Railway vehicle running gear service performance quantitative evaluation method considering life distribution

Abstract

The invention provides a rail vehicle running gear service performance quantitative evaluation method considering life distribution, and belongs to the technical field of rail traffic vehicle state evaluation and reliability engineering, and the method comprises the steps: classifying parts according to a part failure mechanism; respectively selecting corresponding life distribution models for different types of components to obtain time-varying reliability and failure probability of the components; constructing a hierarchical Bayesian network consistent with the walking part in structure, and calculating the posterior fault probability of each layer of nodes in combination with a variable elimination algorithm; based on the part time-varying reliability, the posterior fault probability and the multi-source state monitoring quantity, constructing a part health degree index with a unified dimension; fusing the health degree indexes of all the components from bottom to top to obtain system service performance indexes, and setting grading standards. The method solves the problems that in the prior art, unified evaluation of service performance between parts and system hierarchies is difficult to achieve, the adaptability of life models under different failure mechanisms is insufficient, and quantitative fusion of multi-source state quantities is difficult to achieve.

Description

Technical Field

[0001] This invention relates to the field of rail transit vehicle condition assessment and reliability engineering technology, and in particular to a method for quantitatively assessing the service performance of rail vehicle running gear considering life distribution. Background Technology

[0002] The running gear of a rail vehicle is a core component of a high-speed train, undertaking key functions such as load-bearing, guidance, vibration reduction, traction, and braking. Its service condition directly determines the train's operating speed, running quality, and operational safety. This system consists of key components such as wheelsets, axle boxes, suspension systems, frames, and braking devices, and is subjected to complex loads and vibration impacts over long periods, making it a high-risk area for failures.

[0003] With the rapid development of rail transit, train speeds and carrying capacity continue to increase. The load levels, vibration environments and wheel-rail interactions faced by running components are becoming increasingly complex. Problems such as wheelset wear, axle box abnormalities and suspension component aging occur frequently, leading to a gradual degradation of system performance. Serious faults may cause consequences such as line delays and casualties.

[0004] Existing technologies have significant limitations: on the one hand, they mostly focus on the failure mechanism analysis and reliability assessment at the level of a single component or a single system, making it difficult to reveal the correlation and failure propagation effects between multiple components; on the other hand, the running gear is a typical hierarchical structure system, with significant differences in the failure mechanisms, degradation rates, and safety contributions of different components, making it impossible for traditional single-level reliability indicators to accurately characterize the evolution of the overall system's service performance. Therefore, there is an urgent need for a hierarchical quantitative assessment method that integrates component-level characteristics, system-level structural dependencies, and multi-source condition monitoring quantities. Summary of the Invention

[0005] This invention provides a quantitative evaluation method for the service performance of the running gear of a rail vehicle that takes into account the life distribution, in order to solve the problems in the prior art such as the difficulty in achieving unified evaluation of service performance between component and system levels, insufficient adaptability of life models under different failure mechanisms, and difficulty in quantitatively fusing multi-source state variables.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: Methods for quantitatively evaluating the service performance of rail vehicle running gear considering life distribution include: Historical fault records and failure time data of running gear components are acquired, and the components are classified according to their failure mechanisms. Multi-source status monitoring data during the operation of the running gear are acquired through the sensor acquisition module. For different categories of components, corresponding life distribution models are selected. By minimizing the mean square error between the empirical cumulative distribution and the theoretical distribution and maximizing the coefficient of determination, the adaptive selection and parameter estimation of the life model are achieved, and the time-varying reliability and failure probability of the components are obtained. A hierarchical Bayesian network consistent with the structure of the running section is constructed, and the hierarchical relationship between the root node, intermediate nodes and leaf nodes is defined. The prior fault probability of the leaf node is the failure probability. When the sample is insufficient, the judgment matrix is ​​constructed by the analytic hierarchy process to calculate the node weight. The intermediate nodes adopt the weighted OR gate logic and the posterior fault probability of each layer node is calculated by combining the variable elimination algorithm. Based on component time-varying reliability, posterior failure probability, and multi-source condition monitoring data, a component health index with unified dimensions is constructed. Weights are set according to the structural hierarchy of the running gear, and the health indices of each component are integrated from bottom to top to obtain the system service performance index. Grading standards are set according to the values ​​of the system service performance index to achieve dynamic quantitative grading and trend prediction of the running gear service performance.

[0007] In this specification, the components are classified based on differences in failure mechanisms, into service life type components and full life type components.

