Artificial Intelligence-Based Train Brake Performance Evaluation Method and System

Abstract

This invention provides an artificial intelligence-based method and system for evaluating train brake performance. First, it acquires a dataset of data related to the entire lifecycle of the train brake, encompassing operating parameters, environmental impacts, and historical performance records. Next, it extracts the dynamic relationships between brake components, constructing an operational correlation chain with a timeline and evolving correlation strength across operating stages. Based on the evolution trajectory of this correlation chain, it generates dynamic benchmark features for brake performance, including component correlations and cross-stage collaborative evolution characteristics. A pre-trained artificial intelligence model is then invoked to perform cross-dimensional interactive mapping between real-time operating features and dynamic benchmark features, generating performance evolution evaluation results. Finally, based on the evaluation results and correlation chain trends, it generates a dynamic performance adjustment scheme that includes phased parameter adjustments and cross-stage collaborative optimization suggestions. This invention achieves accurate evaluation and dynamic optimization of train brake performance.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and more specifically, to a method and system for evaluating the performance of train brakes based on artificial intelligence. Background Technology

[0002] In the railway transportation sector, train brakes are a critical component ensuring safe train operation, and their performance stability and reliability are of paramount importance. Traditional methods for evaluating train brake performance mainly rely on periodic manual inspections and static data analysis. Manual inspections require specialized technicians to check each component of the brake one by one, which is not only inefficient but also fails to cover the entire operating cycle of the brake, easily overlooking potential faults.

[0003] In terms of static data analysis, existing technologies typically only perform simple analyses of the operating parameters of the brake at specific times or under specific operating conditions, lacking a comprehensive consideration of the brake's entire lifecycle operating data. Furthermore, existing methods do not fully consider the dynamic interrelationships between the various components of the brake, as well as the changing characteristics of these relationships across different operating stages. For example, under different climatic conditions, operating speeds, and track conditions, the performance and interactions of the various brake components will change, but traditional methods cannot accurately capture these changes. This leads to inaccurate and incomplete assessments of brake performance, making it difficult to predict the evolution trend of brake performance in advance and to take timely and effective maintenance and optimization measures, thereby increasing the safety risks and maintenance costs of train operation. Summary of the Invention

[0004] In view of the aforementioned problems, and in conjunction with the first aspect of the present invention, embodiments of the present invention provide an artificial intelligence-based method for evaluating the performance of train brakes, the method comprising:

[0005] Obtain a set of correlation data for the entire life cycle operation of the train brake, which includes operating parameter data, environmental impact data, and historical performance evolution records of various components of the brake at different operating stages;

[0006] The dynamic correlation between brake components is extracted from the full-cycle operation correlation data set of the train brake, and a brake operation correlation chain is constructed. The brake operation correlation chain connects the correlation relationships of each component with the time axis as the clue, and shows the correlation strength evolution characteristics as the operation stage changes.

[0007] Based on the evolution trajectory of the brake operation correlation chain, dynamic benchmark features of brake performance are generated. These dynamic benchmark features of brake performance include component correlation features at each operation stage and cross-stage collaborative evolution features.

[0008] The pre-trained artificial intelligence model for dynamic evaluation of brake performance is invoked to perform cross-dimensional interactive mapping processing between the real-time collected brake operation characteristics and the dynamic benchmark characteristics of brake performance, thereby generating brake performance evolution evaluation results.

[0009] Based on the brake performance evolution evaluation results and the correlation chain evolution trend, a dynamic adjustment scheme for brake performance is generated. The dynamic adjustment scheme for brake performance includes the direction of component parameter adjustment in stages and cross-stage collaborative optimization suggestions.

[0010] In another aspect, embodiments of the present invention also provide an artificial intelligence-based train brake performance evaluation system, including a processor and a machine-readable storage medium connected to the processor. The machine-readable storage medium is used to store programs, instructions, or code, and the processor is used to execute the programs, instructions, or code in the machine-readable storage medium to implement the above-described method.

[0011] Based on the above, by acquiring a full-cycle operational correlation data set of the train brake, the dynamic correlation relationships between brake components are extracted from this data set, and a brake operation correlation chain is constructed. Using a timeline as a guide, the evolution characteristics of the correlation strength of each component with different operational stages are clearly presented, enabling in-depth insight into the complex internal operating mechanism of the brake. Based on the evolution trajectory of the operational correlation chain, dynamic benchmark features of brake performance are generated, including component correlation features at each operational stage and cross-stage collaborative evolution features. A pre-trained artificial intelligence model is invoked to perform cross-dimensional interactive mapping processing between the real-time collected brake operation features and the dynamic benchmark features, generating accurate brake performance evolution evaluation results, and enabling timely detection of brake performance change trends. Finally, based on the evaluation results and the correlation chain evolution trend, a dynamic performance adjustment scheme is generated, including staged component parameter adjustment directions and cross-stage collaborative optimization suggestions. This achieves accurate evaluation and dynamic optimization of train brake performance, effectively improving the safety and reliability of train operation and reducing maintenance costs. Attached Figure Description

[0012] Figure 1 This is a schematic diagram of the execution flow of the train brake performance evaluation method based on artificial intelligence provided in an embodiment of the present invention.

[0013] Figure 2 This is a schematic diagram of exemplary hardware and software components of the train brake performance evaluation system based on artificial intelligence provided in an embodiment of the present invention. Detailed Implementation

[0014] The present invention will now be described in detail with reference to the accompanying drawings. Figure 1This is a flowchart illustrating an artificial intelligence-based train brake performance evaluation method according to an embodiment of the present invention. The following is a detailed description of the artificial intelligence-based train brake performance evaluation method.

[0015] Step S110: Obtain the train brake full-cycle operation association data set, which includes the operating parameter data, environmental impact data and historical performance evolution record data generated by each component of the brake at different operating stages.

[0016] In this embodiment, a full-cycle operational data set is obtained using the air brake system of a passenger train. Brake system components include air compressors, air reservoirs, brake valves, brake cylinders, release valves, and brake hoses. Different operational phases include the start-up phase, stable operation phase, braking phase, and stopping phase, with the time span of each phase determined according to the train operation plan. Operational parameter data includes indicators such as pressure, flow rate, temperature, operating frequency, and response time for each component; environmental impact data includes external temperature, humidity, atmospheric pressure, and vibration intensity during train operation; historical performance evolution records contain information such as performance status ratings, maintenance times, types of replaced components, and performance degradation trends for each component over multiple past operational cycles.

[0017] For example, step S111: Deploy a full-cycle data acquisition device covering all components of the brake. The full-cycle data acquisition device includes a parameter acquisition module, a stage identification module, and a timestamp marking module. The parameter acquisition module collects the operating parameter data of each component in real time. The stage identification module identifies the operating stage according to the changes in the brake's operating conditions. The timestamp marking module adds timestamps to the collected data.

[0018] Pressure and flow sensors are installed at the output of the air compressor of the brake system; temperature and pressure sensors are installed on the outer wall of the air reservoir; displacement sensors and actuation frequency counters are installed on the brake valve and release valve; pressure sensors and response time timers are installed inside the brake cylinder; and vibration sensors are installed at the brake hose connections. These sensors together constitute a parameter acquisition module, which collects real-time operating parameter data for the corresponding components. The stage identification module identifies the operating stage by receiving signals such as traction, braking, and stopping commands from the train, combined with changes in the parameters of various brake components. For example, when a traction command is received and the air compressor pressure begins to rise while the brake valve is closed, it is determined to be the starting stage; when the train speed stabilizes within the set range and the parameters of each component do not fluctuate drastically, it is determined to be the stable operation stage; when a braking command is received, the brake valve opens and the brake cylinder pressure rises, it is determined to be the braking stage; and when the train speed drops to zero and all braking components are stationary, it is determined to be the stopping stage. The timestamp marking module uses a clock signal synchronized with the train dispatching system to add a timestamp accurate to milliseconds to each collected operating parameter data, ensuring the temporal correlation of the data.

[0019] Step S112: Connect to the train environment monitoring system and obtain environmental impact data during the operation of the brake from the train environment monitoring system. The environmental impact data includes temperature data, humidity data, vibration data and air pressure data for each operating stage. At the same time, add corresponding stage identifiers and timestamps to the environmental impact data.

[0020] The full-cycle data acquisition device is connected to the train environmental monitoring system via an Ethernet interface, and environmental impact data is acquired from the system according to a preset communication protocol (such as TCP / IP). The train environmental monitoring system has environmental sensors installed at the front, rear, and middle of the train to collect environmental data from different locations. For each stage of brake operation, temperature, humidity, vibration, and air pressure data for the corresponding time period are selected from the environmental monitoring system. For example, during the startup phase, environmental data from the start to the end of this phase is acquired, and these data are marked with a "Startup Phase" identifier and a timestamp consistent with the operating parameter data format to ensure the time correspondence between environmental impact data and operating parameter data.

[0021] Step S113: Access the historical performance management database of the train brake and retrieve the historical performance evolution record data of each component of the brake from the historical performance management database of the train brake. The historical performance evolution record data includes the performance status record, maintenance record and performance change trend description of each component in different historical periods and at each operating stage.

[0022] The system connects to the historical performance management database of the train brake system via a database access interface (such as an ODBC interface). This database uses a relational database structure and stores historical data categorized by component type, operating cycle, and operating stage. When retrieving data, based on the component model and number of the current brake system, it filters out performance status records (such as ratings of excellent, good, average, and poor) for each component during the startup, stable operation, braking, and shutdown phases over multiple past operating cycles. It also retrieves maintenance records (such as maintenance date, maintenance content, and maintenance personnel) and descriptions of performance change trends (such as pressure decay rate and response time extension trends). For example, it retrieves the pressure output stability rating, monthly maintenance records, and textual descriptions of pressure decay trends for each stage of the air compressor over the past 12 operating cycles.

[0023] Step S114: Perform stage alignment processing on the collected operating parameter data, environmental impact data, and historical performance evolution record data. Based on the stage identifier and timestamp, match the data from different sources to the corresponding operating stage to form a stage data combination.

[0024] The system extracts operational phase identifiers and timestamps from operational parameter data and environmental impact data, and performs data matching on a phase-by-phase basis. For example, for the startup phase, it filters out all operational parameter data such as air compressor pressure and air receiver temperature data with timestamps between the start and end of the phase, along with environmental impact data such as ambient temperature and humidity for the corresponding time period. This data is then correlated with historical performance evolution records of components such as the air compressor and air receiver during that phase, forming the startup phase data set. Similarly, data sets are formed for the stable operation phase, braking phase, and shutdown phase. During the matching process, if a data point's timestamp falls at the boundary between two phases, it is assigned to the phase based on the operating condition change time point determined by the phase identification module, ensuring that each data point is accurately matched to its corresponding operational phase.

