Hierarchical diagnosis method, system and device for charging module and storage medium

Abstract

The invention discloses a grading diagnosis method, system and device for a charging module and a storage medium, which are used for reducing unnecessary charging interruption and improving the adaptive capacity of the charging module to external disturbance. The method comprises the steps that when it is monitored that the voltage of a direct-current bus drops and the change rate exceeds a preset change rate threshold value, it is judged that a transient event is triggered, a triggering timestamp of the transient event is obtained, and a plurality of operation parameters at the moment before the transient event is triggered are obtained and determined as a stable baseline; evaluating a synchronous change trend of the plurality of operation parameters relative to the stable baseline, and classifying each operation parameter into a preset state label according to an evaluation result; based on the correlation among the plurality of operation parameters and the combination of the state labels, the event source category of the transient event is diagnosed, and the event source category at least comprises a power grid side sag event, a load side sudden load event and a module internal fault event; and executing a hierarchical response strategy matched with the event source category.

Description

Technical Field

[0001] This application relates to the field of power technology, and in particular to a method, system, device and storage medium for graded diagnosis of charging modules. Background Technology

[0002] With the rapid popularization of electric vehicles, energy storage systems, and various power electronic devices, DC fast charging technology has become a critical infrastructure. As an energy conversion unit connecting the grid side and the load side, the charging module is responsible for converting AC grid power into high-voltage DC power to charge these devices. Charging modules typically employ a two-stage or multi-stage topology with a front-end PFC and a rear-end DC-DC isolated converter. In these structures, the internal DC bus acts as an energy buffer between the two stages, and the stability of its voltage directly affects the operational safety of the entire charging module.

[0003] In actual operation, the charging module frequently faces disturbances from both the grid and the load side, both of which can cause instantaneous fluctuations in the internal DC bus voltage. To protect the internal power semiconductor devices, existing technologies typically employ a controller to monitor the DC bus voltage in real time and set fixed overvoltage and undervoltage protection thresholds for it.

[0004] However, the drawback of existing technologies lies in the fact that their protection mechanisms trigger a single and fixed response strategy. Once the controller detects that the DC bus voltage has reached the threshold, regardless of the specific cause of the fluctuation, the controller will execute a uniform protection action: immediately shutting down the upstream PFC and the downstream DC-DC circuit, locking the drive signal, and reporting a fault. This response strategy results in poor adaptability of the charging module to external disturbances, frequently causing unnecessary charging interruptions and severely reducing the reliability and continuity of charging. Summary of the Invention

[0005] This application provides a graded diagnostic method, system, device, and storage medium for charging modules, which reduces unnecessary charging interruptions and improves the adaptability of charging modules to external disturbances.

[0006] The first aspect of this application provides a graded diagnostic method for a charging module, comprising: The charging module's multiple operating parameters are collected synchronously, including AC input voltage, DC bus voltage, output current, and load-side power command. The DC bus voltage is continuously monitored, and the rate of change of the DC bus voltage in each control cycle is calculated. When a drop in the DC bus voltage is detected and the rate of change exceeds a preset rate of change threshold, a transient event is determined to be triggered, and the trigger timestamp of the transient event is obtained. Based on the trigger timestamp, obtain the multiple operating parameters at the moment before the transient event is triggered, and determine them as a stable baseline; Within a preset time window including the trigger timestamp, the synchronous change trend of the multiple operating parameters relative to the stable baseline is evaluated, and each operating parameter is classified into a preset status label based on the evaluation results. Based on the correlation between the multiple operating parameters and the combination of the status labels, the event source category of the transient event is diagnosed. The event source category includes at least grid-side sag events, load-side sudden load events, and module internal fault events. Implement a tiered response strategy that matches the event source category.

[0007] Optionally, the step of evaluating the synchronous change trend of the multiple operating parameters relative to the stable baseline within a preset time window including the trigger timestamp, and classifying each operating parameter into a preset status label based on the evaluation results, includes: Based on the differences between the multiple operating parameters collected within N sampling periods after the trigger timestamp and the stable baseline, a time feature matrix of size N×T is constructed, where N is the number of sampling periods and T is the number of operating parameters. For each element in the time feature matrix, a gradient value is calculated. The gradient value is used to quantify the magnitude and direction of change of each element in adjacent sampling periods. The gradient values ​​are symbolically encoded to form a trend symbol matrix; Based on the gradient value and the trend sign matrix, a change index for each operating parameter during the transient event is calculated. The change index is used to quantify and characterize the synchronous change trend of each operating parameter relative to the stable baseline. The change index is compared with a preset threshold range, and each operating parameter is classified into a preset status label.

[0008] Optionally, the step of constructing a time feature matrix of size N×T based on the differences between the multiple operating parameters collected within N sampling periods after the trigger timestamp and the stable baseline includes: The sampling sequence of the multiple operating parameters within N sampling periods after the trigger timestamp is analyzed to identify the transient response start point of each operating parameter. The transient response start point is the point of maximum gradient change or the point where the stable baseline is first crossed. Using the transient response start point of the DC bus voltage as the reference time point, calculate the time drift value of the transient response start point of other operating parameters relative to the reference time point; The sampling sequence of the multiple operating parameters is shifted and corrected based on the time drift value; Based on the difference between the sampled sequence after translation correction and the stable baseline, a time feature matrix of size N×T is constructed.

[0009] Optionally, diagnosing the event source category of the transient event based on the correlation between the plurality of operating parameters and the combination of the status labels includes: Based on the trend sign matrix, by calculating the forward correlation between the gradient change direction of any operating parameter at the current sampling time and the gradient change direction of another operating parameter at a subsequent sampling time, a directional matrix is ​​obtained to characterize the causal sequence relationship between the operating parameters. The status labels corresponding to the multiple operating parameters are combined with the directional matrix to form a first event feature vector; The first event feature vector is matched with a preset event source feature template, and the confidence score of each event source category is calculated based on the matching result. The category with the highest confidence score is determined as the event source category of the transient event.

[0010] Optionally, diagnosing the event source category of the transient event based on the correlation between the plurality of operating parameters and the combination of the status labels includes: Within a preset time window including the trigger timestamp, the difference between each operating parameter and the stable baseline is calculated to obtain a sequence of change amplitudes for each operating parameter; Calculate the transient correlation coefficient between every two operating parameters based on the change amplitude sequence; The status label corresponding to each operating parameter is combined with the transient correlation coefficient to form a second event feature vector; The second event feature vector is matched with a preset event source feature template, and the confidence score of each event source category is calculated based on the matching result. The category with the highest confidence score is determined as the event source category of the transient event.