[0008] In this specification, the lifetime distribution model includes the Weibull distribution model and the exponential distribution model. The Weibull distribution model is used for life-use components, and the exponential distribution model is used for full-life-use components. The parameters of the Weibull distribution model are estimated by the least squares method, and the parameters of the exponential distribution model are estimated by the maximum likelihood estimation method.

[0009] In this specification, the root node of the hierarchical Bayesian network is the running gear, the intermediate nodes include the load-bearing and guiding system, the traction drive system, and the braking system, and the leaf nodes include wheelsets, axle boxes, suspension devices, and gearboxes. The relationship between the nodes is consistent with the causal relationship of the actual structure of the running gear.

[0010] In this specification, the analytic hierarchy process (AHP) uses a 1-9 scale to construct a judgment matrix. The scale is used to represent the relative importance between different nodes, and the node weights are obtained by calculating the eigenvectors of the judgment matrix.

[0011] In this specification, the variable elimination algorithm is used for probabilistic inference of hierarchical Bayesian networks. By eliminating irrelevant node variables, the posterior fault probability of each layer node is calculated, thereby realizing the hierarchical propagation of fault probability.

[0012] In this specification, the health index is constructed by integrating the time-varying reliability of the component, the posterior failure probability, and the standardized multi-source state monitoring quantity. The weight coefficients of the three are dynamically adjusted according to the operation stage of the rail vehicle, and the sum of the weight coefficients is 1.

[0013] In this specification, multi-source condition monitoring quantities are standardized to be mapped to health-related indicators. The multi-source condition monitoring quantities include vibration monitoring data and temperature monitoring data. The standardization process converts various monitoring data into quantitative indicators within the same range to reflect the impact of real-time operating status on the health level of components.

[0014] In this specification, the system service performance index ranges from 0 to 1. The larger the value, the better the system service performance. The classification standard includes safe zone, early warning zone and dangerous zone. The corresponding area level and operation and maintenance recommendations are determined according to the specific value of the system service performance index.

[0015] In this specification, the threshold values ​​of the grading standards are fine-tuned according to the specific model of the rail vehicle or the environmental conditions of the operating line to adapt to the service performance evaluation requirements under different application scenarios.

[0016] In summary, the present invention has at least the following beneficial effects: It enables quantitative evaluation of the service performance of rail vehicle running gear at all levels, from component level and subsystem level to system level, breaking through the limitations of traditional partial evaluation technology and fully depicting the evolution of the overall service status of the system.

[0017] By adapting lifetime models to different failure mechanisms and using adaptive selection mechanisms, the differentiated characteristics of different components are taken into account, thereby improving the accuracy and adaptability of the evaluation results.

[0018] By integrating multi-source condition monitoring data with probabilistic reasoning logic, dynamic classification and trend prediction of service performance are realized, providing a scientific basis for predictive maintenance and effectively supporting the optimization of operation and maintenance decisions (such as maintenance cycle adjustment and key maintenance planning).

[0019] By combining data-driven approaches with expert knowledge, the technology enhances its engineering practicality and scenario adaptability, making it suitable for evaluating running gear under different vehicle models and route conditions. Attached Figure Description

[0020] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. 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.

[0021] Figure 1 This is a schematic diagram of the method for quantitatively evaluating the service performance of the running gear of a rail vehicle, which takes into account the life distribution, as involved in this invention.

[0022] Figure 2This is a flowchart illustrating the quantitative evaluation method for the service performance of the running gear of a rail vehicle that considers lifespan distribution, as described in this invention.

[0023] Figure 3 This is a schematic diagram of the component life distribution model fitting process involved in this invention, illustrating the process of component data distribution type identification, Weibull / exponential distribution parameter solving, and model optimization.

[0024] Figure 4 This is a schematic diagram of the reliability threshold partitioning model involved in this invention, illustrating the principles for dividing the safe zone, warning zone, and danger zone based on reliability R(t).

[0025] Figure 5 This is a schematic diagram of the hierarchical Bayesian network structure involved in the present invention, showing the causal relationships between the running section, the load-bearing and guiding system, the traction drive system, the braking system, and their key component nodes.

[0026] Figure 6 This diagram illustrates the relationship between health calculation and mapping function involved in this invention, showing the weight structure of reliability, posterior failure probability, and state quantity mapping function in health calculation.

[0027] Figure 7 This is a schematic diagram of the health measurement process involved in this invention. Detailed Implementation

[0028] In the following description, only certain exemplary embodiments are briefly described. As those skilled in the art will recognize, the described embodiments can be modified in various ways without departing from the spirit or scope of the embodiments of the invention. Therefore, the drawings and description are considered to be exemplary in nature and not restrictive.

