A fault detection method for an electric bicycle battery charging cabinet

Abstract

The present application relates to the technical field of charging fault detection, and provides a fault detection method for an electric bicycle battery charging cabinet, which comprises: generating a trace record for a battery connected to a charging grid, performing a diagnostic charge on the grid before normal charging, collecting voltage variation and current variation, and calculating a health value of the grid; collecting a current sequence according to a sliding time window during the normal charging stage, calculating current fluctuation amplitude and average current to evaluate process stability, comparing the process stability with a preset threshold to determine grid fluctuation abnormalities; using the charging trace record with abnormalities as an abnormal sample, calculating an abnormality proportion, calculating a comparison difference, and analyzing and locating the source of abnormalities; performing disposal on the grid according to the abnormal type and the locating result, and recalculating the health value and the process stability as a review when charging is started again. Thus, the grid and battery faults of the charging cabinet can be traced, and self-recovery disposal can be performed in stages.

Description

Technical Field

[0001] This invention relates to the field of charging fault detection technology, and specifically to a fault detection method for an electric bicycle battery charging cabinet. Background Technology

[0002] With the continuous growth in the number of electric bicycles, centralized charging cabinets have been widely deployed due to their small footprint and convenient management. During the charging process, batteries vary in type and aging, and the charging cabinets operate in an environment of high-frequency plugging and unplugging, high humidity and dust, and unstable heat dissipation, which can easily lead to malfunctions such as charging interruption, abnormal overheating, poor contact, and relay failure.

[0003] Chinese patent application number CN202411279173.1, entitled "Fault Detection Method and System for Electric Bicycle Battery Charging Cabinets," discloses a fault detection method for electric bicycle battery charging cabinets. This method includes: dividing a day into four time intervals based on peak and off-peak electricity consumption and setting different weighting coefficients; collecting time-series data on load power, voltage, current, and temperature; calibrating the load power using voltage, current, and temperature; calculating the average power, maximum power, and minimum power for each time interval; collecting load power, voltage, current, and temperature data for the charging cabinet under test; calibrating the load power and calculating its average power, maximum power, and minimum power; generating an abnormal fault index by combining the weighting coefficients, average power, maximum power, and minimum power; comparing this index with a preset threshold to determine whether the charging cabinet has malfunctioned. By collecting and calibrating time-series data of various parameters, and combining the time intervals and weighting coefficients, an abnormal fault index is generated to determine whether the electric bicycle battery charging cabinet has malfunctioned.

[0004] However, in the scenario of centralized charging cabinets for food delivery, the high frequency of plugging and unplugging causes the contact parts to wear out faster, which can easily lead to repeated abnormalities in some compartments. There is a lack of detection methods that can accurately locate the problem to the compartment or component. At the same time, since batteries are transferred between different food delivery riders, incompatible or poor-condition batteries may be mixed in. If the cabinet only judges whether it can be charged, it will be difficult to identify high-risk batteries in time. Summary of the Invention

[0005] To address the aforementioned problems, this invention proposes a fault detection method for electric bicycle battery charging cabinets. This method establishes charging traceability records, calculates health values ​​using diagnostic charging, collects stable charging fluctuation values, calculates process stability, establishes a cross-reference table, analyzes and locates the source of anomalies, implements tiered handling and isolation by compartment, and performs self-recovery for verification.

[0006] The technical solution adopted by this invention to solve the above-mentioned technical problem is: a fault detection method for an electric bicycle battery charging cabinet, comprising:

[0007] A traceability record is generated for the battery connected to the charging port. The record is identified by a combination of port number, start time, and battery QR code. Before normal charging, the port is diagnostically charged, and the voltage and current changes are collected to calculate the health value of the port.

[0008] During the normal charging phase, the current sequence is collected according to the sliding time window, and the current fluctuation amplitude and average current are calculated to evaluate the process stability. The process stability is compared with the preset threshold to determine the abnormal fluctuation of the grid.

[0009] Using charging traceability records with abnormal fluctuations or abnormal health values ​​as abnormal samples, the abnormal proportion of grid slots and the abnormal proportion of batteries in the records are calculated respectively. The difference between the control and control is calculated, and the source of the abnormality is analyzed and located according to the abnormality judgment rules.

[0010] Based on the anomaly type and location results, the grid is handled accordingly. When charging resumes, the health value and process stability are recalculated for verification. If the verification still indicates an anomaly, the handling is escalated. If the verification indicates no anomaly, the alarm is deactivated and the grid service is restored.

[0011] As a preferred embodiment, the specific steps for performing diagnostic charging on the grid port before entering normal charging are as follows:

[0012] Read the QR code on the appearance of the battery to be charged as the battery identifier, and at the same time read the grid number and the current start time. Combine the grid number, start time and battery QR code according to the preset encoding rules to generate a traceability record identifier (RID). Write the RID, the corresponding grid relay status and charging module number into the local storage and report to the operation and maintenance platform.

