A method and system for traceless fault positioning of a charging pile

Abstract

The application discloses a charging pile traceless fault positioning method and system, the method comprises the following steps: obtaining component mapping information and event observation information; determining a target event window according to the occurrence time, and determining a reference window from a candidate window without the target event according to the running condition; performing background elimination and event intensity normalization on the external physical field signals of the target event window and the reference window to obtain an event response kernel; verifying the event allowed response relationship and the cross-physical field response sequence according to the event response kernel, the event type and the component mapping information to obtain a candidate component evidence set; performing spatial inversion constrained by the event allowed response relationship according to the candidate component evidence set and a spatial response dictionary to obtain a fault positioning result; and outputting a fault component, a positioning position and an evidence chain according to the fault positioning result. The application is beneficial to improving the accuracy and reviewability of fault positioning without interfering with normal charging.

Description

Technical Field

[0001] This application relates to the fields of fault detection, non-contact condition monitoring and multi-physics field signal processing technology for new energy vehicle charging facilities, and in particular to a method and system for non-marking fault location of charging piles. Background Technology

[0002] As the charging power of new energy vehicles increases, DC charging piles typically contain power modules, contactors, DC busbars, output terminals, and heat dissipation components. When these components experience early faults such as jamming, increased contact resistance, connector degradation, or bearing wear, fault codes may not be triggered yet, but the electromagnetic, mechanical, acoustic, or thermal conditions of the equipment have already been affected.

[0003] Existing online monitoring mainly reads the voltage, current, temperature, and communication messages inside the charging pile. Although it can detect output anomalies, the same output anomaly may be caused by multiple internal components, making it difficult to pinpoint the specific fault location. Using test loads or active disturbance detection methods would alter normal charging conditions, increasing the risks associated with on-site wiring and high-voltage operations.

[0004] Existing external multimodal detection methods typically involve continuously acquiring different types of signals and performing anomaly assessments separately. However, these signals are easily affected by vehicle operating conditions, environmental noise, the actions of adjacent devices, and load changes, making it difficult to correlate signals to the same internal fault source. Therefore, accurately locating external anomalies to specific internal components without interfering with normal charging has become a technical problem that needs to be solved. Summary of the Invention

[0005] This application proposes a non-marking fault location method and system for charging piles, aiming to locate internal faulty components based on the multi-physics field response of the charging pile without disassembling the machine, modifying the wiring, injecting special detection signals, or changing the normal charging parameters.

[0006] In a first aspect, embodiments of this application provide a method for traceless fault location of charging piles, executed by a fault location device equipped with a processor and a communication interface, wherein the communication interface is connected to external sensors of the charging pile, comprising: acquiring component mapping information and event observation information; the component mapping information including component topology and spatial response dictionary, the event observation information including at least two types of external physical field signals from synchronously acquired magnetic field, acoustic, vibration and thermal fields, as well as the event type, occurrence time, event intensity and operating conditions of passively identified natural charging events; determining a target event window based on the occurrence time, and determining a reference window from candidate windows that do not contain the target event based on the operating conditions; performing background elimination and event intensity normalization on the external physical field signals of the target event window and the reference window to obtain an event response kernel; verifying the permissible event response relationship and cross-physical field response order based on the event response kernel, the event type and the component mapping information to obtain a candidate component evidence set; performing spatial inversion constrained by the permissible event response relationship based on the candidate component evidence set and the spatial response dictionary to obtain a fault location result; and outputting the faulty component, location location and evidence chain based on the fault location result.

[0007] As can be seen, in this embodiment, the fault location device first acquires component mapping information and event observation information; secondly, it determines the target event window and reference window according to the occurrence time and operating conditions, and forms an event response kernel through background elimination and event intensity normalization; then, it verifies the permissible response relationship and cross-physical field response sequence by combining the event type and component mapping information, forming a candidate component evidence set; thirdly, it performs spatial inversion constrained by permissible response relationships on the candidate component evidence set to form the fault location result; finally, it outputs the faulty component, location, and evidence chain. Since normal charging events constitute a common excitation for different physical field responses, various responses can mutually verify each other according to the physical propagation sequence, thereby enabling the convergence of the entire machine's anomaly to a specific internal component without interfering with normal charging.

[0008] In one possible example, the step of verifying the allowed event response relationship and cross-physical field response order based on the event response kernel, the event type, and the component mapping information to obtain a candidate component evidence set includes: extracting a response feature set corresponding to the event type based on the event response kernel, the event type, and the component mapping information; determining the set of allowed event response components and the causal response template based on the event type and the component mapping information; and determining the candidate component evidence set based on the response feature set, the set of allowed event response components, and the causal response template.

[0009] As can be seen, in this example, the scope of response feature extraction is limited by the event type. The event-allowed response component set is used to exclude components that lack physical connection with the target event, and the causal response template is used to verify the order of occurrence of responses in different physical fields. This can reduce erroneous associations caused by environmental disturbances or the actions of adjacent devices.

[0010] In one possible example, determining the candidate component evidence set based on the response feature set, the set of allowed response components for the event, and the causal response template includes: establishing candidate response relationships between response features and candidate components based on the response feature set, the set of allowed response components for the event, and the causal response template; determining prohibited response information for components outside the set of allowed response components for the event, and determining spatial consistency information for components within the set of allowed response components for the event; determining cross-physical field causal temporal consistency information based on the candidate response relationships and the causal response template; and determining the candidate component evidence set based on the prohibited response information, the spatial consistency information, and the cross-physical field causal temporal consistency information.

[0011] As can be seen, in this example, prohibition of response information, spatial consistency information, and cross-physical field causal temporal consistency information constrains candidate components from three dimensions: mechanistic feasibility, response location, and response order, so that the candidate component evidence set can retain positive evidence and exclusionary evidence in the fault location process.

[0012] In one possible example, determining cross-physical field causal timing consistency information based on the candidate response relationship and the causal response template includes: determining the expected physical field response order and the allowable delay interval between adjacent responses based on the candidate response relationship and the causal response template; determining the actual start time and measured response order of each physical field response based on the candidate response relationship; and verifying the expected physical field response order segment by segment based on the actual start time, the measured response order, and the allowable delay interval to obtain the cross-physical field causal timing consistency information.

[0013] As can be seen, in this example, by verifying the expected response sequence, the actual response sequence, and the allowable delay interval segment by segment, it is possible to distinguish between normal delays and fault delays in electromagnetic, mechanical, conduction, and thermal diffusion processes, thus avoiding the judgment of faults based solely on response amplitude.

[0014] In one possible example, determining the actual start time and measured response order of each physical field response based on the candidate response relationship includes: locating the corresponding event response kernel based on the candidate response relationship, extracting the relative event time series of each physical field from the event response kernel; determining the background boundary of each physical field response based on the relative event time series; determining the first moment after the event occurs that continuously exceeds the background boundary and satisfies the persistence condition as the actual start time based on the relative event time series and the background boundary; arranging the corresponding physical fields according to each actual start time, and aggregating each actual start time to obtain the actual start time and measured response order of each physical field response.

[0015] As can be seen, in this example, by using the background boundary and persistence conditions to jointly determine the actual start time, the influence of instantaneous noise peaks and individual abnormal sampling points on the response order can be suppressed, thereby improving the stability of cross-physical field time sequence verification.

[0016] In one possible example, determining the background boundary of each physical field response based on the relative event time series includes: dividing the time series before the event occurs into multiple background sub-windows based on the relative event time series; determining the corresponding central statistic and fluctuation statistic based on the signal amplitude within each background sub-window; removing abnormal background sub-windows based on the central statistic and the fluctuation statistic to obtain effective background statistics; and determining the background boundary of each physical field response based on the effective background statistics.

[0017] As can be seen, in this example, by forming a background boundary through multiple background sub-windows and removing abnormal background sub-windows, the interference of accidental collisions, vehicle noise, or the actions of adjacent charging piles on the response start time can be reduced.

