Online fault early warning method and system suitable for electric vehicle charging facility

Abstract

The invention relates to the field of charging early warning, and discloses an online fault early warning method and system suitable for an electric vehicle charging facility, and the method comprises the steps: carrying out the time rearrangement of voltage, current, interface temperature and communication round-trip delay data according to a stage boundary; aiming at the operation state sampling chain, carrying out cross disassembly on electric parameter transient disturbance, thermal response hysteresis and control signal feedback offset; joint evaluation is carried out on the control response speed, the module output consistency and the interface temperature rise accumulation behavior in the power regulation process through a consistency degradation discrimination model; the potential fault pointing information is traced back to the corresponding charging gun interface, the power module unit and the associated heat dissipation and communication node, and reverse convergence analysis is performed on the propagation path of the abnormal influence; and on the basis of the consistency damage degree of the state change of each node in the suspected fault unit set in time and space, dynamically adjusting the sampling frequency and focus of attention of online monitoring. The method has the advantage of improving the timeliness of fault positioning.

Description

Technical Field

[0001] This invention relates to the field of charging early warning, specifically to an online fault early warning method and system applicable to electric vehicle charging facilities. Background Technology

[0002] In recent years, the number of DC fast charging piles in public parking lots and highway service areas has grown rapidly. However, in high-concurrency charging scenarios, charging facilities frequently experience hidden faults such as poor contact at the charging gun interface, abnormal voltage deviations between modules, and partial blockages in the heat dissipation system. Taking underground parking garages in urban complexes as an example, such scenarios have high humidity, limited ventilation, and high vehicle entry and exit density, leading to problems such as fluctuations in insulation performance, abnormal temperature rise at connection points, or short-term loss of synchronization in control units after long-term operation. However, existing monitoring systems mostly rely on single electrical parameters or alarm events for judgment, making it difficult to collaboratively analyze heterogeneous data from multiple sources such as voltage, current, temperature rise, communication status, and load power, resulting in difficulty in identifying fault characteristics in the early stages. Especially in the early stages of hidden faults, the charging curve often only shows slight disturbances, and traditional threshold-based detection mechanisms cannot effectively capture its gradual evolution trend, leading to delayed fault location, untimely warnings, and even the risk of charging interruption. Therefore, it is necessary to design an online fault early warning method and system suitable for electric vehicle charging facilities to improve the timeliness of fault location. Summary of the Invention

[0003] To address the shortcomings of existing technologies, this invention provides an online fault early warning method and system suitable for electric vehicle charging facilities, which has the advantage of improving the timeliness of fault location and solves the problems mentioned in the background technology.

[0004] To achieve the aforementioned goal of improving the timeliness of fault location, the present invention provides the following technical solution: an online fault early warning method suitable for electric vehicle charging facilities, comprising the following steps: Based on the charging start-up, constant current regulation and power drop operation segments, the voltage, current, interface temperature and communication round-trip delay data are rearranged according to the stage boundaries to form an operation status sampling chain that reflects the actual operating load changes; For the operational status sampling chain, combined with the operational characteristics of charging facilities under different concurrent access conditions, transient disturbances of electrical parameters, thermal response hysteresis and control signal feedback offset are cross-decomposed. Taking the coupling relationship under different time spans as clues, a combined operational description set describing the continuous stability of the charging process is constructed. In the combined operation description set, the control response speed, module output consistency and interface temperature rise accumulation behavior during power regulation are jointly evaluated by the consistency degradation discrimination model. Abnormal segments that have deviated from normal operating conditions but have not triggered the protection threshold are screened out and marked as potential fault pointing information. The potential fault information is traced back to the corresponding charging gun interface, power module unit and associated heat dissipation and communication nodes. Combined with the internal electrical topology and physical layout constraints of the charging pile, the propagation path of the abnormal influence is analyzed in reverse convergence to determine the set of suspected fault units with clear associated boundaries. Based on the degree of temporal and spatial disruption of the consistency of state changes of each node within the suspected fault unit set, the sampling frequency and focus of online monitoring are dynamically adjusted, and early warning information for specific components is generated during the gradual abnormality stage.

[0005] Preferably, the process of forming an operating state sampling chain that reflects changes in actual operating load is as follows: Collect voltage, current, interface temperature and communication round-trip time data of the charging facility during the charging start-up, power ramp-up, constant current maintenance and power fall-off phases; The start and end boundaries of each operating stage are identified based on the charging control command and power change curve. The stage boundaries are used as constraints to perform time alignment and sequence rearrangement of data from different sampling periods and different sources. The rearranged data of various types are connected in chronological order to form a continuous operating status sampling chain that reflects the load change process.

