A method and device for early warning of battery failure and electronic equipment

Abstract

The application discloses a kind of early warning method, device and electronic equipment of battery failure.The method includes: obtaining the historical running data set of target battery and real-time running data;According to historical running data set, the performance benchmark curve corresponding to different aging stages is constructed;And, according to real-time running data, the current aging stage of target battery is determined, and the dynamic performance curve representing current working state is generated;According to aging stage, the deviation feature between dynamic performance curve and performance benchmark curve is calculated, and the initial benchmark threshold value is dynamically corrected based on deviation feature, to obtain the dynamic target threshold range adapted to current operating state;Real-time monitoring value in real-time running data is compared with dynamic target threshold range, if real-time monitoring value exceeds dynamic target threshold range, then determine that target battery exists failure and trigger failure early warning.The method can realize the dynamic, accurate evaluation of battery state, to realize accurate early warning.

Description

Technical Field

[0001] This invention relates to the field of monitoring technology, and in particular to a method, device, and electronic equipment for early warning of battery failure. Background Technology

[0002] Currently, fault monitoring and early warning for batteries mainly rely on two methods: manual inspection and online monitoring based on static thresholds. However, both of these traditional technologies have significant shortcomings and are difficult to meet the needs of modern large-scale, intelligent operation and maintenance.

[0003] First, manual inspection methods suffer from significant timeliness and efficiency bottlenecks. Manual inspections are typically periodic, with long intervals between checks, making it impossible to obtain real-time information on the battery's dynamic operating status and to promptly detect early, subtle fault signals such as leakage, bulging, and overheating. Furthermore, manual inspections rely heavily on manpower, are costly, and are difficult to cover the large-scale base stations and equipment clusters, thus failing to achieve efficient continuous monitoring and fault prevention.

[0004] Secondly, most existing online monitoring systems employ static threshold judgment logic, which determines whether a battery is abnormal by setting a fixed threshold. However, battery performance gradually degrades over time, and the normal ranges of key parameters such as voltage, resistance, and capacity change with the aging process. Fixed static thresholds lack the ability to adapt to such dynamic changes: for new batteries, overly stringent thresholds may lead to false alarms, increasing unnecessary maintenance burdens; for aging batteries, lenient thresholds are prone to missed alarms, failing to provide timely warnings of potential faults and posing safety hazards. Furthermore, the battery operating environment is complex and variable; factors such as temperature, state of charge, and charge / discharge rate all affect its performance. Static thresholds cannot adequately incorporate these dynamic factors, further reducing the accuracy and reliability of fault diagnosis.

[0005] In conclusion, existing technologies cannot achieve dynamic and accurate assessment of battery status. Summary of the Invention

[0006] The technical problem to be solved by this invention is to address the aforementioned shortcomings of the prior art by proposing a method, device, and electronic equipment for early warning of battery faults. This method enables dynamic and accurate assessment of the battery's condition, thereby effectively improving the timeliness and accuracy of fault warnings.

[0007] In a first aspect, the present invention provides a method for early warning of battery malfunctions, the method comprising:

[0008] Obtain the historical and real-time operational data of the target battery;

[0009] Based on historical operating datasets, construct performance benchmark curves corresponding to different aging stages; and determine the current aging stage of the target battery based on real-time operating data, and generate dynamic performance curves characterizing the current working state.

[0010] Based on the aging stage, the deviation characteristics between the dynamic performance curve and the performance baseline curve are calculated. Based on the deviation characteristics, the preset initial baseline threshold is dynamically corrected to obtain a dynamic target threshold range that adapts to the current operating state.

[0011] The real-time monitoring value in the real-time operation data is compared with the dynamic target threshold range. If the real-time monitoring value exceeds the dynamic target threshold range, it is determined that the target battery has a fault and a fault warning is triggered.

[0012] Furthermore, the current aging stage of the target battery is determined based on real-time operating data, specifically including:

[0013] The rate of change of internal resistance of the target battery is calculated based on real-time operating data. The rate of change of internal resistance = (real-time internal resistance - initial internal resistance) / initial internal resistance × 100%;

[0014] The current aging stage is determined by combining the rate of change of internal resistance and the duration of operation.