[0025] Step S115: Summarize the stage data to form a train brake full-cycle operation correlation data set containing operating parameter data, environmental impact data, and historical performance evolution record data generated by each component of the brake at different operating stages.

[0026] Data from the startup, stable operation, braking, and shutdown phases are aggregated chronologically to construct a structured dataset. This dataset is stored in JSON format, with each operational phase as a primary node, each component as a secondary node, and operational parameter data, environmental impact data, and historical performance evolution records as tertiary nodes. For example, the startup phase node includes secondary nodes such as air compressors and air receivers; the air compressor node includes nodes for operational parameter data (pressure, flow rate, etc.), corresponding environmental impact data, and historical performance evolution records. After aggregation, the dataset undergoes an integrity check. If any data of a certain type is missing from a component in a particular phase, the missing location is marked and the missing type is recorded for subsequent processing.

[0027] Step S120: Extract the dynamic correlation between brake components from the train brake full-cycle operation correlation data set, construct the brake operation correlation chain, the brake operation correlation chain connects the correlation relationships of each component with the time axis as the clue, and shows the correlation strength evolution characteristics as the operation stage changes.

[0028] Based on the aforementioned dataset of train brake system's full-cycle operation correlations, the interrelationships between the operating parameters of various components are analyzed to extract dynamic correlations. For example, changes in air compressor pressure data affect air reservoir pressure data, which in turn affects the brake valve's operating state, which in turn affects the brake cylinder's pressure data. These interrelationships constitute dynamic correlations. When constructing the brake system's operational correlation chain, the time axis is used as the horizontal main line, arranging the component correlations at different operational stages in chronological order, while also reflecting the changes in the strength of correlations at different stages. For instance, during the braking stage, the correlation strength between the brake valve and the brake cylinder is higher than during the stable operation stage.

[0029] Step S121: Divide the train brake full-cycle operation associated data set into operation stages. According to the brake's start-up stage, stable operation stage, braking stage, and stop stage, divide the data into stage data units corresponding to each stage. Each stage data unit contains the operating parameter data, environmental impact data, and performance record data of each component in that stage.

[0030] Based on the start and end times of each operating stage determined by the stage identification module in step S111, the data set associated with the entire cycle operation of the train brake is divided. All data belonging to the start-up stage time range is extracted to form a start-up stage data unit; similarly, data for the stable operation stage, braking stage, and stopping stage are extracted to form corresponding stage data units. Within each stage data unit, operating parameter data is organized by component category, and it also includes environmental impact data for that stage, as well as performance record data for each component during that stage (such as performance status rating, whether any abnormalities occurred, etc.). For example, the start-up stage data unit includes the pressure change curve data and flow data of the air compressor during the start-up stage, the temperature and pressure data of the air reservoir, the ambient temperature and humidity data for that stage, and the performance record data of the air compressor and air reservoir during that stage.

[0031] Step S122: For each stage data unit, extract the parameter correlation between any two components within the corresponding stage, record the frequency of occurrence, duration and behavior of the correlation, and form a set of correlations within the stage.

[0032] Taking the start-up phase data unit as an example, we select the air compressor and the air receiver cylinder as two components to analyze the relationship between their pressure parameters. When the air compressor pressure rises, the air receiver cylinder pressure rises accordingly, indicating a parameter correlation between the two. The frequency of this correlation during the start-up phase is recorded; the duration is the time interval from the start to the end of the correlation; and the correlation is described as the correspondence between the rate of change of air compressor pressure and the rate of change of air receiver cylinder pressure. Following this method, we analyze the parameter correlations between all pairs of components during the start-up phase, forming a correlation set for the start-up phase. Using the same method, we form correlation sets for the stable operation phase, braking phase, and shutdown phase, respectively.

[0033] Step S123: Compare the intra-stage association sets of adjacent stage data units, identify the addition, disappearance and intensity changes of associations, and form inter-stage association evolution information. The inter-stage association evolution information includes the type of association change and the triggering conditions for the change.

[0034] The system compares the sets of intra-stage correlations between the start-up phase and the stable operation phase. If a parameter correlation between the brake valve and the release valve, which was not present in the start-up phase, appears in the stable operation phase, it is considered a newly added correlation. If the high-frequency correlation between the air compressor and the air reservoir, which existed in the start-up phase, decreases significantly in frequency during the stable operation phase, it is considered a change in intensity. If a correlation exists in the start-up phase but is completely absent in the stable operation phase, it is considered a lost correlation. The triggering conditions for changes are determined based on the changes in operating conditions and parameters in adjacent phases. For example, the triggering condition for a newly added correlation between the brake valve and the release valve is that the operating frequency of the brake valve stabilizes within a specific range after the train enters the stable operation phase. Following this method, the sets of intra-stage correlations between the stable operation phase and the braking phase, and between the braking phase and the stopping phase, are compared sequentially to identify various changes and form complete inter-stage correlation evolution information.

[0035] Step S124: Using the time axis as the horizontal thread, arrange the sets of relationships within each stage in chronological order, and at the same time, use the inter-stage relationship evolution information as the vertical connection to link the relationships between adjacent stages and form the initial brake operation relationship chain.

[0036] Using a timeline from left to right to represent chronological order, the sets of relationships within the startup phase, stable operation phase, braking phase, and shutdown phase are sequentially arranged below the timeline. Based on the evolution of relationships between phases, lines connect relationships with inherited connections between adjacent phases. For newly added relationships, a line is drawn from the beginning of the corresponding phase to indicate its appearance; for disappearing relationships, the line terminates at the end of the corresponding phase. For relationships with changing intensity, variations in line thickness indicate increases or decreases in intensity. For example, the relationship line between the air compressor and the air receiver cylinder in the startup phase continues into the stable operation phase, with the line becoming thinner indicating a decrease in intensity; the relationship line between the brake valve and the release valve, newly added in the stable operation phase, is drawn starting from the beginning of the stable operation phase. Through this method, the initial braking mechanism operation relationship chain is formed.

[0037] Step S125: Analyze the intensity change curves of each relationship in the initial brake operation association chain over the entire cycle, mark the peak points, valley points and inflection points in the intensity change curves, determine the association intensity evolution characteristics, optimize the structure of the initial brake operation association chain based on the characteristics, and generate the brake operation association chain.

[0038] For each relationship in the initial brake operation chain, such as the relationship between the air compressor and the air receiver cylinder, the intensity values ​​are extracted during the start-up, stable operation, braking, and stopping phases. An intensity variation curve is plotted with time on the horizontal axis and intensity value on the vertical axis. On this curve, the point with the highest intensity value is identified as the peak point, the point with the lowest intensity value as the trough point, and the point where the direction of intensity change changes as the inflection point. For example, in the initial stage of start-up, the intensity of this relationship rapidly rises to the peak point, then gradually decreases to the trough point during stable operation, and rises slightly again after entering the braking phase; the starting point of this rise is the inflection point. Based on the distribution of these peak points, trough points, and inflection points, the evolution characteristics of the relationship intensity are determined, such as large intensity fluctuations during start-up and stable intensity during stable operation. Based on these characteristics, the initial brake operation chain is optimized. For example, relationships with drastic intensity changes are marked with a special color, and the drawing precision of the lines is adjusted to highlight key nodes, thereby generating the final brake operation chain.

[0039] Step S130: Based on the evolution trajectory of the brake operation correlation chain, generate dynamic benchmark features of brake performance. The dynamic benchmark features of brake performance include component correlation features of each operation stage and cross-stage collaborative evolution features.

[0040] The evolution trajectory of the brake's operational correlation chain reflects the changes in various correlations at different operational stages. By extracting and analyzing features from this trajectory, dynamic benchmark features that reflect the normal performance state of the brake are generated. The component correlation features at each operational stage are the core attribute set of the correlations within that stage, while the cross-stage collaborative evolution features reflect the collaborative change patterns of the correlations between different stages. Together, they constitute the dynamic benchmark features of the brake's performance.

[0041] Step S131: Traverse the time axis of the brake operation association chain, extract the association relationship set and association strength evolution characteristics corresponding to each operation stage, and form a stage association data package.

[0042] Following the timeline, the startup phase, stable operation phase, braking phase, and shutdown phase are traversed sequentially. For each operation phase, all relationships within that phase are extracted from the braking mechanism's operational chain to form a relationship set. Simultaneously, the intensity change curves, peak points, trough points, inflection points, and other characteristics of the relationship intensity evolution within that phase are extracted. These relationship sets and intensity evolution characteristics are packaged and stored to form a phase-related data package for each operation phase. For example, the startup phase relationship data package contains all component relationships during the startup phase, as well as the intensity evolution characteristic data of each relationship.

[0043] Step S132: Extract features from the association relationships in the data packets associated with each stage to generate component association features for that stage. The component association features reflect the core performance of the association relationships between components within that stage and their adaptability to the environmental impact data of that stage.

[0044] Taking the startup phase association data package as an example, feature extraction is performed on each association. For the association between the air compressor and the air receiver cylinder, core performance characteristics such as average intensity value, intensity fluctuation amplitude, and peak occurrence time are extracted. Simultaneously, the relationship between the intensity change of this association and the ambient temperature and humidity during startup is analyzed. If the intensity fluctuation amplitude of this association increases when the ambient temperature rises, this adaptability feature is also included. Following the same method, feature extraction is performed on all associations in the startup phase association data package, and the resulting summaries form the component association features for the startup phase. Using the same process, component association features for the stable operation phase, braking phase, and shutdown phase are generated respectively.

[0045] Step S133: Compare and analyze the component association features of adjacent operation stages to identify the inheritance relationship, variation relationship and complementary relationship between features, and form feature association information between stages.

[0046] Compare the component association characteristics between the start-up phase and the stable operation phase. If a component association characteristic in the stable operation phase is essentially consistent with that in the start-up phase, with only slight differences in intensity values, then a relationship of inheritance is determined. If a component association characteristic in the stable operation phase differs significantly from its corresponding characteristic in the start-up phase, such as a change in the intensity variation pattern of the association, then a relationship of variation is determined. If a component association characteristic in the start-up phase and that in the stable operation phase are functionally complementary, jointly ensuring the performance of the brake during phase transitions, then a relationship of complementarity is determined. Record the specific manifestations of these relationships and the types of characteristics involved to form inter-phase feature association information between the start-up phase and the stable operation phase. Following the above method, continue to compare the component association characteristics between the stable operation phase and the braking phase, and between the braking phase and the stopping phase, to form complete inter-phase feature association information.

[0047] Step S134: Based on the inter-stage feature association information, construct a cross-stage co-evolution model. This cross-stage co-evolution model takes the component association features of each stage as input and outputs cross-stage co-evolution features that reflect the continuous change law of features. The cross-stage co-evolution features include the triggering factors of feature changes and the evolution direction.