[0018] In one possible example, after obtaining the candidate component evidence set and before performing the spatial inversion, the method further includes: obtaining similar event response kernels for similar components or historical events of the same component within the same charging pile, based on the event type and operating conditions of the target event; forming similar event clusters based on the similar event response kernels, and determining a real-time reference response kernel from the similar event clusters; determining lateral deviation information for different observation locations and different physical fields based on the event response kernel of the target event and the real-time reference response kernel; updating the candidate component evidence set based on the lateral deviation information, and inputting the updated candidate component evidence set into the spatial inversion.

[0019] As can be seen, in this example, similar components or historical events within the same charging pile can form a natural comparison, so that the fault location process does not need to rely on the fixed threshold shared by different models of charging piles, and can further converge candidate components by utilizing lateral deviation information.

[0020] In one possible example, the step of forming a cluster of similar events based on the similar event response kernels and determining a real-time reference response kernel from the similar event clusters includes: selecting a set of valid similar event response kernels from the similar event response kernels based on the event intensity and operating conditions corresponding to each similar event response kernel; merging the set of valid similar event response kernels according to component type and event type to obtain the similar event clusters; aligning the similar event clusters with relative event times, and performing median aggregation on the aligned responses according to observation location and physical field type to obtain the real-time reference response kernels.

[0021] As can be seen, in this example, by first filtering comparable events according to event intensity and operating conditions, and then performing time alignment and median aggregation, the contamination of the real-time reference response core by single event noise, occasional disturbances, and individual abnormal components can be reduced.

[0022] In one possible example, after outputting the faulty component, the location, and the chain of evidence, the method further includes: acquiring baseline update information; the baseline update information includes verification results of cross-physical field response sequence, prohibited response information, diagnostic results, post-event fault code status, and charging end status; determining a baseline update verification result based on the baseline update information; when the baseline update verification result indicates that the cross-physical field response sequence is normal, no prohibited response occurs, the diagnostic result does not indicate a fault, no fault code is generated after the event, and charging ends normally, adding the event response core and corresponding operating conditions to the healthy baseline; otherwise, prohibiting the event response core from being added to the healthy baseline; and determining a response comparison reference for subsequent natural charging events based on the updated healthy baseline.

[0023] As can be seen in this example, only event response kernels that have undergone multiple verifications of causality, diagnosis, and charging results can enter the health baseline, thereby preventing slowly degrading data or abnormal data that has not yet triggered fault codes from continuously polluting the health baseline.

[0024] Secondly, embodiments of this application provide a charging pile seamless fault location system, including a communication interface connected to external sensors of the charging pile and the following modules implemented by a processor: an information acquisition module, used to acquire component mapping information and event observation information through the communication interface; the component mapping information includes component topology and spatial response dictionary, and the event observation information includes at least two types of synchronous external physical field signals and the event type, occurrence time, event intensity, and operating conditions of natural charging events; a response kernel generation module, used to determine a target event window and a reference window according to the occurrence time and the operating conditions, and to perform background elimination and event intensity normalization on the external physical field signals of the two windows to obtain an event response kernel; a causal verification module, used to verify the permissible event response relationship and cross-physical field response order according to the event response kernel, the event type, and the component mapping information to obtain a candidate component evidence set; a spatial location module, used to perform spatial inversion constrained by the permissible event response relationship according to the candidate component evidence set and the spatial response dictionary to obtain a fault location result; and a diagnostic output module, used to output the faulty component, location, and evidence chain according to the fault location result.

[0025] As can be seen, the functional modules in this embodiment work together in the order of information acquisition, response kernel generation, causal verification, spatial positioning and diagnostic output, so that the external multi-physics response during the natural charging process can be converted into component-level diagnostic results with physical causal relationships.

[0026] Thirdly, embodiments of this application provide an electronic device including a processor, a memory, a communication interface, and one or more programs, the one or more programs being stored in the memory and configured to be executed by the processor, the programs including instructions for performing the steps in the first aspect of embodiments of this application.

[0027] Fourthly, embodiments of this application provide a computer storage medium storing a computer program for electronic data interchange, wherein the computer program causes a computer to perform some or all of the steps described in the first aspect of this embodiment.

[0028] As can be seen, in this embodiment, compared with existing detection methods that continuously acquire signals of different modes and classify faults separately, the fault location device of this application first acquires component mapping information and event observation information; secondly, it determines the target event window and reference window according to the occurrence time and operating conditions, and forms an event response kernel after background elimination and event intensity normalization; then, it verifies the permissible response relationship of the event and the response order across physical fields by combining the event type and component mapping information, forming a candidate component evidence set; thirdly, it performs spatial inversion constrained by permissible response relationships on the candidate component evidence set to form the fault location result; finally, it outputs the faulty component, location, and evidence chain. Since normal charging events constitute the common excitation of multi-physical field responses, and the reference window and event intensity respectively eliminate background interference and differences in operating condition amplitude, the abnormal responses of different physical fields can be mutually verified according to the physical propagation order, thereby converging the overall machine abnormality to a specific internal component, which is beneficial to improve the accuracy and verifiability of fault location without interfering with normal charging. Attached Figure Description

[0029] Figure 1 A schematic diagram of the hardware architecture for traceless fault location of charging piles provided in an embodiment of this application; Figure 2 A schematic diagram of the software architecture for non-marking fault location of charging piles provided in the embodiments of this application; Figure 3 A schematic diagram of an external multiphysics field observation and natural charging event monitoring scenario provided in the embodiments of this application; Figure 4 This is a schematic diagram of a non-marking fault location scenario for the main contactor provided in an embodiment of this application. Figure 5 This is a schematic diagram of a parallel power module anomaly location scenario provided in an embodiment of this application; Figure 6 A flowchart illustrating the non-marking fault location method for charging piles provided in this application embodiment; Figure 7 A flowchart illustrating the verification of event-permitted response relationships and cross-physical field response order provided for embodiments of this application; Figure 8 A schematic diagram illustrating the process of determining the candidate component evidence set provided in the embodiments of this application; Figure 9 A schematic diagram illustrating the process for determining cross-physical field causal timing consistency information provided in an embodiment of this application; Figure 10 A flowchart illustrating the determination of the actual start time and measured response sequence for embodiments of this application; Figure 11 A schematic diagram illustrating the process of determining the background boundary of the physical field response provided in this application embodiment; Figure 12 A flowchart illustrating a horizontal comparison of similar events provided in the embodiments of this application; Figure 13 A schematic diagram of the process for determining the real-time reference response core provided in the embodiments of this application; Figure 14 A schematic diagram illustrating the process of conditional update of health baseline provided in this application embodiment; Figure 15 This is a schematic diagram of the functional modules of the charging pile non-marking fault location system provided in the embodiments of this application. Detailed Implementation

[0030] The technical solutions of the embodiments of this application are clearly and completely described below with reference to the accompanying drawings. The described embodiments are only used to explain this application and are not intended to limit the scope of protection of this application. Where there is no conflict, the following embodiments and their technical features can be combined with each other.

[0031] The term "traceless fault location" as used in this application refers to a testing device located outside the charging pile, or obtaining charging pile status information only through a read-only interface; during the testing process, the cabinet door is not opened, the charging gun, charging gun wire, busbar or power module is not disassembled, test branches are not connected to the high-voltage circuit and control circuit, special pulses or frequency sweep signals are not injected into the circuit under test, and the vehicle requirements or charging pile output parameters are not changed for the purpose of testing.

[0032] Natural charging events are state changes triggered by vehicle demand, charging protocols, the original control logic of the charging pile, or the normal adjustment process of the equipment. Natural charging events can include pre-charge contactor operation, main contactor operation, power module switching, natural changes in output current, transition from constant current to constant voltage stage, fan start / stop, liquid cooling pump start / stop, and normal shutdown. The fault location equipment only identifies and observes natural charging events and does not generate additional detection actions.