[0006] Preferably, the process of cross-decomposing the transient disturbances of electrical parameters, thermal response hysteresis, and control signal feedback offset is as follows: The current concurrent operation level is determined based on the number of charging terminals connected at the same time, and transient fluctuation segments of voltage and current are separated in the operation status sampling chain. Extract the response delay information of interface temperature relative to changes in electrical parameters, and perform statistical analysis on the time offset between the issuance of control commands and the return of feedback signals; The disturbance, hysteresis, and offset information are cross-decomposed according to their corresponding relationships to generate multidimensional operational response components that characterize different operational response dimensions.

[0007] Preferably, the process of constructing a combined operational description set describing the continuous stability of the charging process is as follows: Within a short time span, based on the electrical parameter disturbance component in the multidimensional operating response components, the instantaneous coupling relationship between electrical parameter disturbance and control feedback is analyzed. Over a medium time span, the hysteresis coupling relationship between interface temperature change and power regulation behavior is analyzed based on the thermal response hysteresis component in the multidimensional operating response components. Over a long time span, based on the module consistency component in the multidimensional runtime response components, the trend of module output consistency with runtime is analyzed. The coupling relationships derived from the multidimensional operational response components across different time spans are integrated to form a combined operational description set for describing the continuous stability of the charging process.

[0008] Preferably, the process of jointly evaluating control response speed, module output consistency, and interface temperature rise accumulation behavior during power regulation using a consistency degradation discrimination model is as follows: The combined operational description is used to characterize the input consistency degradation discrimination model of operational description, which represents changes in control commands, power output, and interface temperature. Based on the time difference between the time node when the control command changes and the response time when the power output reaches the corresponding adjustment range, the response delay characteristics in the power adjustment process are quantified to form a response deviation index that characterizes the degree of deviation in control response speed. Under the same regulation conditions, the output voltage, current or power parameters of multiple power modules are synchronously compared. Based on the output deviation between modules and its changing trend during operation, a consistency degradation index reflecting the degree of change in module output consistency is generated. Time correlation analysis was performed on the continuous sampling data of interface temperature. Based on the trend of temperature continuously accumulating and rising with the duration of operation, the temperature rise rate index, which characterizes the degree of heat accumulation at the interface, was extracted. The response deviation index, consistency degradation index, and temperature rise rate index are comprehensively judged to output the judgment result used to characterize the degree of overall operational consistency degradation during power regulation.

[0009] Preferably, the process of marking potential fault indication information is as follows: The output judgment result is compared with the preset normal operation range to filter out the operation segments that do not meet the protection triggering conditions; Determine whether the running segment continuously exhibits a trend of consistency degradation, mark the running segments that meet the conditions, and generate corresponding potential fault indication information.

[0010] Preferably, the process of tracing potential fault information back to the corresponding charging gun interface, power module unit, and associated heat dissipation and communication nodes is as follows: The anomaly occurrence period is determined based on the time stamp contained in the potential fault indication information, and the charging gun interface and power module unit involved in operation are located within the anomaly occurrence period. Identify heat dissipation and communication nodes that have thermal conduction or communication connections with the power module unit, and establish a mapping relationship between potential fault indication information and corresponding physical units.

[0011] Preferably, the process for determining a set of suspected faulty units with clearly defined associated boundaries is as follows: Obtain the electrical connection topology between the charging gun interface, power module unit and control communication unit inside the charging pile, and constrain the path range of abnormal propagation based on the physical layout of each unit in the cabinet; Along the direction of the abnormal influence pointed to by the mapping relationship, reverse convergence is performed on the units with direct electrical connection, thermal coupling or communication association, and the associated nodes that do not meet the topology are eliminated step by step. Units that maintain their correlation after reverse convergence are aggregated to form a set of suspected faulty units with clear correlation boundaries.

[0012] Preferably, the process of generating early warning information for specific components during the anomaly progression phase is as follows: The temporal synchronization degree and spatial distribution characteristics of the state changes of each node in the suspected fault unit set were statistically analyzed. Based on the degree of consistency disruption, key monitoring nodes are identified, the sampling frequency of data corresponding to key monitoring nodes is increased, and the sampling frequency of other nodes is reduced. When an anomaly does not trigger a protective action and is still in the gradual change phase, generate early warning information for specific components.

[0013] An online fault early warning system for electric vehicle charging facilities includes: Stage Reordering Module: Reorders voltage, current, interface temperature, and communication delay data over time to form a sampling chain of operating status that reflects actual load changes; Coupled decomposition module: Cross-decomposes transient disturbances of electrical parameters, thermal response hysteresis, and control signal feedback offset to construct a combined operation description set; Degradation discrimination module: The consistency degradation discrimination model is used to jointly evaluate the control response speed, module output consistency and interface temperature rise accumulation behavior, and filter out abnormal segments and mark them as potential fault information. Fault backtracking module: Backtracks potential fault information to the charging gun interface, power module unit and associated heat dissipation and communication nodes to identify the set of suspected fault units; Early warning and scheduling module: Based on the degree of consistency disruption of the status changes of each node in the suspected fault unit set, it generates early warning information for specific components during the gradual abnormality stage.