[0011] Optionally, the execution of a hierarchical response strategy matching the event source category includes: If the event source category is the grid-side voltage dip event, then the front-end PFC circuit of the charging module is controlled to stop or reduce energy intake from the grid, and the control loop bandwidth of the DC bus voltage is dynamically adjusted to accelerate the stabilization of the DC bus voltage. If the event source category is the load-side sudden event, then the front-end PFC circuit is controlled to actively increase the energy drawn from the grid, and the output current limit value of the charging module is dynamically adjusted to suppress the impact, thereby stabilizing the DC bus voltage. If the event source category is an internal fault event of the module, the operation of the front-end PFC circuit and the rear-end DC-DC circuit shall be stopped immediately, and the connection with the power grid and the load shall be disconnected.

[0012] Optionally, the method further includes: The DC bus energy margin under the current operating condition is calculated based on the current output power of the charging module and the AC side input voltage. The preset rate of change threshold for triggering transient events is dynamically determined based on the DC bus energy margin, and the DC bus energy margin and the preset rate of change threshold are positively correlated.

[0013] A second aspect of this application provides a graded diagnostic system for a charging module, comprising: The acquisition unit is used to synchronously acquire multiple operating parameters of the charging module, including AC side input voltage, DC bus voltage, output current and load side power command. The monitoring unit is used to continuously monitor the DC bus voltage and calculate the rate of change of the DC bus voltage in each control cycle. The determination unit is used to determine that a transient event is triggered when the DC bus voltage drops and the rate of change exceeds a preset rate of change threshold, and to obtain the trigger timestamp of the transient event. The acquisition unit is used to acquire the plurality of operating parameters at the moment before the transient event is triggered, based on the trigger timestamp, and determine them as a stable baseline; The classification unit is used to evaluate the synchronous change trend of the multiple operating parameters relative to the stable baseline within a preset time window including the trigger timestamp, and classify each operating parameter into a preset state label according to the evaluation result. The diagnostic unit is used to diagnose the event source category of the transient event based on the correlation between the multiple operating parameters and the combination of the status labels. The event source category includes at least grid-side sag events, load-side surge events, and module internal fault events. An execution unit is used to execute a hierarchical response strategy that matches the event source category.

[0014] A third aspect of this application provides a graded diagnostic device for a charging module, the device comprising: Processor, memory, input / output units, and bus; The processor is connected to the memory, the input / output unit, and the bus; The memory stores a program, which the processor calls to execute the first aspect and any optional charging module graded diagnostic method in the first aspect.

[0015] The fourth aspect of this application provides a computer-readable storage medium storing a program that, when executed on a computer, performs the graded diagnostic method for a charging module of the first aspect and any optional method of the first aspect.

[0016] As can be seen from the above technical solutions, this application has the following advantages: By synchronously acquiring various operating parameters of the charging module and continuously monitoring changes in the DC bus voltage, a transient event can be accurately identified when the voltage drops rapidly and exceeds a preset rate of change threshold. Simultaneously, the stable state before the event is established as a baseline. Within the transient event trigger window, the changing trends of each operating parameter are evaluated and categorized into status labels. The correlation between parameters is then used to determine the event source category. Finally, different graded response strategies are selected based on the event source category, enabling targeted control actions for grid-side voltage dips, load-side surges, and internal module faults. In this way, the charging module can take different handling measures based on the specific cause of the transient event, rather than uniformly shutting down or locking the output. This reduces unnecessary charging interruptions, improves the charging module's adaptability to external disturbances, and maintains the continuity and stability of the DC bus voltage and output power under different event conditions. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A schematic flowchart of an embodiment of the graded diagnostic method for the charging module provided in this application; Figure 2 A schematic flowchart of an embodiment of the graded diagnosis method for the charging module provided in this application, which classifies each operating parameter into a preset status label; Figure 3 A schematic flowchart of an embodiment of the time alignment correction in the graded diagnostic method for the charging module provided in this application; Figure 4 A schematic flowchart of an embodiment of the graded diagnostic method for charging modules provided in this application for diagnosing the event source category of transient events; Figure 5 A schematic flowchart of another embodiment of the graded diagnostic method for charging modules provided in this application for diagnosing the event source category of the transient event; Figure 6A schematic diagram of an embodiment of the graded diagnostic system for the charging module provided in this application; Figure 7 This is a schematic diagram of an embodiment of the graded diagnostic device for the charging module provided in this application. Detailed Implementation

[0019] This application provides a graded diagnostic method, system, device, and storage medium for charging modules, which reduces unnecessary charging interruptions and improves the adaptability of charging modules to external disturbances.

[0020] Please see Figure 1 , Figure 1 An embodiment of the graded diagnostic method for charging modules provided in this application includes: 101. Synchronously collect multiple operating parameters of the charging module, including AC side input voltage, DC bus voltage, output current and load side power command; The charging module synchronously acquires key parameters during operation through its internal high-speed sampling unit. These parameters include AC input voltage, DC bus voltage, output current, and load-side power command. Specifically, the charging module can simultaneously read the instantaneous values ​​of each of these parameters within a fixed high-speed sampling period, such as 20μs to 50μs. In actual operation, synchronous sampling of these parameters is performed at the arrival of each control cycle to ensure consistent temporal relationships between them.

[0021] The purpose of synchronous sampling is to establish dynamic correspondences between multiple variables. For example, when the DC bus voltage drops rapidly, it can accurately correspond to whether the AC input voltage also drops at the same moment, or whether the output current suddenly increases, thus providing a basis for subsequent transient event identification and event source judgment. For instance, in an electric vehicle charging scenario, when a vehicle is first connected to a charging pile and begins to increase its charging power, the output current will rise rapidly in a short period of time; at the same time, load fluctuations on the grid side may also cause slight changes in the AC input voltage. These changes need to be synchronously collected under the same time reference. If the time references for collecting different operating parameters are inconsistent, errors will occur in subsequent analysis.

[0022] 102. Continuously monitor the DC bus voltage and calculate the rate of change of the DC bus voltage in each control cycle; After completing the synchronous acquisition of multiple operating parameters, the charging module needs to continuously monitor the DC bus voltage. In actual operation, the stability of the DC bus voltage directly reflects the grid input status, load-side power changes, and the internal operating status of the charging module. Taking a practical application scenario as an example, during the charging process of an electric vehicle, if the vehicle's BMS suddenly increases the charging demand, the output current will rise, and the DC bus voltage may drop accordingly. In another scenario, if the grid voltage experiences a temporary drop due to a large load being connected or line disturbances, the DC bus voltage will also be rapidly affected. Both of these events will cause transient changes in the DC bus voltage.

[0023] Therefore, it is necessary to record and process the sampled value of the DC bus voltage in each control cycle. In practice, the rate of change of the DC bus voltage within that control cycle can be calculated by comparing the sampled values ​​of the DC bus voltage in two consecutive adjacent control cycles to reflect the transient trend of voltage changes. Since the judgment of absolute amplitude often has a lag, compared with simply monitoring the absolute amplitude of the DC bus voltage (e.g., judging whether it is below the undervoltage threshold), the rate of change of voltage can more quickly reflect the severity of voltage fluctuations and can capture rapid voltage changes on a microsecond time scale.

