A locomotive brake health state prediction and management method, system and medium

Abstract

This invention discloses a method, system, and medium for predicting and managing the health status of locomotive brake systems based on fault tree analysis and data fusion. The method involves constructing a digital fault prediction model for the locomotive brake system based on fault tree principles; collecting real-time status data from each subsystem of the locomotive and extracting quantitative features to determine the status of each basic event in the model; dynamically calculating the probability of each event using these quantitative features and the digital fault prediction model, and then conducting graded assessments and early warnings accordingly; if an early warning is not triggered, the method continues to collect status data from each subsystem of the locomotive; if an early warning is triggered, an intelligent maintenance work order is generated; maintenance personnel perform on-site repairs on the locomotive brake system according to the intelligent maintenance work order, and the repair results are fed back to the digital fault prediction model to optimize the model and algorithm parameters. This invention, by constructing a fault tree model and fusing multi-source real-time data, achieves dynamic assessment of the brake system's health status, early prediction of fault risks, and intelligent support for maintenance decisions.

Description

Technical Field

[0001] This invention belongs to the field of intelligent operation and maintenance and condition monitoring technology of rail transit equipment. More specifically, it is to design a method, system and medium for predicting and managing the health status of locomotive brakes. Background Technology

[0002] Microcomputer-controlled electro-pneumatic braking systems (such as CCBⅡ brakes) are core components ensuring the safety of modern locomotives, and their performance and stability directly determine the safety and reliability of locomotive operation. With the upgrading of locomotive technology, these systems exhibit high integration and networking characteristics, with close signal interaction and functional coupling between modules. This means that faults are no longer limited to the independent failure of a single component, but rather manifest as complex, multi-factor-induced related faults. For example, "penalty braking" faults may originate from multiple causes such as communication interruption of the integrated processing module (IPM), abnormal signals from the electronic brake valve (EBV), and air circuit faults in the electro-pneumatic control unit (EPCU), significantly increasing the difficulty of fault diagnosis and maintenance.

[0003] Currently, locomotive maintenance departments still mainly rely on a passive response mode after a fault occurs when maintaining microcomputer-controlled electro-pneumatic braking systems. They lack the ability to provide early warning of potential faults and the ability to systematically analyze the root causes of faults. As a result, maintenance work has been in a passive and lagging state for a long time. This not only keeps maintenance costs high, but also makes it difficult to effectively avoid locomotive breakdowns during the journey, which seriously affects locomotive operating efficiency and driving safety.

[0004] Currently, the mainstream maintenance technology solutions adopted in the industry mainly include three categories: First, fault code diagnosis, which reads fault codes (such as 085 and 077 codes) through the brake display screen (LCDM) or locomotive microcomputer screen, and technicians replace the corresponding parts according to the code manual; Second, periodic preventive maintenance, which disassembles, inspects, cleans, and forcibly replaces vulnerable parts according to fixed operating mileage or time, ignoring the actual condition of the parts; Third, simple data recording and analysis, where some high-end systems can record operating parameters, but the analysis is limited to over-limit alarms and simple trend viewing, lacking the ability to deeply integrate and mine multi-source data and intelligently analyze fault correlations.

[0005] A thorough analysis of the existing technical solutions reveals the following significant drawbacks: First, the fault response is passive and lacks early warning capabilities. Existing solutions can only initiate the diagnosis and maintenance process after a fault occurs and triggers a fault code or obvious fault symptoms appear. This makes it impossible to identify potential risks and take early intervention measures in the early stages of a fault, which means that once a fault occurs, it may directly affect the normal operation of the locomotive or even cause a safety accident.

[0006] Secondly, the isolated and one-sided fault diagnosis is prone to misdiagnosis and over-repair. Current diagnostic methods often rely on single fault codes for component location, ignoring the coupling and interrelationships between components in a microcomputer-controlled electro-pneumatic braking system, which easily leads to misdiagnosis. For example, an IPM communication loss fault may be caused by an abnormal power supply to the power module, but technicians often directly replace the IPM module based solely on the fault code, failing to eradicate the root cause of the fault. Furthermore, this isolated diagnostic approach may lead to the forced replacement of unnecessary components, resulting in wasted maintenance resources and over-repair.

[0007] Third, maintenance work is highly dependent on manual experience, resulting in low efficiency and a low degree of standardization. The fault diagnosis and handling process relies heavily on the personal practical experience of technicians. However, the training period for senior technicians is long and costly, and the judgment standards of different technicians vary, making it difficult to form a standardized maintenance operation process. This leads to inconsistent maintenance efficiency and fails to meet the high-efficiency requirements of large-scale locomotive operation and maintenance.