[0014] Compared with the prior art, the present invention provides an online fault early warning method and system suitable for electric vehicle charging facilities, which has the following beneficial effects: This invention reorders multi-source operational data by time and reconstructs load behavior according to operational stages, introducing a cross-analysis mechanism for transient disturbances, thermal response hysteresis, and control feedback offset under concurrent access conditions. This enables the identification of latent anomaly evolution trends in the charging process at an early stage without triggering protection actions. By using a consistency degradation discrimination model to jointly evaluate control response, module output, and interface temperature rise, misjudgments caused by instantaneous fluctuations are effectively avoided, improving the stability and accuracy of anomaly identification. Combined with a reverse convergence analysis method based on electrical topology and physical layout constraints, potential faults can be precisely located to specific charging gun interfaces, power modules, or associated heat dissipation and communication nodes, significantly narrowing the scope of fault investigation. Through dynamic perception of the degree of state consistency disruption of suspected faulty units, adaptive scheduling of monitoring resources is achieved, and early warning information for specific components is output when the anomaly is still in a gradual stage. This guides early maintenance intervention, reduces the risk of charging interruptions and equipment damage, and improves the operational safety and reliability of electric vehicle charging facilities in high-concurrency and complex environments. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the method of the present invention; Figure 2 This is a schematic diagram of the structure of the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Example 1: Please refer to Figure 1 As shown in the figure, an online fault early warning method for electric vehicle charging facilities according to an embodiment of the present invention includes the following steps: S1: Based on the charging start-up, constant current regulation and power drop operation segments, the voltage, current, interface temperature and communication round-trip delay data are rearranged according to the stage boundaries to form an operation status sampling chain that reflects the actual operating load changes.

[0018] The process of forming the operating state sampling chain reflecting the actual operating load changes in S1 is as follows: The system collects voltage, current, interface temperature, and communication round-trip time data of the charging facility during the charging start-up, power ramp-up, constant current maintenance, and power fall-off phases. During the operation of the charging facility, the system acquires operating data for the charging start-up, power ramp-up, constant current maintenance, and power fall-off phases through sensing and monitoring units installed at the charging gun interface, power module output, and control communication link. Voltage and current data are collected in real time by the sampling circuit inside the power module, interface temperature is obtained by temperature sensors located near the gun head terminals or connecting copper busbars, and communication round-trip time is calculated by recording the control command issuance time and status feedback return time. Based on the charging control commands and power change curves, the start and end boundaries of each operating stage are identified. Using the stage boundaries as constraints, data from different sampling periods and sources are time-aligned and rearranged. Based on the charging mode switching commands output by the charging control unit and the inflection point characteristics of the power change curve, the collected data stream is divided into operating stages. The start and end boundaries of stages such as charging start-up, power ramp-up, constant current maintenance, and power fall-off are identified. Using the identified stage boundaries as constraints, data from different sampling periods and different hardware modules are corrected to a unified time reference. The sampling frequency differences are eliminated through interpolation, padding, or alignment. The data entries are reordered according to the actual occurrence order to ensure that various operating parameters are correlated on the same time axis. The rearranged data of various types are connected in chronological order to form an operating status sampling chain that continuously reflects the load change process. After time alignment and sequence rearrangement, the voltage, current, interface temperature and communication round-trip delay and other types of operating data are connected in series according to their time sequence to form a continuous data sequence covering the entire charging process. The operating status sampling chain can continuously reflect the evolution of load changes, thermal response and control interaction behavior.

[0019] S2: For the operating state sampling chain, combined with the operating characteristics of the charging facility under different concurrent access conditions, the transient disturbances of electrical parameters, thermal response hysteresis and control signal feedback offset are cross-decomposed. Taking the coupling relationship under different time spans as the clue, a combined operating description set describing the continuous stability of the charging process is constructed.