[0024] 103. When a drop in DC bus voltage is detected and the rate of change exceeds a preset rate of change threshold, a transient event is determined to be triggered, and the trigger timestamp of the transient event is obtained. Based on the continuous monitoring of the DC bus voltage and its rate of change in step 102, when the system detects a significant drop in the DC bus voltage, and the rate of change of this drop exceeds a preset rate of change threshold, a transient event is determined to have been triggered in the current operating state. This transient event typically represents a sudden disturbance in the system. The causes of transient events are varied and not specifically limited here. It should be noted that this preset rate of change threshold is used to distinguish between normal minor fluctuations and true abnormal drops, filtering out voltage drift caused by conventional voltage ripple or slow load adjustments. This preset rate of change threshold can be specifically determined based on the structure of the charging module, the energy storage characteristics of the capacitor, and the target control performance.

[0025] To facilitate subsequent analysis, the trigger timestamp of a transient event needs to be recorded when it is determined to have occurred. This trigger timestamp marks the instant when the system switches from steady-state operation to transient disturbance, serving as the time reference point for all subsequent diagnostic analyses.

[0026] In some specific embodiments, the preset rate of change threshold can be set as a dynamic value, determined as follows: The DC bus energy margin under the current operating condition is calculated based on the current output power of the charging module and the AC side input voltage; the preset rate of change threshold for triggering transient events is dynamically determined based on the DC bus energy margin, and the DC bus energy margin and the preset rate of change threshold are positively correlated. In practical applications, when the output power is high or the input voltage is low, the bus energy margin is relatively small; while under low power or high input voltage conditions, the energy margin is large. By using a dynamic value that shows a positive correlation between the preset rate of change threshold and the bus energy margin, when the bus energy margin is large, large transient changes in the DC bus voltage can be allowed without misjudgment, therefore the preset rate of change threshold is relatively high; when the bus energy margin is small, the preset rate of change threshold is relatively lower to more sensitively capture voltage drops. This dynamic threshold setting can adapt to the transient event triggering conditions under different operating conditions, avoiding false triggering under high energy margin conditions and ensuring rapid response under low energy margin conditions, thereby improving the reliability and accuracy of transient event detection.

[0027] 104. Based on the trigger timestamp, obtain multiple running parameters at the moment before the transient event is triggered, and determine them as the stable baseline; When a transient event occurs, the recorded trigger timestamp is used to trace back one control cycle from the corresponding trigger timestamp. All operating parameters from the most recent stable control cycle before the transient event are extracted to ensure that the acquired operating parameter data is prior to the transient event and has not been affected by the transient disturbance. The acquired operating parameters represent the system's true operating state under normal steady-state conditions, and therefore can be used as the steady-state baseline. In practice, the charging module's controller typically maintains a circular data buffer to store historical sampling data from the most recent control cycles in real time. The controller indexes backward in the buffer based on the recorded trigger timestamp, reading the AC input voltage, DC bus voltage, output current, and load power command corresponding to the control cycle preceding the trigger time, and uses this set of historical data as the stable baseline.

[0028] The purpose of establishing a stable baseline is to provide an undisturbed reference standard for subsequent diagnostic analysis. When a transient event is determined to be triggered, the current operating parameters have actually undergone a sudden change, and the sampled values ​​at this time cannot represent the normal operating point of the system.

[0029] 105. Within a preset time window that includes the trigger timestamp, evaluate the synchronous change trend of multiple operating parameters relative to a stable baseline, and classify each operating parameter into a preset status label based on the evaluation results. After obtaining the trigger timestamp of the transient event and the stable baseline of the previous control cycle, a systematic analysis of the changes in operating parameters needs to be performed over a period of time before and after the trigger time. First, a preset time window is set around the trigger timestamp. This preset time window can cover several control cycles after the transient event, for example, 5 to 10 sampling points. By reading the real-time sampled values ​​of the AC input voltage, DC bus voltage, output current, and load power command within this preset time window, the deviation of these sampled values ​​from the stable baseline determined in step 104 can be calculated, i.e., the synchronous change trend of multiple operating parameters relative to the stable baseline can be evaluated.

[0030] Based on the calculated deviation and direction of change, the dynamic performance of each operating parameter is mapped to discrete state labels. In practice, each operating parameter can be evaluated through methods such as abrupt change detection, slope assessment, waveform morphology matching, or correlation judgment. Abrupt change detection detects whether the parameter's jump amplitude exceeds a set threshold within a very short time, identifying rapidly changing quantities such as inductor current surges or sudden drops in bus voltage. Slope assessment calculates the rate of change trend of the parameter within a time window, determining whether the change is gradually increasing, gradually decreasing, or remaining stable. Waveform morphology matching compares the shapes of voltage and current waveforms before and after disturbances, matching them with typical disturbance patterns such as "step drop," "sawtooth fluctuation," and "gradual rise and fall." Correlation judgment assesses whether there is a synchronous offset in the operating parameter, identifying group disturbances (such as grid flicker) and individual disturbances (such as single module failures). Subsequently, based on the evaluation results, each operating parameter is mapped to a predefined state label system. The preset state labels may include, but are not limited to, "stable," "rising," "falling," "abrupt," and "abnormal." This processing method can transform continuous analog numerical values ​​into qualitative states that are easy for computer logic to process.

[0031] 106. Based on the correlation between multiple operating parameters and the combination of status labels, diagnose the event source category of transient events. The event source category includes at least grid-side sag events, load-side sudden load events, and module internal fault events. Transient events may originate from external grid disturbances to the charging module, sudden power changes on the load side, or control anomalies or device failures within the module itself. A single parameter cannot distinguish the fault source. A DC bus voltage drop is merely a consequence of a disruption in the system's energy balance. This result could be caused by either a reduction in input energy supply or excessively rapid output energy consumption. Monitoring data of the DC bus voltage alone cannot differentiate between these two factors; therefore, multi-parameter correlation analysis is necessary to achieve this distinction.

[0032] After classifying the status labels of each operating parameter in step 105, the specific source of the transient event can be determined based on the correlation between multiple operating parameters and the combination of these status labels, thereby identifying the root cause of the DC bus voltage drop. Specifically, the controller has pre-stored feature logic tables or matching rules for different event sources. The multiple status labels generated in step 105 are compared as a combined feature with the pre-stored rules. During this process, the correlation between different parameters at the moment of change can also be combined, such as the synchronization degree between the AC input voltage change trend and the DC bus voltage change trend on the time axis, to help confirm the causal relationship of the transient event. By introducing status labels for the AC side input voltage and output current, a complete logic chain containing input state, intermediate state, and output state can be constructed, thereby decoupling grid-side disturbances and load-side disturbances and achieving accurate fault source location. The following is an example of a specific scenario: A grid-side transient event occurs when the AC input voltage suddenly changes or drops sharply, and its status label shows "sudden change" or "drop," while the DC bus voltage also drops synchronously with a similar slope. If the output current and load power command do not show a positive correlation increase at the moment the event is triggered, it can be inferred that the transient event is caused by a disturbance at the grid input. Such events typically exhibit the characteristic of "synchronous decrease in input voltage and bus voltage."