[0029] The following disclosure provides many different implementations or examples for carrying out different structures of the embodiments of the present invention. To simplify the disclosure of the embodiments of the present invention, specific examples of components and arrangements are described below. Of course, these are merely examples and are not intended to limit the embodiments of the present invention. Furthermore, reference numerals and / or reference letters may be repeated in different examples of the embodiments of the present invention; such repetition is for simplification and clarity and does not in itself indicate a relationship between the various implementations and / or arrangements discussed.

[0030] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0031] like Figure 1 As shown, this embodiment provides a method for quantitatively evaluating the service performance of the running gear of a rail vehicle considering life distribution, including: Acquire historical fault records and failure time data of the running gear components, and classify the components according to their failure mechanisms; acquire multi-source status monitoring data during the operation of the running gear through the sensor acquisition module; For different types of components, corresponding lifetime distribution models are selected. By minimizing the mean square error between the empirical cumulative distribution and the theoretical distribution and maximizing the coefficient of determination, the adaptive selection of lifetime models and parameter estimation are achieved, and the time-varying reliability and failure probability of the components are obtained. A hierarchical Bayesian network consistent with the structure of the running section is constructed, and the hierarchical relationship between the root node, intermediate nodes and leaf nodes is defined. The prior fault probability of the leaf node is the failure probability. When the sample is insufficient, the judgment matrix is ​​constructed by the analytic hierarchy process to calculate the node weight. The intermediate nodes adopt the weighted OR gate logic and the posterior fault probability of each layer node is calculated by combining the variable elimination algorithm. Based on component time-varying reliability, posterior failure probability, and multi-source condition monitoring data, a component health index with unified dimensions is constructed. Weights are set according to the structural hierarchy of the running gear, and the health indices of each component are integrated from bottom to top to obtain the system service performance index. Grading standards are set according to the values ​​of the system service performance index to achieve dynamic quantitative grading and trend prediction of the running gear service performance.

[0032] In some embodiments, the components are classified based on differences in failure mechanisms, into service life-type components and full-life-type components.

[0033] In some embodiments, the lifetime distribution model includes a Weibull distribution model and an exponential distribution model. The Weibull distribution model is used for life-use components, and the exponential distribution model is used for full-life-use components. The parameters of the Weibull distribution model are estimated by the least squares method, and the parameters of the exponential distribution model are estimated by the maximum likelihood estimation method.

[0034] In some embodiments, the root node of the hierarchical Bayesian network is the running gear, the intermediate nodes include the load-bearing and guiding system, the traction drive system, and the braking system, and the leaf nodes include wheelsets, axle boxes, suspension devices, and gearboxes. The relationship between the nodes is consistent with the causal relationship of the actual structure of the running gear.

[0035] In some embodiments, the analytic hierarchy process (AHP) constructs a judgment matrix using a 1-9 scale, where the scale represents the relative importance of different nodes, and the node weights are obtained by calculating the eigenvectors of the judgment matrix.

[0036] In some embodiments, the variable elimination algorithm is used for probabilistic inference of hierarchical Bayesian networks. By eliminating irrelevant node variables, the posterior fault probability of each layer node is calculated, thereby realizing the hierarchical propagation of fault probability.

[0037] In some embodiments, the health index is constructed by integrating the time-varying reliability of components, the posterior failure probability, and the standardized multi-source state monitoring quantity. The weight coefficients of the three are dynamically adjusted according to the operation stage of the rail vehicle, and the sum of the weight coefficients is 1.

[0038] In some embodiments, multi-source condition monitoring quantities are standardized to map them into health-related indicators. The multi-source condition monitoring quantities include vibration monitoring data and temperature monitoring data. The standardization process converts various types of monitoring data into quantitative indicators within the same range to reflect the impact of real-time operating status on the health level of components.

[0039] In some embodiments, the system service performance index ranges from 0 to 1, with a larger value indicating better system service performance. The grading criteria include safe zones, warning zones, and danger zones, and the corresponding area level and maintenance recommendations are determined based on the specific value of the system service performance index.

[0040] In some embodiments, the threshold of the grading standard is fine-tuned according to the specific model of the rail vehicle or the environmental conditions of the operating line to adapt to the service performance evaluation requirements under different application scenarios.