[0020] The process of cross-decomposing transient disturbances in electrical parameters, thermal response hysteresis, and control signal feedback offset in S2 is as follows: The current concurrent operation level is determined based on the number of charging terminals connected simultaneously. Transient fluctuation segments of voltage and current are separated in the operation status sampling chain. During the operation of the charging facility, the number of charging terminals participating in charging simultaneously is obtained by reading the real-time access status information recorded by the charging control unit, and the concurrent operation level of the charging facility is determined. Based on the concurrent operation level, local waveform analysis is performed on the voltage and current data in the operation status sampling chain. The focus is on identifying short-term sudden changes, oscillations or offset segments that occur during power regulation. The electrical parameter fluctuations with short duration and significant amplitude changes are separated from the overall data sequence to form transient disturbance segments corresponding to different concurrent conditions. The response delay information of interface temperature relative to changes in electrical parameters is extracted, and the time offset between the issuance of control commands and the return of feedback signals is statistically analyzed. For the separated electrical parameter change segments, the corresponding data of interface temperature change over time is extracted, and the time delay required for the interface temperature to reach a stable change trend is calculated to characterize the response hysteresis characteristics of temperature to changes in electrical parameters. By recording the issuance time of charging control commands and the return time of power module or interface status feedback signals, the round-trip delay of control signals in the communication link and execution link is statistically analyzed to obtain the time offset information between control commands and actual execution effects, which is used to reflect the response efficiency of the control link. Disturbance, hysteresis, and offset information are cross-decomposed according to their correspondence to generate multi-dimensional operational response components representing different operational response dimensions. After obtaining transient disturbances in electrical parameters, interface temperature response hysteresis, and control signal feedback offset information, the three types of information are cross-correlated and decomposed according to their correspondence on the time axis and their objects of action. Taking electrical parameter disturbances as the triggering benchmark, the associated temperature hysteresis characteristics and control offset characteristics are matched and classified to generate operational response components representing electrical response, thermal response, and control response, respectively.

[0021] The process of constructing the combined runtime description set describing the continuous stability of the charging process in S2 is as follows: Within a short time span, based on the electrical parameter disturbance component in the multidimensional operating response components, the instantaneous coupling relationship between electrical parameter disturbance and control feedback is analyzed. Within a short time span, using millisecond to second-level time windows as the analysis unit, the electrical parameter disturbance component in the multidimensional operating response components is selected as the main analysis object. By aligning the voltage and current transient fluctuation segments with the corresponding control command feedback data in time, the response amplitude and response timing relationship of electrical parameter changes after the power adjustment command is issued are analyzed, thereby determining the instantaneous coupling characteristics between electrical parameter disturbance and control feedback. This process is used to reflect the control system's ability to follow load changes and its stability level during the rapid adjustment phase. Within a medium time span, the hysteresis component of the thermal response in the multidimensional operating response components is used to analyze the hysteresis coupling relationship between interface temperature change and power regulation behavior. Within a medium time span, with the analysis interval from several seconds to several minutes, the relationship between interface temperature change and power regulation behavior is analyzed based on the hysteresis component of the thermal response in the multidimensional operating response components. By matching the power change process with the interface temperature rise or fall process, the delay time and trend of temperature response relative to power change are evaluated to characterize the heat conduction efficiency and heat dissipation state of the interface and connection parts under continuous load, thereby reflecting the thermal stability characteristics during the charging process. Over a long time span, based on the module consistency component in the multidimensional operating response components, the trend of module output consistency with operating time is analyzed. Over a long time span, taking a complete charging cycle or multiple consecutive charging cycles as the analysis range, the module consistency component in the multidimensional operating response components is selected as the analysis basis to conduct trend analysis on the changes of output parameters of multiple power modules with operating time. By comparing the output voltage, current or power deviation of each power module under the same operating conditions, it is identified whether the module output consistency gradually weakens with operating time, which reflects the impact of module aging, changes in connection status or internal performance deviation on overall stability. The coupling relationships derived from the multidimensional operational response components at different time spans are integrated to form a combined operational description set for describing the continuous stability of the charging process. After obtaining the coupling relationships at short, medium, and long time spans, the various coupling features derived from the multidimensional operational response components are uniformly organized and integrated. By associating the description results reflecting the control response, thermal response, and module consistency at different time scales with the operational object in chronological order, a combined operational description set that can continuously characterize the stability evolution of the charging process is formed.

[0022] S3: In the combined operation description set, the control response speed, module output consistency and interface temperature rise accumulation behavior during the power regulation process are jointly evaluated by the consistency degradation discrimination model. Abnormal segments that have deviated from normal operating conditions but have not triggered the protection threshold are screened out and marked as potential fault pointing information.