[0033] A load-side power surge event occurs when the vehicle's BMS or other load equipment suddenly increases its power demand, causing a rapid rise in output current. The status label is typically "rising" or "sudden change." Simultaneously, the DC bus voltage usually drops to some extent due to the load pull, but the AC input voltage mostly remains stable without significant fluctuations. If a significant jump in the load power command is detected before and after the event triggers, and the change in output current is highly correlated with the change in this command, the event source can be identified as a load-side power surge event.

[0034] Internal module fault events, when transient events occur, if the AC input voltage and load power command remain stable, but the changes in internal parameters such as DC bus voltage, output current, or duty cycle exhibit abnormal patterns—such as no sudden changes in external commands, parameter changes that do not conform to electrical laws, or only a single module's internal parameters showing a drastic shift—can be inferred to originate from internal faults such as component failure, clock synchronization instability, or abnormal drive triggering. These events typically exhibit the characteristic of stable external variables while internal variables are abnormal.

[0035] By identifying the above-mentioned combination of features, the system can classify transient events into preset event source categories. It should be noted that this application is not limited to the three event source categories mentioned above. In practical applications, the event source types can be expanded according to requirements, such as grid-connected harmonic impulses, EMI-triggered disturbances, inter-module coupling interference, and other event modes. The core of the above diagnostic mechanism lies in utilizing the dual information of status labels and parameter correlations, so that the diagnostic process no longer relies on the absolute change of a single variable, but rather recovers the true source of the event through the synergistic relationship between variables, thereby significantly improving the accuracy and reliability of event source identification.

[0036] 107. Implement a tiered response strategy that matches the event source category.

[0037] Different event source categories imply different physical mechanisms of DC bus voltage drops, thus requiring differentiated control measures to maintain stable system operation and ensure safety. Once the charging module diagnoses a specific event source category, it can immediately invoke preset control logic to execute a tiered response strategy corresponding to that event source category. This avoids a one-size-fits-all approach of complete shutdown, enabling the charging module to respond more flexibly and intelligently to different types of transient events, thereby improving system availability and operational continuity.

[0038] In this embodiment, by synchronously acquiring various operating parameters of the charging module and continuously monitoring changes in the DC bus voltage, a transient event can be accurately determined when the voltage drops rapidly and exceeds a preset rate of change threshold. Simultaneously, the stable state before the event is established as a baseline. Within the transient event trigger window, the changing trends of each operating parameter are evaluated and categorized into status labels. The correlation between parameters is then used to determine the event source category. Finally, different graded response strategies are selected based on the event source category, enabling targeted control actions to be executed for grid-side voltage dips, load-side surges, and internal module faults. In this way, the charging module can take different handling measures based on the specific cause of the transient event, rather than uniformly shutting down or locking the output. This reduces unnecessary charging interruptions, improves the charging module's adaptability to external disturbances, and maintains the continuity and stability of the DC bus voltage and output power under different event conditions.

[0039] In some specific embodiments, the hierarchical response strategy can be set as follows: In response to grid-side voltage dips, charging modules typically do not need to shut down immediately. Instead, they employ mild intervention strategies, such as appropriately reducing power output, delaying dynamic adjustments in the control loop, and temporarily increasing the tolerance of the bus voltage loop to enhance the module's adaptability to grid disturbances. In practical applications, when a momentary drop in grid voltage occurs at an electric vehicle charging station, the charging module can maintain stable output for 10–20 ms, weathering the brief disturbance by adjusting the boost duty cycle, thus avoiding frequent entry into protection mode due to external grid fluctuations. Specifically, if the event source is a grid-side voltage dip, the front-end PFC circuit controlling the charging module stops or reduces power intake from the grid and dynamically adjusts the control loop bandwidth of the DC bus voltage to accelerate DC bus voltage stabilization.

[0040] In response to sudden load surges, the system should prioritize ensuring the responsiveness to the demands of the vehicle or downstream load. Therefore, moderate adjustment strategies can be implemented, such as rapidly adjusting the output current control loop, reallocating output power limits, and enhancing transient compensation capabilities, to ensure the module can smoothly handle power surges on the load side. For example, in scenarios where electric vehicles suddenly increase their charging power, the system can automatically increase the bandwidth of the current loop or enable additional feedforward compensation strategies to quickly track load changes and reduce disturbances to the bus voltage. Specifically, if the event source is a sudden load surge, the system controls the upstream PFC circuit to actively increase power intake from the grid and dynamically adjusts the output current limit of the charging module to suppress the surge, thereby stabilizing the DC bus voltage.

[0041] For internal module fault events, which may involve device failure, drive anomalies, control loop divergence, etc., the highest level of safety response strategy must be implemented. This includes immediately reducing the module's output power, shutting down high-risk switching devices, switching to degraded operation mode, and even stopping the module operation and reporting the fault information to the superior system if necessary. For example, in a real-world scenario, when abnormal jitter in the drive signal or an abnormal drop in bus voltage despite a stable input is detected, the system can immediately enter a safety protection mode to prevent further damage to power devices or bus oscillation. Specifically, if the event source is an internal module fault event, the operation of the upstream PFC circuit and the downstream DC-DC circuit will be immediately stopped, and the connection to the power grid and load will be disconnected.

[0042] In step 105, state label classification can be achieved by constructing a temporal feature matrix and performing gradient calculations. Please refer to [link to relevant documentation]. Figure 2 , Figure 2 A schematic flowchart of an embodiment of the graded diagnostic method for charging modules provided in this application, which categorizes each operating parameter into a preset status label, is shown. The method includes: 201. Based on the differences between multiple operating parameters collected within N sampling periods after the trigger timestamp and the stable baseline, construct a time feature matrix of size N×T, where N is the number of sampling periods and T is the number of operating parameters; After determining the trigger timestamp and stable baseline, the charging module standardizes the raw sampled data. Specifically, it extracts multiple operating parameter data within N sampling cycles (e.g., 5 consecutive cycles) starting from the trigger timestamp. Then, it subtracts the stable baseline value determined in step 104 from each of these real-time sampled data points, obtaining a set of relative differences. Finally, it uses these differences to construct an N-row, T-column two-dimensional data structure, namely the time feature matrix. Here, N represents the length of the time dimension, and T represents the number of parameter dimensions, for example, four dimensions: AC voltage, bus voltage, output current, and power command.

[0043] The purpose of constructing this time characteristic matrix is ​​to eliminate the interference of the static operating point on transient analysis. For example, regardless of whether the charging module is currently operating at 380V or 400V input, as long as a drop of the same magnitude occurs, its difference relative to the baseline should be comparable. By converting absolute values ​​into relative deviations, this time characteristic matrix can purely reflect the dynamic deviation of each operating parameter during a transient event.