[0041] The technical concept of this invention is as follows: like Figure 1 and Figure 2 As shown, based on the failure mechanism, key components of the running gear are divided into service-life type and full-life type. Life models are constructed using Weibull distribution and exponential distribution, respectively. In the model optimization process, the minimum mean square error (MSE) is first used as the primary criterion. Minimizing the MSE ensures that the fitting deviation between the empirical cumulative distribution function and the theoretical distribution function is minimized. When the MSE differences between different models are not significant (e.g., the relative difference is less than 5%), the coefficient of determination R is then used... 2 As a secondary criterion, R is preferred. 2 Larger models are used to further improve the model's fit consistency and stability, thereby achieving reliable adaptive model selection.

[0042] The time-varying reliability and failure probability of the components are obtained; a hierarchical Bayesian network consistent with the structure of the running gear is constructed, and the posterior failure probability of each layer is calculated using weighted OR gate logic (weights determined by AHP) and variable elimination algorithm; a health index with unified dimensions is constructed. (in , w 1. w 2. w 3 represents the weighting coefficient, with the preferred initial value. , , ),in Standardized mapping functions for multi-source state variables such as vibration and temperature; For reliability function; The system service performance index is obtained by bottom-up fusion, representing the posterior failure probability. (Range [0,1]), and set thresholds: SPI≥0.90 is the safe zone, 0.70≤SPI<0.90 is the warning zone, and SPI<0.70 is the danger zone, supporting dynamic quantitative grading and trend prediction of service performance. The weights are used to determine the levels of service performance. This invention enables hierarchical quantitative assessment and risk classification of service performance from components to systems, supporting predictive maintenance.

[0043] The first aspect: Construction of the lifespan model and adaptive parameter estimation Historical fault records and failure time data of key components in the running gear were obtained. Based on the failure mechanism, these components were classified into service-life-dependent components (such as wheelsets and axle boxes) and life-cycle-dependent components (such as gearboxes). The former are mainly affected by wear or fatigue accumulation, while the latter exhibit memoryless characteristics and an approximately constant failure rate during service. For different component categories, Weibull life distribution models and exponential distribution models were used respectively, and the mean square error (MSE) and coefficient of determination (R²) between the empirical and theoretical distribution functions were used as the basis for the classification. 2 Using this as the selection criterion, the optimal lifetime distribution model and its parameters are determined, enabling adaptive identification of the component-level reliability model.

[0044] The second aspect: Reliability analysis and adaptive construction of lifetime models: For life-cycle components, a two-parameter or three-parameter Weibull distribution is used; for full-life-cycle components, an exponential distribution is used; this is achieved by minimizing the empirical cumulative distribution function. With the theoretical cumulative distribution function The mean squared error (MSE), combined with the coefficient of determination (R²), 2 Model optimization and parameter estimation are performed to obtain the time-varying reliability of the component. With failure probability .

[0045] Third aspect: Construction of hierarchical Bayesian network model for the traversing part A fault tree model consistent with the running gear structure is established, with the running gear as the root node, the load-bearing and guidance system, traction drive system, and braking system as intermediate nodes, and wheelsets, axle boxes, suspension devices, etc., as leaf nodes. The prior fault probability of the leaf nodes is calculated in step 2, and the intermediate nodes use weighted OR gate logic. ; Among them, weight When the sample size is insufficient, the Analytic Hierarchy Process (AHP) is used to determine the eigenvectors. A judgment matrix is ​​constructed and eigenvectors are calculated using a 1-9 scale (1 indicates equal importance, 9 indicates extreme importance, and the reciprocal indicates reverse comparison). The failure probability of child node i is obtained by fitting its lifetime distribution model. The posterior failure probability of parent node j is obtained through the propagation reasoning of the failure probabilities of child nodes in the network.

[0046] Fourth aspect: Integration of health model construction and system performance Based on reliability function Posterior probability of failure In addition to monitoring state indicators such as vibration and temperature, a component health index with unified dimensions is constructed. ; in, This is used to standardize, normalize, and convert multidimensional state parameters into a health index in the [0,1] interval to reflect the impact of real-time operating status on the health level of components. , , The weighting coefficients can be dynamically adjusted based on expert experience or the operational phase of the rail vehicle, and Preferred initial value , and .

[0047] By using weighted fusion to calculate the system-level service performance index, the quantitative transfer of component health to system service performance is achieved.