[0023] The process of jointly evaluating control response speed, module output consistency, and interface temperature rise accumulation behavior during power regulation using a consistency degradation discrimination model in S3 is as follows: The consistency degradation discrimination model inputs operation descriptions that characterize changes in control commands, power output, and interface temperature into the combined operation description set. The operation descriptions that characterize changes in control commands, power output, and interface temperature separately in the combined operation description set are then uniformly organized and aligned according to timestamps to form an input data sequence suitable for model analysis. Data from different sources are resampled and standardized using a unified time base to ensure comparability of control commands, power output, and temperature changes on the same time axis, thereby ensuring that the consistency degradation discrimination model can perform joint analysis of different operation descriptions. Based on the time difference between the control command change point and the power output reaching the corresponding adjustment range, the response delay characteristics in the power adjustment process are quantified to form a response deviation index characterizing the degree of deviation of the control response speed. Based on the control command change point, the starting moment when the power output enters the corresponding adjustment range is located. By calculating the time difference between the two, the response delay characteristics in the power adjustment process are quantified. In multiple consecutive adjustment events, the time difference is statistically analyzed to extract its mean, fluctuation range and trend, which are used to form a response deviation index characterizing the degree of deviation of the control response speed from the normal state, thereby reflecting the change in the control system's ability to follow load changes. Under the same regulation conditions, the output voltage, current, or power parameters of multiple power modules are synchronously compared. Based on the deviation of output between modules and its changing trend during operation, a consistency degradation index reflecting the degree of change in module output consistency is generated. Under the same power regulation conditions, the output voltage, current, or power parameters of multiple power modules participating in operation are synchronously collected and compared. By calculating the deviation of each module's output parameters relative to the group average or reference module, and combining the changing trend of the deviation with the operating time, it is assessed whether the module output consistency has gradually weakened, thereby generating a consistency degradation index reflecting the degree of change in the cooperative stability between modules. A time correlation analysis was performed on the continuous sampling data of the interface temperature. Based on the trend of continuous temperature increase with the duration of operation, the temperature rise rate index, which characterizes the degree of heat accumulation at the interface, was extracted. The continuous sampling data of the interface temperature was also analyzed for time correlation. Combined with the charging duration, the temperature change curve was piecewise fitted and the trend was extracted. By identifying whether there is a continuous cumulative increase in temperature during the stable operation phase and calculating the temperature rise per unit time, the temperature rise rate index, which characterizes the degree of heat accumulation at the interface, was extracted to reflect the thermal bearing status of the interface and connection parts under load. The response deviation index, consistency degradation index, and temperature rise rate index are comprehensively judged to output a judgment result that characterizes the degree of overall operational consistency degradation during power regulation. The response deviation index, consistency degradation index, and temperature rise rate index are used as joint inputs and introduced into the consistency degradation judgment model for comprehensive analysis. By weighted fusion of each index and threshold comparison, a judgment result reflecting the degree of overall operational consistency degradation during power regulation is output. This judgment result is used to distinguish between normal fluctuations and potential abnormal states.

[0024] The process of marking potential fault information in S3 is as follows: The output judgment results are compared with the preset normal operation range to filter out the operation segments that do not meet the protection triggering conditions; the judgment results output by the consistency degradation judgment model are compared with the pre-established normal operation range item by item. The normal operation range is obtained based on the statistical data of the same type of charging facility under standard operating conditions, long-term operation test and historical stable operation data. By setting multi-level judgment thresholds, the completely normal state, slight deviation state and protection triggering state are distinguished, and the operation segments that are in the slight deviation state and do not meet the protection action triggering conditions are filtered out. The system determines whether a running segment continuously exhibits a trend of consistency degradation. Running segments that meet the criteria are marked, and corresponding potential fault indication information is generated. For the selected running segments, a time continuity analysis is performed on their corresponding response deviation index, consistency degradation index, and temperature rise rate index to determine whether the indicators show a continuous deviation or gradually worsening trend within multiple adjacent time windows. By setting minimum duration and trend stability conditions, short-term fluctuations caused by occasional load changes or environmental factors are excluded, retaining only running segments that are continuous in time and directional in amplitude. For running segments that meet the criteria for determining the continuity of consistency degradation trend, corresponding potential fault indication information is generated, and the marking information is associated with the time range of occurrence, the types of running indicators involved, and the direction of degradation characteristics. This potential fault indication information serves as input for subsequent fault location and propagation path analysis, indicating the range of components that may have performance degradation or hidden anomalies, thereby achieving early detection and location of fault risks before triggering protection actions.

[0025] S4: Trace the potential fault information back to the corresponding charging gun interface, power module unit and associated heat dissipation and communication nodes. Combine the internal electrical topology and physical layout constraints of the charging pile to perform reverse convergence analysis on the propagation path of abnormal effects and determine the set of suspected fault units with clear associated boundaries.