[0008] Fourth, the lack of systematic root cause analysis capabilities makes it difficult to support continuous improvement of the maintenance system. Existing technical solutions can only achieve superficial handling of "failure occurrence - component replacement," and cannot delve into the root causes behind the failures. For example, they cannot answer deeper questions such as "why does a certain component fail frequently" or "is the failure related to design flaws or unreasonable maintenance procedures?" This makes it impossible to make targeted improvements from the aspects of design optimization and maintenance system improvement, and it is difficult to reduce the failure rate from the root cause.

[0009] In summary, current maintenance technologies for locomotive microcomputer-controlled electro-pneumatic braking systems suffer from numerous shortcomings, including passive response, limited diagnostics, low efficiency, and a lack of root cause analysis. These shortcomings make it difficult to meet the high reliability and low maintenance costs required by modern locomotives for braking systems. Therefore, developing a technical solution capable of predicting health status, accurately diagnosing faults, and systematically managing them has become an urgent technical problem to be solved in this field. Summary of the Invention

[0010] To overcome the shortcomings of existing technologies, this invention proposes a method, system, and medium for predicting and managing the health status of locomotive brakes based on fault tree and data fusion. It constructs a digital fault prediction model based on the fault tree principle and integrates real-time data from multiple sources to achieve dynamic assessment of the health status of the brake system, early prediction of fault risks, and intelligent support for maintenance decisions.

[0011] The objective of this invention can be achieved through the following technical solutions.

[0012] A method for predicting and managing the health status of locomotive brakes includes the following steps: S1. Based on the fault tree principle, a digital fault prediction model for the locomotive braking system is constructed. S2 collects the status data of each subsystem of the locomotive in real time and extracts quantitative features from them to determine the status of each basic event in the digital fault prediction model. S3. Based on the quantitative characteristics, the probability of occurrence of each event is dynamically calculated using the above-mentioned digital fault prediction model, and graded evaluation and early warning are performed based on the probability calculation results. S4. If the warning is not triggered, the process loops to step S2. If the warning is triggered, an intelligent maintenance work order is generated. S5: After the maintenance personnel complete the on-site maintenance of the locomotive braking system according to the intelligent maintenance work order, they feed the processing results back to the digital fault prediction model in step S1 to optimize the model and algorithm parameters and form a closed loop.

[0013] Furthermore, the construction process of the digital fault prediction model of the locomotive braking system described in step S1 is as follows: Based on the fault tree principle, with "locomotive braking system functional failure" as the top event, and according to the engineering principles of the locomotive braking system, all intermediate events and basic events are decomposed layer by layer downwards. Each basic event is associated with one or more measurable quantitative features, and finally a structured logical model is formed, which is the digital fault prediction model of the locomotive braking system.

[0014] Furthermore, the status data of each subsystem of the locomotive mentioned in step S2 includes the event log of the IPM system, the node pressure data of the EPCU system, the handle position signal of the EBV system, the pressure data of each major module provided by the 6A system, and the environmental data provided by the TCMS system.

[0015] Further, the specific process of step S3 is as follows: The quantitative features extracted in real time are input into the digital fault prediction model. The probability of occurrence of the top event, intermediate events, and basic events under the current data is dynamically calculated using the Bayesian network inference algorithm. The probability of occurrence of each event is compared with the corresponding multi-level early warning threshold, and the health status is assessed in a graded manner, including normal, observation, early warning, and alarm. When the probability of occurrence of a certain event exceeds the observation threshold, an early warning is issued to the operation and maintenance personnel.

[0016] Furthermore, the intelligent maintenance work order described in step S4 clearly points out the risk points and possible cause chains of the locomotive braking system. This work order not only includes early warning information, but also includes the most likely root cause of the failure, specific maintenance procedure suggestions, and a list of recommended spare parts and tools.

[0017] The objective of this invention can also be achieved through the following technical solutions.

[0018] A locomotive brake health status prediction and management system, comprising: Digital Fault Prediction Model Construction Module: Based on the fault tree principle, a digital fault prediction model for the locomotive braking system is constructed. Multi-source data acquisition and fusion module: Real-time acquisition of status data of various subsystems of locomotive, and extraction of quantitative features from them to determine the status of each basic event in the digital fault prediction model; Health status assessment and risk prediction module: Based on the quantitative characteristics, the probability of occurrence of each event is dynamically calculated using the above-mentioned digital fault prediction model, and graded assessment and early warning are performed based on the probability calculation results; Intelligent maintenance decision support module: If the warning is not triggered, the multi-source data acquisition and fusion module continues to collect the status data of each subsystem of the locomotive in real time. If the warning is triggered, an intelligent maintenance work order is generated. After the maintenance personnel complete the on-site maintenance of the locomotive braking system according to the intelligent maintenance work order, they will feed the processing results back to the digital fault prediction model construction module to optimize the digital fault prediction model and algorithm parameters.