[0013] Before starting the normal charging process, diagnostic charging is performed on the target grid port. First, the output current is set to 0, and the grid port output circuit is kept in a controllable state. Then, after the relay is closed, a current step command is issued to directly switch the output current from 0 to the diagnostic current and maintain it. The sampling time before the current step is recorded. Sampling time during the hold phase The voltage and output current of the target grid terminal are collected respectively to obtain the voltage change and current change, and the health value of the target grid is calculated based on these values. The formula is as follows:

[0014] ,

[0015] in Indicates the sampling time The voltage of the grid terminals, Indicates the sampling time ; output current;

[0016] When the diagnostic charging is finished, set the output current back to 0, and confirm that the current has dropped to 0 before controlling the relay to disconnect;

[0017] When the health value meets the preset unqualified judgment conditions, the traceability record is marked as an interface health abnormality and the port is prohibited from entering normal charging. At the same time, port isolation and prompt for port replacement are performed. When the health value meets the preset qualified judgment conditions, the health value is written into the traceability record and the port is allowed to enter the normal charging process.

[0018] As a preferred embodiment, the specific steps for calculating the current fluctuation amplitude and average current to evaluate process stability are as follows:

[0019] Once the health value of the grid port is deemed acceptable, it enters the normal charging process, and the output current of the port is continuously collected. A sliding time window with a window length of 10 seconds is constructed. A current sequence is formed for the current sampling points within each sliding time window. The current fluctuation amplitude and average current within each window are calculated at each sliding time to obtain a process stability index reflecting the stability of the grid port's charging process. The formula is:

[0020] ,

[0021] Where stable represents the process stability index. This indicates the maximum current within the sliding window. This represents the minimum current within the sliding window. This represents the average current within the sliding window. represents a constant to prevent the average current from approaching zero and causing division by zero, and max represents the maximum value function.

[0022] As a preferred embodiment, the specific steps for comparing process stability with a preset threshold to determine abnormal fluctuations in the grid are as follows:

[0023] The process stability of multiple consecutive sliding time windows is used for judgment. When the process stability corresponding to three consecutive adjacent sliding time windows is greater than or equal to the stability threshold, it is determined that there is an abnormal fluctuation in the grid, and the abnormal fluctuation is marked and written into the traceability record corresponding to the grid. At the same time, the graded handling strategy is triggered. When the process stability does not meet the continuous over-threshold condition, normal charging is maintained and rolling evaluation continues.

[0024] As a preferred embodiment, the specific steps for calculating the grid anomaly ratio and the battery anomaly ratio in the records, and calculating the comparison difference, are as follows:

[0025] A cross-reference table is constructed using abnormal charging traceability records as abnormal samples, and the source of the abnormality is determined based on the difference in the abnormality ratio. First, a set of traceability records is maintained in local storage. Each traceability record includes a record identifier (RID), grid number, battery identifier, start time, and abnormality flag. The abnormality flag is written by the diagnostic process and includes interface health abnormality and fluctuation abnormality. Rolling statistics are performed on the traceability records of the most recent 7 days. The total number of charging times and the number of abnormalities for each grid are accumulated by grid number. The abnormality ratio of the grid is obtained by dividing the number of abnormalities by the total number of charging times. At the same time, the total number of charging times and the number of abnormalities for each battery are accumulated by battery identifier. The abnormality ratio of the battery is obtained by dividing the number of abnormalities by the total number of charging times.

[0026] When any traceability record is marked as an anomaly, the counters for the corresponding grid number and battery identifier are updated. The total number of charging times for that grid is accumulated at the grid level, and the anomaly count is accumulated when an anomaly occurs. The total number of charging times for that battery is accumulated at the battery level, and the anomaly count is accumulated when an anomaly occurs. These are then written into the corresponding fields of the cross-reference table. The grid number and battery identifier of the anomaly traceability record are associated and written into the association list as an association item, so that the anomaly distribution of the same battery in different grids and the same grid for different batteries can be directly compared in the same table.

[0027] As a preferred implementation, the specific steps for analyzing and locating the source of the anomaly based on the anomaly determination rules are as follows:

[0028] For the same abnormal sample record, calculate the comparison difference between the corresponding grid abnormality ratio and the battery abnormality ratio, and compare the comparison difference with a preset threshold to locate the source of the abnormality. When the comparison difference meets the grid-side judgment rule, determine that the source of the abnormality is a grid abnormality, locate the faulty object to that grid, trigger the isolation of that grid, and allow other grids to continue to provide services. When the comparison difference meets the battery-side judgment rule, determine that the source of the abnormality is a battery abnormality, locate the faulty object to that battery, and trigger the restriction of charging for that battery.

[0029] As a preferred implementation, the specific steps for handling the grid based on the anomaly type and location result are as follows:

[0030] After obtaining the anomaly type and location result, the target grid is subjected to graded handling, and a review is performed when charging starts again to achieve self-recovery closed loop. When the anomaly type corresponding to the trace record is interface health anomaly, that is, the grid health value obtained from the diagnostic charging exceeds the threshold, or fluctuation anomaly, that is, the stability of the sliding window process exceeds the threshold, and the location result is grid anomaly, only the target grid is handled without stopping other grid services.

[0031] The handling is divided into three levels: Level 1, Level 2, and Level 3. Level 1 involves displaying a prompt to the user and the cabinet screen to replace the compartment, limiting the output current of the compartment, and ensuring that the compartment can be restarted for charging. Level 2 involves locking the compartment and preventing it from entering the normal charging process, while keeping the cabinet door open to remove the battery, and marking the compartment as pending maintenance. Level 3 involves immediately setting the output current of the compartment to 0, disconnecting the compartment relay after confirming that the current is 0, triggering an audible and visual alarm, and reporting to the maintenance platform to generate a work order.