[0026] In S4, the process of tracing potential fault information back to the corresponding charging gun interface, power module unit, and associated heat dissipation and communication nodes is as follows: The abnormal occurrence period is determined based on the time identifier contained in the potential fault indication information. During the abnormal occurrence period, the charging gun interface and power module unit involved in operation are located. The corresponding time identifier and duration range are extracted from the potential fault indication information. Using this time range as a constraint, the charging gun interface and power module unit in working state are searched in the charging facility operation log and status record. By comparing the connection status, power distribution record and module start-stop information of each charging gun, the charging gun interface and power module unit that actually participated in energy output and control regulation during the abnormal occurrence period are screened out, thereby excluding units that did not participate in operation or were in standby state. Identify heat dissipation and communication nodes that have thermal conduction or communication connections with the power module unit, and establish a mapping relationship between potential fault indication information and corresponding physical units. For the located power module unit, based on the internal structural design information and equipment configuration table of the charging pile, identify heat dissipation nodes that have direct or indirect thermal conduction relationships with the power module unit, including air duct inlets, cooling fans, heat sinks, or temperature acquisition points. Based on the control and monitoring communication topology, determine communication nodes that have data interaction relationships with the power module unit, such as control boards, acquisition interfaces, or bus nodes. By associating and labeling potential fault indication information with the charging gun interface, power module unit, and corresponding heat dissipation and communication nodes, establish a mapping relationship between abnormal features and specific physical units.

[0027] The process of determining the set of suspected faulty units with clearly defined associated boundaries in S4 is as follows: The electrical connection topology between the charging gun interface, power module unit, and control communication unit inside the charging pile is obtained. Combined with the physical layout of each unit in the cabinet, the path range of anomaly propagation is constrained. The electrical connection topology between the charging gun interface, power module unit, and control communication unit inside the charging pile is obtained, including the DC bus connection relationship, the parallel or series connection of power modules, and the communication link of control and acquisition signals. Combined with the actual physical layout of each unit in the cabinet, such as the distribution between upper and lower layers, the adjacent distance, and the shared heat dissipation channel, the path range of possible anomaly propagation is spatially constrained. By superimposing the electrical topology constraints and the physical layout constraints, the impact of anomalies is limited to propagation between units with actual conduction or interaction conditions, avoiding the mistaken inclusion of units that do not have the structural conditions for propagation in the analysis scope. Along the direction of the anomaly's impact indicated by the mapping relationship, reverse convergence is performed on units with direct electrical connections, thermal coupling, or communication associations, eliminating associated nodes that do not meet the topology requirements level by level. Starting from the established mapping relationship between potential fault-pointing information and physical units, reverse tracing analysis is performed along the possible propagation direction of the anomaly's impact. Adjacent units with direct electrical connections, thermal coupling, or communication associations to the target unit are verified level by level. Only units with actual electrical connections, thermal conduction conditions, or data interaction on communication links are retained. Nodes that do not meet any of the association conditions are determined to lack the basis for anomaly propagation and are eliminated. Through this step-by-step convergence process, the scope of anomaly associations gradually concentrates from a dispersed state towards the core unit. Units that maintain their correlation after reverse convergence are aggregated to form a set of suspected faulty units with clear correlation boundaries. Charging gun interfaces, power module units, and control communication units that maintain their correlation after reverse convergence analysis are also aggregated, and the integrity of their connection relationships is verified to ensure that the resulting set is topologically closed and physically continuous. Boundary markings are applied to the edge units of the set to clarify the start and end range of abnormal influences, thus forming a set of suspected faulty units with clear correlation boundaries.

[0028] S5: Based on the degree of temporal and spatial inconsistency in the state changes of each node within the suspected fault unit set, the sampling frequency and focus of online monitoring are dynamically adjusted, and early warning information for specific components is generated during the gradual abnormality stage.

[0029] In S5, the process of generating early warning information for specific components during the anomaly progression phase is as follows: The statistical analysis examines the temporal synchronization and spatial distribution characteristics of state changes of each node in the suspected fault unit set. It performs statistical analysis on the state changes of each node in the suspected fault unit set within the same time window, and calculates the temporal synchronization of state changes between different nodes and the distribution characteristics of abnormal features in physical space. The temporal synchronization is used to measure whether abnormal changes of multiple nodes occur in close time, and the spatial distribution characteristics are used to determine whether the abnormality is concentrated in a local area or spreads along a specific structural path. Through this statistical process, it identifies node combinations that exhibit consistent destructive characteristics in both time and space. Based on the degree of consistency disruption, key monitoring nodes are identified, and the sampling frequency of data corresponding to these key nodes is increased while the sampling frequency of other nodes is decreased. Nodes within the suspected fault unit set are managed hierarchically based on the degree of consistency disruption calculated from time synchronization and spatial distribution characteristics. For nodes with a high degree of consistency disruption and more concentrated abnormal feature changes, the sampling density or refresh frequency of their corresponding monitoring data is dynamically increased to enhance the ability to capture details of abnormal evolution. For nodes with a low degree of consistency disruption and no signs of abnormal spread, the original sampling strategy is maintained or appropriately slowed down to ensure that the overall monitoring load remains within a controllable range. This reallocation of monitoring resources aims to focus on potentially risky nodes rather than simply reducing the monitoring scope. When an anomaly does not trigger a protection action and is still in a gradual change phase, a warning message for a specific component is generated. When an anomaly does not meet the protection action triggering conditions, and time series analysis confirms that the anomaly change is still in a slow evolution phase, a warning message for a specific component is generated based on the operating status of key monitoring nodes and their corresponding physical component information. This warning message includes, but is not limited to, the anomaly-related component identifier, anomaly characteristic type, change trend description, and suggested attention level. This is used to prompt maintenance personnel to intervene in advance for inspection or maintenance without interrupting charging services, thereby preventing the anomaly from further evolving into a visible fault.