[0015] The current aging stage is determined based on a combination of the rate of change in internal resistance and the duration of operation, specifically including:

[0016] When the rate of change of internal resistance is in the first rate of change range and the running time is in the first time range, it is determined to be a new commissioning phase.

[0017] When the rate of change of internal resistance is in the second rate of change range and the running time is in the second time range, it is determined to be in the stable operation phase.

[0018] When the rate of change of internal resistance is in the third rate of change range or the running time is in the third time range, it is determined to be in the aging stage.

[0019] Specifically, the upper limit of the first rate of change interval is less than the lower limit of the second rate of change interval, and the upper limit of the second rate of change interval is less than the lower limit of the third rate of change interval; the upper limit of the first duration interval is less than the lower limit of the second duration interval, and the upper limit of the second duration interval is less than the lower limit of the third duration interval.

[0020] Furthermore, based on the aging stage, the deviation characteristics between the dynamic performance curve and the performance baseline curve corresponding to the aging stage are calculated, including:

[0021] The dynamic performance curve and the performance benchmark curve corresponding to the aging stage are synchronously segmented according to the preset time window.

[0022] Within each time window, calculate the average deviation, maximum deviation, and duration of deviation of the dynamic performance curve relative to the performance baseline curve corresponding to the aging stage.

[0023] The average deviation is the difference between the average value of the dynamic performance curve within the time window and the average value of the performance reference curve corresponding to the aging stage; the maximum deviation is the difference between the peak value of the dynamic performance curve within the time window and the peak value of the performance reference curve corresponding to the aging stage; the duration of deviation is the cumulative duration for which the value of the dynamic performance curve exceeds the upper limit of the fluctuation range threshold or falls below the lower limit of the fluctuation range threshold of the performance reference curve corresponding to the aging stage.

[0024] The average deviation value, maximum deviation value, and deviation duration are defined as the deviation characteristics between the dynamic performance curve and the performance reference curve corresponding to the aging stage.

[0025] Furthermore, based on the deviation characteristics, the preset initial benchmark threshold is dynamically corrected to obtain a dynamic target threshold range that adapts to the current operating state, specifically including:

[0026] Retrieve the initial baseline threshold corresponding to the current aging stage;

[0027] Based on the preset deviation and threshold mapping model, the threshold correction coefficient is calculated according to the average deviation value, maximum deviation value and deviation duration in the deviation characteristics;

[0028] The initial reference threshold is calculated using the corresponding correction formula selected according to the parameter type. For temperature parameters, the corrected threshold = initial reference threshold × (1 + threshold correction coefficient); for voltage parameters, the corrected threshold = initial reference threshold × (1 - threshold correction coefficient).

[0029] The calculated corrected threshold is subject to boundary constraints to ensure it does not exceed the physical tolerance limit of the target battery, and the corrected threshold after boundary constraint correction is used as the range value of the dynamic target threshold.

[0030] Furthermore, the method also includes:

[0031] The corrected thresholds are indexed in a multidimensional way according to the battery cell number, parameter type, aging stage and correction time, and stored centrally to form the corresponding threshold files.

[0032] Based on the stored threshold files, the dynamic changes of deviation characteristics are continuously tracked. When the change of deviation characteristics exceeds the preset ratio or the current aging stage changes, the threshold re-correction process is automatically triggered. The recalculated threshold is used to overwrite and store the threshold files and push them synchronously.

[0033] Furthermore, fault warning specifically includes:

[0034] Cross-validation using multi-dimensional operational data is used to determine the fault point and fault type. Among them, abnormal temperature accompanied by voltage drop is identified as a heat fault, voltage drop accompanied by sudden increase in internal resistance is identified as a liquid leakage fault, and temperature rise accompanied by shell deformation data is identified as a bulging fault.

[0035] A three-level early warning mechanism is adopted for graded response based on the type and severity of the fault. The first-level early warning sends an SMS notification to the maintenance personnel and is applicable to minor anomalies in the early stage. The second-level early warning sends a pop-up warning on the platform and automatically generates a maintenance work order, which is applicable to confirmed medium risks. The third-level early warning triggers an on-site audible and visual warning and links the energy storage system to cut off the circuit of the faulty battery pack, which is applicable to serious faults.