[0033] Event observation information is a set of information associated with a natural charging event, which may include event type, occurrence time, event intensity, operating conditions, and synchronous external physical field signals. Operating conditions may include output power, output current, bus voltage, ambient temperature, number of modules in operation, fan status, liquid cooling pump status, and charging stage. Event intensity characterizes the magnitude of energy or state changes caused by the natural charging event, and can be determined by changes in power, current, external magnetic field step at the charging gun, integral of the contactor coil magnetic field, or change in the rotational speed of the heat dissipation components.

[0034] The target event window is a signal interval extending from before to after a natural charging event, using the time of the event as the time reference. The reference window is a signal interval that does not include the target event but has similar operating conditions to the target event window. The reference window can be drawn from the same charging process or from historical normal charging processes of the same charging station.

[0035] An event response kernel is a dynamic response formed by performing time correction, background removal, and event intensity normalization on the external physical field signals of the target event window and the reference window. The event response kernel preserves the response changes relative to the moment the event occurs and reduces the incomparability caused by different vehicles, states of charge, and power variation amplitudes.

[0036] Event-permissible response relationships describe the internal components that are theoretically allowed to produce a synchronous response when a certain event type occurs. The event-permissible response component set is the set of components in the event-permissible response relationship that correspond to the target event. For example, in a main contactor closure event, the main contactor coil, armature, contacts, adjacent DC bus, and output terminals can produce a response, while other unconnected power modules should generally not produce a significant synchronous response.

[0037] The causal response template describes the expected physical field response order, allowable time delay interval, spatial location relationship, and attenuation relationship corresponding to a certain event type. The spatial response dictionary describes the typical response distribution of internal components to different external observation locations and different physical fields under the action of a specific event. The event-permissible response relationship focuses on limiting whether a component can participate in the target event, while the spatial response dictionary focuses on describing how the component's response is transmitted to the external observation location.

[0038] The candidate component evidence set is a collection of diagnostic evidence organized according to candidate components. The evidence for each candidate component can include response characteristics, prohibited response information, spatial consistency information, cross-physical field causal-temporal consistency information, lateral deviation information from similar events, and historical baseline deviation information. The candidate component evidence set can store both positive evidence supporting a particular component as a fault source and negative evidence excluding that component.

[0039] See Figure 1The charging pile non-marking fault location hardware system 100 may include a charging pile 110, a charging gun 120, a charging gun cable 130, a magnetic field sensor array 140, a microphone array 150, a laser vibration measurement module 160, an infrared thermal imaging module 170, a visible light positioning module 180, and a fault location device 190. The charging pile 110 may internally include a power module 111, a main contactor 112, a pre-charge contactor 113, a DC busbar 114, a fan 115, a liquid cooling pump 116, and a charging gun cable output terminal 117. The fault location device 190 may include a processor 191, a memory 192, a communication interface 193, a time synchronization component 194, and a read-only operating interface 195.

[0040] The magnetic field sensor array 140 can be installed in the external area of ​​the cabinet corresponding to the power module 111, main contactor 112, DC busbar 114, and charging gun line 130 to collect near-field magnetic responses generated by coil excitation, current establishment, module switching, and high-frequency ringing. For metal cabinets, low-frequency and mid-low-frequency near-field magnetic fields can be prioritized for collection, and cabinet door seams, vents, and cable outlets can be used as windows for magnetic field propagation.

[0041] The microphone array 150 can be positioned in front of the cabinet door of the charging pile 110 to collect airborne acoustic signals generated by contactor impact, fan bearing, liquid cooling pump cavitation, and component collisions. The microphone array 150 can also determine the direction of the sound source based on the time difference of arrival of multiple channels, thereby suppressing non-target sound sources generated by adjacent charging piles or vehicles.

[0042] The laser vibration measurement module 160 can perform non-contact vibration measurement on the outer surface of the cabinet corresponding to the main contactor 112, power module 111, and heat dissipation components. The infrared thermal imaging module 170 can cover the corresponding external areas of the main contactor 112, DC busbar 114, gun line output terminal 117, and power module 111 to form characteristics such as the center of thermal gradient, direction of thermal diffusion, and temperature rise per unit event.

[0043] The visible light positioning module 180 is used to identify the cabinet outline, cabinet door seams, ventilation openings, and sensor locations, and maps the magnetic field observation point, sound source direction, laser vibration measurement point, and infrared pixel area to a unified external coordinate system. Once the charging pile structure and sensor locations have been calibrated, the visible light positioning module 180 can be omitted.

[0044] The time synchronization component 194 is used to establish a unified time axis for sensors with different sampling frequencies and transmission delays. The magnetic field, acoustic, and vibration channels can share a hardware clock, and the infrared thermal image can be mapped to this time axis based on the frame timestamp. As an example implementation, the sampling frequency of the magnetic field and vibration signals can be set to 50 kHz, the sampling frequency of the acoustic signal can be set to 96 kHz, and the frame rate of the infrared thermal image can be set to 50 Hz.

[0045] The read-only operating interface 195 can read contactor status, module activation number, output power, output current, and heat dissipation equipment status from the charging pile controller or station-level platform. The read-only operating interface 195 does not write control commands to the charging pile. When read-only operating data is unavailable, the processor 191 can also identify natural charging events based on external physical field signals.

[0046] See Figure 2 The software architecture 200 for seamless fault location of charging piles may include an acquisition and synchronization layer 210, a natural event recognition layer 220, a reference window matching layer 230, a response kernel processing layer 240, a causal response analysis layer 250, a similar event comparison layer 260, a spatial inversion layer 270, a baseline management layer 280, and a diagnostic output layer 290.

[0047] The acquisition and synchronization layer 210 is used to receive external physical field signals and read-only running data, and to perform timestamp unification, sampling rate adaptation, missing frame marking, and channel status checks on each data stream. The natural event recognition layer 220 is used to determine the event type, occurrence time, event intensity, and operating conditions, and to associate them with the corresponding external physical field signals.

[0048] The reference window matching layer 230 is used to filter reference windows with similar operating conditions from candidate windows that do not contain the target event. The response kernel processing layer 240 is used to perform time correction, background removal, and event intensity normalization on the target event window and the reference window, and extract the response feature set from the formed event response kernel.

[0049] The causal response analysis layer 250 is used to manage permissible event response relationships and causal response templates, and to generate prohibited response information, spatial consistency information, and cross-physical field causal temporal consistency information. The similar event comparison layer 260 is used to form similar event clusters, determine real-time reference response kernels, and calculate lateral deviation information.

[0050] The spatial inversion layer 270 is used to invoke the spatial response dictionary to converge at least a minimum number of candidate components to anomalous responses from different observation locations and different physical fields. The baseline management layer 280 is used to save the healthy baseline and perform conditional updates. The diagnostic output layer 290 is used to generate faulty components, location information, fault type candidates, confidence states, and evidence chains.

[0051] Each software layer can run on edge computing devices at charging stations or on diagnostic servers that communicate with multiple charging piles. Event identification and rapid response analysis with high real-time requirements can be completed at the edge; aggregation of similar historical events and cross-device dictionary maintenance can be completed on the server side.

[0052] See Figure 3The external multiphysics field observation scenario can include an external observation coordinate system 310, a power module observation area 320, a contactor and busbar observation area 330, a heat dissipation component observation area 340, a cabinet propagation window 350, a gun line observation area 360, a natural event timeline 370, and an operating condition label 380. A magnetic field sensor array 140, a microphone array 150, a laser vibration measurement module 160, and an infrared thermal imaging module 170 are located outside the charging pile 110. A visible light positioning module 180 is used to map each observation position to the external observation coordinate system 310.