[0019] An electronic device includes 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 steps of the above-described locomotive brake health status prediction and management method.

[0020] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described locomotive brake health status prediction and management method.

[0021] Compared with the prior art, the beneficial effects of the technical solution of the present invention are: (1) Change from passive to active: The operation and maintenance mode is completely transformed from "fault repair" to "predictive condition repair", which can issue early warnings before the fault occurs, greatly reduce mechanical damage and emergency repairs, and ensure driving safety.

[0022] (2) Improve diagnostic accuracy: Through systematic fault tree model and multivariate data correlation analysis, the root cause of the fault can be accurately located, avoiding "misjudgment" and "repeated repair", significantly improving the first-time repair success rate and reducing spare parts and labor costs.

[0023] (3) Realize knowledge accumulation and standardization: Solidify various possible fault events of locomotive brakes in the fault tree model, reduce the dependence on individual senior technicians, make maintenance operation process streamlined and standardized, which is conducive to technology transfer and new employee training.

[0024] (4) Forming a closed-loop optimization ecosystem: By continuously learning maintenance feedback data, it can optimize and update itself, making the fault prediction model more and more accurate, forming a continuously evolving intelligent operation and maintenance ecosystem. Attached Figure Description

[0025] Figure 1 This is a flowchart of the locomotive brake health status prediction and management method of the present invention.

[0026] Figure 2 This is an example fault tree diagram of the locomotive braking system in this invention. Detailed Implementation

[0027] The present invention will now be further described with reference to the accompanying drawings.

[0028] like Figure 1 As shown, the locomotive brake health status prediction and management method of the present invention mainly includes the following steps: S1. Based on the fault tree principle, a digital fault prediction model for the locomotive braking system is constructed. S2 collects the status data of each subsystem of the locomotive in real time and extracts quantitative features from them to determine the status of each basic event in the digital fault prediction model. S3. Based on the quantitative characteristics, the probability of occurrence of each event is dynamically calculated using the above-mentioned digital fault prediction model, and graded evaluation and early warning are performed based on the probability calculation results. S4. If the warning is not triggered, the process loops to step S2. If the warning is triggered, an intelligent maintenance work order is generated. S5: After the maintenance personnel complete the on-site maintenance of the locomotive braking system according to the intelligent maintenance work order, they feed back the processing results (such as information on the replaced parts) to the digital fault prediction model in step S1, optimize the model and algorithm parameters, and form a closed-loop prediction and management.

[0029] In one possible implementation, preferably, the specific process of step S1 above is as follows: based on the fault tree principle, with "locomotive brake system functional failure" as the top event, and according to the locomotive brake system engineering principle, all intermediate events and basic events are decomposed layer by layer downwards. Each basic event is associated with one or more measurable quantitative features, and finally a structured logical model is formed, that is, the digital fault prediction model of the locomotive brake system is obtained.

[0030] The basic events mentioned above are the lowest-level events in the fault tree, which do not require or cannot be further decomposed. They typically represent a fault, failure, or abnormal state of a specific component or function in the locomotive braking system. Basic events can be directly associated with specific measurable performance indicators or physical parameters and are the basic units for constructing the logical reasoning of the fault tree. Examples include "IPM communication interruption" and "EBV potentiometer signal out of tolerance".

[0031] The intermediate event refers to an event located between the top event and the basic events. It represents a fault manifestation of one or more intermediate states that lead to the occurrence of the top event. It is usually an anomaly of a subsystem or functional module composed of several basic events combined through logical relationships (such as "AND gate" or "OR gate"). For example, "penalty braking cannot be relieved" is an intermediate event, which may be caused by multiple basic events.

[0032] In the context of locomotive braking systems, the quantified characteristics refer to physical or logical parameters that can be measured numerically or statistically, reflecting component performance degradation or fault conditions. They serve as a crucial bridge for transforming abstract fault phenomena into calculable and monitorable data. For example, the "IPM pin wear" basic event. Monitor "IPM communication port signal strength" and "bit error rate". Basic event: "EBV potentiometer poor contact". Monitor the "variance of potentiometer output voltage" and the "resistance value drift trend".