[0048] The cross-stage co-evolution model employs a multilayer perceptron structure. The input layer contains all dimensions of the component association features at each stage, such as the average strength value of each association, the intensity fluctuation amplitude, and environmental adaptability parameters. Three hidden layers are configured: the first layer performs preliminary nonlinear transformation on the input feature data; the second layer fuses the transformed data; and the third layer further refines the fused features and uncovers the co-evolutionary relationships between features. The output layer outputs cross-stage co-evolutionary features, which include triggering factors for feature changes (such as changes in environmental temperature and component movement frequency) and evolutionary directions (such as increased intensity, decreased intensity, and increased fluctuation amplitude). During model training, historical inter-stage feature association information and corresponding cross-stage evolution results are used as training data. The weight parameters of each layer are adjusted using a backpropagation algorithm until the error between the model's output cross-stage co-evolutionary features and the actual evolution results is within a preset range.

[0049] Step S135: Integrate the component association features and cross-stage co-evolution features of all operating stages in chronological order, label the operating stage identifiers and evolution nodes corresponding to each feature, and form the dynamic benchmark features of brake performance.

[0050] Following the timeline from the start-up phase to the stop phase, the component-related features of each phase are arranged sequentially. Then, cross-phase co-evolution features are inserted into the corresponding phase transition positions. For example, after the component-related features of the start-up phase and before the component-related features of the stable operation phase, cross-phase co-evolution features from the start-up phase to the stable operation phase are inserted. Each component-related feature is labeled with a corresponding operation phase identifier (such as "start-up phase," "stable operation phase," etc.), and evolution nodes are labeled for the cross-phase co-evolution features. These evolution nodes represent the starting time points and key turning points of feature changes. After integration, a dynamic benchmark feature for brake performance is formed. This dynamic benchmark feature for brake performance is stored in sequence and can completely reflect the normal performance evolution law of the brake throughout its entire life cycle.

[0051] Step S140: Call the pre-trained artificial intelligence model for dynamic evaluation of brake performance, and perform cross-dimensional interactive mapping processing on the real-time collected brake operation features and the dynamic benchmark features of brake performance to generate brake performance evolution evaluation results.

[0052] In this embodiment, the pre-trained AI model for dynamic evaluation of brake performance is a sequence prediction model based on the Transformer architecture. This sequence prediction model has been trained using a large amount of historical brake full-cycle operation data, corresponding dynamic benchmark features, and performance evaluation result labels. Real-time collected brake operation features are acquired using the same full-cycle data acquisition device as in step S111, covering parameter association data of each component in the current operation phase, environmental adaptation data, and phase transition data. After inputting the real-time operation features and the brake performance dynamic benchmark features into the model, the model constructs a cross-dimensional mapping relationship between the two through a feature interaction layer, explores the matching and deviation between real-time features and benchmark features, and then integrates and analyzes the results through a result generation layer, outputting an evaluation result that includes the performance status of each component and the cross-phase collaborative status.

[0053] Step S141: Obtain the real-time collected brake operation characteristics, which include the component association characteristics, real-time environment adaptation characteristics, and real-time stage transition characteristics during the real-time operation phase.

[0054] Sensors deployed on various components of the brake system collect real-time parameter data of the current operating state, such as the current pressure output data of the air compressor, the operating frequency data of the brake valve, and the response time data of the brake cylinder. Based on this parameter data, the parameter correlation between any two components in the current operating phase is analyzed, and the component correlation characteristics of the real-time operating phase are extracted. These component correlation characteristics include attributes such as the correlated component pairs, correlation strength, and correlation stability. Real-time environmental adaptation characteristics are generated by combining the current environmental data such as temperature, humidity, and air pressure obtained from the train environmental monitoring system with the degree of adaptation between the current component correlation characteristics and the environmental data. For example, the adaptation of the correlation strength between the brake hose and the brake cylinder when the ambient temperature rises. Real-time phase transition characteristics are generated by analyzing the change law, change rate, and transition triggering conditions of component correlation characteristics during the transition process between the current operating phase and the previous operating phase. For example, the change of correlation characteristics between the brake valve and the air reservoir when transitioning from the stable operating phase to the braking phase.

[0055] Step S142: Input the dynamic benchmark features of brake performance and the real-time collected brake operation features into the feature interaction layer of the pre-trained artificial intelligence model for dynamic evaluation of brake performance.

[0056] The feature interaction layer of the AI ​​model for dynamic evaluation of brake performance includes a feature standardization module, a feature alignment module, and an interaction mapping module. The dynamic benchmark features of brake performance and the real-time acquired brake operation features are first input into the feature standardization module. This module uses a min-max standardization method to transform the data of each dimension of the two types of features to the same data range, eliminating the influence of differences in dimensions. The standardized features then enter the feature alignment module. This module aligns the real-time operation features with the corresponding features in the dynamic benchmark features along the time dimension, based on the feature's operational stage identifier and timestamp, ensuring that features at the same time point can be effectively compared. The aligned features are then input into the interaction mapping module, which calculates the correlation weights between the real-time features and the benchmark features across each dimension using an attention mechanism.

[0057] Step S143: Construct cross-dimensional mapping rules between dynamic baseline features and real-time running features through the feature interaction layer. The cross-dimensional mapping rules include stage matching rules, feature type correspondence rules, and evolution trend comparison rules.

[0058] The interaction mapping module of the feature interaction layer first constructs stage matching rules based on the aligned feature data, then constructs feature type corresponding rules based on feature attributes, and finally constructs evolution trend comparison rules by combining the time series changes of features, forming a complete cross-dimensional mapping rule system. The above rules will serve as the basis for interactive comparison between dynamic benchmark features and real-time running features, ensuring the standardization and accuracy of the comparison process and avoiding evaluation errors caused by feature dimension mismatch or inconsistent comparison logic.

[0059] Step S1431: Analyze the operating stage identifier in the dynamic benchmark characteristics of the brake performance, determine the time range, environmental characteristics and core correlation types of each stage, and construct a stage feature library.

[0060] The system extracts operational stage identifiers from the dynamic baseline characteristics of brake performance, including component correlation features and cross-stage collaborative evolution features, such as "start-up stage" and "stable operation stage." Based on the timestamps corresponding to these stage identifiers, the start and end times of each stage are determined, forming the time range for each stage. Environmental impact data corresponding to each stage is screened from the dynamic baseline characteristics, and the environmental characteristics of each stage are extracted, such as the ambient temperature range for the start-up stage and the vibration intensity range for the stable operation stage. Simultaneously, the frequency and intensity of correlations within each stage are analyzed to determine the core correlation type for each stage. For example, the core correlation for the start-up stage is the pressure correlation between the air compressor and the air receiver cylinder, while the core correlation for the braking stage is the action correlation between the brake valve and the brake cylinder. The time ranges, environmental characteristics, and core correlation types for each stage are categorized and stored by stage, constructing a stage feature library.

[0061] Step S1432: Extract the real-time operation stage information from the real-time collected brake operation features, compare the real-time operation stage information with the stage features in the stage feature library, determine the benchmark stage corresponding to the real-time operation features, and form a stage matching rule. The stage matching rule includes stage time overlap determination, environmental feature similarity determination, and core association relationship type matching determination.

[0062] Real-time operation phase information is extracted from the real-time collected brake operation features, including the start time of the current operation phase, current environmental data, and current core correlation type. The overlap ratio between the time range of the current operation phase and the time ranges of each phase in the phase feature library is calculated; the higher the overlap ratio, the higher the time matching degree. The cosine similarity algorithm is used to calculate the similarity between the current environmental data and the environmental features of each phase in the phase feature library, obtaining the environmental feature similarity. The consistency between the current core correlation type and the core correlation types of each phase in the phase feature library is compared to determine the core correlation type matching. Based on the combined results of these three determinations, when the phase time overlap, environmental feature similarity, and core correlation type matching degree all reach a preset threshold, the phase is determined to be the baseline phase corresponding to the real-time operation feature, thus forming the phase matching rule.

[0063] Step S1433: Type labeling of component association features and cross-stage co-evolution features in dynamic baseline features, classifying feature types into parameter association type, environment adaptation type, and evolution trend type.

[0064] Analyze the attributes of component-related features in the dynamic baseline features. If a feature mainly reflects the parameter interaction relationship between components, such as the correspondence between air compressor pressure and air receiver cylinder pressure, it is labeled as parameter-related. If a feature mainly reflects the adaptation of component relationships to environmental factors, such as the adaptation between brake valve operating frequency and ambient temperature, it is labeled as environmentally adaptable. For cross-stage co-evolution features, if they mainly reflect the changing pattern of related features between different stages, such as the changing trend of brake cylinder pressure related features from the stable operation stage to the braking stage, they are labeled as evolution trend type. According to the above standards, all features in the dynamic baseline features are labeled with a type to ensure that each feature corresponds to a unique feature type.

[0065] Step S1434: Type labeling is performed on the component association features, real-time environment adaptation features, and real-time stage transition features in the real-time operation features to make them consistent with the feature type classification standard of the dynamic benchmark features, forming feature type correspondence rules. The feature type correspondence rules include feature type definition, type identification basis, and feature conversion method when types do not match.

[0066] Referring to the feature type classification standard of dynamic benchmark features, real-time operation features are categorized. Component-related features reflecting the mutual influence of component parameters are categorized as parameter-related, real-time environment adaptation features as environment adaptation, and real-time stage transition features reflecting the feature change patterns during stage transitions are categorized as evolution trend. Based on this, feature type definitions are clarified: parameter-related features reflect the interaction relationship between component parameters, environment adaptation features reflect the relationship and environmental adaptation, and evolution trend features reflect the temporal evolution pattern of features. Simultaneously, the criteria for type identification are determined, namely, type judgment is based on the core attributes and descriptive content of the feature. If a feature in the real-time operation features cannot directly match the above types, it is processed through feature transformation, such as decomposing a feature that simultaneously contains both parameter-related and environment adaptation attributes into two independent features, corresponding to parameter-related and environment adaptation types respectively, thus forming a feature type correspondence rule.

[0067] Step S1435: Analyze the evolution curves of cross-stage co-evolutionary features in the dynamic baseline features, extract the direction of change, rate of change and period of change of the evolution curves, and construct an evolution trend template.

[0068] For each cross-stage co-evolutionary feature in the dynamic baseline features, its corresponding evolution curve data is extracted, with time as the horizontal axis and feature intensity as the vertical axis. The direction of change is determined by analyzing the slope of the evolution curve: a positive slope indicates an upward trend, a negative slope indicates a downward trend, and a zero slope indicates a stationary trend. The rate of change is determined by calculating the change in feature intensity per unit time; a larger change indicates a faster rate of change. The change period is determined by identifying recurring feature intensity fluctuation patterns in the evolution curve; the time interval between two identical fluctuation patterns is the change period. The extracted directions, rates, and periods of change are categorized and organized according to the type of cross-stage co-evolutionary feature to construct evolution trend templates. Each template corresponds to a typical evolutionary pattern of a type of cross-stage co-evolutionary feature.