[0053] The power module observation area 320 includes multiple magnetic field observation points and multiple thermal imaging areas arranged along the power module arrangement direction. The contactor and busbar observation area 330 covers the cabinet surface corresponding to the main contactor, pre-charge contactor, and adjacent DC busbars. The heat dissipation component observation area 340 covers the areas corresponding to the fan, liquid cooling pump, and liquid cooling path. The cabinet propagation window 350 includes cabinet door seams, vents, and structural openings. The charging gun wire observation area 360 is located at the charging gun wire outlet and the base of the charging gun.

[0054] The natural event timeline 370 can sequentially record events such as contactor closure, power module activation, output power changes, and heat dissipation component startup (fan or liquid cooling pump startup). The operating condition label 380 is associated with each event in the natural event timeline 370 and is used to record the output current, ambient temperature, number of activated modules, and heat dissipation status when the event occurs.

[0055] For different natural charging events, the fault location device can select different analysis windows. The magnetic field, acoustic, and vibration responses corresponding to contactor operation and module switching can be analyzed using a fast analysis window at the millisecond level; the thermal response can be analyzed using a slow analysis window lasting several seconds or tens of seconds. Different analysis windows all use the time of the same event as the time reference, but different physical fields are not required to have the same sampling frequency.

[0056] See Figure 4 In the main contactor fault location scenario, the responses include coil magnetic field response 410, armature mechanical response 420, load magnetic field response 430, contact area thermal response 440, contact bounce evidence chain 450, increased contact resistance evidence chain 460, and contactor location result 470. When the main contactor 112 closes naturally, the coil magnetic field response 410, armature mechanical response 420, load magnetic field response 430, and contact area thermal response 440 appear in the order of electromagnetic excitation, mechanical movement, main circuit conduction, and thermal diffusion.

[0057] In a specific example, the object under test is a DC charging pile with a rated power of 120 kW. The system sets up three triaxial magnetic field observation points in the cabinet door area corresponding to the main contactor, two triaxial magnetic field observation points in the area corresponding to the DC busbar, and one magnetic field observation point near the charging gun line outlet. Other physical field responses are collected through microphone array, laser vibration measurement, and infrared thermography.

[0058] Fault location equipment can first determine the causal response template of the device from multiple health closure events. For example, the moment when the coil magnetic field first continuously exceeds the background boundary can be taken as the event occurrence time. The first obvious mechanical shock in a health event usually occurs later than the occurrence of the coil magnetic field, the stable establishment of the load magnetic field occurs later than the mechanical shock, and the thermal response gradually appears thereafter. The allowable time delay interval is preferably obtained from the statistics of health events of the same charging pile, rather than using a fixed time delay shared by all models.

[0059] In the contact bounce example, the coil magnetic field response 410 shows only one complete energization, but the armature mechanical response 420 exhibits multiple impact peaks, and the load magnetic field response 430 experiences a short-term drop and re-establishment between adjacent impact peaks. The resulting contact bounce evidence chain 450 characterizes "single coil energization—multiple mechanical impacts—repeated main circuit conduction." Although external impact noise may cause mechanical shocks, it will not consistently cause synchronous interruption of the load magnetic field, and therefore can be excluded by the causal response template.

[0060] In the example of increased contact resistance, the order of occurrence of the coil magnetic field response 410, the armature mechanical response 420, and the load magnetic field response 430 is within the healthy range. However, the unit event intensity temperature rise of the contact area thermal response 440 is continuously higher than the healthy baseline, and the center of the thermal response stably falls near the output terminal of the main contactor. The resulting chain of evidence 460 for increased contact resistance can narrow down candidate faults to contact wear, contact surface oxidation, reduced contact pressure, or loose terminal connections.

[0061] The contactor location result 470 may include candidate fault types, response start times of each physical field, spatial consistency status, and evidence chains, and may also include the main contactor and the corresponding external area. This scenario illustrates that mechanical faults and conduction-heating faults of the same component can be distinguished by different cross-physical field response sequences and response intensity relationships.

[0062] See Figure 5 , Figure 5 This is a schematic diagram of an abnormal positioning scenario for parallel power modules provided in an embodiment of this application. The abnormal positioning scenario for parallel power modules may include a group of parallel power modules 510, a target module 520, a cluster of similar events 530, a low-frequency magnetic response deviation 540, a high-frequency ringing deviation 550, a thermal response deviation 560, a compensation response from adjacent modules 570, and a module positioning result 580.

[0063] In a specific example, the charging pile has a rated power of 180 kW, and the parallel power module group 510 includes six power modules with a rated power of 30 kW each. The fault location equipment waits for the vehicle's power demand to increase naturally before activating the modules using the original scheduling strategy of the charging pile, without requiring the control system to activate the modules in turn according to the detection order.

[0064] The fault location device can filter module activation events with similar power change amplitudes, ambient temperatures, and fan speeds, forming a cluster of similar events 530. The event response core of the target module 520 is compared with the real-time reference response core of the cluster of similar events 530 to determine the low-frequency magnetic response deviation 540, the high-frequency ringing deviation 550, and the thermal response deviation 560.

[0065] As a concrete data example, the low-frequency magnetic response peak of the target module 520 when it is put into operation is approximately 0.72 times that of the real-time reference response core, the high-frequency ringing energy is approximately 1.85 times that of the real-time reference response core, and the unit power thermal response is approximately 1.42 times that of the real-time reference response core. Meanwhile, the adjacent module compensation response 570 shows an increase in the output undertaken by the neighboring module. This combination of responses can improve the candidate for power transmission path degradation of the target module and reduce the candidate for simple fan failure.

[0066] When the low-frequency magnetic response and high-frequency ringing of the target module are within the normal range, but the bearing characteristic frequency and vibration amplitude increase, and the cooling rate of the heat dissipation area decreases during the fan start-up event, the module location result 580 can point to the cooling fan of the target module, rather than the power device. Therefore, the power transmission path fault and the cooling mechanical fault can be distinguished based on the combined relationship between electromagnetic, mechanical, and thermal responses.

[0067] See Figure 6 The method for locating faults in charging piles without leaving a trace can be achieved through the following steps S101 to S105.

[0068] S101. Obtain component mapping information and event observation information; the component mapping information includes component topology and spatial response dictionary, and the event observation information includes at least two types of external physical field signals from synchronously acquired magnetic field, acoustic, vibration and thermal fields, as well as the event type, occurrence time, event intensity and operating conditions of passively identified natural charging events.

[0069] The component topology is used to describe the type, location, and connection relationships of the internal components of the charging pile. For example, the component topology can record the connection relationships between the main contactor and the precharge contactor, the DC bus, and the gun output terminals, and it can also record the correspondence between multiple power modules and the DC bus and heat dissipation components.

[0070] The spatial response dictionary can include multiple response mapping entries. Each response mapping entry can be associated with event type, internal components, external observation location, physical field type, typical response distribution, and propagation direction. The spatial response dictionary can be built from natural event data of a healthy prototype, historical health data of a charging pile of the same model, or simulation results of electromagnetic, acoustic, vibration, and heat transfer.

[0071] Natural charging events can be identified through read-only operating data or through external physical field signals. For example, a sudden change in the magnetic field in the contactor coil area can be used to identify contactor operation, a magnetic field step near the charging gun line can be used to identify output current establishment, and the appearance of narrowband acoustic and vibration frequencies can be used to identify the start-up of a fan or liquid-cooled pump. When using external signals to identify events, the timing of the event can be confirmed by cross-checking at least two signal channels.

[0072] S102. Determine the target event window based on the occurrence time, and determine the reference window from the candidate windows that do not contain the target event based on the operating conditions; perform background elimination and event intensity normalization on the external physical field signals of the target event window and the reference window to obtain the event response kernel.

[0073] The target event window can include the background interval before the event and the response interval after the event. Candidate windows can be extracted from the same charging process or historical normal charging processes, and windows containing the target event, other strong transient events, or abnormal sensor states are excluded.