[0048] ; The SPI range is [0,1], with larger values ​​indicating better system performance. Preferably, a threshold is set to implement hierarchical classification. When the SPI is not less than 0.90, the system is in the safe zone with excellent service performance and can maintain the existing maintenance cycle and continue to operate normally. When the SPI is greater than or equal to 0.70 and less than 0.90, the system is in the warning zone and its service performance begins to degrade. It is recommended to strengthen condition monitoring, appropriately shorten the preventive maintenance cycle, or arrange key inspections. When the SPI is less than 0.70, the system is in the danger zone with a high risk of failure. It is recommended to immediately arrange key maintenance or limit the speed of operation until the SPI recovers to above 0.90.

[0049] By adopting the above technical solution, the present invention realizes hierarchical modeling, quantitative evaluation and adaptive fusion of service performance at the component level, subsystem level and system level of the running gear. It can simultaneously take into account the differentiated failure mechanisms of different components and the service performance change law of the overall system, thereby improving the accuracy, real-time performance and engineering application value of running gear service performance evaluation.

[0050] In one specific embodiment: Step 1: Fault Data Acquisition and Preprocessing Failure times and fault data for key components such as wheelsets, axle boxes, suspension systems, braking systems, gearboxes, and couplings are acquired through the rail vehicle operation and maintenance management system, historical maintenance records, and sensor acquisition modules. A timestamp parsing and regular expression cleaning algorithm is used to unify the data format, and the service days for each sample are calculated using a fixed base date. Records are automatically categorized into the corresponding component datasets based on fault description keyword matching rules.

[0051] Time is defined as: ; T j For the first j The time difference between the secondary failure or malfunction and the base date. Based on the base date, For the first j The moment when the fault or failure occurs.

[0052] Sort by time in ascending order Cumulative distribution of experience: ; The empirical cumulative distribution function; This represents the value of the empirical cumulative distribution function corresponding to the j-th sample, which is the proportion of observations in the sample that are less than or equal to the value at time point j. .

[0053] Based on differences in their service mechanisms, components are classified into: service-life components (primarily failing due to wear, fatigue, or aging) and life-cycle components (existing with no memory and a constant failure rate). For each category, service-life components are modeled using the Weibull parameter model, while life-cycle components are modeled using the exponential distribution model. Specifically, wheelsets, axle boxes, couplings, suspension systems, and braking systems are modeled using the Weibull distribution, while gearboxes are modeled using the exponential distribution model.

[0054] The Weibull parametric model and the exponential parametric model are shown below: ; ; in, For scale parameters, For shape parameters, The constant failure rate characterizes the failure probability per unit time of a component with a full lifespan. This is the cumulative failure distribution function of the component modeled using the Weibull distribution, which represents the probability of failure of this type of cumulatively worn component. This is the cumulative distribution function for component failures using an exponential distribution.

[0055] For the Weibull parameter model, the least squares method is used, which essentially minimizes the sum of squared errors to obtain the optimal function for data matching. For the exponential distribution model, the memoryless maximum likelihood estimation method is used.

[0056] By using the least squares method, a set of parameters θ is found that makes the theoretical distribution curve best approximate the empirical distribution curve. The CDF curve represents the empirical cumulative distribution. Compared with theoretical cumulative distribution The sum of squared errors between them is minimized: ; in, For empirical cumulative distribution, This represents the theoretical cumulative distribution.

[0057] Calculate the mean square error (MSE) and coefficient of determination (R²) between the empirical cumulative distribution function (CDF) and the theoretical distribution function. 2 ), and with the minimum MSE and R 2 The optimal lifetime model and its parameters are determined by using the maximum as the selection criterion.

[0058] ; ; Step 3: Reliability and Failure Probability Calculation. Based on the selected life distribution model, calculate the reliability and failure probability of each component during its operating time. t Reliability function and failure probability under the following conditions : ; Simultaneously set a reliability partitioning threshold: Safe Zone: ; Warning zone: ; Danger Zone: ; Reliability curves and comparison charts are plotted using time-series calculations to analyze the dynamic trends in component service performance. The component lifespan distribution model fitting process is as follows: Figure 3 As shown. The reliability threshold partitioning model is as follows. Figure 4 As shown.

[0059] Step 4: Bayesian Network Hierarchical Modeling and Inference A fault tree model of the running gear is constructed. The purpose of constructing the fault tree is to analyze the causes and influencing factors of the top-level events. The analysis proceeds layer by layer from top to bottom to identify all possible causes, i.e., all bottom-level events. The fault tree model of the running gear is constructed as follows: Figure 5 As shown.

[0060] Fault trees provide the structural foundation, and Bayesian networks inject probabilistic inference. A hierarchical Bayesian network model consistent with the traverse structure is established, such as... Figure 6 As shown. The running gear is set as the root node, the load-bearing and guiding system, the traction drive system, and the braking system are the intermediate nodes, and the wheelset, axle box, frame, gearbox, and traction device are the leaf nodes.