[0069] Step S1436: Determine the change pattern of the transition features in the real-time operation features, form a real-time evolution trend description, and construct an evolution trend comparison rule. The evolution trend comparison rule includes the consistency judgment of evolution direction, the setting of the deviation range of change rate, and the determination of the overlap of change cycle.

[0070] Analyze the parameter changes of the transition features in the real-time operation characteristics to determine their direction, rate, and period of change, forming a real-time evolution trend description. Compare the real-time evolution trend description with the evolution trend template to determine if their change directions are consistent. If the real-time change direction is the same as the template change direction, the evolution direction consistency is determined to be compliant; otherwise, it is determined to be non-compliant. Set a range for the deviation of the change rate. When the difference between the real-time change rate and the template change rate is within this range, it is determined to be compliant; otherwise, it is determined to be non-compliant. Calculate the overlap ratio between the real-time change period and the template change period, i.e., the change period overlap degree. When the overlap ratio reaches a preset threshold, it is determined to be compliant. Combining the above three judgment criteria, an evolution trend comparison rule is constructed.

[0071] Step S144: Based on the cross-dimensional mapping rule, perform interactive comparison between the component association features of each stage in the dynamic benchmark features, the cross-stage collaborative evolution features, and the corresponding features in the real-time operation features to mine the evolution matching relationship and deviation association relationship between features.

[0072] After determining the baseline stage corresponding to the real-time running features based on the stage matching rules in the cross-dimensional mapping rules, the dynamic baseline features are matched with similar features of the real-time running features according to the feature type matching rules. Then, evolutionary trend features are specifically compared using evolutionary trend comparison rules. During the comparison process, the matching status of feature attributes is recorded. Features with consistent attributes or differences within the allowable range are identified as evolutionary matching relationships. For features with attribute differences exceeding the allowable range, the causes and effects of the deviations are analyzed to identify deviation correlation relationships, thereby completing a comprehensive interactive comparison of the two types of features.

[0073] Step S1441: According to the stage matching rules, the real-time running features are grouped with the dynamic benchmark features of the corresponding benchmark stage to form feature comparison groups. Each feature comparison group contains the component association features of the benchmark stage, the cross-stage co-evolution features, and the corresponding features in the real-time running features.

[0074] Based on the stage matching rules formed in step S1432, if the benchmark stage corresponding to the real-time operating feature is determined to be a stable operating stage, then the component association features of the stable operating stage and the cross-stage co-evolution features between the stable operating stage and adjacent stages are extracted from the dynamic benchmark features. These are then combined with the component association features of the stable operating stage, the real-time environment adaptation features, and the real-time stage transition features in the real-time operating features to form a stable operating stage feature comparison group. Using the same method, if the real-time operating feature corresponds to other benchmark stages, then feature comparison groups for the corresponding stages are formed respectively, ensuring that the benchmark features and real-time features within each feature comparison group belong to the same operating stage or stage transition process.

[0075] Step S1442: Based on the feature type correspondence rule, within each feature comparison group, match the parameter-related features of the dynamic benchmark feature with the parameter-related features of the real-time running feature, match the environment-adaptive features with the environment-adaptive features, and match the evolutionary trend features with the evolutionary trend features.

[0076] Within the stable operation phase feature comparison group, following the feature type correspondence rules, parameter-related component association features (such as the parameter interaction features between the brake valve and the release valve) in the stable operation phase of the dynamic benchmark features are matched with parameter-related component association features in the stable operation phase of the real-time operation features; environmental adaptation component association features (such as the adaptation features of brake hose association strength and humidity) in the stable operation phase of the dynamic benchmark features are matched with real-time environmental adaptation features in the real-time operation features; and cross-phase co-evolutionary trend features between the stable operation phase and the braking phase in the dynamic benchmark features are matched with real-time phase transition features (features transitioning from the stable operation phase to the braking phase) in the real-time operation features. Other feature comparison groups also complete the matching of similar features in the same way.

[0077] Step S1443: Perform attribute comparison on the corresponding type of features. For parameter-related features, compare the related component pairs and related performance. For environmental adaptation features, compare the adaptation performance of environmental factors and related relationships. For evolutionary trend features, compare the direction of change and change nodes.

[0078] For the corresponding group of parameter-related features, compare whether the related component pairs in the dynamic baseline features and real-time operation features are consistent. For example, if both are related to the brake valve and the release valve, then the component pairs are consistent. Then compare the correlation performance, such as whether the response mode of the release valve is the same when the brake valve is activated. For the corresponding group of environmentally adaptable features, compare the adaptation performance of environmental factors (such as temperature and humidity) and the correlation relationship, such as whether the change range of the brake hose correlation strength is consistent under the same humidity conditions. For the corresponding group of evolution trend features, compare whether the direction of change is the same, such as whether the correlation strength increases when transitioning from the stable operation stage to the braking stage. Then compare the change nodes, such as whether the time point when the correlation strength begins to increase is consistent.

[0079] Step S1444: If the attribute comparison result of the corresponding type feature meets the preset matching threshold range, then the feature relationship of the feature comparison group is marked as an evolutionary matching relationship, and the matched feature attributes and matching degree description are recorded.

[0080] The matching thresholds for preset parameter-related features are: the related components are completely identical and the similarity of their related performance reaches a preset ratio; the matching thresholds for environmental adaptation features are: the deviation of environmental factor adaptation performance is within a preset range; and the matching thresholds for evolutionary trend features are: the change direction is consistent and the time difference between the change nodes is within a preset interval. If, in the feature comparison group during the stable operation phase, the parameter-related feature attribute comparison results of the brake valve and the relief valve meet the above-mentioned matching thresholds for parameter-related features, then the feature relationship of this feature comparison group is marked as an evolutionary matching relationship, and the matching attribute is recorded as "related component pairs are identical, and the similarity of related performance meets the requirements," and the matching degree is described as "highly matched."

[0081] Step S1445: If the attribute comparison result of the corresponding type feature exceeds the preset matching threshold range, analyze the stage position, feature type and associated evolution node of the deviation, determine the influence range of the deviation and possible triggering factors, mark the feature relationship of the feature comparison group as the deviation relationship, and record the deviation attribute, deviation degree and deviation triggering factor description.

[0082] If, in the stable operation phase feature comparison group, the comparison results of the brake hose correlation strength and humidity environmental adaptation features show that the change in the brake hose correlation strength in real-time features exceeds the deviation range of the dynamic benchmark features, then the stage where the deviation occurred is determined to be the stable operation phase, and the feature type is environmental adaptation. Analysis shows that the evolution node of this deviation correlation is in the middle of the stable operation phase, and the impact range involves the correlation stability between the brake hose and the brake cylinder. Combining current environmental data and component parameters, the possible triggering factors are determined to be that the actual humidity exceeds the humidity range of the benchmark environment and that the brake hose exhibits slight aging. The feature relationship of this feature comparison group is labeled as a deviation correlation relationship, and the deviation attribute is recorded as "Brake hose correlation strength and humidity adaptation deviation," the deviation degree as "moderate deviation," and the deviation triggering factor described as "actual humidity exceeding the standard and slight aging of the brake hose."

[0083] Step S1446: Traverse all feature comparison groups, summarize evolutionary matching relationships and deviation correlation relationships, and form a complete set of feature interaction comparison results.

[0084] The above attribute comparison, relationship labeling and recording operations are performed sequentially on the feature comparison groups corresponding to the startup phase, stable operation phase, braking phase and shutdown phase. All feature groups labeled with evolutionary matching relationship and deviation association relationship and their detailed information are summarized to form a complete result set containing the feature interaction situation of each phase.

[0085] Step S145: Through the result generation layer of the brake performance dynamic evaluation artificial intelligence model, the evolution matching relationship and deviation correlation relationship are integrated, the operation stage and evolution node corresponding to the deviation are marked, and the brake performance evolution evaluation result including component performance evolution status evaluation and cross-stage collaborative status evaluation is generated.

[0086] The result generation layer of the AI ​​model for dynamic evaluation of brake performance includes a feature relationship integration module, a deviation localization module, and an evaluation report generation module. The feature relationship integration module receives a set of feature interaction comparison results, categorizes and integrates evolutionary matching relationships by operational stage to form a feature matching summary for each stage, and categorizes and integrates deviation relationships by deviation type and scope of influence to form a deviation summary. The deviation localization module, based on the records of deviation relationships, marks the specific operational stage (e.g., mid-stage of stable operation) and evolutionary nodes (e.g., key time points of change in correlation strength) corresponding to each deviation. The evaluation report generation module, based on the feature matching summary and deviation localization results, generates a component performance evolution status assessment and a cross-stage collaborative status assessment. The component performance evolution status assessment, for each component, describes whether its performance evolution conforms to the baseline trend by combining the evolutionary matching and deviation of its associated features; the cross-stage collaborative status assessment, for the stage transition process, describes whether cross-stage collaboration is normal by combining the matching and deviation of evolutionary trend features, ultimately forming a complete brake performance evolution evaluation result.

[0087] Step S150: Based on the brake performance evolution evaluation results and the correlation chain evolution trend, generate a dynamic adjustment scheme for brake performance. The dynamic adjustment scheme for brake performance includes the direction of component parameter adjustment in stages and cross-stage collaborative optimization suggestions.

[0088] Based on the component performance deviations and cross-stage collaborative deviations identified in the brake performance evolution evaluation results, and combined with the evolution trend of the brake operation correlation chain (such as the changing pattern of correlation strength and the evolution direction of correlation relationship), adjustment directions for component parameters are formulated for different operating stages. At the same time, collaborative optimization suggestions are formulated for the stage transition process. The above adjustment directions and optimization suggestions need to consider the correlation between components and the influence of environmental factors to ensure that the scheme can effectively correct performance deviations and make the brake performance return to the baseline evolution trend.

[0089] Step S151: Analyze the component performance evolution status assessment in the brake performance evolution assessment results, identify the components whose performance evolution deviates from the baseline trend and the corresponding deviation stages, and form a phased list of components to be adjusted.

[0090] The component performance evolution status assessment in the brake performance evolution evaluation results is analyzed to extract the content describing the performance deviation trend from the baseline. If the evaluation results show that the performance evolution of the brake hose deviates from the baseline trend in the middle of the stable operation phase, and the performance evolution of the brake cylinder deviates from the baseline trend in the early braking phase, then the brake hose and brake cylinder are identified as deviating components, corresponding to the middle of the stable operation phase and the early braking phase, respectively. The above information is categorized and organized according to the deviation phase to form a phased list of components to be adjusted. For example, the component to be adjusted in the stable operation phase is the brake hose, and the component to be adjusted in the early braking phase is the brake cylinder.