[0074] Fault location equipment can compare the average power, average current, ambient temperature, number of deployed modules, and status of heat dissipation equipment between the target event window and candidate windows. Operating conditions with different dimensions can be standardized first, and then the differences in operating conditions can be formed according to a preset importance. Candidate windows with small differences in operating conditions and that do not contain the target event can be determined as reference windows, or multiple candidate windows can be selected to form a combined reference window.

[0075] Background removal can include time alignment, trend correction, and window signal subtraction. Event intensity normalization can select the corresponding intensity information based on the event type. Power change events can use the power change or current change, contactor events can use the coil magnetic field change, and heat dissipation equipment events can use the speed change or drive magnetic field change. The normalized event response kernel is used to represent the external dynamic response corresponding to a unit event intensity.

[0076] S103. Based on the event response kernel, the event type, and the component mapping information, verify the event-permitted response relationship and the cross-physical field response order to obtain a candidate component evidence set.

[0077] For a target event, the fault location device can first obtain the corresponding set of event-permitted response components from the event-permitted response relationship, and then call the causal response template. For example, the response sequence corresponding to the main contactor closing event may include the coil magnetic field response, armature mechanical response, load magnetic field response, and contact area thermal response. The response sequence corresponding to the fan starting event may include the fan drive response, mechanical rotation response, and temperature change in the heat dissipation area.

[0078] Verification of the sequence of responses across physical fields does not require all responses to occur simultaneously. Fault location equipment can determine whether expected responses are missing, premature, delayed, or repeated, and interpret different forms of deviation as different evidence of faults. For example, a missing mechanical response can support armature jamming, a repeated mechanical response can support contact bounce, and normal electromagnetic and mechanical responses but enhanced thermal response can support increased contact resistance.

[0079] S104. Perform spatial inversion constrained by the event-permitted response relationship based on the candidate component evidence set and the spatial response dictionary to obtain the fault location result.

[0080] Fault location equipment can combine response characteristics from different observation locations and different physical fields into external observation information, and obtain response mapping entries corresponding to each internal component from a spatial response dictionary. Spatial inversion is used to determine the response combinations of a small number of internal components, so that the external distribution corresponding to the response combination matches the actual external observation information.

[0081] Spatial inversion can employ nonnegativity and sparsity constraints to concentrate anomaly intensity on a small number of candidate components. For components outside the set of allowed response components, a penalty for prohibiting responses can be added. For components that are spatially adjacent and have similar response distributions, further differentiation can be achieved by combining component topology, thermal diffusion direction, and multiphysics spatial center.

[0082] S105. Output the faulty component, location, and evidence chain based on the fault location result.

[0083] The location can include internal component identifiers and corresponding external observation areas. The chain of evidence can be organized according to the event occurrence time, including event type, response start time, response peak, spatial center, prohibited response information, causal time sequence consistency information, and historical deviation information. The fault location device can also output fault type candidates, confidence levels, and suggested handling levels for maintenance personnel to review.

[0084] As can be seen, in this example, information acquisition, response kernel generation, causal verification, spatial inversion and diagnostic output are completed sequentially through steps S101 to S105, so that different physical field responses share the same time reference of the natural charging event, and the positioning results are constrained by the component mechanism and spatial transmission relationship, thereby improving the reliability of component-level fault positioning.

[0085] See Figure 7 The specific implementation process of step S103 can be achieved through the following steps S201 to S203.

[0086] S201. Based on the event response kernel, the event type, and the component mapping information, extract the response feature set corresponding to the event type.

[0087] The response feature set is a set of features selected from the event response kernel that characterize the propagation process of the target event. Response features may include the response start time, peak time, rise time, decay time, dominant frequency, harmonic energy, high-frequency ringing, sound source direction, location of maximum vibration, center of thermal gradient, direction of thermal diffusion, and temperature rise per unit event intensity.

[0088] Different event types can correspond to different combinations of response characteristics. For example, when a contactor operates, priority is given to extracting the initiation time of the coil magnetic field, the number of mechanical impacts, the establishment time of the load magnetic field, and the thermal response of the contact area; when a power module is engaged, priority is given to extracting the low-frequency magnetic field increment, high-frequency ringing, compensation response of adjacent modules, and thermal response per unit power.

[0089] S202. Based on the event type and the component mapping information, determine the set of components that can respond to the event and the causal response template.

[0090] The set of components allowed to respond to an event can be determined by the records corresponding to the target event type in the event allowed response relationship. The causal response template can be determined by a healthy prototype, a healthy baseline of the same charging pile, or the working mechanism of the components, and can include the expected response sequence, the allowable delay interval, the allowable spatial region, and the response missing conditions.

[0091] In one possible embodiment of this application, the fault location device can also dynamically adjust the set of components allowed to respond to events based on operating conditions. For example, when operating conditions indicate that a certain power module is not in operation, the allowable level for that power module to generate a synchronous response can be reduced; when the liquid cooling pump starts but the output current remains stable, the liquid cooling pump and related components of the liquid cooling path can be retained, while non-target power modules related to current changes can be excluded. This embodiment solves the problem that fixed allowable relationships cannot adapt to changes in module scheduling status, which is beneficial for narrowing down the range of candidate components.

[0092] S203. Determine the candidate component evidence set based on the response feature set, the event-allowed response component set, and the causal response template.

[0093] The fault location device can match the response feature set with the set of components that can respond to the event and the causal response template, and organize the matching results according to the candidate components. For responses that meet the expected response order and spatial range, supporting evidence can be formed; for cases where the expected response is missing, the response order is abnormal, or components outside the set respond synchronously, abnormal evidence or exclusionary evidence can be formed.

[0094] As can be seen, this embodiment ensures that the selection of response features, the limitation of candidate components, and the cross-physical field time sequence verification all correspond to the event type, thereby reducing the influence of irrelevant features on fault judgment and improving the correspondence between the candidate component evidence set and the actual fault mechanism.

[0095] See Figure 8 The specific implementation process of step S203 can be achieved through the following steps S301 to S304.

[0096] S301. Based on the response feature set, the event-allowed response component set, and the causal response template, establish a candidate response relationship between the response features and the candidate components.

[0097] The candidate response relationship is used to record the internal components that may correspond to a response feature and the reasons for that correspondence. The fault location device can associate a response feature with one or more candidate components based on the external spatial center, physical field type, event type, and component topology of the response feature.

[0098] For example, when a mechanical impact occurs in the area corresponding to the main contactor during a main contactor closing event, and a load magnetic field is established in the gun line observation area at the same time, the two response characteristics can be associated with the main contactor mechanical mechanism and the main circuit conduction path, respectively, and the transmission relationship between the two response characteristics can be established through the component topology.

[0099] S302. Based on the candidate response relationship, determine prohibited response information for components outside the set of allowed response components for the event, and determine spatial consistency information for components within the set of allowed response components for the event.

[0100] The prohibited response information is used to describe situations where components outside the permitted set of responding components respond synchronously to an event. This situation can characterize both environmental interference and anomalous coupling. Fault location equipment can distinguish between sporadic interference and stable anomalous coupling based on whether the response recurs in multiple similar events.

[0101] Spatial consistency information is used to describe whether the external spatial centers of different physical field responses point to the same candidate component. Since magnetic, acoustic, vibration, and thermal fields have different propagation ranges, spatial tolerances can be determined separately based on the corresponding spatial response dictionaries, without requiring the spatial centers of different physical fields to completely coincide.

[0102] In one possible implementation of this application, a channel reliability state can be determined for different observation channels, and the contribution of spatial consistency information can be adjusted based on sensor signal-to-noise ratio, occlusion status, or propagation path stability. For example, when the infrared observation area is briefly occluded by a vehicle or other object, the spatial contribution of the thermal response can be reduced while preserving the magnetic field and vibration responses. This implementation can solve the problem of spatial judgment distortion caused by the degradation of individual modal quality.

[0103] S303. Determine cross-physical field causal timing consistency information based on the candidate response relationship and the causal response template.