[0032] If the location result indicates a battery abnormality, the battery identifier is added to the restricted list, and charging is refused for the current charge associated with that battery, while the target cell is not locked. When a trigger event is detected that the same cell has started charging again, a verification mode is entered. The trigger event is the successful reading of the battery QR code and the generation of a new traceability record. In verification mode, diagnostic charging is repeatedly performed on the cell, and the health value is recalculated. At the same time, after entering normal charging, the process stability is calculated according to a fixed window length to verify whether the abnormality judgment conditions are still met. If the verification result is still abnormal, the handling level is upgraded step by step according to the number of consecutive abnormalities and written to the corresponding handling field of the traceability record. If the verification result is not abnormal, the alarm mark of the cell is cleared and the lock is released, restoring the cell's ability to enter normal charging. At the same time, the verification pass result is written to the traceability record and used as a normal sample for subsequent statistics.

[0033] Beneficial effects

[0034] Compared with the prior art, the present invention has the following advantages:

[0035] 1. By diagnosing charging health values ​​and assessing process stability, and combining the differences between grid and battery cross-comparison values, the source of anomalies can be accurately located to the grid side and the battery side.

[0036] 2. A graded handling and self-recovery mechanism based on grid isolation and recharging verification is adopted. Only abnormal grids are isolated or abnormal batteries are restricted, while other grids continue to provide charging services. This improves the efficiency of the charging cabinet and reduces the occurrence of accidental shutdowns that affect charging. Attached Figure Description

[0037] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation on the scope of this application.

[0038] Figure 1 This is a flowchart illustrating the present invention;

[0039] Figure 2This is a comparison diagram of the effects of the present invention and the prior art, where gray bars represent the prior art and black bars represent the present invention. Detailed Implementation

[0040] To make the technical means, creative features, achieved objectives, and effects of this invention easier to understand, the invention is further described below with reference to specific embodiments. However, the following embodiments are merely preferred embodiments of this invention and not all of them. Other embodiments obtained by those skilled in the art based on the embodiments described herein without creative effort are all within the protection scope of this invention.

[0041] Example 1:

[0042] To achieve the above objectives, please refer to Figures 1 to 2 This invention provides a fault detection method for electric bicycle battery charging cabinets, the method comprising the following steps:

[0043] Step S1: Establish charging traceability records and diagnose charging health values;

[0044] Step S2: Collect stable charging fluctuation values ​​and calculate process stability;

[0045] Step S3: Establish a cross-reference table to analyze and locate the source of the anomaly;

[0046] Step S4: Classify and isolate according to grid, and perform self-recovery verification.

[0047] This method is implemented in the order of S1–S4, and its overall process is as follows:

[0048] A traceability record is generated for the battery connected to the charging port. The record is identified by a combination of port number, start time, and battery QR code. Before normal charging, the port is diagnostically charged, and the voltage and current changes are collected to calculate the health value of the port.

[0049] During the normal charging phase, the current sequence is collected according to the sliding time window, and the current fluctuation amplitude and average current are calculated to evaluate the process stability. The process stability is compared with the preset threshold to determine the abnormal fluctuation of the grid.

[0050] Using charging traceability records with abnormal fluctuations or abnormal health values ​​as abnormal samples, the abnormal proportion of grid slots and the abnormal proportion of batteries in the records are calculated respectively. The difference between the control and control is calculated, and the source of the abnormality is analyzed and located according to the abnormality judgment rules.

[0051] Based on the anomaly type and location results, the grid is handled accordingly. When charging resumes, the health value and process stability are recalculated for verification. If the verification still indicates an anomaly, the handling is escalated. If the verification indicates no anomaly, the alarm is deactivated and the grid service is restored.

[0052] The specific steps for establishing charging traceability records and diagnosing charging health values ​​are as follows:

[0053] When a delivery rider inserts a battery into the target charging slot and the cabinet door is closed, the controller in the charging cabinet reads the QR code on the battery's exterior as the battery identifier, and at the same time reads the slot number and the current start time. According to the preset coding rules, the slot number, start time, and battery QR code are combined to generate a traceability record identifier (RID). The RID, along with the corresponding slot's relay status and charging module number, is written to local storage and reported to the operation and maintenance platform, so that any subsequent anomaly can be traced back to the specific slot and the specific battery.

[0054] Specifically, the controller reads the QR code on the battery's exterior as the battery identifier. When the rider opens the compartment door and places the battery inside, the door's magnetic switch detects the change in the closed state and outputs a trigger signal to the controller. Upon receiving the trigger signal, the controller activates the QR code reading component corresponding to that compartment, such as a fixed scanning module. The controller controls the QR code reading component to complete one QR code acquisition and decoding operation within a preset time window: the controller drives the auxiliary light to illuminate, acquires the image frame of the QR code on the battery surface, performs QR code positioning and decoding on the image frame, and obtains the QR code string data; when decoding is successful, the QR code string is used as the battery identifier BID, and combined with the compartment number SID and the start time. The system combines and generates a traceability record identifier (RID), writes it to local storage, and enters the subsequent diagnostic charging process. If decoding fails within the preset time window, the controller performs a resampling once. If it still fails, the record is marked as unrecognized QR code, and the rider is prompted to adjust the battery position and close the door again to trigger reading.