[0061] Preferably, the prior failure probability of the leaf node is calculated by the lifetime distribution model in steps 2 and 3, i.e., the failure probability of the component. For intermediate nodes, when sample data is insufficient, the expert experience method or the judgment matrix method (AHP) is introduced to compare the importance of the parent nodes in pairs to obtain the weight coefficients between nodes. .

[0062] For example, taking the "bearing system" and its sub-nodes (wheelset assembly, frame, axle box assembly) as an example, the AHP judgment matrix is ​​constructed as shown in Table 1 (using a scale of 1-9 to evaluate relative importance, 1 indicates that the two elements are equally important, 3 indicates that element i is slightly more important than element j, "2, 4" are the intermediate values ​​of adjacent judgments, such as 2 indicating that it is between "slightly important" and "equally important". If the importance of element i to j is 2, then j is 1 / 2 more important than i. The remaining values ​​are 5 indicating that element i is significantly more important than element j, 7 indicating that element i is extremely important than element j, and 9 indicating that element i is strongly more important than element j): Table 1. AHP Judgment Matrix ; By calculating the eigenvectors, the weighting coefficients are obtained: wheelset assembly 0.545, frame 0.167, and axle box assembly 0.318. The consistency check CR < 0.1, meeting the requirements. These coefficients are used to construct the CPD of the weighted OR gate logic and calculate the failure probability of intermediate nodes.

[0063] ; Variable elimination algorithm is used for network inference: ; The posterior failure probability of each node is obtained.

[0064] Preferably, the system can dynamically update prior probabilities and weights according to the operational phase, thereby achieving the integration of data-driven approaches and expert knowledge.

[0065] Step 5: Health Calculation and Mapping Function Design. Based on component reliability... Posterior probability of failure and monitoring status quantities Construct a health index with unified dimensions: ; in, For reliability, For the posterior probability of failure, For state variable mapping function, , , These are the weighting coefficients, and their sum is 1.

[0066] Preferably, Standardize the multi-source monitoring data to the [0,1] interval: Regarding vibration monitoring data, ,in As a safety threshold, the horizontal speed is 0.7 m / s. 2 Longitudinal 1.25m / s 2 Vertical 2.0m / s 2 .

[0067] Regarding temperature monitoring data, ,in For ambient temperature, This is the alarm threshold (axle box 100℃). Based on safety threshold data defined by relevant industry standards, temperature data is as follows: around 80℃: attention / warning; around 90℃: alarm; ≥100℃: danger / must limit speed or stop operation for inspection.

[0068] Multi-source data fusion (where weight) The weight is 0.6. With a weight of 0.4, this function ensures the quantification of degradation levels and supports real-time early warning. Based on expert experience in reliability engineering and rail transit, the following reasonable weight allocation is recommended. These recommendations consider the following factors: The value should be relatively high because R(t) is a long-term reliability index based on historical failure data, reflecting the component life mechanism and playing a core role in predictive maintenance. The probability of failure is moderate because the posterior failure probability is captured by Bayesian networks and updated in real time, but it depends on the quality of evidence. Lower, because This is a short-term supplementary monitoring method, susceptible to noise, and requires effective and efficient pretreatment. Taking all the above considerations into account, and based on expert experience, the following approach was adopted: , and This method ensures long-term reliability and is suitable for quantitative assessment scenarios at the running gear level of rail vehicles. The health measurement process is as follows: Figure 7 As shown.

[0069] Step 6: Hierarchical integration and service performance evaluation.

[0070] Weights are set based on structural hierarchy and equivalent failure rate. The system integrates component health data from the bottom up and calculates the system service performance index. Characterizes the distribution of system health and the time evolution trend of SPI.

[0071] ; Preferably, the System Service Performance Index (SPI) ranges from [0,1], and its hierarchical quantitative evaluation criteria are as follows: When the SPI is not less than 0.90, the system is in the safe zone with excellent service performance and can maintain the existing maintenance cycle and continue to operate normally. When the SPI is greater than or equal to 0.70 and less than 0.90, the system is in the warning zone and its service performance begins to degrade. It is recommended to strengthen condition monitoring, appropriately shorten the preventive maintenance cycle, or arrange key inspections. When the SPI is less than 0.70, the system is in the danger zone with a high risk of failure. It is recommended to immediately arrange key maintenance or limit the speed of operation until the SPI recovers to above 0.90.