[0091] Step S152: For each component and corresponding stage in the list of components to be adjusted in stages, analyze the reasons why the performance evolution of the component deviates from the baseline trend. The reasons are determined based on the deviation relationship between the component's associated characteristics and the baseline characteristics in that stage, as well as the evolution trend of the associated chain.

[0092] For brake hoses that require adjustment during the stable operation phase, component correlation features of the brake hose in the middle of the stable operation phase are extracted. Simultaneously, benchmark component correlation features in the middle of the stable operation phase are extracted from the dynamic benchmark features of the brake performance. The component correlation features of the brake hose are compared with those of the benchmark components, and the deviation correlation between the two is used to determine the characteristic attributes with deviations, forming a list of deviation attributes. This list of deviation attributes may include deviations in the correlation strength between the brake hose and the brake valve, deviations in brake hose pressure transmission delay, etc.

[0093] Analyze each deviation attribute in the deviation attribute list, and combine it with environmental impact data during the mid-term of stable operation to determine whether environmental factors caused the deviation attribute. If the ambient humidity during this stage is higher than the baseline ambient humidity, and there is a corresponding relationship between the pressure transmission delay deviation of the brake hose and the increase in humidity, then environmental adaptation deviation is considered one of the causes. Retrieve the evolution characteristics of the correlation relationships before and after the mid-term of stable operation in the brake system's operational correlation chain, and analyze the impact of the correlation relationships in the previous stage (early stage of stable operation) on the correlation characteristics of the brake hose in the current stage. If the correlation strength between the brake valve and the brake hose fluctuates abnormally in the early stage of stable operation, and this fluctuation continues to affect the mid-term, then the evolutionary transmission deviation in the previous stage is considered one of the causes. Analyze the correlation relationships between the brake hose and other components (such as the brake cylinder and air reservoir) during the mid-term of stable operation. If the pressure output of the brake cylinder is unstable during this stage, leading to abnormal load on the brake hose and subsequently causing correlation characteristic deviation, then the influence deviation of the related components is considered one of the causes. By considering the environmental adaptation deviation, the evolutionary transmission deviation in the previous stage, and the influence deviation of related components, and taking into account the amplifying effect of the continuous decline in the correlation strength of the brake hose in the evolution trend of the correlation chain, a complete description of the reasons why the performance evolution of the brake hose deviates from the baseline trend in the middle stage of stable operation is formed.

[0094] For brake cylinders requiring adjustment at the initial stage of braking, the same analysis process is employed. The component correlation characteristics of the brake cylinder at the initial stage of braking and the corresponding baseline component correlation characteristics are extracted. After comparison, a list of deviation attributes is determined. Then, combined with environmental impact data for this stage, correlation evolution characteristics from the previous stage (end of the stable operation stage), and correlations with other components (such as brake valves and release valves), the reasons for brake cylinder performance deviations are comprehensively determined. These may include environmental adaptation deviations caused by excessive vibration intensity at the initial stage of braking, evolution transmission deviations from the previous stage caused by unstable air reservoir pressure at the end of the stable operation stage, and correlation deviations caused by delayed brake valve response.

[0095] Step S1521: Extract the component association features of the component to be adjusted in the corresponding stage, and at the same time extract the reference component association features of the brake performance dynamic reference features in that stage.

[0096] For each component in the phased list of components to be adjusted, based on its corresponding deviation phase, all associated feature data of that component in that phase are located and extracted from the brake operation correlation data. This data covers the correlation strength, correlation response time, and collaborative change pattern of correlation parameters between that component and all other related components. Simultaneously, based on the identification of the deviation phase, the correlation features of the benchmark component corresponding to the component to be adjusted within that phase are extracted from the brake performance dynamic benchmark features. These benchmark component correlation features include the same-dimensional correlation strength benchmark value, correlation response time benchmark range, and collaborative change pattern of correlation parameters, ensuring that the extracted correlation features of the component to be adjusted and the benchmark component correlation features are completely matched in feature dimension and phase range.

[0097] Step S1522: Compare the component association features of the component to be adjusted with the reference component association features, and combine the deviation association relationship to determine the feature attributes with deviations, and form a deviation attribute list.

[0098] The component association features of the component to be adjusted are compared with the baseline component association features one by one according to feature dimensions. Referring to the established deviation relationships between the two, all feature dimensions that do not meet the baseline feature requirements are screened out; these feature dimensions are the feature attributes with deviations. These feature attributes with deviations are sorted by importance, and the deviation manifestation of each attribute is recorded sequentially, such as association strength lower than the baseline value, association response time exceeding the baseline range, and association parameter cooperative change pattern inconsistent with the baseline pattern, etc., ultimately forming a structured list of deviation attributes.

[0099] Step S1523: Analyze each deviation attribute in the deviation attribute list, and combine it with the environmental impact data of this stage to determine whether environmental factors cause the deviation attribute to occur. If there is a correlation between environmental factors and deviation attributes, then environmental adaptation deviation is taken as one of the causes.

[0100] Each deviation attribute in the deviation attribute list is analyzed one by one, and complete environmental impact data for that deviation stage is retrieved, including the change curves of specific environmental parameters such as temperature, humidity, vibration, and air pressure. By analyzing the correlation between the occurrence time and change trend of the deviation attribute and the change time and change trend of the environmental parameters, if the occurrence time of a certain deviation attribute coincides with the abnormal change time of the environmental parameter and the change trend is consistent, it is determined that the environmental factor caused the deviation attribute, and the environmental adaptation deviation is included in the category of causes of the component's performance deviation.

[0101] Step S1524: Retrieve the evolution characteristics of the correlation relationship before and after this stage in the brake operation correlation chain, analyze the influence of the correlation relationship of the previous stage on the correlation characteristics of the component to be adjusted in the current stage, and determine whether the current deviation is caused by the abnormal correlation evolution of the previous stage. If there is an influence, the deviation of the evolution transmission of the previous stage is taken as one of the reasons.

[0102] Extract the evolutionary characteristics of the correlation between the component to be adjusted and the preceding and following stages of the deviation stage from the operational chain of the brake system. Focus on the intensity changes, stability, and abnormal fluctuations of all correlations related to the component to be adjusted in the preceding stage. Analyze whether these correlation evolutionary characteristics of the preceding stage are abnormal, and whether such abnormalities are transmitted to the current stage through the correlations between components, thus causing the correlation characteristics of the component to be adjusted to deviate. If a clear transmission path and evidence of influence exist, the deviation in the evolutionary transmission of the preceding stage is listed as one of the causes.

[0103] Step S1525: Analyze the relationship between the component to be adjusted and other components at this stage, and determine whether the deviation of the related characteristics of the component to be adjusted is caused by the abnormal performance of other components. If so, the deviation caused by the related components is taken as one of the reasons.

[0104] The process involves identifying the relationship network between the component to be adjusted and all related components (such as upstream supply components, downstream execution components, and collaborative control components) during the deviation phase, and extracting the performance data and correlation characteristics of these related components at that stage. Each related component is then examined for performance anomalies, such as parameters exceeding normal ranges, delayed action response, or unstable correlation strength, and it is analyzed whether these anomalies directly or indirectly affect the correlation characteristics of the component to be adjusted. If a performance anomaly of a related component has a clear causal relationship with the deviation in the correlation characteristics of the component to be adjusted, then the influence of the related component on the deviation is considered one of the causes.

[0105] Step S1526: Combining environmental adaptation deviation, previous stage evolution transmission deviation, related component influence deviation and other deviation types, and taking into account the amplification or suppression effect of the related chain evolution trend on the deviation, form a complete description of the reasons why the performance evolution of the component deviates from the baseline trend in the corresponding stage.

[0106] Summarize the various causes of deviation identified in the above analysis, and identify any other unaddressed types of deviation (such as inherent deviations caused by component aging, historical deviations left over from improper maintenance, etc.). Consider the impact of the deviations in the related characteristics of the components to be adjusted on the evolutionary trend of the brake operation chain, and determine whether the evolutionary trend amplifies or inhibits the development of deviations. Sort the various causes of deviation according to their degree of impact, and describe in detail the mechanism of action, manifestations, and correlation with the deviation for each cause, ultimately forming a logically complete and clearly hierarchical description of the causes.

[0107] Step S153: Based on the reasons stated above, and considering the operating parameter adaptation range of the component in the corresponding stage and its relationship with other components within the stage, determine the parameter adjustment direction of the component in that stage, forming a phased component parameter adjustment direction. The phased component parameter adjustment direction includes the timing of parameter adjustment in the stage, the adjustment trend, and the adaptation requirements with the stage environment.

[0108] For brake hoses in the mid-stage of stable operation, based on the causes of performance deviations, the operating parameter adaptation range of the brake hoses in the mid-stage of stable operation is obtained. This operating parameter adaptation range includes the allowable range of working pressure, the allowable range of pressure transmission delay, and the adaptation range of the correlation strength with the brake valve. For environmental adaptation deviations (pressure transmission delay caused by excessive humidity), the pressure compensation parameters of the brake hoses need to be adjusted, with the adjustment trend being to increase the pressure compensation coefficient. For transmission deviations in the previous stage (abnormal fluctuations in the correlation strength of the brake valve), the coupling parameters of the brake hoses and brake valves need to be adjusted simultaneously, with the adjustment trend being to optimize the coupling frequency. For deviations caused by the influence of related components (abnormal brake cylinder load), the flow regulation parameters of the brake hoses need to be adjusted, with the adjustment trend being to increase the flow threshold.

[0109] Analyze whether the adjustment trends of these parameters conform to the operating parameter adaptation range of the brake hose in the middle of the stable operation phase. If it remains within the allowable range after increasing the pressure compensation coefficient, the adjustment trend is confirmed as feasible. Simultaneously, analyze whether the parameter adjustments will affect the inter-phase relationship between the brake hose and other components (such as the brake cylinder and air reservoir). If adjusting the flow regulation parameters may disrupt the pressure balance with the air reservoir, then the pressure adaptation auxiliary parameters for the air reservoir need to be set simultaneously while adjusting the flow parameters. The optimal timing for parameter adjustments is determined to be during the period of decreasing ambient humidity in the middle of the stable operation phase to reduce the interference of environmental factors on the adjustment effect. It is also clarified that the adjustments must adapt to the environmental conditions such as temperature and humidity during this phase, thus establishing the parameter adjustment direction for the brake hose in the middle of the stable operation phase.

[0110] For the brake cylinder in the initial stage of braking, based on the cause of deviation, the suitable range of operating parameters for the brake cylinder in the initial stage of braking is obtained, including the braking pressure output range, response time range, and the threshold for coordinated action with the brake valve. For environmental adaptation deviations caused by excessive vibration intensity, the vibration damping parameters of the brake cylinder are adjusted, with the adjustment trend being to increase the damping coefficient. For the evolution and transmission deviations in the early stage caused by unstable air reservoir pressure, the pressure receiving sensitivity parameters of the brake cylinder are adjusted, with the adjustment trend being to decrease sensitivity. For the deviations in the influence of related components caused by brake valve response delay, the action triggering parameters of the brake cylinder are adjusted, with the adjustment trend being to advance the triggering time. These adjustment trends are verified to be within the suitable range, and the impact on the relationship with other components is assessed. The adjustment timing is determined to be after the vibration peak in the initial stage of braking, while simultaneously meeting the air pressure conditions for this stage, thus forming the phased parameter adjustment direction for the brake cylinder.