[0055] Before initiating the normal charging process, the controller performs diagnostic charging on the target grid port. Specifically, it first sets the output current of the charging module to 0 and keeps the grid port output circuit in a controllable state. Then, after the relay closes, it sends a current step command to the charging module, causing the output current to switch directly from 0 to the preset diagnostic current and maintain it for a preset holding time. Specifically, the preset diagnostic current is set according to 10% of the channel's rated charging current and is limited to between 0.5A and 3A. The controller samples the current before the step command. Sampling time during the hold phase The voltage and output current of the target grid terminal are collected respectively to obtain the voltage change and current change, and the health value of the target grid is calculated based on these values. The formula is as follows:

[0056] ,

[0057] in Indicates the sampling time The voltage of the grid terminals, Indicates the sampling time ; output current;

[0058] Specifically, the calculation of the health value of the target cell, taking a deployed 24-cell charging cabinet as an example, involves a normal charging rated current of 10A for each cell. After the battery is inserted and a traceability record (RID) is generated, the charging cabinet controller performs a step-excitation diagnostic charging on the target cell before entering constant current charging. This step-excites the output current from 0A to a diagnostic current of 1.0A, holding it for 1.0s. The sampling frequency is 100Hz. To eliminate sampling noise and transient oscillations, the controller performs window averaging on the voltage and current before and after the step. Specifically, it takes 5 sampling points within a 50ms window before the step and averages them as the result. and Five sampling points are taken within a window of 200ms to 250ms after the step jump, and the average is taken as the result. and The controller calculates the voltage and current changes accordingly and obtains the grid health value. Simultaneously, it writes the value (H) into the traceability record for subsequent fault location and maintenance tracing. To ensure clear judgment, the health value threshold is set to 0.12. When the value is greater than or equal to 0.12, the grid interface contact status is deemed unqualified, and normal charging is prohibited. When the value is less than 0.12, normal charging is allowed. In actual implementation, the health value threshold can be determined by adding the upper limit of the grid health value obtained from diagnostic charging of each grid under standard load at the factory stage, along with environmental drift margin and sampling measurement error margin, and then fixing it.

[0059] When the diagnostic charging is finished, the controller sets the output current back to 0 and controls the relay to disconnect after confirming that the current has dropped to 0. When the health value meets the preset unqualified judgment condition, the controller marks the trace record as an interface health abnormality and prohibits the interface from entering normal charging. At the same time, it performs interface isolation and prompts to change the interface. When the health value meets the preset qualified judgment condition, the controller writes the health value into the trace record and allows the interface to enter the normal charging process.

[0060] By tracing and recording the RID identifier, the association between the battery and the grid is realized. The grid interface is quantitatively pre-checked by using current step diagnostic charging. The voltage response to current step directly reflects the interface contact status, enabling fault detection to be brought forward to before charging begins and the anomaly to be located to the specific grid, reducing false alarms and shortening the maintenance and troubleshooting path.

[0061] The specific steps for collecting stable charging fluctuation values ​​and calculating process stability are as follows:

[0062] Once the health value of the grid port is deemed acceptable, it enters the normal charging process, and the output current of the port is continuously collected. A sliding time window with a window length of 10 seconds is constructed. A current sequence is formed for the current sampling points within each sliding time window. The current fluctuation amplitude and average current within each window are calculated at each sliding time to obtain a process stability index reflecting the stability of the grid port's charging process. The formula is:

[0063] ,

[0064] Where stable represents the process stability index. This indicates the maximum current within the sliding window. This represents the minimum current within the sliding window. This represents the average current within the sliding window. The constant represents the function to prevent the average current from approaching zero and causing division by zero; max represents the maximum value function.

[0065] Specifically, the current fluctuation amplitude is used to characterize the difference between the maximum and minimum current values ​​within the window, the average current is used to characterize the overall current level within the window, and the process stability index is the normalized ratio of the current fluctuation amplitude to the average current. The process stability is determined by combining multiple consecutive sliding time windows. When the process stability corresponding to three consecutive adjacent sliding time windows is greater than or equal to the stability threshold, the controller determines that there is an abnormal fluctuation in the grid and writes the abnormal fluctuation mark into the traceability record corresponding to the grid. At the same time, a graded handling strategy is triggered, such as current limiting for the grid and prompting the rider to replace the grid. When the process stability does not meet the continuous over-threshold condition, the controller maintains normal charging and continues to perform rolling evaluation.

[0066] Specifically, the stability threshold is determined by factory calibration. Normal charging sampling is performed on each grid under standard load, and the process stability is calculated according to a fixed window length. The maximum value of all grids under normal operating conditions is taken as the benchmark, and then the margin formed by charging module ripple, ambient temperature and sampling error is added.

[0067] By combining the sliding time window with stability indicators and continuous over-threshold judgment, this technology differs from the existing technology's judgment method for single sampling points. It can effectively identify continuous fluctuations caused by intermittent terminal contact, micro-discontinuity of wiring harness, and relay jitter, thereby improving the ability to detect early hidden dangers and reducing the false shutdown rate.