[0072] The thresholds of 0.90 and 0.70 are determined with reference to the 90% and 70% reliability margins commonly used in rail vehicle reliability engineering, and have been verified in a large amount of simulation and actual application data. In practical applications, the thresholds can be fine-tuned by ±0.05 according to the specific vehicle model (such as the CR400 series) or track conditions (such as high temperature and high humidity environment).

[0073] The embodiments described above are for illustrative purposes only and are not intended to limit the invention. Therefore, any changes in numerical values ​​or substitutions of equivalent elements should still fall within the scope of this invention.

[0074] The above detailed description will enable those skilled in the art to understand that the present invention can indeed achieve the aforementioned objectives and has complied with the provisions of the Patent Law.

[0075] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the invention. The above descriptions are merely preferred embodiments of the invention and are not intended to limit the invention. It should be noted that any modifications, equivalent substitutions, and improvements made within the spirit and principles of the invention should be included within the scope of protection of the invention.

[0076] It should be noted that the above description of the process is for illustrative purposes only and does not limit the scope of this specification. Those skilled in the art can make various modifications and changes to the process under the guidance of this specification. However, these modifications and changes remain within the scope of this specification.

[0077] The basic concepts have been described above. Obviously, for those skilled in the art who have read this application, the above disclosure is merely illustrative and does not constitute a limitation of this application. Although not explicitly stated herein, those skilled in the art may make various modifications, improvements, and corrections to this application. Such modifications, improvements, and corrections are suggested in this application, and therefore, such modifications, improvements, and corrections still fall within the spirit and scope of the exemplary embodiments of this application.

[0078] Furthermore, this application uses specific terms to describe its embodiments. For example, "an embodiment," "one embodiment," and / or "some embodiments" refer to a particular feature, structure, or characteristic related to at least one embodiment of this application. Therefore, it should be emphasized and noted that "an embodiment," "one embodiment," or "an alternative embodiment" mentioned twice or more in different positions in this specification do not necessarily refer to the same embodiment. In addition, certain features, structures, or characteristics in one or more embodiments of this application can be appropriately combined.

[0079] Furthermore, those skilled in the art will understand that aspects of this application can be described and illustrated through several patentable types or situations, including any new and useful combination of processes, machines, products, or substances, or any new and useful improvements thereof. Therefore, aspects of this application can be implemented entirely in hardware, entirely in software (including firmware, resident software, microcode, etc.), or a combination of hardware and software. All of the above hardware or software can be referred to as a “unit,” “module,” or “system.” Furthermore, aspects of this application can take the form of a computer program product embodied in one or more computer-readable media, wherein computer-readable program code is contained therein.

[0080] The computer program code required for the operation of each part of this application can be written in any one or more programming languages, including object-oriented programming languages ​​such as Java, Scala, Smalltalk, Eiffel, JADE, Emerald, C++, C#, VB.NET, and Python; general programming languages ​​such as C; Visual Basic, Fortran2103, Perl, COBOL2102, PHP, and ABAP; dynamic programming languages ​​such as Python, Ruby, and Groovy; or other programming languages. This program code can run entirely on the user's computer, or as a standalone software package on the user's computer, or partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the latter case, the remote computer can be connected to the user's computer via any network, such as a local area network (LAN) or wide area network (WAN), or connected to an external computer (e.g., via the Internet), or in a cloud computing environment, or used as a service such as Software as a Service (SaaS).

[0081] Furthermore, unless expressly stated in the claims, the order of processing elements and sequences, the use of numbers and letters, or other names described in this application are not intended to limit the order of the processes and methods of this application. Although some currently considered useful embodiments of the invention have been discussed in the foregoing disclosure by way of various examples, it should be understood that such details are for illustrative purposes only, and the appended claims are not limited to the disclosed embodiments; rather, the claims are intended to cover all modifications and equivalent combinations that conform to the substance and scope of the embodiments of this application. For example, although the implementation of the various components described above can be embodied in a hardware device, it can also be implemented as a purely software solution, such as an installation on an existing server or mobile device.

[0082] Similarly, it should be noted that, in order to simplify the description of the present application and thus aid in the understanding of one or more embodiments of the invention, the foregoing description of the embodiments of the present application sometimes combines multiple features into a single embodiment, drawing, or description thereof. However, this approach of the present application should not be construed as reflecting an intention that the claimed subject matter requires more features than expressly recited in each claim. Rather, the subject of the invention should possess fewer features than in any single embodiment described above.