[0036] Furthermore, performance benchmark curves corresponding to different aging stages are constructed based on historical runtime datasets, specifically including:

[0037] Based on historical operation datasets, the life cycle of the target battery is divided into the new commissioning stage, the stable operation stage, and the aging stage. The historical operation datasets include the factory test data and annual operation and maintenance data of the target battery. The timestamp of the factory test data is the factory time of the target battery, and the time span of the annual operation and maintenance data covers the entire life cycle from the new commissioning of the target battery to the aging and scrapping.

[0038] For the new commissioning stage, stable operation stage, and aging stage, respectively, the fluctuation range curves of voltage parameters, current parameters, temperature parameters, and internal resistance parameters of the target battery are fitted to form the performance benchmark curve of the target battery throughout its entire life cycle.

[0039] Furthermore, for the new commissioning stage, stable operation stage, and aging stage, respectively, the fluctuation range curves of the target battery's voltage, current, temperature, and internal resistance parameters are fitted, specifically including:

[0040] For the new commissioning stage, stable operation stage, and aging stage, voltage, current, temperature, and internal resistance data of the target battery are extracted at different time points or under the corresponding state of charge within the corresponding stage.

[0041] For each type of parameter data in each stage, calculate the mean and standard deviation of the data at the same time node or under the same state of charge; and use the time node or state of charge as the horizontal axis, the corresponding mean of the data as the median, and the standard deviation of the data at a preset multiple as the fluctuation boundary to fit and generate the corresponding fluctuation range curve for each type of parameter in each stage.

[0042] Secondly, the present invention provides a battery fault early warning device, the device comprising:

[0043] The acquisition unit is used to acquire the historical operating dataset and real-time operating data of the target battery.

[0044] The processing unit, connected to the acquisition unit, is used to construct performance benchmark curves corresponding to different aging stages based on historical operating datasets; and to determine the current aging stage of the target battery based on real-time operating data, and generate dynamic performance curves characterizing the current working state.

[0045] The calculation unit, connected to the processing unit, is used to calculate the deviation characteristics between the dynamic performance curve and the performance reference curve based on the aging stage.

[0046] The correction unit, connected to the calculation unit, is used to dynamically correct the preset initial benchmark threshold based on the deviation characteristics to obtain a dynamic target threshold range that adapts to the current operating state.

[0047] The comparison unit, connected to the correction unit, is used to compare the real-time monitoring value in the real-time operation data with the dynamic target threshold range. If the real-time monitoring value exceeds the dynamic target threshold range, it is determined that the target battery has a fault and a fault warning is triggered.

[0048] Thirdly, the present invention provides an electronic device including a memory and a processor, wherein the memory stores a computer program, and when the processor runs the computer program stored in the memory, the processor executes the battery fault warning method according to the first aspect.

[0049] This invention acquires historical and real-time operational data of the target battery to construct performance benchmark curves corresponding to different aging stages. It then combines real-time operational data to determine the current aging stage and generate a dynamic performance curve. Furthermore, it calculates deviation characteristics to dynamically correct the initial benchmark threshold. Finally, it compares the real-time monitoring values ​​with the dynamic target threshold range to trigger an early warning. This method enables dynamic and accurate assessment of battery status, with the following specific benefits:

[0050] (1) Achieving adaptive dynamic adjustment of the threshold, effectively solving the limitations of static threshold judgment: This invention can accurately identify the current life cycle stage of the battery by constructing performance benchmark curves for different aging stages. Based on this, the threshold is dynamically corrected according to the deviation characteristics between the dynamic performance curve and the benchmark curve, so that the warning threshold can adaptively change as the battery aging degree deepens. This effectively overcomes the defects of traditional static thresholds being too strict for new batteries (prone to false alarms) and too lenient for aged batteries (prone to missed alarms), significantly improving the accuracy and adaptability of fault judgment.