[0033] Examples of the logical relationships between the top event, intermediate events, and basic events mentioned above are shown in Table 1 and... Figure 2 As shown.

[0034] Table 1 shows the logical relationships between top events, intermediate events, and basic events.

[0035] In one possible implementation, preferably, step S2 above involves the following steps: Real-time acquisition of status data from each subsystem via the vehicle network. This data primarily includes event logs from the IPM (Integrated Processing Module) system, node pressure data from the EPCU (Electro-Pneumatic Control Unit) system, handle position signals from the EBV (Electronic Brake Valve) system, pressure data from each major module provided by the 6A system, and environmental data such as locomotive vibration and grid voltage provided by the TCMS (Locomotive Control and Monitoring System). From this status data, quantitative features for determining the status of each basic event in the digital fault prediction model are extracted, such as the "correlation characteristics between LON network load rate and bit error rate" and "power supply ripple amplitude."

[0036] In one possible implementation, preferably, the specific process of step S3 above is as follows: the quantitative features extracted in real time are input into the digital fault prediction model of the locomotive braking system. The model uses Bayesian network inference algorithms and other methods to dynamically calculate the probability of occurrence of the top event, intermediate events, and basic events under the current data. The probability of occurrence of each event is compared with the corresponding multi-level warning thresholds to assess the health status in a graded manner, including normal, observation, warning, and alarm.

[0037] The multi-level early warning thresholds corresponding to the top, intermediate, and basic events mentioned above can be customized according to actual conditions. The customization format for each event is as follows: If the probability of an event occurring is less than or equal to the normal threshold, the event is considered to be in a normal state; if the normal threshold is less than the probability of an event occurring and less than or equal to the observation threshold, the event is considered to be in an observation state; if the observation threshold is less than the probability of an event occurring and less than or equal to the early warning threshold, the event is considered to be in an early warning state; if the early warning threshold is less than the probability of an event occurring, the event is considered to be in an alarm state. For example: Normal ≤ 5%, Observation 5%-20%, Early Warning 20%-50%, Alarm > 50%. When the probability of an event occurring exceeds the observation threshold, an early warning is issued to the operations and maintenance personnel.

[0038] In one possible implementation, preferably, the specific process of step S4 is as follows: if the warning is not triggered, the process loops back to step S2 to continue collecting status data of each subsystem of the locomotive in real time; if the warning is triggered, an intelligent maintenance work order is generated. The intelligent maintenance work order clearly points out the risk points and possible cause chains of the locomotive braking system. The work order not only includes warning information, but also the most likely root cause of the fault (based on fault tree deduction), specific maintenance step suggestions (such as "prioritize checking the PJB power line connectors"), and a recommended list of spare parts and tools.

[0039] The core of this invention lies in constructing a "model-driven, data-empowered" intelligent operation and maintenance closed loop. This invention transforms the complex fault logic of locomotive braking systems into a computable digital fault tree, and effectively correlates abstract fault phenomena with specific, monitorable quantitative characteristics. Quantitative characteristics capable of effectively characterizing the states of each basic event are extracted from data of different sources and formats, such as IPM, EBV, EPCU, TCMS, and 6A systems. Using the input quantitative characteristic data, rapid and accurate logical reasoning and probability calculations are performed in the fault tree model to achieve quantitative risk assessment.

[0040] Based on the principles of the above-mentioned locomotive brake health status prediction and management method, this invention also proposes a locomotive brake health status prediction and management system, which mainly includes the following modules: Digital Fault Prediction Model Construction Module: Based on the fault tree principle, a digital fault prediction model for the locomotive braking system is constructed. Multi-source data acquisition and fusion module: Real-time acquisition of status data of various subsystems of locomotive, and extraction of quantitative features from them to determine the status of each basic event in the digital fault prediction model; Health status assessment and risk prediction module: Based on the quantitative characteristics, the probability of occurrence of each event is dynamically calculated using the above-mentioned digital fault prediction model, and graded assessment and early warning are performed based on the probability calculation results; Intelligent maintenance decision support module: If the warning is not triggered, the multi-source data acquisition and fusion module continues to collect the status data of each subsystem of the locomotive in real time. If the warning is triggered, an intelligent maintenance work order is generated. After the maintenance personnel complete the on-site maintenance of the locomotive braking system according to the intelligent maintenance work order, they will feed the processing results back to the digital fault prediction model construction module to optimize the digital fault prediction model and algorithm parameters.