[0104] Cross-physics causal temporal consistency information can record whether the expected response occurs, whether the order of response occurrence is consistent, and whether the time delay of adjacent responses is within the allowable range. For physics fields with different time scales, fast response segments and slow response segments can be identified separately, and then the two response segments can be associated as the same natural event.

[0105] S304. Determine the candidate component evidence set based on the prohibited response information, the spatial consistency information, and the cross-physical field causal temporal consistency information.

[0106] The fault location device can store prohibited response information, spatial consistency information, and cross-physics causal temporal consistency information for each candidate component, and mark the source observation location and physics type of each piece of information. The candidate component evidence set can be sorted according to the degree of primacy, but there is no need to directly output the final fault conclusion at this step.

[0107] As can be seen, this embodiment organizes candidate component evidence from three aspects: permissive mechanism, spatial location, and response timing, so that subsequent spatial inversion can use structured and traceable inputs, avoiding the direct input of unverified multiphysics features into the fault classification model.

[0108] See Figure 9 The specific implementation process of step S303 can be achieved through the following steps S401 to S403.

[0109] S401. Based on the candidate response relationship and the causal response template, determine the expected physical field response order and the allowable time delay interval between adjacent responses.

[0110] The expected order of physical field responses can be represented as a set of response nodes with a sequential relationship. Taking the main contactor closing event as an example, the coil magnetic field response precedes the armature mechanical response, the armature mechanical response precedes the load magnetic field response, and the contact area thermal response follows the load magnetic field response.

[0111] The allowable time delay interval can be obtained from health event statistics of the same charging pile, or it can be determined based on the operating mechanism of the same model prototype and components. For slow processes such as thermal response, the allowable time delay interval can adopt a time scale independent of fast electromagnetic and mechanical processes.

[0112] S402. Determine the actual start time and measured response sequence of each physical field response based on the candidate response relationship.

[0113] The actual start time is the moment when the event response of the corresponding physical field continuously breaks away from the background state after the event occurs. The measured response sequence is determined based on the actual start time of each physical field's response. For expected responses that are not detected, they can be recorded as missing responses without artificially supplementing the start time.

[0114] S403. Based on the actual start time, the measured response sequence, and the allowable delay interval, the expected physical field response sequence is verified segment by segment to obtain the cross-physical field causal timing consistency information.

[0115] Segment-by-segment verification compares the order and delay of adjacent response nodes sequentially. If the response order is consistent and the delay is within the allowable range, the corresponding causal segment is recorded as passed; if there are reversed responses, missing responses, delays exceeding limits, or multiple responses corresponding to a single stimulus, the corresponding anomaly is recorded.

[0116] In one possible implementation of this application, the causal response template can employ a partial sequential relationship, without requiring all responses to form a single linear sequence. For example, both the load magnetic field response and the acoustic response can follow the armature mechanical response, but a fixed sequential relationship is not required between them. The fault location device can verify only the pairs of response nodes with a clear physical causal relationship, thereby resolving unnecessary sequence conflicts caused by differences in delays or propagation paths between different sensors.

[0117] As can be seen, this embodiment converts cross-physical field correlation into verifiable physical propagation relationships by verifying each segment between response nodes, which helps to distinguish between the real response caused by the same event and external interference that is accidentally adjacent in time.

[0118] See Figure 10 The specific implementation process of step S402 can be achieved through the following steps S501 to S504.

[0119] S501. Locate the corresponding event response kernel based on the candidate response relationship, and extract the relative event time sequence of each physical field from the event response kernel.

[0120] The relative event time series uses the occurrence time of the natural charging event as the time reference, including background data before the event and response data after the event. The fault location device can select the corresponding observation location, physical field type, and time scale according to the candidate response relationship, avoiding the inclusion of channels unrelated to the candidate component in the initial time detection.

[0121] S502. Determine the background boundary of each physical field response based on the relative event time series.

[0122] Background boundaries are used to distinguish between background fluctuations and event responses, and can correspond to signal amplitude, frequency band energy, temperature gradient, or other response quantities, respectively. Different physical fields and different observation channels can have their own background boundaries.

[0123] S503. Based on the relative event time series and the background boundary, determine the first moment after the event occurs that continuously exceeds the background boundary and satisfies the continuity condition as the actual start moment.

[0124] The duration of the sampling condition can include the number of consecutive sampling points extending beyond the background boundary, the duration of the sampling, and the effective sampling ratio. For infrared thermal images with low sampling frequencies, this can be determined by the number of consecutive effective frames; for high-frequency magnetic field, acoustic, and vibration signals, it can be determined by the consecutive sampling intervals.

[0125] In one possible implementation of this application, a first background boundary for entering the response state and a second background boundary for exiting the response state can be set, and the actual start time is only confirmed after the response has been maintained outside the first background boundary for a preset duration. This implementation can reduce the start time jitter caused by repeated fluctuations in the response quantity near a single boundary, and solve the problem of difficulty in stable identification when weak responses are close to background fluctuations.

[0126] S504. Arrange the corresponding physical fields according to each actual start time, and collect each actual start time to obtain the actual start time of the response of each physical field and the measured response order.

[0127] For multiple physical fields whose initial time differences are less than the time synchronization error, they can be marked as responding simultaneously, without being required to determine the order of their responses. For physical fields with missing responses, the missing marker is retained in the measured response sequence for subsequent causal verification.

[0128] As can be seen, this embodiment determines the actual start time by combining the relative event time series, background boundary and duration conditions, which can reduce response misjudgment caused by isolated noise peaks and provide stable time series input for cross-physical field causal time series verification.

[0129] See Figure 11The specific implementation process of step S502 can be achieved through the following steps S601 to S604.

[0130] S601. Based on the relative event time series, the time series before the event occurs is divided into multiple background sub-windows.

[0131] The length of the background sub-window can be determined based on the physical field type and sampling frequency. Adjacent background sub-windows can partially overlap or be arranged continuously. Preferably, the background sub-windows do not contain other identified natural charging events.

[0132] S602. Determine the corresponding central statistic and fluctuation statistic based on the signal amplitude in each background sub-window.

[0133] Centrality statistics are used to represent the stable position of the background signal, while fluctuation statistics are used to represent the normal range of variation of the background signal. Centrality statistics can be the median, truncated mean, or other anomaly-resistant statistics; fluctuation statistics can be quantile ranges or absolute deviation statistics.

[0134] S603. Based on the central statistic and the fluctuation statistic, remove the abnormal background sub-windows to obtain the effective background statistic.

[0135] When the central statistic of a background sub-window deviates significantly from that of other background sub-windows, or when strong transient disturbances occur within the window represented by its fluctuation statistic, that background sub-window can be marked as an anomaly. The central and fluctuation statistics of the remaining background sub-windows constitute the valid background statistics.

[0136] S604. Determine the background boundary of each physical field response based on the effective background statistics.

[0137] Fault location equipment can save background boundaries according to physical field type, observation location, and event type. The background boundaries can be updated slowly with environmental conditions, but the background boundaries of the current event are not directly adjusted using abnormal response data after the event occurs.

[0138] As can be seen, this embodiment forms a background boundary by removing multiple background sub-windows and abnormal windows, which can avoid the pre-event interval being contaminated by temporary sound sources, mechanical collisions or the actions of adjacent devices, and improve the reliability of weak fault response start time identification.

[0139] See Figure 12 After obtaining the evidence set of candidate components and before performing spatial inversion, a horizontal comparison of similar events can be performed through the following steps S701 to S704.

[0140] S701. Based on the event type and operating conditions of the target event, obtain the event response core of the same type of component within the same charging pile or the historical events of the same component.

[0141] Components of the same type may include multiple power modules connected in parallel, multiple fans, or contactors with identical functions and structures. Historical events for the same component may include multiple natural closing events of the same main contactor or multiple natural starting events of the same liquid-cooled pump.