[0051] (2) Improving the accuracy and reliability of fault early warning: Compared with the judgment method that relies solely on a single fixed threshold, this invention introduces a dynamic performance curve that characterizes the current working state and calculates its deviation characteristics from the baseline curve. This method based on curve trends and deviation analysis can more sensitively capture subtle changes in battery performance and early fault signals, comprehensively reflecting the true health status of the battery in specific aging stages and complex operating environments, thereby significantly reducing the probability of missed and false alarms and improving the accuracy of early warning.

[0052] (3) Overcoming the shortcomings of poor timeliness of manual inspections and realizing all-weather intelligent monitoring: This invention performs dynamic evaluation and comparison based on real-time operating data, eliminating the reliance on periodic manual inspections. This not only solves the problems of long time intervals and inability to obtain dynamic status in real time in manual inspections, but also fills the monitoring blind spots between inspections, enabling timely detection of potential hazards such as leakage, swelling, and overheating, and realizing real-time, continuous, and efficient monitoring of battery status, meeting the needs of large-scale intelligent operation and maintenance.

[0053] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0054] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof. The above and other features and advantages will become more apparent to those skilled in the art from the detailed description of exemplary embodiments with reference to the accompanying drawings, in which:

[0055] Figure 1 A schematic diagram of a battery fault early warning method provided in an embodiment of the present invention;

[0056] Figure 2 A schematic diagram of a battery fault early warning device provided in an embodiment of the present invention;

[0057] Figure 3 A framework diagram of an electronic device provided in an embodiment of the present invention.

[0058] Reference numerals: 10, acquisition unit; 20, processing unit; 30, calculation unit; 40, correction unit; 50, comparison unit; 100, processor; 200, memory. Detailed Implementation

[0059] It is understood that the specific embodiments and accompanying drawings described herein are merely for explaining the invention and are not intended to limit the invention.

[0060] It is understood that, without conflict, the various embodiments and features in the embodiments of the present invention can be combined with each other.

[0061] It is understood that, for ease of description, only the parts related to the present invention are shown in the accompanying drawings, while the parts unrelated to the present invention are not shown in the drawings.

[0062] It is understood that each unit or module involved in the embodiments of the present invention may correspond to only one entity structure, or may be composed of multiple entity structures, or multiple units or modules may be integrated into one entity structure.

[0063] It is understood that, without conflict, the functions and steps marked in the flowcharts and block diagrams of this invention may occur in a different order than that marked in the accompanying drawings.

[0064] It is understood that the flowcharts and block diagrams of this invention illustrate the possible architecture, functions, and operations of systems, apparatuses, devices, and methods according to various embodiments of this invention. Each block in the flowchart or block diagram may represent a unit, module, program segment, or code, containing executable instructions for implementing the specified function. Furthermore, each block or combination of blocks in the block diagram and flowchart can be implemented using a hardware-based system to achieve the specified function, or using a combination of hardware and computer instructions.

[0065] It is understood that the units and modules involved in the embodiments of the present invention can be implemented by software or by hardware. For example, the units and modules can be located in a processor.

[0066] Example 1:

[0067] This embodiment provides a method for early warning of battery failures, applicable to backup power management for critical infrastructure such as communication base stations, data centers, rail transit, and power grid substations. In such scenarios, battery banks are often in a float-charge state for extended periods, exhibiting a significantly non-linear aging process. Traditional monitoring methods rely on fixed thresholds, making it difficult to distinguish between normal aging degradation and sudden performance deterioration, easily leading to false alarms or missed alarms. Taking communication base stations as an example, battery banks are distributed outdoors, affected by temperature fluctuations and differences in charging and discharging, resulting in uneven aging processes for individual cells. Using a uniform static threshold may frequently trigger invalid alarms when batteries are healthy, unnecessarily increasing the maintenance burden; it may also fail to provide timely warnings when batteries are severely degraded and backup time is insufficient, creating a potential for power outages and system crashes. By adopting the method in this embodiment, a phased performance benchmark curve adapted to the aging trajectory can be constructed based on the historical operating data of a single battery, the aging stage can be determined in real time, and the threshold range can be adaptively corrected according to the deviation characteristics between the dynamic performance curve and the benchmark curve, thereby accurately identifying abnormal states and providing early warnings before actual faults occur. This not only avoids interference from invalid alarms but also provides sufficient processing window for on-site maintenance, significantly improving the reliability of power supply for critical facilities.