[0044] 202. Calculate the gradient value for each element in the time feature matrix. The gradient value is used to quantify the magnitude and direction of change of each element in adjacent sampling periods. To more accurately describe the rate of change of operating parameters during transient events, gradient values ​​can be calculated from the differences between adjacent sampling periods in the time feature matrix. These gradient values ​​have both magnitude and sign: the magnitude represents the severity of the change, and the sign indicates the direction of change (increasing or decreasing). In the early stages of a transient event, although the absolute deviation of some parameters may not yet have accumulated significantly, their rate of change (gradient) is often already very significant. For example, at the instant of a sudden load increase, the absolute value of the output current may only increase by 5A, but its rise rate on the microsecond scale is extremely high. By quantifying this characteristic, the system can capture the kinetic energy of transient changes and identify whether the signal is in an accelerating deviation or an oscillating recovery phase.

[0045] 203. Encode the gradient values ​​with symbols to form a trend symbol matrix; In power electronic systems, high-frequency switching noise is unavoidable, and directly using the raw gradient values ​​can easily lead to misjudgments. To further improve computational efficiency and simplify subsequent trend judgment, gradient values ​​are symbolized according to their direction. Specifically, positive gradients are encoded as positive symbols, negative gradients as negative symbols, and gradients with minimal changes are encoded as stable symbols. By symbolizing the gradient values ​​of all operating parameters, a trend symbol matrix can be formed to describe the change direction pattern of each parameter at different times during transient events. In specific implementation, a dead zone threshold can be preset to filter out inherent white noise or small ripple fluctuations in the sensor. When the calculated gradient value falls within this dead zone range, it is encoded as "0" (stable); when the gradient value exceeds the positive threshold, it is encoded as "+1" (rising trend); and when the gradient value exceeds the negative threshold, it is encoded as "-1" (falling trend). After this processing, the originally complex floating-point gradient matrix can be transformed into a trend symbol matrix composed of {-1, 0, +1}. Furthermore, the trend symbol matrix also helps to identify typical change patterns in a shorter time. For example, a continuous negative sign sequence of AC input voltage often corresponds to a grid sag event, while a continuous positive sign sequence of output current is highly correlated with a sudden load increase.

[0046] 204. Based on the gradient value and trend sign matrix, calculate the change index of each operating parameter during the transient event. The change index is used to quantify and characterize the synchronous change trend of each operating parameter relative to the stable baseline. After completing gradient calculation and trend symbolization, it is necessary to combine the gradient magnitude and trend direction to further calculate the change index of each operating parameter during the entire transient event. The change index can comprehensively reflect the cumulative degree of parameter change, speed characteristics, and trend stability. The purpose of calculating the change index is to compress the complex waveform characteristics within a time window into a single quantitative score. Specifically, the change index can be calculated by associating the cumulative absolute value of the gradient, the persistence of the trend symbol, and the peak magnitude of the difference. For example, for AC input voltage, if a trend symbol of "-1" appears consecutively within N cycles and the gradient value is large, the cumulatively calculated change index will be a large negative value, intuitively reflecting that the voltage has undergone a sharp and continuous decline.

[0047] This change index not only contains information about how much it has changed, but also implies information about how it changed. Therefore, in practical applications, the change index can effectively describe the intensity of the impact of transient events on various parameters. For example, in an internal fault scenario, the DC bus voltage may continuously decrease and exhibit abnormally large gradients and negative sign sequences, resulting in a change index for this parameter that is significantly higher than other event modes.

[0048] 205. Compare the changing indicators with the preset threshold range, and classify each operating parameter into a preset status label.

[0049] Finally, the calculated change index for each operating parameter is compared with a preset threshold range to output a final qualitative conclusion, i.e., a status label. The preset threshold range defines the numerical boundaries of different states. For example, if the change index of AC voltage is below a certain negative threshold, it is labeled "decreasing"; if it is within a small range near zero, it is labeled "stable"; and if it is above a certain positive threshold, it is labeled "increasing". These status labels provide a clear and structured basis for subsequent event source diagnosis, enabling the determination of the source of transient events based on the joint change pattern between parameters.

[0050] In this embodiment, after a transient event occurs, dynamic features with temporal correlation are extracted from real-time data of multiple operating parameters. These features are then compressed through gradient analysis and trend symbolization to ultimately form highly discriminative change indicators and status labels. Transforming the originally complex and noisy transient waveform into structured parameter trend features enables rapid determination of the change patterns of different operating parameters during transient events without relying on large amounts of computational resources. This provides a highly reliable data foundation and judgment basis for the entire hierarchical diagnostic method.

[0051] In step 201 above, time alignment correction can be performed before constructing the time feature matrix to eliminate potential start-up delays or phase differences that may occur in the transient response of different operating parameters, thereby ensuring that the dynamic changes of each parameter are aligned in the time dimension. Please refer to [link to relevant documentation]. Figure 3 , Figure 3 A schematic flowchart of an embodiment of the graded diagnostic method for the charging module provided in this application for time alignment correction is shown. The method includes: 301. Analyze the sampling sequences of multiple operating parameters within N sampling periods after the trigger timestamp to identify the transient response start point of each operating parameter. The transient response start point is the point of maximum gradient change or the point where the stable baseline is first crossed. After a transient event is triggered, each operating parameter may begin to respond with a different delay or rate. To accurately align the transient behavior of each parameter, the system needs to analyze a sequence of N sampling periods after the trigger timestamp to identify the transient response start point for each parameter. The transient response start point can be defined as the point where the parameter's gradient change is largest in the time series (indicating the maximum acceleration of change) or the point where it first crosses the stable baseline (indicating the first significant deviation from the steady state). For example, in a sudden load increase scenario, the output current usually responds immediately, while the bus voltage may lag slightly due to the energy storage characteristics of the capacitor. This step can identify the initial response time of each parameter, providing a benchmark for subsequent time alignment.

[0052] 302. Using the transient response start point of the DC bus voltage as the reference time point, calculate the time drift value of the transient response start point of other operating parameters relative to the reference time point; Among the determined transient response start points for each parameter, the start point of the DC bus voltage is selected as the reference time point. This is because the bus voltage typically has a global representativeness of the transient behavior of the entire module and is the source signal that triggers transient events; its timestamp is closest to the actual physical moment of the event. The time difference between the transient response start points of other operating parameters and the reference point of the bus voltage is the time drift value. This time drift value quantifies the relative lag or lead of different signals during transmission and response. For example, if the DC bus voltage starts to drop at the 0th sampling point, while the output current only begins to rise significantly at the 3rd sampling point, then the time drift value of the output current is +3 cycles.

[0053] 303. Perform translation correction on the sampling sequence of multiple operating parameters based on the time drift value; After obtaining the time drift values ​​of each operating parameter, a translation correction is performed on each sampled sequence to align the transient response start point of all operating parameters with the reference time point of the DC bus voltage. Specifically, this can be achieved through methods such as sequence interpolation, cyclic buffering, or zero-filling to adjust the ahead or behind sequences to a unified start point in the time dimension.