[0068] The specific steps for establishing a cross-reference table and analyzing and locating the source of the anomaly are as follows:

[0069] The charging cabinet controller constructs a cross-reference table using abnormal charging traceability records as abnormal samples, and determines the source of the abnormality based on the difference in the abnormality ratio. First, a set of traceability records is maintained in local storage. Each traceability record includes a record identifier (RID), a compartment number, a battery identifier, a start time, and an abnormality marker. The abnormality marker is written by the diagnostic process and includes interface health abnormalities and fluctuation abnormalities. Rolling statistics are performed on the traceability records of the most recent 7 days. The total number of charging times and the number of abnormalities for each compartment are accumulated by the compartment number. The abnormality ratio of the compartment is obtained by dividing the number of abnormalities by the total number of charging times. At the same time, the total number of charging times and the number of abnormalities for each battery are accumulated by the battery identifier. The abnormality ratio of the battery is obtained by dividing the number of abnormalities by the total number of charging times.

[0070] Specifically, the controller maintains a cross-reference table in local storage that associates grid number with battery identifier. Using the charging traceability records of the most recent 7 days as the data source, when any traceability record is marked as an anomaly, the counters for the corresponding grid number and battery identifier are updated. Specifically, for the grid dimension, the total number of charging times for that grid is incremented, and the anomaly count is incremented when an anomaly occurs; for the battery dimension, the total number of charging times for that battery is incremented, and the anomaly count is incremented when an anomaly occurs. These are then written to the corresponding fields in the cross-reference table. The grid number and battery identifier association of the anomaly traceability record are combined as an association item and written to the association list. This allows for direct comparison of the anomaly distribution of the same battery across different grids, and the same grid for different batteries, within the same table. For example, if grid 05 shows 20 anomalies out of 100 charging cycles in the statistics window, then the corresponding field for grid 05 in the cross-reference table is recorded as follows: =100、 =20、 =0.20; Battery B-901 experienced two anomalies during 30 charging cycles, so the corresponding field record for battery B-901 in the cross-reference table is: =30、 =2、 =0.0667, where Indicates the number of times the grid port has been charged. Indicates the number of grid anomalies. Indicates the percentage of grid anomalies. Indicates the number of times the battery has been charged. Indicates the number of battery malfunctions. Indicates the percentage of abnormal battery capacity;

[0071] For the same abnormal sample record, calculate the comparison difference between the corresponding grid anomaly ratio and the battery anomaly ratio, and compare the comparison difference with a preset threshold to locate the source of the anomaly. When the comparison difference meets the grid-side judgment rule, the source of the anomaly is determined to be a grid anomaly, the faulty object is located to that grid, isolation measures are triggered for that grid, and other grids are allowed to continue service. When the comparison difference meets the battery-side judgment rule, the source of the anomaly is determined to be a battery anomaly, the faulty object is located to that battery, and charging restriction measures are triggered for that battery. The formula for calculating the comparison difference is:

[0072] ,

[0073] in Indicates the percentage of grid anomalies. Indicates the percentage of abnormal battery capacity;

[0074] Specifically, the method of comparing the control difference with a preset threshold to locate the source of the anomaly involves setting the preset threshold through a combination of online calibration period and fixed threshold. In the initial stage of online operation, the calibration period is set to the most recent 7 days. Anomaly tracing records within this period are collected, and the root cause of the anomaly is confirmed through a closed-loop operation and maintenance system. When the anomaly disappears after terminal tightening or relay replacement of a certain grid, the corresponding record is marked as a grid root cause sample, and its control difference is calculated to form a grid root cause difference set. When the anomaly disappears after battery replacement, the corresponding record is marked as a battery root cause sample, and its control difference is calculated to form a battery root cause difference set. The minimum value of the grid root cause difference set is taken as the grid threshold, and the minimum absolute value of the battery root cause difference set is taken as the battery threshold. The grid threshold and battery threshold are written into a parameter table for fixed operation. During operation, if the control difference is greater than or equal to the grid threshold, the grid anomaly is located; if the control difference is less than or equal to the negative battery threshold, the battery anomaly is located.

[0075] Compared with existing technologies that only provide instantaneous threshold alarms and output channel anomalies for single charging data, this implementation method establishes a cross-reference table by tracking anomaly records, calculates the anomaly ratio difference, and combines the technical features of judgment rules to transform anomalies from one-off phenomena into statistically attributable results. This allows the source of anomalies to be clearly located on both the grid side and the battery side, making it suitable for high-concurrency mixed-use operating conditions of centralized charging cabinets for food delivery, reducing accidental grid shutdowns and shortening the maintenance and troubleshooting path.

[0076] The specific steps for tiered handling and isolation based on designated grid areas, along with self-recovery verification, are as follows:

