakage rate; The leakage level is determined based on the percentage range of the relative leakage rate.

[0037] According to an embodiment of the present invention, preferably, it further includes a multi-model ensemble decision-making step: Based on the first unsupervised learning model, the second unsupervised learning model, and the diagnostic model based on physical rules, respectively, the first anomaly determination result, the second anomaly determination result, and the third anomaly determination result are generated; The first anomaly determination result, the second anomaly determination result, and the third anomaly determination result are weighted and voted or fused with confidence to obtain the final leak diagnosis result; The first unsupervised learning model and the second unsupervised learning model are different types of models.

[0038] As can be seen from the above description, this application achieves the following technical effects: In this embodiment, an unsupervised machine learning model is used to collect pressure data from multiple brake cylinders in the braking system. This data includes at least first brake cylinder pressure data, second brake cylinder pressure data, and target brake cylinder pressure data. Based on the target brake cylinder pressure data, continuous data segments indicating a steady-state pressure-holding condition in the braking circuit are identified. The first and second brake cylinder pressure data from these continuous data segments are input into a trained unsupervised machine learning model. Based on the output of the unsupervised machine learning model, an anomaly index characterizing the leakage state of the braking circuit is generated. In response to the anomaly index exceeding a preset threshold, a leakage is determined, and the relationship between the first and second brake cylinder pressure data is considered. By identifying leakage characteristics and distinguishing between internal and external leaks, the system achieves high-frequency real-time monitoring of brake circuit leaks, thus realizing intelligent monitoring and diagnosis of brake circuit leaks. This overcomes the shortcomings of traditional detection methods in terms of real-time performance and accuracy. Furthermore, it solves the current problem that brake circuit leak detection mainly relies on periodic manual inspections or simple pressure threshold judgments. However, manual inspections are inefficient, lack real-time performance, and depend on the experience of maintenance personnel, making it difficult to detect early or intermittent leaks. Although electronic control detection methods based on fixed thresholds have achieved a certain degree of automation, their sensitivity and accuracy are limited, they are easily affected by operating condition fluctuations and sensor noise, and they cannot effectively distinguish between leak types (such as internal and external leaks), let alone quantify the degree of leakage.

[0039] Compared with the prior art, the present invention also has the following beneficial effects: 1. This application features real-time and intelligent capabilities. By collecting data in real time and using an unsupervised machine learning model for online analysis, it achieves uninterrupted intelligent monitoring of brake circuit leakage, overcoming the lag of manual inspection and periodic testing.

[0040] 2. This application has high accuracy and strong anti-interference ability. It adopts advanced algorithms such as GAN and Auto-Encoder collaborative model and RNN-Conceptor time series model, which can keenly capture the weak abnormal fluctuation characteristics of leakage from complex noise background, and further reduce the false alarm rate through multi-model integrated decision-making.

[0041] 3. This application has the ability to perform detailed fault diagnosis. It can not only detect the occurrence of leaks, but also effectively distinguish between the two key fault types of internal and external leaks. It can also quantitatively assess the leakage rate and level, providing accurate decision support for maintenance personnel.

[0042] 4. This application has the capability of proactive maintenance, i.e., predictive capability. By analyzing the long-term drift of model characteristics, it can perceive the slow degradation trend of system performance and realize the transformation from "post-failure maintenance" to "predictive maintenance".

[0043] This application also relates to another aspect, providing a brake circuit leakage early warning device. For example... Figure 6 As shown, the brake circuit leakage early warning device includes: The data acquisition module 601 is used to acquire multi-channel brake cylinder pressure data in the braking system; wherein, the multi-channel brake cylinder pressure data includes at least first brake cylinder pressure data, second brake cylinder pressure data and target brake cylinder pressure data; The working condition identification module 602 is used to identify continuous data segments in which the braking circuit is in a steady-state pressure-holding condition based on the target brake cylinder pressure data. Edge computing module 603 is used to input the first brake cylinder pressure data and the second brake cylinder pressure data in the continuous data segment into a trained unsupervised machine learning model; Anomaly indicator generation module 604 is used to generate anomaly indicators characterizing the leakage state of the braking circuit based on the output of the unsupervised machine learning model. The leak type determination and differentiation module 605 is used to determine that a leak has occurred in response to the abnormal indicator exceeding a preset threshold, and to differentiate the leak type as internal or external based on the correlation characteristics between the first brake cylinder pressure data and the second brake cylinder pressure data.