[0054] Translation correction mathematically eliminates the impact of hardware latency, forcing the starting points of all operating parameters to coincide on the time axis. For example, after translation correction, the rising edge of the output current will align with the falling edge of the DC bus voltage on the same sampling index. This allows subsequent analysis to focus on the shape and amplitude correlation of the waveform without being misled by phase misalignment, greatly improving the reliability of multivariate correlation analysis.

[0055] 304. Based on the difference between the shift-corrected sampling sequence and the stable baseline, construct a time feature matrix of size N×T.

[0056] After completing the translation correction, the difference between each parameter sequence and the stable baseline is calculated again to obtain the corrected relative deviation. A new time feature matrix is ​​then constructed according to the structure of N sampling periods × T parameter dimensions. Compared with the original matrix in step 201, this matrix obtained in step 304 can more accurately reflect the synchronous dynamic deviation of each operating parameter during transient events, providing a more accurate data foundation for subsequent gradient analysis, trend symbolization, and calculation of change indicators. The results show that when analyzing grid sag events, the time-aligned AC input voltage and bus voltage curves can simultaneously show a downward trend, avoiding misjudgment or confusion caused by response lag.

[0057] In this embodiment, a timing alignment correction mechanism for multiple physical quantities is introduced. In practical power electronic systems, due to limitations in sensor bandwidth, ADC sampling channel sequence, and the inductive or capacitive inertia of the circuit itself, the responses of different operating parameters to the same transient disturbance often exhibit asynchronous physical time, i.e., a phase difference. This embodiment effectively eliminates this hardware-introduced physical time delay at the data processing level by identifying the transient response start point of each operating parameter and performing translation correction based on the DC bus voltage reference. This ensures that the final constructed time feature matrix is ​​strictly aligned based on event causal logic, thereby enabling subsequent gradient calculation, trend symbolization, and change index analysis to accurately reflect the synchronous change characteristics of each parameter.

[0058] In some specific embodiments, a more accurate event source diagnosis result can be obtained by analyzing the causal sequence between operating parameters, constructing a first event feature vector based on status labels, and matching it with a preset event source feature template. Please refer to [link to relevant documentation]. Figure 4 , Figure 4 A schematic flowchart of an embodiment of the graded diagnostic method for charging modules provided in this application for diagnosing the event source category of transient events, the method including: 401. Based on the trend sign matrix, by calculating the forward correlation between the gradient change direction of any operating parameter at the current sampling time and the gradient change direction of another operating parameter at subsequent sampling times, a directional matrix is ​​obtained to characterize the causal sequence relationship between the operating parameters. After obtaining the trend sign matrix generated in step 203, the trend sign matrix can be used as input data to calculate the forward correlation between the gradient direction (rising / falling / stationary) of any parameter at the current sampling time and the gradient direction of another parameter at subsequent sampling times. This calculation quantifies the causal sequence of changes between parameters, i.e., determining which parameter changes before another and may affect it. Subsequently, the causal relationships between all operating parameters are summarized to form a directional matrix. In practical applications, this directional matrix can reveal the temporal logic of parameter changes in transient events. For example, in a load surge event, the change in power command usually precedes the change in output current; while in a grid sag event, the change in AC input voltage precedes the decrease in bus voltage.

[0059] In its implementation, the trend sign matrix consists of a series of discrete signs (e.g., +1, 0, -1), describing the trajectory of each parameter along the time axis. Based on this, pairwise analysis is performed on each pair of operating parameters P_i and P_j, where P_i is the reference parameter and P_j is the parameter to be analyzed. For each sampling time t_k (belonging to the sequence of P_i), a certain number of sampling points forward or backward are selected as the analysis window. For example, P_j can be selected from t_k+1 to t_k+m subsequent sampling points to capture potential hysteresis responses. The window size m can be determined based on the sampling period and the duration of the transient characteristic. Within each window, the sign of P_i(t_k) is compared with the sign of P_j(t_k+l) (l=1,...,m), and the proportion of identical signs is counted. Identical signs are defined as both being +1 or both being -1; inconsistent signs are one positive and one negative; a stationary sign of 0 is considered neutral and not counted in positive matching. A high consistency ratio indicates that the change in P_i may precede and influence the change in P_j; a low consistency ratio indicates that the changes in the two parameters are not temporally correlated or have a weak causal relationship. Finally, the forward correlation values ​​of all parameter pairs are summarized to form a directionality matrix D. Each element Dij in the matrix represents the probability or causal strength that the change in parameter P_i precedes that of P_j, and its value range can be normalized to 0–1, with the closer to 1 indicating a stronger causal relationship.

[0060] 402. Combine the status labels corresponding to multiple operating parameters with the directional matrix to form the first event feature vector; After obtaining the directional matrix, the system combines the state label (rising / falling / stable) of each operating parameter with its causal relationship information in the directional matrix to form a high-dimensional event feature vector, namely the first event feature vector. This first event feature vector not only includes the magnitude and direction of change of each operating parameter but also incorporates temporal sequence and causal logic, making the feature representation of transient events richer and more structured. For example, for the cases of grid voltage dips and load surges, even though their state labels may be similar under certain extreme conditions (e.g., both showing a bus voltage drop), their causal orientations in the directional matrix are completely opposite: the former is the input guiding bus, and the latter is the output guiding bus. This combination of first event feature vectors can completely eliminate the ambiguity that may be caused by single-dimensional features.

[0061] 403. Match the first event feature vector with the preset event source feature template, and calculate the confidence score of each event source category based on the matching results. The category with the highest confidence score is determined as the event source category of the transient event.

[0062] After constructing the event feature vector, it is matched against a pre-established event source feature template. The event source feature template contains typical feature patterns corresponding to different event source types (such as grid sags, load surges, and internal module faults). Specific matching methods can include cosine similarity, Euclidean distance, or other similarity metrics; no specific method is limited here. Based on the matching results, a confidence score is calculated for each event source category to quantify the degree of conformity between that category and the current event features. Finally, the category with the highest confidence score is identified as the event source of the transient event. This feature vector matching method is more robust than ordinary hard logic judgment. Even if the field signal has certain distortions or atypical features, as long as its core causal orientation and state combination are consistent with the template in the general direction, a high-confidence correct diagnosis can still be obtained, thus adapting to complex and ever-changing actual power grid and load environments.

[0063] Traditional correlation analysis focuses only on the synchronicity of numerical changes, easily confusing cause and effect. In this embodiment, by calculating forward correlation based on a trend sign matrix and constructing a directional matrix, the sequence of physical changes is introduced into the diagnostic logic, clearly identifying the transmission path of disturbances. This is then combined with a high-dimensional feature vector formed by state labels, enabling accurate identification even when faced with complex faults with similar characteristics, significantly improving diagnostic accuracy and anti-interference capabilities.