[0077] After obtaining the anomaly type and location result, the target cell is subjected to tiered handling, and a review is performed when charging resumes to achieve a self-recovery closed loop. When the anomaly type corresponding to the trace record is interface health anomaly (i.e., the cell health value obtained from the diagnostic charging exceeds the threshold) or fluctuation anomaly (i.e., the stability of the sliding window process exceeds the threshold), and the location result is a cell anomaly, the controller only handles the target cell without stopping other cell services. The handling is divided into first-level handling, second-level handling, and third-level handling. Specifically, first-level handling outputs a prompt to replace the cell to the user terminal and the cabinet screen, and limits the output current of the cell, while keeping the cell able to start charging again; second-level handling locks the cell and prohibits entry into the normal charging process, the controller keeps the cabinet door open to remove the battery, and marks the cell as pending maintenance; third-level handling immediately sets the output current of the cell to 0 and disconnects the cell relay after confirming that the current is 0, while triggering an audible and visual alarm. An alarm is triggered and a work order is generated on the operation and maintenance platform. If the location result indicates a battery abnormality, the battery identifier is added to the restricted list, and charging is refused for the current charging associated with that battery, while the target cell is not locked. To enable verification and escalation, the controller enters verification mode when it detects a trigger event that the same cell has started charging again. The trigger event is the successful reading of the battery QR code and the generation of a new traceability record. In verification mode, diagnostic charging is repeatedly performed on the cell, and the health value is recalculated. At the same time, after entering normal charging, the process stability is calculated according to a fixed window length to verify whether the abnormality judgment conditions are still met. When the verification result is still abnormal, the handling level is upgraded step by step according to the number of consecutive abnormalities and written to the corresponding handling field of the traceability record. When the verification result is not abnormal, the alarm mark of the cell is cleared and the lock is released, restoring the cell's ability to enter normal charging. At the same time, the verification pass result is written to the traceability record and used as a normal sample for subsequent statistics.

[0078] By employing a tiered approach to handling issues based on the number of charging slots, a re-charging-triggered verification process, and a combination of verification failure escalation and successful self-recovery technologies, the system can differentiate between transient disturbances caused by high-frequency plugging and unplugging in centralized food delivery scenarios and actual faults without altering the hardware structure of the charging cabinet. This avoids the current technology's approach of shutting down the entire cabinet to handle service interruptions caused by single-slot anomalies, and ensures that the handling decisions have repeatable judgment criteria and a traceable record link.

[0079] like Figure 2 The graph shows the effect comparison of a fault detection method for an electric bicycle battery charging cabinet. The horizontal axis lists the key performance indicators, and the vertical axis represents the exemplified performance scores, ranging from 0 to 100. The higher the value, the better the performance. The purpose is to intuitively demonstrate the expected improvement of the present invention in key capabilities compared to typical prior art.

[0080] Example 2:

[0081] Based on the above Embodiment 1, a fault detection method for an electric bicycle battery charging cabinet in the scenario of a centralized charging cabinet for takeaway is specifically the following solution:

[0082] Step 1, when the takeaway rider inserts the battery into the target charging slot and the cabinet door is closed, the controller activates the communication interface of this slot, sends a handshake request to the battery BMS (Battery Management System) through the communication line of the charging connector, and reads the battery serial number as the battery identifier; the controller simultaneously reads the slot number and the start time, combines the battery identifier, the slot number and the start time according to the preset coding rule to generate a trace record identifier, and writes the trace record identifier, the relay status of this slot and the charging module number into the local storage and reports to the operation and maintenance platform; before entering normal charging, the controller performs a dual-level pulse diagnostic charge on the target slot, sequentially outputs the first diagnostic current and the second diagnostic current after the relay is closed and maintains the preset duration respectively, where the first diagnostic current is 10% of the channel rated charging current and the second diagnostic current is 20% of the channel rated charging current; collect the terminal voltage and current in the stable section, calculate the slot health value, when it exceeds the health value threshold, mark the trace record as interface health abnormality and isolate this slot, when it is less than the health value threshold, allow normal charging to enter, and when the diagnosis ends, first drop the output current to 0 and then disconnect the relay;

[0083] Step 2, when the slot health value is determined to be qualified, the controller makes this slot enter normal charging and continuously collects the output current on the charging output side; adopt a stability measurement method based on adjacent sampling difference, construct a sliding time window with a window length of 10s, form a current sequence within each sliding time window, calculate the adjacent difference mean within the window, and calculate the average current within the window to obtain the process stability index; when the process stability indexes corresponding to 3 consecutive adjacent sliding time windows are all greater than or equal to the stability threshold, the controller determines that there is a fluctuation abnormality in this slot, writes the fluctuation abnormality into the trace record and triggers hierarchical handling, and prompts the rider to replace the slot for this slot; when the condition of continuous exceeding the threshold is not met, maintain normal charging and continue rolling evaluation. The stability threshold is固化 through factory calibration, sample each slot under the standard load and calculate the stability according to the difference index, and take the maximum value of all slots under normal working conditions as the benchmark and superimpose the ripple, temperature and sampling error margin to determine;

[0084] Step 3: The controller maintains a local traceability record set. Each traceability record includes a traceability record identifier, grid number, battery identifier, start time, and anomaly flag. The controller uses the most recent 200 traceability records as a scrolling window to maintain a weighted anomaly score and a weighted total score for each grid and each battery. Interface health anomalies are assigned a weight of 1, and fluctuation anomalies are assigned a weight of 0.6. Each time an anomaly occurs, the corresponding object is added to the weighted score. At the same time, each time a charging occurs, the corresponding object is added to the total score by 1. Based on this, the grid weighted anomaly rate and the battery weighted anomaly rate are obtained. The controller calculates the comparison difference and compares it with a fixed threshold. When the difference is greater than or equal to the fixed threshold, it is identified as a grid anomaly, and the grid is isolated. When the difference is less than the fixed threshold, it is identified as a battery anomaly, and charging of the battery is restricted. The battery threshold and grid threshold are determined through a closed-loop operation and maintenance process within the online calibration period. The minimum value of the grid root cause and the minimum value of the absolute value of the battery root cause are then fixed and written into the parameter table.