[0142] S702. Form a cluster of similar events based on the similar event response cores, and determine a real-time reference response core from the similar event clusters.

[0143] Events within a cluster of similar events should have the same or comparable event types, and their intensity and operating conditions should be within permissible ranges of variation. The real-time reference response core can be updated as more similar events become available during the current charging process.

[0144] S703. Based on the event response kernel of the target event and the real-time reference response kernel, determine the lateral deviation information of different observation positions and different physical fields.

[0145] Lateral deviation information describes the direction and degree of deviation of a target event from the real-time reference response kernel in terms of response amplitude, start time, decay time, frequency distribution, spatial center, and temperature rise per unit event intensity. Preferably, the lateral deviation information retains the corresponding observation location and physical field type.

[0146] S704. Update the candidate component evidence set according to the lateral deviation information, and input the updated candidate component evidence set into the spatial inversion.

[0147] When multiple physical field responses of a target component deviate along anomaly directions, while other similar components remain close to the real-time reference response kernel, the candidate strength of the target component can be increased. When multiple similar components simultaneously deviate in the same direction, the candidate strength of a single component failure can be reduced, while the candidate strength of common busbar, environmental conditions, or system-level causes can be increased.

[0148] As can be seen, this embodiment utilizes the natural similarity comparison relationship within the same charging pile to reduce the absolute threshold differences caused by different models, different sites and different sensor installation conditions, which is beneficial for locating the continuous degradation of a single parallel module or a single heat dissipation component.

[0149] See Figure 13 The specific implementation process of step S702 can be achieved through the following steps S801 to S803.

[0150] S801. Based on the event intensity and operating conditions corresponding to each type of event response core, select a set of valid type-of-the-event response cores from the type-of-the-event response cores.

[0151] The fault location device can compare the power change amplitude, current change amplitude, ambient temperature, number of activated modules, and status of heat dissipation equipment for each event. Response cores that lack critical channels, have time synchronization anomalies, or contain other strong transient events may not be included in the set of valid response cores for the same type of event.

[0152] S802. Merge the set of valid similar event response kernels according to component type and event type to obtain the similar event cluster.

[0153] For multiple parallel power modules, they can be grouped according to the module activation event; for fans or liquid-cooled pumps, they can be grouped according to the start, speed regulation, or stop events; for the historical actions of the same contactor, they can be grouped according to the closing and opening events.

[0154] S803. Align the relative event times of the clusters of similar events, and perform median aggregation on the aligned responses according to the observation location and physical field type to obtain the real-time reference response kernel.

[0155] Relative event time alignment can unify the occurrence times of different events as a comparison starting point. For channels with different sampling rates, they can be first mapped to a unified relative time coordinate of the corresponding physical field before aggregation. Median aggregation can reduce the impact of transient noise or weak anomalies in individual events on the real-time reference response kernel.

[0156] When the number of similar events is insufficient, the historical reference response core in the health baseline can be used temporarily; once the number of valid similar events reaches a preset number, it can be switched to the real-time reference response core. This process can support both the cold start of the equipment in its initial stage and gradually adapt to the current operating status of the equipment.

[0157] As can be seen, this embodiment forms a real-time reference response kernel by filtering operating conditions, merging events, aligning time, and aggregating medians, which makes the horizontal comparison have clear comparable conditions and reduces the diagnostic bias caused by the random fluctuation of a single reference event.

[0158] See Figure 14 After outputting the faulty component, location, and chain of evidence, the health baseline can be updated through the following steps S901 to S904.

[0159] S901. Obtain baseline update information; the baseline update information includes the verification result of the response order across physical fields, prohibited response information, diagnostic results, post-event fault code status, and charging end status.

[0160] The post-event fault code status can be continuously acquired during the observation period following the natural charging event. The charging completion status is used to determine whether the charging process ended normally, or whether it was prematurely terminated due to equipment malfunction, vehicle malfunction, or communication malfunction.

[0161] S902. Determine the baseline update verification result based on the baseline update information.

[0162] Baseline update verification results can include allow update and prohibit update. The fault location device can verify the response sequence across physical fields, prohibited response information, current diagnostic results, subsequent fault codes, and charging end status separately. Only when all verification items meet the health conditions will the update be allowed.

[0163] S903. When the baseline update verification result indicates that the cross-physical field response sequence is normal, no prohibited response occurs, the diagnostic result does not indicate a fault, no fault code is generated after the event, and the charging ends normally, the event response core and the corresponding operating conditions are added to the healthy baseline; otherwise, the event response core is prohibited from being added to the healthy baseline.

[0164] For event response kernels that have not yet completed post-event observation or charging, they can be saved as data to be verified instead of being immediately written to the health baseline. If a fault code or abnormal termination occurs subsequently, the data to be verified should be deleted or transferred to the abnormal sample set.

[0165] S904. Determine a comparative reference for the response to subsequent natural charging events based on the updated health baseline.

[0166] The updated health baseline can be saved separately according to event type, operating condition range, and component type. In subsequent event processing, the fault location device can select the response comparison reference closest to the current event from the health baseline and use it for background boundary determination, causal response template adjustment, or historical deviation judgment.

[0167] As can be seen, this embodiment controls the health baseline update by jointly controlling the event causal state, the prohibited response state, the diagnostic state, and the charging result, so that slow degradation events that have not yet triggered fault codes will not be mistakenly written as health data, thereby maintaining the reliability of the baseline used for long-term trend diagnosis.

[0168] See Figure 15 The charging pile traceless fault location system 1500 may include a communication interface 193, a processor 191, and an information acquisition module 1510, a response kernel generation module 1520, a causal verification module 1530, a spatial positioning module 1540, and a diagnostic output module 1550 implemented by the processor 191. The communication interface 193 is connected to external sensors of the charging pile for receiving external physical field signals.

[0169] The information acquisition module 1510 is used to acquire component mapping information and event observation information through the communication interface 193. The component mapping information includes component topology and spatial response dictionary, and the event observation information includes at least two types of synchronous external physical field signals and the event type, occurrence time, event intensity and operating conditions of natural charging events.

[0170] The information acquisition module 1510 can call pre-saved charging pile structure data to generate component topology, and obtain a spatial response dictionary through health data of the same model, external measurement data, or simulation data. The information acquisition module 1510 can also check sensor channel status, time synchronization status, and event recognition confidence status.

[0171] The response kernel generation module 1520 is used to determine the target event window and reference window based on the occurrence time and operating conditions, and to perform background elimination and event intensity normalization on the external physical field signals of the two windows to obtain the event response kernel.

[0172] The response kernel generation module 1520 can process physical field signals at different time scales. For fast responses such as magnetic fields, acoustics, and vibrations, it can retain millisecond-level dynamic changes; for thermal responses, it can retain thermal gradients and thermal diffusion changes from several seconds to tens of seconds after the event occurs.

[0173] The causal verification module 1530 is used to verify the permissible response relationship and cross-physical field response order based on the event response kernel, event type and component mapping information, and obtain the candidate component evidence set.

[0174] The causal verification module 1530 can call the set of allowed response components and causal response templates according to the event type, and generate prohibited response information, spatial consistency information, and cross-physical field causal temporal consistency information respectively. After the information is collected according to the candidate components, a candidate component evidence set is formed.

[0175] The spatial positioning module 1540 is used to perform spatial inversion constrained by the event-permitted response relationship based on the candidate component evidence set and the spatial response dictionary to obtain the fault location result.

[0176] The spatial positioning module 1540 can prioritize a small number of candidate components that can jointly interpret the external observation response and reduce the positioning priority of components outside the set of components that can respond to the event. When the response distributions of adjacent components are similar, the spatial positioning module 1540 can further distinguish them by combining the component topology and the spatial center across the physical field.

[0177] The diagnostic output module 1550 is used to output the faulty component, location, and evidence chain based on the fault location results.