[0068] like Figure 1 As shown, this embodiment specifically includes steps S1 to S4.

[0069] Step S1: Obtain the historical operating dataset and real-time operating data of the target battery.

[0070] Specifically, the construction of historical operational datasets relies on the integration and governance of data assets throughout their entire lifecycle, including:

[0071] First, complete archive information of the batteries since they left the factory is collected, including the initial performance benchmarks in the type test report, factory acceptance test data (covering core parameters such as open-circuit voltage, rated capacity, and initial internal resistance), and structured and unstructured data such as inspection records, verification discharge test data, and fault handling records accumulated during years of operation and maintenance. This forms a longitudinal dataset covering the entire time span of "new commissioning stage (0-2 years)," "stable operation stage (2-5 years)," and "aging stage (5 years and above)." To address the characteristics of multi-source heterogeneous data, unified data cleaning and standardization rules are established. Missing values ​​are imputed using the mean interpolation method based on a group of devices of the same model, and abnormal jump data is smoothed using Kalman filtering. Finally, a high-quality historical operation database is constructed, providing reliable training samples for the piecewise fitting algorithm.

[0072] Real-time operational data is acquired through a multi-parameter sensor network deployed in individual battery cells and battery pack circuits. Voltage sensors continuously monitor the float charge voltage and terminal voltage fluctuations of each cell with an accuracy of ±0.01V; Hall current sensors track and record the direction and amplitude changes of charging and discharging current with an accuracy of ±0.1A; distributed temperature sensing nodes collect the surface temperature field distribution of cells and the ambient temperature with an accuracy of ±0.5℃; and AC injection-type internal resistance detectors periodically scan the trend of internal resistance changes in cells with an accuracy of ±5mΩ. Each sensing node collects data synchronously at a sampling frequency of 1 time per minute. After preliminary data verification and compression by edge computing nodes, the data is transmitted to the intelligent operation and management platform via dual-channel redundant transmission through a 4G / 5G wireless private network or industrial Ethernet. Simultaneously, a local embedded database implements a 72-hour rolling caching mechanism to ensure zero data loss in abnormal operating conditions such as network interruptions. Once the network is restored, the data is automatically resumed and replenished, ensuring the integrity and continuity of the real-time data stream.

[0073] Historical and real-time data are integrated and converged at the data hub of the intelligent operation and management platform. Through unified timestamp alignment and association with equipment codes, a panoramic data view is formed that has both vertical historical depth and horizontal real-time breadth.

[0074] Step S2: Construct performance benchmark curves corresponding to different aging stages based on historical operating datasets; and determine the current aging stage of the target battery based on real-time operating data, and generate dynamic performance curves characterizing the current working state.

[0075] As a specific implementation method, the current aging stage of the target battery is determined based on real-time operating data, specifically including:

[0076] The rate of change of internal resistance of the target battery is calculated based on real-time operating data. The rate of change of internal resistance = (real-time internal resistance - initial internal resistance) / initial internal resistance × 100%;

[0077] The current aging stage is determined by combining the rate of change of internal resistance and the duration of operation.

[0078] The current aging stage is determined based on a combination of the rate of change in internal resistance and the duration of operation, specifically including:

[0079] When the rate of change of internal resistance is in the first rate of change range and the running time is in the first time range, it is determined to be a new commissioning phase.

[0080] When the rate of change of internal resistance is in the second rate of change range and the running time is in the second time range, it is determined to be in the stable operation phase.

[0081] When the rate of change of internal resistance is in the third rate of change range or the running time is in the third time range, it is determined to be in the aging stage.

[0082] Specifically, the upper limit of the first rate of change interval is less than the lower limit of the second rate of change interval, and the upper limit of the second rate of change interval is less than the lower limit of the third rate of change interval; the upper limit of the first duration interval is less than the lower limit of the second duration interval, and the upper limit of the second duration interval is less than the lower limit of the third duration interval.