[0041] Based on the principle of the above-mentioned locomotive brake health status prediction and management method, the present invention also proposes an electronic 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 steps of the above-mentioned locomotive brake health status prediction and management method.

[0042] Based on the principle of the above-mentioned locomotive brake health status prediction and management method, the present invention also proposes a computer-readable storage medium storing a computer program thereon. When the computer program is executed by a processor, the steps of the above-mentioned locomotive brake health status prediction and management method are implemented.

[0043] Although the functions and working processes of the present invention have been described above in conjunction with the accompanying drawings, the present invention is not limited to the specific functions and working processes described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims, and all of these are within the protection scope of the present invention.

Claims

1. A method for predicting and managing the health status of locomotive brakes, characterized in that, Includes the following steps: S1. Based on the fault tree principle, a digital fault prediction model for the locomotive braking system is constructed. S2 collects the status data of each subsystem of the locomotive in real time and extracts quantitative features from them to determine the status of each basic event in the digital fault prediction model. S3. Based on the quantitative characteristics, the probability of occurrence of each event is dynamically calculated using the above-mentioned digital fault prediction model, and graded evaluation and early warning are performed based on the probability calculation results. S4. If the warning is not triggered, the process loops to step S2. If the warning is triggered, an intelligent maintenance work order is generated. S5: After the maintenance personnel complete the on-site maintenance of the locomotive braking system according to the intelligent maintenance work order, they feed the processing results back to the digital fault prediction model in step S1 to optimize the model and algorithm parameters and form a closed loop.

2. The method for predicting and managing the health status of locomotive brakes according to claim 1, characterized in that, The construction process of the digital fault prediction model of the locomotive braking system described in step S1 is as follows: Based on the fault tree principle, with "locomotive braking system functional failure" as the top event, and according to the engineering principles of the locomotive braking system, all intermediate events and basic events are decomposed layer by layer downwards. Each basic event is associated with one or more measurable quantitative features, and finally a structured logical model is formed, which is the digital fault prediction model of the locomotive braking system.

3. The method for predicting and managing the health status of locomotive brakes according to claim 1, characterized in that, The status data of each subsystem of the locomotive mentioned in step S2 includes the event log of the IPM system, the node pressure data of the EPCU system, the handle position signal of the EBV system, the pressure data of each major module provided by the 6A system, and the environmental data provided by the TCMS system.

4. The method for predicting and managing the health status of locomotive brakes according to claim 1, characterized in that, The specific process of step S3 is as follows: The quantitative features extracted in real time are input into the digital fault prediction model. The Bayesian network inference algorithm is used to dynamically calculate the probability of occurrence of the top event, intermediate events, and basic events under the current data. The probability of occurrence of each event is compared with the corresponding multi-level early warning threshold, and the health status is assessed in a graded manner, including normal, observation, early warning, and alarm. When the probability of occurrence of a certain event exceeds the observation threshold, an early warning is issued to the operation and maintenance personnel.

5. The method for predicting and managing the health status of locomotive brakes according to claim 1, characterized in that, The intelligent maintenance work order described in step S4 clearly points out the risk points and possible cause chains of the locomotive braking system. The work order not only includes early warning information, but also the most likely root cause of the failure, specific maintenance procedure suggestions, and a list of recommended spare parts and tools.

6. A locomotive brake health status prediction and management system based on the locomotive brake health status prediction and management method according to any one of claims 1 to 5, characterized in that, include: Digital Fault Prediction Model Construction Module: Based on the fault tree principle, a digital fault prediction model for the locomotive braking system is constructed. Multi-source data acquisition and fusion module: Real-time acquisition of status data of various subsystems of locomotive, and extraction of quantitative features from them to determine the status of each basic event in the digital fault prediction model; Health status assessment and risk prediction module: Based on the quantitative characteristics, the probability of occurrence of each event is dynamically calculated using the above-mentioned digital fault prediction model, and graded assessment and early warning are performed based on the probability calculation results; Intelligent maintenance decision support module: If the warning is not triggered, the multi-source data acquisition and fusion module continues to collect the status data of each subsystem of the locomotive in real time. If the warning is triggered, an intelligent maintenance work order is generated. After the maintenance personnel complete the on-site maintenance of the locomotive braking system according to the intelligent maintenance work order, they will feed the processing results back to the digital fault prediction model construction module to optimize the digital fault prediction model and algorithm parameters.

7. An electronic 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 steps of the locomotive brake health status prediction and management method as described in any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the locomotive brake health status prediction and management method as described in any one of claims 1 to 5.

Images