[0085] Step 4: After obtaining the anomaly type and location results, perform graded handling and self-recovery verification on the target object. A scoring-based handling level is adopted. When the anomaly type is interface health anomaly or fluctuation anomaly and the location is grid anomaly, the controller only performs handling on the target grid and keeps other grids continuously in service. Specifically, a risk score P is set for the grid, with 2 points for interface health anomaly and 1 point for fluctuation anomaly. After accumulation, the handling action is selected according to the threshold range: when P is between 1 and 2, a grid switching prompt is output and the output current is limited; when P is between 3 and 4, the grid is paused and the door is allowed to be opened to remove the battery, while marking it as pending maintenance; when P is greater than or equal to 5, the output current of the grid is reduced to zero, and after confirmation that it is 0, the relay is disconnected, triggering an audible and visual alarm and reporting a work order; if the location is a battery anomaly, the battery identifier is added to the restricted list and charging is refused for the battery. When the same grid is detected to initiate charging again and the QR code recognition is successful to generate a new traceability identifier, the verification mode is entered. In the verification mode, the diagnostic charging calculation of health value is repeated, and the process stability is calculated after normal charging. If the review is still abnormal, P is incremented and the handling range is automatically increased. If the review is not abnormal, P is decayed to 0, the alarm is cleared, the grid service is restored, and the review result is written to the traceability record as a normal sample.

[0086] The embodiments of the present invention described above are subject to modification and change of method by those skilled in the art without departing from the embodiments and broader aspects of the present invention. The appended claims are intended to include all such modifications and changes of method that do not depart from the present invention.

Claims

1. A fault detection method for an electric bicycle battery charging cabinet, characterized in that, include: A traceability record is generated for the battery connected to the charging port. The record is identified by a combination of port number, start time, and battery QR code. Before normal charging, the port is diagnostically charged, and the voltage and current changes are collected to calculate the health value of the port. During the normal charging phase, the current sequence is collected according to the sliding time window, and the current fluctuation amplitude and average current are calculated to evaluate the process stability. The process stability is compared with the preset threshold to determine the abnormal fluctuation of the grid. Using charging traceability records with abnormal fluctuations or abnormal health values ​​as abnormal samples, the abnormal proportion of grid slots and the abnormal proportion of batteries in the records are calculated respectively. The difference between the control and control is calculated, and the source of the abnormality is analyzed and located according to the abnormality judgment rules. Based on the anomaly type and location results, the grid is handled accordingly. When charging resumes, the health value and process stability are recalculated for verification. If the verification still indicates an anomaly, the handling is escalated. If the verification indicates no anomaly, the alarm is deactivated and the grid service is restored.

2. The fault detection method for an electric bicycle battery charging cabinet according to claim 1, characterized in that, The specific steps for performing diagnostic charging on the grid port before entering normal charging are as follows: Read the QR code on the appearance of the battery to be charged as the battery identifier, and at the same time read the grid number and the current start time. Combine the grid number, start time and battery QR code according to the preset encoding rules to generate a traceability record identifier (RID). Write the RID, the corresponding grid relay status and charging module number into the local storage and report to the operation and maintenance platform. Before starting the normal charging process, diagnostic charging is performed on the target grid port. First, the output current is set to 0, and the grid port output circuit is kept in a controllable state. Then, after the relay is closed, a current step command is issued to directly switch the output current from 0 to the diagnostic current and maintain it. The sampling time before the current step is recorded. Sampling time during the hold phase The voltage and output current of the target grid terminal are collected respectively to obtain the voltage change and current change, and the health value of the target grid is calculated based on these values. The formula is as follows: ， in Indicates the sampling time The voltage of the grid terminals, Indicates the sampling time ; output current; When the diagnostic charging is complete, set the output current back to 0, and confirm that the current has dropped to 0 before controlling the relay to disconnect.

3. The fault detection method for an electric bicycle battery charging cabinet according to claim 2, characterized in that, The specific steps for performing diagnostic charging on the grid port before entering normal charging also include: When the health value meets the preset unqualified judgment conditions, the traceability record is marked as an interface health abnormality and the port is prohibited from entering normal charging. At the same time, port isolation and prompt for port replacement are performed. When the health value meets the preset qualified judgment conditions, the health value is written into the traceability record and the port is allowed to enter the normal charging process.

4. The fault detection method for an electric bicycle battery charging cabinet according to claim 1, characterized in that, The specific steps for calculating the current fluctuation amplitude and average current to assess process stability are as follows: Once the health value of the grid port is deemed acceptable, it enters the normal charging process, and the output current of the port is continuously collected. A sliding time window with a window length of 10 seconds is constructed. A current sequence is formed for the current sampling points within each sliding time window. The current fluctuation amplitude and average current within each window are calculated at each sliding time to obtain a process stability index reflecting the stability of the grid port's charging process. The formula is: ， Where stable represents the process stability index. This indicates the maximum current within the sliding window. This represents the minimum current within the sliding window. This represents the average current within the sliding window. represents a constant to prevent the average current from approaching zero and causing division by zero, and max represents the maximum value function.