[0083] As a specific implementation method, performance benchmark curves corresponding to different aging stages are constructed based on historical runtime datasets, specifically including:

[0084] Based on historical operation datasets, the life cycle of the target battery is divided into the new commissioning stage, the stable operation stage, and the aging stage. The historical operation datasets include the factory test data and annual operation and maintenance data of the target battery. The timestamp of the factory test data is the factory time of the target battery, and the time span of the annual operation and maintenance data covers the entire life cycle from the new commissioning of the target battery to the aging and scrapping.

[0085] For the new commissioning stage, stable operation stage, and aging stage, respectively, the fluctuation range curves of voltage parameters, current parameters, temperature parameters, and internal resistance parameters of the target battery are fitted to form the performance benchmark curve of the target battery throughout its entire life cycle.

[0086] As a specific implementation method, for the new commissioning stage, stable operation stage, and aging stage, the fluctuation range curves of the target battery's voltage parameters, current parameters, temperature parameters, and internal resistance parameters are fitted, specifically including:

[0087] For the new commissioning stage, stable operation stage, and aging stage, voltage, current, temperature, and internal resistance data of the target battery are extracted at different time points or under the corresponding state of charge within the corresponding stage.

[0088] For each type of parameter data in each stage, calculate the mean and standard deviation of the data at the same time node or under the same state of charge; and use the time node or state of charge as the horizontal axis, the corresponding mean of the data as the median, and the standard deviation of the data at a preset multiple as the fluctuation boundary to fit and generate the corresponding fluctuation range curve for each type of parameter in each stage.

[0089] This embodiment uses the battery internal resistance change rate as the core judgment indicator and the operating time as an auxiliary verification method. Specifically, it acquires the uploaded real-time internal resistance data and compares it with the initial internal resistance data of the battery model at the time of manufacture, and calculates it according to the formula "Internal resistance change rate = (real-time internal resistance - initial internal resistance) / initial internal resistance × 100%"; at the same time, it reviews the commissioning time in the equipment operation and maintenance records. Based on the above two data, the current aging stage is comprehensively evaluated: if the internal resistance change rate is ≤10% and the operating time is ≤2 years, it is judged as "new commissioning stage"; if the internal resistance change rate is between 10% and 20% and the operating time is between 2 and 5 years, it is judged as "stable operation stage"; if the internal resistance change rate is >20% or the operating time is >5 years, it is judged as "aging stage". The final judgment result is synchronized in real time to the deviation feature calculation and threshold correction algorithm.

[0090] Step S3: Based on the aging stage, calculate the deviation characteristics between the dynamic performance curve and the performance baseline curve, and dynamically correct the preset initial baseline threshold based on the deviation characteristics to obtain a dynamic target threshold range that adapts to the current operating state.

[0091] As a specific implementation method, based on the aging stage, the deviation characteristics between the dynamic performance curve and the performance reference curve corresponding to the aging stage are calculated, including:

[0092] The dynamic performance curve and the performance benchmark curve corresponding to the aging stage are synchronously segmented according to the preset time window.

[0093] Within each time window, calculate the average deviation, maximum deviation, and duration of deviation of the dynamic performance curve relative to the performance baseline curve corresponding to the aging stage.

[0094] The average deviation is the difference between the average value of the dynamic performance curve within the time window and the average value of the performance reference curve corresponding to the aging stage; the maximum deviation is the difference between the peak value of the dynamic performance curve within the time window and the peak value of the performance reference curve corresponding to the aging stage; the duration of deviation is the cumulative duration for which the value of the dynamic performance curve exceeds the upper limit of the fluctuation range threshold or falls below the lower limit of the fluctuation range threshold of the performance reference curve corresponding to the aging stage.

[0095] The average deviation value, maximum deviation value, and deviation duration are defined as the deviation characteristics between the dynamic performance curve and the performance reference curve corresponding to the aging stage.

[0096] As a specific implementation method, a preset initial benchmark threshold is dynamically corrected based on deviation characteristics to obtain a dynamic target threshold range that adapts to the current operating state, specifically including:

[0097] Retrieve the initial baseline threshold corresponding to the current aging stage;

[0098] Based on the preset deviation and threshold mapping model, the threshold correction coefficient is calculated according to the average deviation value, maximum deviation value and deviation duration in the deviation characteristics;

[0099] The initial reference threshold is calculated using the corresponding correction formula selected according to the parameter type. For temperature parameters, the corrected threshold = initial reference threshold × (1 + threshold correction coefficient); for voltage parameters, the corrected threshold = initial reference threshold × (1 - threshold correction coefficient).