[0111] Step S1531: Obtain the operating parameter adaptation range of the component in the corresponding stage. The operating parameter adaptation range includes the allowable fluctuation range of each operating parameter of the component in the stage and the adaptation range with environmental characteristics.

[0112] By consulting the brake component design manual, historical operating data statistics, and the manufacturer's technical specifications, the operating parameter adaptation range of the component to be adjusted during the corresponding deviation stage is obtained. This operating parameter adaptation range defines the allowable fluctuation range for each operating parameter of the component (such as pressure, flow rate, temperature, response time, and operating frequency), i.e., the upper and lower limits. Simultaneously, considering the typical environmental characteristics of this stage (such as temperature range, humidity range, and vibration level), the adaptation range for each operating parameter corresponding to different environmental characteristics is determined, ensuring that the parameters maintain normal function under specific environmental conditions.

[0113] Step S1532: For each reason why the performance evolution of the component deviates from the baseline trend, determine the corresponding type of parameter that needs to be adjusted. The type of parameter that needs to be adjusted includes environmental adaptation parameters, correlation and coordination parameters, and stage transition parameters.

[0114] For each specific cause of component performance deviation, analyze the category of operating parameters affected by that cause, and then determine the corresponding parameter type that needs adjustment. If the cause is related to environmental factors (such as deviations caused by environmental humidity or temperature), the parameter type that needs adjustment is environmental adaptation parameter, which is used to optimize the component's adaptability to environmental changes. If the cause involves abnormal correlation with other components (such as insufficient correlation strength or misalignment of coordinated actions), the parameter type that needs adjustment is correlation and coordination parameter, which is used to improve the collaborative working effect between components. If the cause is related to improper parameter connection during the phase transition process, the parameter type that needs adjustment is phase transition parameter, which is used to ensure a smooth transition of parameters during phase switching.

[0115] Step S1533: Analyze the difference between the current parameter performance and the baseline parameter performance of the parameter type to be adjusted. Combined with the operating parameter adaptation range at this stage, preliminarily determine the parameter adjustment trend. If the current parameter performance exceeds the upper limit of the adaptation range, the preliminary adjustment trend is to reduce the parameter. If it is below the lower limit of the adaptation range, the preliminary adjustment trend is to increase the parameter.

[0116] For each identified parameter type requiring adjustment, extract its current parameter performance data (e.g., current pressure value, current response time, current correlation strength value, etc.) and compare it with the baseline parameter performance (e.g., baseline pressure value, baseline response time, baseline correlation strength value), calculating the degree of difference between the two. Considering the parameter's operating parameter adaptation range at the corresponding stage, determine whether the current parameter performance exceeds the range. If the current parameter performance is higher than the upper limit of the adaptation range, the initial adjustment trend is determined to be decreasing the parameter; if the current parameter performance is lower than the lower limit of the adaptation range, the initial adjustment trend is to increase the parameter; if the current parameter performance is within the adaptation range but deviates from the baseline performance, the adjustment trend towards the baseline value is determined based on the direction of the deviation.

[0117] Step S1534: Analyze the relationship between this component and other components at this stage, and determine whether the initially determined parameter adjustment trend will affect the parameter performance of other components beyond their adaptation range. If there is an impact, analyze the parameter sensitivity of the related components and adjust the magnitude of the initial adjustment trend.

[0118] The connection relationships and parameter interaction logic between the component to be adjusted and all related components at the corresponding stage are analyzed. The potential changes in the parameter performance of related components after the initial parameter adjustment trend is implemented are simulated. If the simulation results show that the parameter performance of a certain related component may exceed its own operating parameter adaptation range, the sensitivity of that related component to the adjusted parameters is further analyzed, i.e., the magnitude of the parameter change of the related component when the adjusted parameter changes by one unit. Based on the sensitivity analysis results, the magnitude of the initial adjustment trend is appropriately reduced or increased to ensure that the parameters of the component to be adjusted meet the requirements after adjustment, while the parameters of related components also remain within their adaptation range.

[0119] Step S1535: Based on the environmental characteristics of this stage, determine whether the adjusted parameter trend is compatible with the environmental characteristics. If it is not compatible, further correct the parameter adjustment trend so that the adjusted parameters can be compatible with the environment of this stage.

[0120] Retrieve environmental characteristic data for the corresponding stage, including the specific values, frequency of change, and amplitude of fluctuations of environmental parameters. Substitute the parameter values ​​corresponding to the adjusted parameter trends into the component operation simulation model under this environmental characteristic to determine whether the parameters can function normally under this environmental condition and whether new deviations will occur due to insufficient environmental adaptability. If the simulation results show that the parameter adjustment trend is not compatible with the environmental characteristics, for example, the adjusted pressure parameters still have a transmission delay in a high humidity environment, then further correct the parameter adjustment trend according to the specific impact of the environmental characteristics, such as increasing the pressure compensation amplitude or adjusting the dynamic response speed of the parameters, until the adjusted parameters can adapt to the environmental characteristics of this stage.

[0121] Step S1536: Mark the timing of parameter adjustment stages, the gradual requirements of adjustment range, and the coordinated adjustment sequence with related components to form the direction of component parameter adjustment in stages.

[0122] Based on the operational characteristics and environmental change patterns of the corresponding stage, determine the optimal timing for parameter adjustment. For example, select periods with relatively stable environmental parameters, periods when component load is at a moderate level, or non-critical operating periods within the stage. For the adjustment magnitude, set gradual adjustment requirements to avoid sudden parameter changes impacting components and related systems. For example, stipulate that adjustments be made gradually to the target value in multiple steps within a certain timeframe, with each adjustment not exceeding a preset threshold. Analyze the dependencies of related components on the adjustment and determine the coordinated adjustment sequence. For example, adjust the parameters of upstream supply components first, then adjust the parameters of the component to be adjusted, and finally adjust the parameters of downstream execution components to ensure the coordination and stability of parameter adjustments throughout the entire related system. Integrate the stage timing, gradual adjustment requirements, and coordinated adjustment sequence with the parameter adjustment trend to form a complete phased component parameter adjustment direction.

[0123] Step S154: Analyze the cross-stage collaborative state assessment in the brake performance evolution assessment results, identify the collaborative links that deviate from the baseline trend in cross-stage collaborative evolution and the corresponding transition stages, and form a list of cross-stage collaborative links to be optimized.

[0124] A detailed analysis of the cross-stage collaborative state assessment content in the brake performance evolution evaluation results is conducted, extracting the assessment descriptions of collaborative relationships during the transition between different stages. If the evaluation results show that, during the transition from the stable operation stage to the braking stage, the collaborative link between the air reservoir and the brake valve deviates from the baseline trend—specifically, the pressure supply rate from the air reservoir to the brake valve during the transition stage is lower than the baseline pressure supply rate; or during the transition from the braking stage to the stopping stage, there is a delay in the collaborative action between the brake cylinder and the release valve, deviating from the baseline collaborative timing—then the collaborative links between the air reservoir and the brake valve, and between the brake cylinder and the release valve are identified as collaborative links deviating from the baseline trend, corresponding to the stable operation-braking transition stage and the braking-stopping transition stage, respectively. These collaborative links and their corresponding transition stages are then organized in transition order to form a list of cross-stage collaborative links to be optimized.

[0125] Step S155: For each collaborative link and corresponding transition stage in the list of cross-stage collaborative links to be optimized, analyze the reasons for the deviation of the collaborative evolution from the baseline trend. The reasons are determined based on the deviation correlation between the cross-stage collaborative evolution characteristics and the baseline characteristics and the evolution trend of the correlation chain.

[0126] For the coordinated operation of the air reservoir and brake valve during the stable operation-braking transition phase, the cross-stage coordinated evolution characteristics of this coordinated link during this transition phase are extracted. Simultaneously, the benchmark cross-stage coordinated evolution characteristics of the corresponding transition phase are extracted from the dynamic benchmark characteristics of the brake performance. By comparing the two, deviation attributes are determined, such as pressure supply response time deviation and pressure coordination accuracy deviation. Combined with environmental impact data for this transition phase (such as vibration and temperature data), it is determined whether environmental factors cause deviations. If the vibration intensity during the transition phase is higher than the benchmark value, and the pressure supply response time deviation increases with increasing vibration intensity, then environmental adaptation deviation is one of the causes. The correlation evolution characteristics before and after this transition phase in the brake operation correlation chain are retrieved. If the pressure stability deviation of the air reservoir at the end of the stable operation phase is transmitted to the transition phase, then the evolution transmission deviation in the previous stage is one of the causes. The correlation between the air reservoir and brake valve within the coordinated link is analyzed. If abnormal valve opening adjustment of the brake valve leads to increased pressure supply resistance, then the influence deviation of the related components is one of the causes. Based on these factors, and considering the amplifying effect of the continuous weakening of the correlation strength of this collaborative link in the evolution trend of the correlation chain, the reasons for the deviation of this collaborative link are described.

[0127] For the coordination between the brake cylinder and the release valve during the braking-stop transition phase, the same process was used to analyze the causes. The cross-stage coordinated evolution characteristics and baseline characteristics of this phase were extracted to determine deviation attributes (such as action coordination timing deviation and pressure release coordination deviation). Then, combined with environmental data from the transition phase, the related evolution characteristics of the previous stage (end of the braking phase), and the relationships between internal components, the possible causes were identified as environmental adaptation deviations due to a sudden drop in air pressure during the transition phase, transmission deviations from the previous stage due to residual pressure in the brake cylinder at the end of the braking phase, and deviations in the influence of related components due to insufficient opening speed of the release valve.

[0128] Step S156: Based on the reasons stated above, and in conjunction with the direction of parameter adjustment for components in each stage and the cross-stage correlation, formulate cross-stage collaborative optimization suggestions for this collaborative link. The cross-stage collaborative optimization suggestions include the parameter collaborative adjustment method for the transition stage, the coordination adjustment requirements for related components, and the direction of evolution trend correction.