Claims

1. A method for quantitatively evaluating the service performance of the running gear of a rail vehicle considering lifespan distribution, characterized in that, include: Acquire historical fault records and failure time data of the running gear components, and classify the components according to their failure mechanisms; acquire multi-source status monitoring data during the operation of the running gear through the sensor acquisition module; For different types of components, corresponding lifetime distribution models are selected. By minimizing the mean square error between the empirical cumulative distribution and the theoretical distribution and maximizing the coefficient of determination, the adaptive selection of lifetime models and parameter estimation are achieved, and the time-varying reliability and failure probability of the components are obtained. A hierarchical Bayesian network consistent with the structure of the running section is constructed, and the hierarchical relationship between the root node, intermediate nodes and leaf nodes is defined. The prior fault probability of the leaf node is the failure probability. When the sample is insufficient, the judgment matrix is ​​constructed by the analytic hierarchy process to calculate the node weight. The intermediate nodes adopt the weighted OR gate logic and the posterior fault probability of each layer node is calculated by combining the variable elimination algorithm. Based on component time-varying reliability, posterior failure probability, and multi-source state monitoring data, a component health index with unified dimensions is constructed. Weights are set according to the structural hierarchy of the running gear, and the health indices of each component are integrated from bottom to top to obtain the system service performance index. Grading standards are set according to the values ​​of the system service performance index to achieve dynamic quantitative grading and trend prediction of the running gear service performance.

2. The method for quantitatively evaluating the service performance of rail vehicle running gear considering life distribution according to claim 1, characterized in that, The components are classified based on differences in failure mechanisms, into service life-type components and full-life-type components.

3. The method for quantitatively evaluating the service performance of rail vehicle running gear considering life distribution according to claim 2, characterized in that, The lifetime distribution model includes the Weibull distribution model and the exponential distribution model. The Weibull distribution model is used for life-use components, and the exponential distribution model is used for full-life-use components. The parameters of the Weibull distribution model are estimated by the least squares method, and the parameters of the exponential distribution model are estimated by the maximum likelihood estimation method.

4. The method for quantitatively evaluating the service performance of rail vehicle running gear considering life distribution according to claim 1, characterized in that, The root node of the hierarchical Bayesian network is the running gear, the intermediate nodes include the load-bearing and guiding system, the traction drive system, and the braking system, and the leaf nodes include wheelsets, axle boxes, suspension devices, and gearboxes. The relationships between the nodes are consistent with the causal relationships of the actual structure of the running gear.

5. The method for quantitatively evaluating the service performance of rail vehicle running gear considering life distribution according to claim 1, characterized in that, The analytic hierarchy process (AHP) uses a 1-9 scale to construct a judgment matrix. The scale is used to represent the relative importance between different nodes, and the node weights are obtained by calculating the eigenvectors of the judgment matrix.

6. The method for quantitatively evaluating the service performance of rail vehicle running gear considering life distribution according to claim 1, characterized in that, The variable elimination algorithm is used for probabilistic inference in hierarchical Bayesian networks. By eliminating irrelevant node variables, it calculates the posterior fault probability of each node and realizes the hierarchical propagation of fault probability.

7. The method for quantitatively evaluating the service performance of rail vehicle running gear considering life distribution according to claim 1, characterized in that, The health index is constructed by integrating the time-varying reliability of components, the posterior failure probability, and the standardized multi-source state monitoring quantities. The weight coefficients of the three are dynamically adjusted according to the operation stage of the rail vehicle, and the sum of the weight coefficients is 1.

8. The method for quantitatively evaluating the service performance of rail vehicle running gear considering life distribution according to claim 1, characterized in that, Multi-source condition monitoring data are standardized to be mapped to health-related indicators. These multi-source condition monitoring data include vibration monitoring data and temperature monitoring data. The standardization process converts various types of monitoring data into quantitative indicators within the same range to reflect the impact of real-time operating status on the health level of components.

9. The method for quantitatively evaluating the service performance of rail vehicle running gear considering life distribution according to claim 1, characterized in that, The system service performance index ranges from 0 to 1, with a higher value indicating better system service performance. The grading criteria include safe zones, warning zones, and danger zones. The corresponding area level and maintenance recommendations are determined based on the specific value of the system service performance index.

10. The method for quantitatively evaluating the service performance of a rail vehicle running gear considering life distribution according to claim 9, characterized in that, The threshold values ​​of the grading standard are fine-tuned based on the specific model of the rail vehicle or the environmental conditions of the operating line to adapt to the service performance evaluation requirements under different application scenarios.

Images