[0064] In other specific embodiments, event source diagnosis can also be achieved by analyzing the transient amplitude and correlation of operating parameters and constructing a second event feature vector in conjunction with status labels. Please refer to [link to relevant documentation]. Figure 5 , Figure 5 A flowchart illustrating another embodiment of the graded diagnostic method for charging modules provided in this application for diagnosing the event source category of transient events, the method comprising: 501. Within a preset time window including the trigger timestamp, calculate the difference between each operating parameter and the stable baseline to obtain a sequence of change amplitudes for each operating parameter; After a transient event occurs, a preset time window is first determined, starting from the trigger timestamp, for example, several sampling periods before or after the trigger time. Then, for each operating parameter, the difference between its instantaneous sampled value within the time window and the stable baseline determined in step 104 is calculated. These difference sequences reflect the amplitude deviation of each parameter during the transient event. For example, in a scenario of sudden load increase, the amplitude difference sequence of the output current will rise rapidly, while the amplitude difference sequence of the bus voltage may show a decreasing trend. By recording the difference for each sampling period, the dynamic amplitude changes of the parameters during the transient event can be fully described, providing a quantitative basis for subsequent analysis.

[0065] 502. Calculate the transient correlation coefficient between every two operating parameters based on the change amplitude sequence; Subsequently, the transient correlation coefficient between each pair of operating parameters is calculated based on the amplitude change sequence. This transient correlation coefficient is a value between -1 and +1, used to quantify the similarity in shape and the synchronicity in the direction of change of the two parameter waveforms. Specifically, it can be calculated using Pearson correlation coefficient, Spearman rank correlation coefficient, or other correlation measurement methods suitable for short-time series; no specific limitation is made here. A correlation coefficient close to +1 indicates that the amplitudes of the two parameters change synchronously; for example, a decrease in bus voltage and a decrease in AC input voltage may be highly correlated. A correlation coefficient close to -1 indicates that the amplitudes change in opposite directions; for example, an increase in output current is accompanied by a decrease in bus voltage. A correlation coefficient close to 0 indicates that the two parameters have no obvious synchronicity.

[0066] In power electronic circuits, specific fault sources can induce waveforms with specific coupling characteristics across different parameters. For example, when a grid voltage dip occurs, the AC input voltage dip waveform is highly similar in shape and synchronous with the DC bus voltage dip waveform, resulting in a transient correlation coefficient close to +1. Conversely, when a sudden load increase causes a bus voltage dip, the output current exhibits an increasing waveform, while the bus voltage exhibits a decreasing waveform, showing an inverse relationship, leading to a transient correlation coefficient close to -1. By calculating the transient correlation coefficient, this physical coupling characteristic can be captured numerically.

[0067] 503. Combine the status label corresponding to each operating parameter with the transient correlation coefficient to form the second event feature vector; After obtaining the transient correlation coefficient, it is combined with the state label to construct a high-dimensional event feature vector, namely the second event feature vector. Specifically, the state label can be encoded as discrete values ​​(such as -1, 0, +1), and then concatenated with the correlation coefficient matrix between each operating parameter to form a vector or expanded matrix. This allows the second event feature vector to simultaneously contain information on the direction and degree of amplitude change, as well as the transient correlation between parameters.

[0068] 504. Match the second event feature vector with the preset event source feature template, and calculate the confidence score of each event source category based on the matching results. Determine the category with the highest confidence score as the event source category of the transient event.

[0069] The matching method in step 504 is similar to that in step 403 above, and will not be repeated here.

[0070] It should be noted that the first event feature vector is mainly constructed based on the trend symbol matrix and the state labels of operating parameters. It emphasizes the directionality and causal sequence of parameter changes in transient events, accurately capturing the time-series logic of parameter changes, such as which parameter responds first and which follows, thus reflecting the dynamic causal pattern of transient events in event source determination. In contrast, the second event feature vector focuses on the amplitude changes of parameters and the correlation between amplitudes. Combined with state labels to form a feature vector, it emphasizes the synchronicity or inverse relationship of the amplitudes of parameters when the event occurs. The advantage of the first event feature vector is that it can identify the time-series characteristics of events under conditions of small changes or delayed responses, making it suitable for distinguishing time-sequence sensitive event types, such as the causal response patterns of grid sags and load surges. The advantage of the second event feature vector is that it can quantify the intensity of amplitude changes and the amplitude relationship between parameters, making it suitable for identifying typical event sources through amplitude patterns and correlations in the presence of noise or switching fluctuations. Therefore, the two feature vectors have different focuses: the first feature vector emphasizes the temporal causal sequence, while the second feature vector emphasizes the synchronous relationship of amplitudes. In practical applications, the two can also be combined. First, the first event feature vector and the second event feature vector are matched with their respective event source feature templates to calculate the confidence score. Then, the confidence scores obtained from the two types of feature vectors are weighted and fused or logically determined. The event source category with the highest comprehensive fusion confidence score is determined as the event source of the transient event.

[0071] The graded diagnostic system for the charging module provided in this application is described in detail below. Please refer to [link / reference]. Figure 6 , Figure 6 Another embodiment of the graded diagnostic system for the charging module provided in this application includes: The acquisition unit 601 is used to synchronously acquire multiple operating parameters of the charging module, including AC side input voltage, DC bus voltage, output current and load side power command. The monitoring unit 602 is used to continuously monitor the DC bus voltage and calculate the rate of change of the DC bus voltage in each control cycle. The determination unit 603 is used to determine the triggering of a transient event when a drop in DC bus voltage is detected and the rate of change exceeds a preset rate of change threshold, and to obtain the triggering timestamp of the transient event. The acquisition unit 604 is used to acquire multiple running parameters at the moment before the transient event is triggered based on the trigger timestamp, and determine them as a stable baseline; The classification unit 605 is used to evaluate the synchronous change trend of multiple operating parameters relative to a stable baseline within a preset time window containing a trigger timestamp, and classify each operating parameter into a preset status label based on the evaluation results. The diagnostic unit 606 is used to diagnose the event source category of transient events based on the correlation between multiple operating parameters and the combination of status labels. The event source category includes at least grid-side sag events, load-side surge events, and module internal fault events. Execution unit 607 is used to execute a hierarchical response strategy that matches the event source category.

[0072] In this embodiment, the functions of each unit are the same as described above. Figures 1 to 5 The steps in the method embodiments shown correspond to those in the examples, and will not be repeated here.

[0073] This application also provides a graded diagnostic device for a charging module; please refer to [link / reference]. Figure 7 , Figure 7 One embodiment of the graded diagnostic device for the charging module provided in this application includes: Processor 701, memory 702, input / output unit 703, bus 704; The processor 701 is connected to the memory 702, the input / output unit 703, and the bus 704; The memory 702 stores a program, and the processor 701 calls the program to execute the graded diagnostic method for any of the charging modules described above.

[0074] This application also relates to a computer-readable storage medium on which a program is stored, which, when run on a computer, causes the computer to perform the graded diagnostic method for any of the charging modules described above.