[0129] For the coordination between the air reservoir and the brake valve during the stable operation-braking transition phase, cross-phase coordination optimization suggestions are formulated based on the causes of deviations and the adjustment directions of the respective phased component parameters of the air reservoir and brake valve. Regarding the parameter coordination adjustment method, given the existence of pressure supply response time deviations and pressure coordination accuracy deviations, a coordination adjustment method of "air reservoir pre-pressurization + brake valve early opening" is adopted. That is, the air reservoir increases its pre-pressurization parameter at the end of the stable operation phase according to the phased adjustment direction, and the brake valve opens its valve port early at the beginning of the transition phase. The adjustment time difference between the two is set to one-fifth of the transition phase duration. Regarding the coordination adjustment requirements of related components, the upstream air compressor is required to maintain stable pressure supply during the transition phase, and the downstream brake cylinder must be prepared to receive pressure. Its parameter adjustment must be synchronized with the coordinated adjustment of the air reservoir and brake valve. For example, the adjustment of the brake cylinder's pressure receiving sensitivity must be initiated after the air reservoir pre-pressurization is completed. In terms of evolution trend correction, in response to the continuous weakening of the correlation strength of this collaborative link, a phased target threshold for the correlation strength between the air storage cylinder and the brake valve is set. Combined with the actual correlation strength monitoring results after parameter coordination adjustment, the pre-pressurization parameters and valve opening amplitude are dynamically fine-tuned to gradually bring the correlation strength back to the baseline evolution trend, ensuring that the pressure supply speed and pressure accuracy in the transition phase meet the baseline requirements.

[0130] For the coordination between the brake cylinder and the release valve during the braking-stop transition phase, corresponding cross-stage coordination optimization suggestions are formulated based on the causes of deviation and the phased parameter adjustment directions of both. Regarding the parameter coordination adjustment method, due to the existence of action coordination timing deviations and pressure release coordination deviations, a coordination adjustment method of "brake cylinder pressure pre-attenuation + release valve gradient opening" is adopted. That is, the brake cylinder initiates the pressure pre-attenuation program in advance according to the phased adjustment direction at the end of the braking phase, and the release valve gradually increases its opening degree according to a preset gradient during the transition phase. The gradient interval is set to one-sixth of the transition phase duration to achieve precise matching between pressure release and release action. Regarding the adjustment requirements of related components, it is clearly stated that the air reservoir must maintain a stable low-pressure output during the transition phase, and the brake valve must send a signal to trigger the release valve action when the brake cylinder pressure pre-attenuates to a preset value. The parameter adjustment sequence of both must be consistent with the coordinated adjustment of the brake cylinder and the release valve. For example, the low-pressure stability parameter setting of the air reservoir must be completed after the brake cylinder pressure pre-attenuation parameter is determined. In terms of evolution trend correction, in response to the trend of gradually increasing coordination delay in this coordinated link, a coordination timing deviation early warning mechanism is established to monitor the matching degree between the brake cylinder pressure decay rate and the relief valve opening progress in real time. Based on the early warning signal, the pressure pre-decay rate and the relief valve opening gradient are dynamically adjusted to promote the coordination timing to return to the baseline evolution trajectory.

[0131] Step S157: Integrate the phased component parameter adjustment direction and cross-phase collaborative optimization suggestions according to the time axis and related chain evolution nodes, mark the priority and implementation conditions of the adjustment, and generate a dynamic adjustment scheme for brake performance.

[0132] Using the timeline as the core thread, the parameter adjustment directions of all components in each stage, such as the brake hose in the middle of the stable operation phase and the brake cylinder in the early braking phase, are arranged on the timeline according to the corresponding operating phase sequence. Simultaneously, cross-stage collaborative optimization suggestions, such as the collaborative optimization suggestions for the air reservoir and brake valve in the stable operation-braking transition phase and the collaborative optimization suggestions for the brake cylinder and release valve in the braking-stop transition phase, are inserted into the corresponding transition phase positions on the timeline to ensure the temporal continuity of phased adjustments and cross-stage optimizations.

[0133] By linking each adjustment direction and optimization suggestion to the corresponding evolution node in the brake operation chain, the specific evolution node targeted by the adjustment measures is clearly identified. For example, the parameter adjustment direction of the brake hose is linked to the inflection point of the correlation strength in the middle of the stable operation stage, and the collaborative optimization suggestion of the air reservoir and brake valve is linked to the starting node of the stable operation-brake transition stage.

[0134] Adjustment measures are prioritized based on their impact on brake performance and their urgency. Adjustments that directly affect braking safety (such as parameter adjustments to the brake cylinder during the initial braking phase) are given the highest priority; adjustments that improve performance stability (such as parameter adjustments to the brake hose) are given medium priority; and adjustments that optimize coordination efficiency (such as coordination optimization suggestions during the transition phase) are given the next highest priority.

[0135] For each adjustment measure, the implementation conditions are clearly defined. Implementation conditions include environmental conditions (such as ambient temperature and humidity being within preset ranges), operating conditions (such as train speed and load being within specific ranges), and conditions for completion of prior adjustments (such as prior adjustments of related components having been performed and achieving the expected results). For example, the implementation conditions for adjusting the parameters of the brake cylinder in the initial braking phase are set as follows: the ambient vibration intensity is below a preset threshold, the train load is between 60% and 90% of the rated load, and the pre-pressurization adjustment of the air reservoir has been completed.

[0136] The integrated phased component parameter adjustment directions, cross-phase collaborative optimization suggestions, and corresponding priorities and implementation conditions are organized in a structured format to form a complete dynamic adjustment plan for brake performance. This dynamic adjustment plan for brake performance is presented in document form, including a list of adjustment measures, a timetable, a priority ranking table, and an implementation condition description table, to facilitate subsequent execution and monitoring.

[0137] Figure 2 The diagram illustrates exemplary hardware and software components of an AI-based train brake performance evaluation system 100 that can implement the ideas of this application, according to some embodiments of this application. For example, a processor 120 can be used in the AI-based train brake performance evaluation system 100 and to perform the functions described in this application.

[0138] The AI-based train brake performance evaluation system 100 can be a general-purpose server or a special-purpose server; both can be used to implement the AI-based train brake performance evaluation method of this application. Although only one server is shown in this application, for convenience, the functions described in this application can be implemented in a distributed manner on multiple similar platforms to balance the load.

[0139] For example, the AI-based train brake performance evaluation system 100 may include a network port 110 connected to a network, one or more processors 120 for executing program instructions, a communication bus 130, and various forms of storage media 140, such as a disk, ROM, or RAM, or any combination thereof. Exemplarily, the AI-based train brake performance evaluation system 100 may also include program instructions stored in ROM, RAM, or other types of non-transitory storage media, or any combination thereof. The methods of this application can be implemented according to these program instructions. The AI-based train brake performance evaluation system 100 also includes an I / O interface 150 between the computer and other input / output devices.

[0140] For ease of explanation, only one processor is described in the AI-based train brake performance evaluation system 100. However, it should be noted that the AI-based train brake performance evaluation system 100 of this application may also include multiple processors. Therefore, the steps executed by one processor as described in this application may also be executed jointly by multiple processors or individually. For example, if the processor of the AI-based train brake performance evaluation system 100 executes steps A and B, it should be understood that steps A and B may also be executed jointly by two different processors or individually by one processor. For example, the first processor executes step A, the second processor executes step B, or the first processor and the second processor jointly execute steps A and B.

[0141] Furthermore, this embodiment of the invention also provides a readable storage medium, wherein computer-executable instructions are preset in the readable storage medium, and when the processor executes the computer-executable instructions, the above-mentioned artificial intelligence-based train brake performance evaluation method is implemented.

[0142] It should be noted that, in order to simplify the description of the present invention and thus help to understand one or more embodiments of the invention, multiple features may sometimes be grouped into one embodiment, drawing or description thereof in the foregoing description of the embodiments of the present invention.

Claims

1. A method for evaluating the performance of train brakes based on artificial intelligence, characterized in that, The method includes: Obtain a set of correlation data for the entire life cycle operation of the train brake, which includes operating parameter data, environmental impact data, and historical performance evolution records of various components of the brake at different operating stages; The dynamic correlation between brake components is extracted from the full-cycle operation correlation data set of the train brake, and a brake operation correlation chain is constructed. The brake operation correlation chain connects the correlation relationships of each component with the time axis as the clue, and shows the correlation strength evolution characteristics as the operation stage changes. Based on the evolution trajectory of the brake operation correlation chain, dynamic benchmark features of brake performance are generated. These dynamic benchmark features of brake performance include component correlation features at each operation stage and cross-stage collaborative evolution features. The pre-trained artificial intelligence model for dynamic evaluation of brake performance is invoked to perform cross-dimensional interactive mapping processing between the real-time collected brake operation characteristics and the dynamic benchmark characteristics of brake performance, thereby generating brake performance evolution evaluation results. Based on the brake performance evolution evaluation results and the correlation chain evolution trend, a dynamic adjustment scheme for brake performance is generated. The dynamic adjustment scheme for brake performance includes the adjustment direction of component parameters in stages and cross-stage collaborative optimization suggestions. The step of extracting the dynamic relationships between brake components from the train brake's full-cycle operation correlation data set and constructing a brake operation correlation chain includes: The train brake's full-cycle operation data set is divided into operation stages. Based on the brake's start-up stage, stable operation stage, braking stage, and stop stage, the data is divided into stage data units corresponding to each stage. Each stage data unit contains the operating parameter data, environmental impact data, and performance record data of each component within that stage. For each stage data unit, extract the parameter correlation between any two components within the corresponding stage, record the frequency of occurrence, duration and behavior of the correlation, and form a set of correlations within the stage; By comparing the set of intra-stage associations of adjacent stage data units, the addition, disappearance and intensity changes of associations are identified to form inter-stage association evolution information, which includes the type of association change and the triggering conditions for the change. Using the timeline as the horizontal thread, the sets of relationships within each stage are arranged in chronological order. At the same time, using the evolutionary information between stages as the vertical connection, the relationships between adjacent stages are linked together to form the initial brake operation relationship chain. The intensity change curves of each relationship in the initial brake operation correlation chain are analyzed over the entire cycle. The peak points, valley points and inflection points in the intensity change curves are marked to determine the correlation intensity evolution characteristics. Based on these characteristics, the structure of the initial brake operation correlation chain is optimized to generate the brake operation correlation chain.

2. The train brake performance evaluation method based on artificial intelligence according to claim 1, characterized in that, The generation of dynamic benchmark features for brake performance based on the evolution trajectory of the brake operation correlation chain includes: Traverse the timeline of the brake operation association chain, extract the association relationship set and association strength evolution characteristics corresponding to each operation stage, and form a stage association data package; Features are extracted from the relationships in the data packets associated with each stage to generate the component association features for that stage. These component association features reflect the core characteristics of the relationships between components within that stage and their adaptability to the environmental impact data for that stage. By comparing and analyzing the associated features of components in adjacent operating phases, the inheritance, variation and complementarity relationships between features are identified, forming feature association information between phases; Based on the inter-stage feature correlation information, a cross-stage co-evolution model is constructed. This cross-stage co-evolution model takes the component correlation features of each stage as input and outputs cross-stage co-evolution features that reflect the continuous change law of features. The cross-stage co-evolution features include the triggering factors and evolution direction of feature changes. All component-related features and cross-stage co-evolution features of all operating stages are integrated in chronological order, and the corresponding operating stage identifiers and evolution nodes of each feature are marked to form dynamic benchmark features of brake performance.