[0178] The diagnostic output module 1550 can output the name of the faulty component, its internal location, the corresponding external observation area, candidate fault types, confidence level, critical response time, and recommended handling level to the operation and maintenance terminal. Each piece of evidence in the chain of evidence can be traced back to the corresponding natural charging event, physical field type, and external observation location.

[0179] The charging pile seamless fault location system 1500 can be deployed on edge computing devices near a single charging pile or on a charging station diagnostic server. Each functional module can be implemented using software, firmware, or a combination of hardware and software, and can be distributed on the edge and server sides according to real-time processing capabilities. Through data transfer between these modules, the charging pile seamless fault location system 1500 can generate component-level fault location results and a verifiable chain of diagnostic evidence without altering the normal charging control process.

[0180] This application also provides a computer storage medium storing a computer program for electronic data interchange, which causes a computer to perform some or all of the steps of any of the methods described in the above method embodiments, wherein the computer includes an electronic device.

[0181] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0182] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0183] In the several embodiments provided in this application, it should be understood that the disclosed apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above 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; the indirect coupling or communication connection between devices or units may be electrical or other forms.

[0184] The units described above 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.

[0185] 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.

[0186] If the aforementioned integrated units are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage device (CMD). 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 memory and includes several instructions to enable a computer device, such as a personal computer, electronic device, or network device, to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0187] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, which may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, etc.

[0188] The embodiments of this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for traceless fault location of charging piles, executed by a fault location device equipped with a processor and a communication interface, wherein the communication interface is connected to an external sensor of the charging pile, characterized in that, include: Obtain component mapping information and event observation information; The component mapping information includes component topology and spatial response dictionary, and the event observation information includes at least two types of external physical field signals from synchronously acquired magnetic, acoustic, vibration and thermal fields, as well as the event type, occurrence time, event intensity and operating conditions of passively identified natural charging events. The target event window is determined based on the occurrence time, and the reference window is determined from the candidate windows that do not contain the target event based on the operating conditions. Background removal and event intensity normalization are performed on the external physical field signals of the target event window and the reference window to obtain the event response kernel; Based on the event response kernel, the event type, and the component mapping information, the event-permitted response relationship and the cross-physical field response order are verified to obtain a candidate component evidence set. Based on the candidate component evidence set and the spatial response dictionary, perform spatial inversion constrained by the event-permissible response relationship to obtain the fault location result; Based on the fault location results, the faulty component, its location, and the chain of evidence are output.

2. The method according to claim 1, characterized in that, The step involves verifying the permissible event response relationship and cross-physics response order based on the event response kernel, the event type, and the component mapping information, to obtain a candidate component evidence set, including: Based on the event response kernel, the event type, and the component mapping information, extract the response feature set corresponding to the event type; Based on the event type and the component mapping information, determine the set of components that can respond to the event and the causal response template; The candidate component evidence set is determined based on the response feature set, the event-allowed response component set, and the causal response template.

3. The method according to claim 2, characterized in that, The step of determining the candidate component evidence set based on the response feature set, the event-allowed response component set, and the causal response template includes: Based on the response feature set, the event-allowed response component set, and the causal response template, establish candidate response relationships between response features and candidate components; Based on the candidate response relationship, prohibition response information is determined for components outside the set of allowed response components for the event, and spatial consistency information is determined for components within the set of allowed response components for the event. Based on the candidate response relationship and the causal response template, determine cross-physical field causal temporal consistency information; The candidate component evidence set is determined based on the prohibited response information, the spatial consistency information, and the cross-physical field causal temporal consistency information.

4. The method according to claim 3, characterized in that, The step of determining cross-physical field causal temporal consistency information based on the candidate response relationship and the causal response template includes: Based on the candidate response relationship and the causal response template, the expected physical field response order and the allowable time delay interval between adjacent responses are determined; The actual start time and measured response sequence of each physical field response are determined based on the candidate response relationship. Based on the actual start time, the measured response sequence, and the allowable delay interval, the expected physical field response sequence is verified segment by segment to obtain the cross-physical field causal timing consistency information.

5. The method according to claim 4, characterized in that, The step of determining the actual start time and measured response order of each physical field response based on the candidate response relationship includes: Based on the candidate response relationship, locate the corresponding event response kernel, and extract the relative event time series of each physical field from the event response kernel; The background boundaries of each physical field response are determined based on the relative event time series. Based on the relative event time series and the background boundary, the first moment after the event occurs that continuously exceeds the background boundary and satisfies the continuity condition is determined as the actual start time; Arrange the corresponding physical fields according to each actual start time, and collect each actual start time to obtain the actual start time of the response of each physical field and the measured response order.

6. The method according to claim 5, characterized in that, The step of determining the background boundary of each physical field response based on the relative event time series includes: Based on the relative event time sequence, the time sequence before the event occurs is divided into multiple background sub-windows; Determine the corresponding central statistic and fluctuation statistic based on the signal amplitude in each background sub-window; Based on the central statistic and the fluctuation statistic, abnormal background sub-windows are removed to obtain the effective background statistic; The background boundaries of each physical field response are determined based on the effective background statistics.

7. The method according to any one of claims 1 to 6, characterized in that, After obtaining the candidate component evidence set and before performing the spatial inversion, the method further includes: Based on the event type and operating conditions of the target event, obtain the event response kernel of the same type of component within the same charging pile or the historical events of the same component; Based on the similar event response kernels, similar event clusters are formed, and real-time reference response kernels are determined from the similar event clusters; Based on the event response kernel of the target event and the real-time reference response kernel, determine the lateral deviation information for different observation locations and different physical fields; The candidate component evidence set is updated based on the lateral deviation information, and the updated candidate component evidence set is input into the spatial inversion.

8. The method according to claim 7, characterized in that, The step of forming a cluster of similar events based on the similar event response cores, and determining a real-time reference response core from the similar event clusters, includes: Based on the event intensity and operating conditions corresponding to each type of event response kernel, a set of valid type-of-the-event response kernels is selected from the type-of-the-event response kernels. The set of valid similar event response kernels is merged according to component type and event type to obtain the similar event cluster; The event clusters of the same type are aligned relative to each other at specific times, and the aligned responses are then aggregated at the median according to the observation location and the type of physical field to obtain the real-time reference response kernel.

9. The method according to any one of claims 1 to 6, characterized in that, After outputting the faulty component, the location, and the chain of evidence, the method further includes: Obtain baseline update information; the baseline update information includes verification results of response order across physical fields, prohibited response information, diagnostic results, post-event fault code status, and charging end status; The baseline update verification result is determined based on the baseline update information; When the baseline update verification result indicates that the cross-physical field response sequence is normal, no prohibited response occurs, the diagnostic result does not indicate a fault, no fault code is generated after the event, and charging ends normally, the event response core and corresponding operating conditions are added to the healthy baseline; otherwise, the event response core is prohibited from being added to the healthy baseline. The updated health baseline is used to determine the response comparison for subsequent natural charging events.

10. A traceless fault location system for charging piles, characterized in that, It includes a communication interface for connecting to external sensors of the charging pile, and the following modules implemented by a processor: The information acquisition module is used to acquire component mapping information and event observation information through the communication interface; the component mapping information includes component topology and spatial response dictionary, and the event observation information includes at least two types of synchronous external physical field signals and the event type, occurrence time, event intensity and operating conditions of natural charging events; The response kernel generation module is used to determine the target event window and the reference window based on the occurrence time and the operating conditions, and to perform background elimination and event intensity normalization on the external physical field signals of the two windows to obtain the event response kernel. The causal verification module is used to verify the permissible response relationship and cross-physical field response order of the event based on the event response kernel, the event type and the component mapping information, and to obtain a candidate component evidence set; The spatial positioning module is used to perform spatial inversion constrained by the event-permitted response relationship based on the candidate component evidence set and the spatial response dictionary to obtain the fault location result; The diagnostic output module is used to output the faulty component, location, and evidence chain based on the fault location results.

Images