[0030] Example 2: Figure 2 As shown, an online fault early warning system suitable for electric vehicle charging facilities includes: Stage Reordering Module: Reorders voltage, current, interface temperature, and communication delay data over time to form a sampling chain of operating status that reflects actual load changes; Coupled decomposition module: Cross-decomposes transient disturbances of electrical parameters, thermal response hysteresis, and control signal feedback offset to construct a combined operation description set; Degradation discrimination module: The consistency degradation discrimination model is used to jointly evaluate the control response speed, module output consistency and interface temperature rise accumulation behavior, and filter out abnormal segments and mark them as potential fault information. Fault backtracking module: Backtracks potential fault information to the charging gun interface, power module unit and associated heat dissipation and communication nodes to identify the set of suspected fault units; Early warning and scheduling module: Based on the degree of consistency disruption of the status changes of each node in the suspected fault unit set, it generates early warning information for specific components during the gradual abnormality stage.

[0031] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0032] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An online fault early warning method suitable for electric vehicle charging facilities, characterized in that, Includes the following steps: Based on the charging start-up, constant current regulation and power drop operation segments, the voltage, current, interface temperature and communication round-trip delay data are rearranged according to the stage boundaries to form an operation status sampling chain that reflects the actual operating load changes; For the operational status sampling chain, combined with the operational characteristics of charging facilities under different concurrent access conditions, transient disturbances of electrical parameters, thermal response hysteresis and control signal feedback offset are cross-decomposed. Taking the coupling relationship under different time spans as clues, a combined operational description set describing the continuous stability of the charging process is constructed. In the combined operation description set, the control response speed, module output consistency and interface temperature rise accumulation behavior during power regulation are jointly evaluated by the consistency degradation discrimination model. Abnormal segments that have deviated from normal operating conditions but have not triggered the protection threshold are screened out and marked as potential fault pointing information. The potential fault information is traced back to the corresponding charging gun interface, power module unit and associated heat dissipation and communication nodes. Combined with the internal electrical topology and physical layout constraints of the charging pile, the propagation path of the abnormal influence is analyzed in reverse convergence to determine the set of suspected fault units with clear associated boundaries. Based on the degree of temporal and spatial disruption of the consistency of state changes of each node within the suspected fault unit set, the sampling frequency and focus of online monitoring are dynamically adjusted, and early warning information for specific components is generated during the gradual abnormality stage.

2. The online fault early warning method for electric vehicle charging facilities according to claim 1, characterized in that, The process of forming an operating state sampling chain that reflects changes in actual operating load is as follows: Collect voltage, current, interface temperature and communication round-trip time data of the charging facility during the charging start-up, power ramp-up, constant current maintenance and power fall-off phases; The start and end boundaries of each operating stage are identified based on the charging control command and power change curve. The stage boundaries are used as constraints to perform time alignment and sequence rearrangement of data from different sampling periods and different sources. The rearranged data of various types are connected in chronological order to form a continuous operating status sampling chain that reflects the load change process.

3. The online fault early warning method for electric vehicle charging facilities according to claim 2, characterized in that, The process of cross-decomposing the transient disturbances of electrical parameters, thermal response hysteresis, and control signal feedback offset is as follows: The current concurrent operation level is determined based on the number of charging terminals connected at the same time, and transient fluctuation segments of voltage and current are separated in the operation status sampling chain. Extract the response delay information of interface temperature relative to changes in electrical parameters, and perform statistical analysis on the time offset between the issuance of control commands and the return of feedback signals; The disturbance, hysteresis, and offset information are cross-decomposed according to their corresponding relationships to generate multidimensional operational response components that characterize different operational response dimensions.