3. The train brake performance evaluation method based on artificial intelligence according to claim 2, characterized in that, The process involves calling a pre-trained AI model for dynamic evaluation of brake performance, performing cross-dimensional interactive mapping between the real-time collected brake operating characteristics and the dynamic benchmark characteristics of brake performance, and generating brake performance evolution evaluation results, including: The brake operating characteristics are acquired in real time, including component association characteristics, real-time environment adaptation characteristics, and real-time stage transition characteristics during the real-time operation phase. The dynamic benchmark features of brake performance and the real-time collected brake operation features are input into the feature interaction layer of the pre-trained artificial intelligence model for dynamic evaluation of brake performance. The feature interaction layer constructs a cross-dimensional mapping rule between dynamic baseline features and real-time running features. The cross-dimensional mapping rule includes stage matching rules, feature type correspondence rules, and evolution trend comparison rules. Based on the cross-dimensional mapping rules, the corresponding features in the dynamic benchmark features, the component association features at each stage, the cross-stage collaborative evolution features, and the real-time operation features are interactively compared to mine the evolutionary matching relationship and deviation association relationship between features. The result generation layer of the artificial intelligence model for dynamic evaluation of brake performance integrates the evolution matching relationship and the deviation correlation relationship, marks the operating stage and evolution node corresponding to the deviation, and generates the brake performance evolution evaluation result including component performance evolution status evaluation and cross-stage collaborative status evaluation.

4. The train brake performance evaluation method based on artificial intelligence according to claim 3, characterized in that, The construction of cross-dimensional mapping rules between dynamic baseline features and real-time running features through the feature interaction layer includes: Analyze the operational stage identifiers in the dynamic benchmark features of the brake performance, determine the time range, environmental characteristics, and core correlation types of each stage, and construct a stage feature library; Extract the real-time operation stage information from the real-time collected brake operation features, compare the real-time operation stage information with the stage features in the stage feature library, determine the benchmark stage corresponding to the real-time operation features, and form a stage matching rule. The stage matching rule includes stage time overlap determination, environmental feature similarity determination, and core association type matching determination. The component association features and cross-stage co-evolution features in the dynamic benchmark features are labeled with types, and the feature types are divided into parameter association type, environment adaptation type, and evolution trend type. The component association features, real-time environment adaptation features, and real-time stage transition features in the real-time operation features are labeled with types to ensure consistency with the feature type classification standard of the dynamic benchmark features, forming feature type correspondence rules. The feature type correspondence rules include feature type definition, type identification basis, and feature conversion method when types do not match. Analyze the evolution curves of cross-stage co-evolutionary features in dynamic benchmark features, extract the direction of change, rate of change and period of change of the evolution curves, and construct an evolution trend template; The changing patterns of transitional features in the real-time operation are determined to form a real-time evolution trend description. Evolution trend comparison rules are constructed, which include consistency determination of evolution direction, setting of deviation range of change rate, and determination of overlap of change cycle.

5. The train brake performance evaluation method based on artificial intelligence according to claim 3, characterized in that, Based on the cross-dimensional mapping rule, the corresponding features in the dynamic benchmark features, such as the component association features at each stage, the cross-stage collaborative evolution features, and the real-time operation features, are interactively compared to mine the evolutionary matching relationships and deviation correlation relationships between features, including: According to the stage matching rules, the real-time running features are grouped with the dynamic benchmark features of the corresponding benchmark stage to form feature comparison groups. Each feature comparison group contains the component association features of the benchmark stage, the cross-stage co-evolution features, and the corresponding features in the real-time running features. Based on the feature type correspondence rule, within each feature comparison group, the parameter-related features of the dynamic benchmark feature are matched with the parameter-related features of the real-time running feature, the environment-adaptive features are matched with the environment-adaptive features, and the evolutionary trend features are matched with the evolutionary trend features. For features of the corresponding type, attribute comparison is performed: parameter correlation feature comparison is performed on related component pairs and correlation performance; environmental adaptation feature comparison is performed on the adaptation performance of environmental factors and correlation relationships; and evolutionary trend feature comparison is performed on the direction of change and change nodes. If the attribute comparison results of the corresponding type feature meet the preset matching threshold range, then the feature relationship of the feature comparison group is marked as an evolutionary matching relationship, and the matched feature attributes and matching degree description are recorded. If the attribute comparison results of the corresponding type of feature exceed the preset matching threshold range, analyze the stage position, feature type and associated evolution node of the deviation, determine the scope of influence of the deviation and possible triggering factors, mark the feature relationship of the feature comparison group as the deviation relationship, and record the deviation attribute, deviation degree and deviation triggering factor description. Traverse all feature comparison groups, summarize evolutionary matching relationships and deviation correlation relationships, and form a complete set of feature interaction comparison results.

6. The train brake performance evaluation method based on artificial intelligence according to claim 3, characterized in that, The step of generating a dynamic adjustment scheme for brake performance based on the brake performance evolution evaluation results and the correlation chain evolution trend includes: The component performance evolution status assessment in the brake performance evolution assessment results is analyzed to identify components whose performance evolution deviates from the baseline trend and the corresponding deviation stages, forming a phased list of components to be adjusted. For each component and corresponding stage in the list of components to be adjusted in stages, the reasons for the deviation of the component's performance evolution from the baseline trend are analyzed. The reasons are determined based on the deviation relationship between the component's associated characteristics and the baseline characteristics in that stage, as well as the evolution trend of the associated chain. Based on the reasons stated above, and considering the operating parameter adaptation range of the component in the corresponding stage and its relationship with other components within the stage, the parameter adjustment direction of the component in that stage is determined, forming a phased component parameter adjustment direction. The phased component parameter adjustment direction includes the timing of parameter adjustment in the stage, the adjustment trend, and the adaptation requirements with the stage environment. The cross-stage collaborative state assessment in the brake performance evolution evaluation results is analyzed to identify the collaborative links that deviate from the baseline trend in cross-stage collaborative evolution and the corresponding transition stages, forming a list of cross-stage collaborative links to be optimized. For each collaborative link and corresponding transition stage in the list of cross-stage collaborative links to be optimized, the reasons for the deviation of the collaborative evolution from the baseline trend are analyzed. The reasons are determined based on the deviation correlation between the cross-stage collaborative evolution characteristics and the baseline characteristics and the evolution trend of the correlation chain. Based on the reasons stated above, and considering the direction of parameter adjustment for components in each stage and the cross-stage relationships, cross-stage collaborative optimization suggestions are formulated for this collaborative link. These cross-stage collaborative optimization suggestions include the parameter collaborative adjustment method for the transition stage, the coordination adjustment requirements for related components, and the direction of evolution trend correction. The phased component parameter adjustment directions and cross-phase collaborative optimization suggestions are integrated according to the time axis and the evolution nodes of the related chain, and the priority and implementation conditions of the adjustment are marked to generate a dynamic adjustment scheme for brake performance.

7. The train brake performance evaluation method based on artificial intelligence according to claim 6, characterized in that, The analysis of the reasons why the performance evolution of each component in the phased adjustment list deviates from the baseline trend includes: Extract the component association features of the component to be adjusted in the corresponding stage, and at the same time extract the benchmark component association features of the brake performance dynamic benchmark features for that stage. By comparing the component association features of the component to be adjusted with the baseline component association features, and combining the deviation association relationships, the characteristic attributes with deviations are determined, and a list of deviation attributes is formed. Analyze each deviation attribute in the deviation attribute list, and combine it with the environmental impact data of this stage to determine whether environmental factors cause the deviation attribute. If there is a correlation between environmental factors and deviation attributes, then environmental adaptation deviation is taken as one of the causes. Retrieve the evolution characteristics of the correlation relationship before and after this stage in the brake operation correlation chain, analyze the influence of the correlation relationship of the previous stage on the correlation characteristics of the component to be adjusted in the current stage, and determine whether the current deviation is caused by the abnormal correlation evolution of the previous stage. If there is an influence, the deviation of the evolution transmission of the previous stage is taken as one of the reasons. Analyze the relationship between the component to be adjusted and other components at this stage, and determine whether the deviation of the related characteristics of the component to be adjusted is caused by the abnormal performance of other components. If so, the deviation caused by the related components is taken as one of the reasons. By considering environmental adaptation deviations, previous stage evolution transmission deviations, related component influence deviations, and other deviation types, and combining the amplification or suppression effect of the related chain evolution trend on the deviations, a complete description of the reasons why the performance evolution of the component deviates from the baseline trend in the corresponding stage is formed.

8. The train brake performance evaluation method based on artificial intelligence according to claim 6, characterized in that, Based on the aforementioned reasons, and considering the component's operating parameter adaptation range in the corresponding stage and its inter-stage relationship with other components, the parameter adjustment direction of the component in that stage is determined, forming a phased component parameter adjustment direction, including: Obtain the operating parameter adaptation range of the component in the corresponding stage. The operating parameter adaptation range includes the allowable fluctuation range of each operating parameter of the component in the stage and the adaptation range with environmental characteristics. For each reason why the performance evolution of this component deviates from the baseline trend, the corresponding type of parameter that needs to be adjusted is determined. The type of parameter that needs to be adjusted includes environmental adaptation parameters, correlation and coordination parameters, and stage transition parameters. Analyze the difference between the current parameter performance and the baseline parameter performance of the parameter type to be adjusted. Combined with the operating parameter adaptation range at this stage, the initial trend of parameter adjustment is determined. If the current parameter performance exceeds the upper limit of the adaptation range, the initial adjustment trend is to reduce the parameter. If it is below the lower limit of the adaptation range, the initial adjustment trend is to increase the parameter. Analyze the relationship between this component and other components at this stage, and determine whether the initially determined parameter adjustment trend will affect the parameter performance of other components beyond their adaptation range. If there is an impact, analyze the parameter sensitivity of the related components and adjust the magnitude of the initial adjustment trend. Based on the environmental characteristics of this stage, determine whether the adjusted parameter trend is compatible with the environmental characteristics. If not, further correct the parameter adjustment trend so that the adjusted parameters can be adapted to the environment of this stage. The timing of parameter adjustments, the gradual requirements for adjustment magnitude, and the order of coordinated adjustments with related components are used to form the direction of component parameter adjustments in stages.

9. A train brake performance evaluation system based on artificial intelligence, characterized in that, The system includes a processor and a memory, the memory being connected to the processor. The memory is used to store programs, instructions, or code, and the processor is used to execute the programs, instructions, or code in the memory to implement the artificial intelligence-based train brake performance evaluation method as described in any one of claims 1-8.

Images