[0044] As can be seen from the above description, this application achieves the following technical effects: In this embodiment, an unsupervised machine learning model is used to collect pressure data from multiple brake cylinders in the braking system. This data includes at least first brake cylinder pressure data, second brake cylinder pressure data, and target brake cylinder pressure data. Based on the target brake cylinder pressure data, continuous data segments indicating a steady-state pressure-holding condition in the braking circuit are identified. The first and second brake cylinder pressure data from these continuous data segments are input into a trained unsupervised machine learning model. Based on the output of the unsupervised machine learning model, an anomaly index characterizing the leakage state of the braking circuit is generated. In response to the anomaly index exceeding a preset threshold, a leakage is determined, and the relationship between the first and second brake cylinder pressure data is considered. By identifying leakage characteristics and distinguishing between internal and external leaks, the system achieves high-frequency real-time monitoring of brake circuit leaks, thus realizing intelligent monitoring and diagnosis of brake circuit leaks. This overcomes the shortcomings of traditional detection methods in terms of real-time performance and accuracy. Furthermore, it solves the current problem that brake circuit leak detection mainly relies on periodic manual inspections or simple pressure threshold judgments. However, manual inspections are inefficient, lack real-time performance, and depend on the experience of maintenance personnel, making it difficult to detect early or intermittent leaks. Although electronic control detection methods based on fixed thresholds have achieved a certain degree of automation, their sensitivity and accuracy are limited, they are easily affected by operating condition fluctuations and sensor noise, and they cannot effectively distinguish between leak types (such as internal and external leaks), let alone quantify the degree of leakage.

[0045] like Figure 7 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded into the RAM 13 from storage unit 18. The RAM 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0046] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0047] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as a braking circuit leakage early warning method.

[0048] In some embodiments, the brake circuit leakage warning method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the brake circuit leakage warning method described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to perform the brake circuit leakage warning method by any other suitable means (e.g., by means of firmware).

[0049] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0050] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0051] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0052] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0053] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A brake circuit leak early warning method, characterized by, include: Collect pressure data from multiple brake cylinders in the braking system; wherein the pressure data from multiple brake cylinders includes at least pressure data from a first brake cylinder, pressure data from a second brake cylinder, and pressure data from a target brake cylinder. Based on the target brake cylinder pressure data, identify continuous data segments in which the brake circuit is in a steady-state pressure-holding condition; The first brake cylinder pressure data and the second brake cylinder pressure data in the continuous data segment are input into the trained unsupervised machine learning model; Based on the output of the unsupervised machine learning model, anomaly indicators characterizing the leakage state of the braking circuit are generated. In response to the abnormal index exceeding a preset threshold, a leak is determined to have occurred, and based on the correlation characteristics between the first brake cylinder pressure data and the second brake cylinder pressure data, the leak type is distinguished as internal or external leakage.

2. The method of claim 1, wherein, Based on the target brake cylinder pressure data, identify continuous data segments where the brake circuit is in a steady-state pressure-holding condition, including: Filter out data segments in which the target brake cylinder pressure data remains stable for a duration exceeding a first preset time. The pressure data of the first brake cylinder and the pressure data of the second brake cylinder in the selected data segment are filtered to remove interference data with fluctuation amplitude less than or equal to a preset fluctuation threshold.