4. The online fault early warning method for electric vehicle charging facilities according to claim 3, characterized in that, The process of constructing a combinatorial runtime description set describing the continuous stability of the charging process is as follows: Within a short time span, based on the electrical parameter disturbance component in the multidimensional operating response components, the instantaneous coupling relationship between electrical parameter disturbance and control feedback is analyzed. Over a medium time span, the hysteresis coupling relationship between interface temperature change and power regulation behavior is analyzed based on the thermal response hysteresis component in the multidimensional operating response components. Over a long time span, based on the module consistency component in the multidimensional runtime response components, the trend of module output consistency with runtime is analyzed. The coupling relationships derived from the multidimensional operational response components across different time spans are integrated to form a combined operational description set for describing the continuous stability of the charging process.

5. The online fault early warning method for electric vehicle charging facilities according to claim 4, characterized in that, The process of jointly evaluating control response speed, module output consistency, and interface temperature rise accumulation behavior during power regulation using a consistency degradation discrimination model is as follows: The combined operational description is used to characterize the input consistency degradation discrimination model of operational description, which represents changes in control commands, power output, and interface temperature. Based on the time difference between the time node when the control command changes and the response time when the power output reaches the corresponding adjustment range, the response delay characteristics in the power adjustment process are quantified to form a response deviation index that characterizes the degree of deviation in control response speed. Under the same regulation conditions, the output voltage, current or power parameters of multiple power modules are synchronously compared. Based on the output deviation between modules and its changing trend during operation, a consistency degradation index reflecting the degree of change in module output consistency is generated. Time correlation analysis was performed on the continuous sampling data of interface temperature. Based on the trend of temperature continuously accumulating and rising with the duration of operation, the temperature rise rate index, which characterizes the degree of heat accumulation at the interface, was extracted. The response deviation index, consistency degradation index, and temperature rise rate index are comprehensively judged to output the judgment result used to characterize the degree of overall operational consistency degradation during power regulation.

6. The online fault early warning method for electric vehicle charging facilities according to claim 5, characterized in that, The process of marking potential fault information is as follows: The output judgment result is compared with the preset normal operation range to filter out the operation segments that do not meet the protection triggering conditions; Determine whether the running segment continuously exhibits a trend of consistency degradation, mark the running segments that meet the conditions, and generate corresponding potential fault indication information.

7. The online fault early warning method for electric vehicle charging facilities according to claim 6, characterized in that, The process of tracing potential fault information back to the corresponding charging gun interface, power module unit, and associated heat dissipation and communication nodes is as follows: The anomaly occurrence period is determined based on the time stamp contained in the potential fault indication information, and the charging gun interface and power module unit involved in operation are located within the anomaly occurrence period. Identify heat dissipation and communication nodes that have thermal conduction or communication connections with the power module unit, and establish a mapping relationship between potential fault indication information and corresponding physical units.

8. The online fault early warning method for electric vehicle charging facilities according to claim 7, characterized in that, The process of identifying a set of suspected faulty units with clearly defined associated boundaries is as follows: Obtain the electrical connection topology between the charging gun interface, power module unit and control communication unit inside the charging pile, and constrain the path range of abnormal propagation based on the physical layout of each unit in the cabinet; Along the direction of the abnormal influence pointed to by the mapping relationship, reverse convergence is performed on the units with direct electrical connection, thermal coupling or communication association, and the associated nodes that do not meet the topology are eliminated step by step. Units that maintain their correlation after reverse convergence are aggregated to form a set of suspected faulty units with clear correlation boundaries.

9. The online fault early warning method for electric vehicle charging facilities according to claim 8, characterized in that, The process of generating early warning information for specific components during the anomaly progression phase is as follows: The temporal synchronization degree and spatial distribution characteristics of the state changes of each node in the suspected fault unit set were statistically analyzed. Based on the degree of consistency disruption, key monitoring nodes are identified, the sampling frequency of data corresponding to key monitoring nodes is increased, and the sampling frequency of other nodes is reduced. When an anomaly does not trigger a protective action and is still in the gradual change phase, generate early warning information for specific components.

10. An online fault early warning system for electric vehicle charging facilities, applied to the method described in any one of claims 1-9, characterized in that, include: Stage Reordering Module: Reorders voltage, current, interface temperature, and communication delay data over time to form a sampling chain of operating status that reflects actual load changes; Coupled decomposition module: Cross-decomposes transient disturbances of electrical parameters, thermal response hysteresis, and control signal feedback offset to construct a combined operation description set; Degradation discrimination module: The consistency degradation discrimination model is used to jointly evaluate the control response speed, module output consistency and interface temperature rise accumulation behavior, and filter out abnormal segments and mark them as potential fault information. Fault backtracking module: Backtracks potential fault information to the charging gun interface, power module unit and associated heat dissipation and communication nodes to identify the set of suspected fault units; Early warning and scheduling module: Based on the degree of consistency disruption of the status changes of each node in the suspected fault unit set, it generates early warning information for specific components during the gradual abnormality stage.

Images