5. The fault detection method for an electric bicycle battery charging cabinet according to claim 1, characterized in that, The specific steps for comparing process stability with a preset threshold to determine abnormal fluctuations in the grid are as follows: The process stability of multiple consecutive sliding time windows is used for judgment. When the process stability corresponding to three consecutive adjacent sliding time windows is greater than or equal to the stability threshold, it is determined that there is an abnormal fluctuation in the grid, and the abnormal fluctuation is marked and written into the traceability record corresponding to the grid. At the same time, the graded handling strategy is triggered. When the process stability does not meet the continuous over-threshold condition, normal charging is maintained and rolling evaluation continues.

6. The fault detection method for an electric bicycle battery charging cabinet according to claim 1, characterized in that, The specific steps for calculating the abnormal proportion of grid slots and the abnormal proportion of batteries in the records, and calculating the difference between the two, are as follows: A cross-reference table is constructed using abnormal charging traceability records as abnormal samples, and the source of the abnormality is determined based on the difference in the abnormality ratio. First, a set of traceability records is maintained in local storage. Each traceability record contains a record identifier (RID), grid number, battery identifier, start time, and abnormality flag. The abnormality flag is written by the diagnostic process and includes interface health abnormality and fluctuation abnormality. Rolling statistics are performed based on the traceability records of the most recent 7 days. The total number of charging times and the number of abnormalities for each grid are accumulated by grid number. The number of abnormalities is divided by the total number of charging times to obtain the grid abnormality ratio. At the same time, the total number of charging times and the number of abnormalities for each battery are accumulated by battery identifier. The number of abnormalities is divided by the total number of charging times to obtain the battery abnormality ratio.

7. A fault detection method for an electric bicycle battery charging cabinet according to claim 6, characterized in that, The specific steps for calculating the grid anomaly ratio and the battery anomaly ratio in the records, and calculating the control difference, further include: When any traceability record is marked as an anomaly, the counters for the corresponding grid number and battery identifier are updated. The total number of charging times for that grid is accumulated at the grid level, and the anomaly count is accumulated when an anomaly occurs. The total number of charging times for that battery is accumulated at the battery level, and the anomaly count is accumulated when an anomaly occurs. These are then written into the corresponding fields of the cross-reference table. The grid number and battery identifier of the anomaly traceability record are associated and written into the association list as an association item, so that the anomaly distribution of the same battery in different grids and the same grid for different batteries can be directly compared in the same table.

8. A fault detection method for an electric bicycle battery charging cabinet according to claim 1, characterized in that, The specific steps for analyzing and locating the source of an anomaly based on the anomaly determination rules are as follows: For the same abnormal sample record, calculate the comparison difference between the corresponding grid abnormality ratio and the battery abnormality ratio, and compare the comparison difference with a preset threshold to locate the source of the abnormality. When the comparison difference meets the grid-side judgment rule, determine that the source of the abnormality is a grid abnormality, locate the faulty object to that grid, trigger the isolation of that grid, and allow other grids to continue to provide services. When the comparison difference meets the battery-side judgment rule, determine that the source of the abnormality is a battery abnormality, locate the faulty object to that battery, and trigger the restriction of charging for that battery.

9. A fault detection method for an electric bicycle battery charging cabinet according to claim 1, characterized in that, The specific steps for handling the grid based on the anomaly type and location result are as follows: After obtaining the anomaly type and location result, the target grid is subjected to graded handling, and a review is performed when charging starts again to achieve self-recovery closed loop. When the anomaly type corresponding to the trace record is interface health anomaly, that is, the grid health value obtained from the diagnostic charging exceeds the threshold, or fluctuation anomaly, that is, the stability of the sliding window process exceeds the threshold, and the location result is grid anomaly, only the target grid is handled without stopping other grid services.

10. A fault detection method for an electric bicycle battery charging cabinet according to claim 9, characterized in that, The specific steps for handling the grid based on the anomaly type and location result also include: The handling is divided into three levels: Level 1, Level 2, and Level 3. Level 1 involves displaying a prompt to the user and the cabinet screen to replace the compartment, limiting the output current of the compartment, and ensuring that the compartment can be restarted for charging. Level 2 involves locking the compartment and preventing it from entering the normal charging process, while keeping the cabinet door open to remove the battery, and marking the compartment as pending maintenance. Level 3 involves immediately setting the output current of the compartment to 0, disconnecting the compartment relay after confirming that the current is 0, triggering an audible and visual alarm, and reporting to the maintenance platform to generate a work order. If the location result indicates a battery abnormality, the battery identifier is added to the restricted list, and charging is refused for the current charge associated with that battery, while the target cell is not locked. When a trigger event is detected that the same cell has started charging again, a verification mode is entered. The trigger event is the successful reading of the battery QR code and the generation of a new traceability record. In verification mode, diagnostic charging is repeatedly performed on the cell, and the health value is recalculated. At the same time, after entering normal charging, the process stability is calculated according to a fixed window length to verify whether the abnormality judgment conditions are still met. If the verification result is still abnormal, the handling level is upgraded step by step according to the number of consecutive abnormalities and written to the corresponding handling field of the traceability record. If the verification result is not abnormal, the alarm mark of the cell is cleared and the lock is released, restoring the cell's ability to enter normal charging. At the same time, the verification pass result is written to the traceability record and used as a normal sample for subsequent statistics.

Images