3. The method according to claim 1, characterized in that, The unsupervised machine learning model includes a collaborative model of generative adversarial networks and autoencoders; The generation of abnormal indicators characterizing the leakage state of the braking circuit based on the output of the unsupervised machine learning model includes: The reconstructed data of the input pressure data is obtained through the autoencoder model, and the first reconstruction error is calculated; The probability score of input stress data being judged as normal data is obtained through the discriminator of the generative adversarial network; The anomaly index is calculated based on the first reconstruction error and the probability score.

4. The method according to claim 3, characterized in that, The generative adversarial network is a conditional generative adversarial network, whose conditional information includes the target brake cylinder pressure value and / or braking command level; And / or, The distinction between leakage types includes: Calculate the real-time pressure difference and the rate of change of pressure difference between the pressure data of the first brake cylinder and the pressure data of the second brake cylinder; When the differential pressure remains below the first threshold and the abnormal indicator remains excessively high, it is determined to be an internal leak. When the pressure data of the first brake cylinder and the pressure data of the second brake cylinder show the same deviation in direction and have similar fluctuation patterns, it is determined to be an external leakage.

5. The method according to claim 1, characterized in that, The unsupervised machine learning model also includes a time series analysis model based on RNN-Conceptor; The generation of anomaly indicators characterizing the leakage state of the braking circuit based on the output of the unsupervised machine learning model further includes: The RNN-Conceptor model is used to extract the dynamic features of the input pressure data over time and generate the corresponding Conceptor matrix. Calculate the similarity between the Conceptor matrix and the pre-stored normal operating condition baseline Conceptor matrix; The anomaly index is generated based on the similarity.

6. The method according to claim 5, characterized in that, Also includes: Monitor the long-term drift trend of the Conceptor matrix relative to a historical benchmark; Based on the long-term drift trend, the performance degradation status of the braking circuit is assessed, and predictive maintenance prompts are generated.

7. The method according to claim 1, characterized in that, In response to the abnormal indicator exceeding a preset threshold, a leak is determined to have occurred. Based on the correlation characteristics between the first brake cylinder pressure data and the second brake cylinder pressure data, the leak type is distinguished as internal or external leakage. The process further includes: Once a leak is detected, the leak rate is calculated based on the changing trends of the first brake cylinder pressure data and the second brake cylinder pressure data within the continuous data segment. The leakage level is determined based on the numerical range of the leakage rate; wherein the leakage level includes at least minor leakage, moderate leakage, and severe leakage.

8. The method according to claim 7, characterized in that, Determining the leakage level includes: The leakage rate is normalized with the total air pressure of the current braking system to obtain the relative leakage rate; The leakage level is determined based on the percentage range of the relative leakage rate.

9. The method according to claim 1, characterized in that, It also includes multi-model ensemble decision-making steps: Based on the first unsupervised learning model, the second unsupervised learning model, and the diagnostic model based on physical rules, respectively, the first anomaly determination result, the second anomaly determination result, and the third anomaly determination result are generated; The first anomaly determination result, the second anomaly determination result, and the third anomaly determination result are weighted and voted or fused with confidence to obtain the final leak diagnosis result; The first unsupervised learning model and the second unsupervised learning model are different types of models.

10. A brake circuit leakage early warning device, characterized in that, include: The data acquisition module is used to acquire pressure data of multiple brake cylinders in the braking system; wherein, the pressure data of multiple brake cylinders includes at least the pressure data of the first brake cylinder, the pressure data of the second brake cylinder, and the pressure data of the target brake cylinder; The operating condition identification module is used to identify continuous data segments in which the braking circuit is in a steady-state pressure-holding condition based on the target brake cylinder pressure data. An edge computing module is used to input the first brake cylinder pressure data and the second brake cylinder pressure data in the continuous data segment into a trained unsupervised machine learning model; An anomaly indicator generation module is used to generate anomaly indicators characterizing the leakage state of the braking circuit based on the output of the unsupervised machine learning model. The leak type determination and differentiation module is used to determine that a leak has occurred in response to the abnormal index exceeding a preset threshold, and to differentiate between internal and external leaks based on the correlation characteristics between the first brake cylinder pressure data and the second brake cylinder pressure data.

Images