[0075] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0076] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, or indirect coupling or communication connection between apparatuses or units, and may be electrical, mechanical, or other forms.

[0077] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0078] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0079] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A graded diagnostic method for a charging module, characterized in that, The graded diagnostic method includes: The charging module's multiple operating parameters are collected synchronously, including AC input voltage, DC bus voltage, output current, and load-side power command. The DC bus voltage is continuously monitored, and the rate of change of the DC bus voltage in each control cycle is calculated. When a drop in the DC bus voltage is detected and the rate of change exceeds a preset rate of change threshold, a transient event is determined to be triggered, and the trigger timestamp of the transient event is obtained. Based on the trigger timestamp, obtain the multiple operating parameters at the moment before the transient event is triggered, and determine them as a stable baseline; Within a preset time window including the trigger timestamp, the synchronous change trend of the multiple operating parameters relative to the stable baseline is evaluated, and each operating parameter is classified into a preset status label based on the evaluation results. Based on the correlation between the multiple operating parameters and the combination of the status labels, the event source category of the transient event is diagnosed. The event source category includes at least grid-side sag events, load-side sudden load events, and module internal fault events. Implement a tiered response strategy that matches the event source category.

2. The grading diagnostic method according to claim 1, characterized in that, Within a preset time window including the trigger timestamp, the synchronous change trend of the multiple operating parameters relative to the stable baseline is evaluated, and each operating parameter is classified into a preset status label based on the evaluation results, including: Based on the differences between the multiple operating parameters collected within N sampling periods after the trigger timestamp and the stable baseline, a time feature matrix of size N×T is constructed, where N is the number of sampling periods and T is the number of operating parameters. For each element in the time feature matrix, a gradient value is calculated. The gradient value is used to quantify the magnitude and direction of change of each element in adjacent sampling periods. The gradient values ​​are symbolically encoded to form a trend symbol matrix; Based on the gradient value and the trend sign matrix, a change index for each operating parameter during the transient event is calculated. The change index is used to quantify and characterize the synchronous change trend of each operating parameter relative to the stable baseline. The change index is compared with a preset threshold range, and each operating parameter is classified into a preset status label.

3. The graded diagnostic method according to claim 2, characterized in that, The step of constructing a time feature matrix of size N×T based on the differences between the multiple operating parameters collected within N sampling periods after the trigger timestamp and the stable baseline includes: The sampling sequence of the multiple operating parameters within N sampling periods after the trigger timestamp is analyzed to identify the transient response start point of each operating parameter. The transient response start point is the point of maximum gradient change or the point where the stable baseline is first crossed. Using the transient response start point of the DC bus voltage as the reference time point, calculate the time drift value of the transient response start point of other operating parameters relative to the reference time point; The sampling sequence of the multiple operating parameters is shifted and corrected based on the time drift value; Based on the difference between the sampled sequence after translation correction and the stable baseline, a time feature matrix of size N×T is constructed.

4. The grading diagnostic method according to claim 2, characterized in that, The method of diagnosing the event source category of the transient event based on the correlation between the multiple operating parameters and the combination of the status labels includes: Based on the trend sign matrix, by calculating the forward correlation between the gradient change direction of any operating parameter at the current sampling time and the gradient change direction of another operating parameter at a subsequent sampling time, a directional matrix is ​​obtained to characterize the causal sequence relationship between the operating parameters. The status labels corresponding to the multiple operating parameters are combined with the directional matrix to form a first event feature vector; The first event feature vector is matched with a preset event source feature template, and the confidence score of each event source category is calculated based on the matching result. The category with the highest confidence score is determined as the event source category of the transient event.

5. The grading diagnostic method according to claim 1, characterized in that, The method of diagnosing the event source category of the transient event based on the correlation between the multiple operating parameters and the combination of the status labels includes: Within a preset time window including the trigger timestamp, the difference between each operating parameter and the stable baseline is calculated to obtain a sequence of change amplitudes for each operating parameter; Calculate the transient correlation coefficient between every two operating parameters based on the change amplitude sequence; The status label corresponding to each operating parameter is combined with the transient correlation coefficient to form a second event feature vector; The second event feature vector is matched with a preset event source feature template, and the confidence score of each event source category is calculated based on the matching result. The category with the highest confidence score is determined as the event source category of the transient event.

6. The grading diagnostic method according to claim 1, characterized in that, The execution of the hierarchical response strategy matching the event source category includes: If the event source category is the grid-side voltage dip event, then the front-end PFC circuit of the charging module is controlled to stop or reduce energy intake from the grid, and the control loop bandwidth of the DC bus voltage is dynamically adjusted to accelerate the stabilization of the DC bus voltage. If the event source category is the load-side sudden event, then the front-end PFC circuit is controlled to actively increase the energy drawn from the grid, and the output current limit value of the charging module is dynamically adjusted to suppress the impact, thereby stabilizing the DC bus voltage. If the event source category is an internal fault event of the module, the operation of the front-end PFC circuit and the rear-end DC-DC circuit shall be stopped immediately, and the connection with the power grid and the load shall be disconnected.

7. The grading diagnostic method according to any one of claims 1 to 6, characterized in that, The method further includes: The DC bus energy margin under the current operating condition is calculated based on the current output power of the charging module and the AC side input voltage. The preset rate of change threshold for triggering transient events is dynamically determined based on the DC bus energy margin, and the DC bus energy margin and the preset rate of change threshold are positively correlated.

8. A graded diagnostic system for a charging module, characterized in that, The graded diagnostic system includes: The acquisition unit is used to synchronously acquire multiple operating parameters of the charging module, including AC side input voltage, DC bus voltage, output current and load side power command. The monitoring unit is used to continuously monitor the DC bus voltage and calculate the rate of change of the DC bus voltage in each control cycle. The determination unit is used to determine that a transient event is triggered when the DC bus voltage drops and the rate of change exceeds a preset rate of change threshold, and to obtain the trigger timestamp of the transient event. The acquisition unit is used to acquire the plurality of operating parameters at the moment before the transient event is triggered, based on the trigger timestamp, and determine them as a stable baseline; The classification unit is used to evaluate the synchronous change trend of the multiple operating parameters relative to the stable baseline within a preset time window including the trigger timestamp, and classify each operating parameter into a preset state label according to the evaluation result. The diagnostic unit is used to diagnose the event source category of the transient event based on the correlation between the multiple operating parameters and the combination of the status labels. The event source category includes at least grid-side sag events, load-side surge events, and module internal fault events. An execution unit is used to execute a hierarchical response strategy that matches the event source category.

9. A graded diagnostic device for a charging module, characterized in that, The graded diagnostic device includes: Processor, memory, input / output units, and bus; The processor is connected to the memory, the input / output unit, and the bus; The memory stores a program, which the processor invokes to perform the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium contains a program that, when executed on a computer, performs the method as described in any one of claims 1 to 7.

Images