[0100] The calculated corrected threshold is subject to boundary constraints to ensure it does not exceed the physical tolerance limit of the target battery, and the corrected threshold after boundary constraint correction is used as the range value of the dynamic target threshold.

[0101] In this embodiment, after obtaining the baseline curve and dynamic curve obtained from the performance curve analysis, a "time window segmented comparison" method is used to extract deviation features. Specifically, the curve is divided into continuous time windows with a 12-hour interval, and three types of feature parameters between the dynamic curve and the baseline curve are calculated in each window: first, the "average deviation value," which is the deviation of the average value of the dynamic curve relative to the average value of the baseline curve within the 12-hour window (e.g., the average value of the dynamic temperature curve is 2.5℃ higher than the baseline curve); second, the "maximum deviation value," which is the deviation of the extreme value of the dynamic curve relative to the extreme value of the baseline curve within the window (e.g., the peak value of the dynamic temperature is 4℃ lower than the peak value of the baseline curve); and third, the "duration of deviation," which is the cumulative length of time that the dynamic curve exceeds the normal fluctuation range of the baseline curve (e.g., a continuous deviation of up to 3 hours). Finally, these three types of deviation feature parameters are integrated and sent to the threshold correction algorithm.

[0102] This embodiment achieves threshold correction through a built-in "deviation-threshold mapping model." This model is trained and generated based on massive historical battery operation and maintenance data, and has a pre-defined correspondence between "average deviation value, maximum deviation value, deviation duration" and "threshold correction coefficient" at different aging stages (for example, at a specific aging stage, the threshold correction coefficient increases by 0.05 for every 1°C increase in the average temperature deviation value). The specific execution process is as follows:

[0103] (1) Obtaining benchmark and deviation data: First, receive the current stage information output by the aging degree determination link and retrieve the initial benchmark threshold corresponding to the stage (e.g., the initial benchmark threshold for temperature is 38℃ and the initial benchmark threshold for internal resistance is 75mΩ); Second, combine the three types of deviation parameters (average deviation, maximum deviation, and duration) output by the deviation characteristic calculation link and calculate the corresponding threshold correction coefficient through the "deviation-threshold mapping model" (e.g., the average temperature deviation is 2.5℃, the maximum deviation is 4℃, and the deviation duration is 3 hours, corresponding to a temperature threshold correction coefficient of 0.08).

[0104] (2) Calculate the dynamic correction threshold: Based on the parameter characteristics, calculate the dynamic correction threshold for each parameter using the following formula:

[0105] Temperature parameters (threshold needs to be adjusted upwards due to decreased heat dissipation capacity after aging): Corrected threshold = Initial baseline threshold × (1 + threshold correction factor)

[0106] Voltage parameters (thresholds need to be lowered due to decreased voltage stability after aging): Corrected threshold = Initial reference threshold × (1 - threshold correction factor)

[0107] Calculation example: The threshold after temperature correction during the aging stage = 38℃ × (1 + 0.08) = 41.04℃, rounded to 41℃; The threshold after internal resistance correction = 75mΩ × (1 + 0.06) = 79.5mΩ, rounded to 80mΩ.

[0108] (3) Set boundary limits: At the same time, set “corrected threshold boundary limits” – the corrected threshold must not exceed the physical tolerance limit of the battery model (for example, the temperature threshold should not exceed 50°C and the voltage threshold should not be lower than 10V) to avoid threshold loss of control due to excessive deviation.

[0109] Step S4: Compare the real-time monitoring value in the real-time operation data with the dynamic target threshold range. If the real-time monitoring value exceeds the dynamic target threshold range, it is determined that the target battery has a fault and a fault warning is triggered.

[0110] In this embodiment, the corrected threshold calculated by the threshold correction algorithm is indexed in the format of "battery cell number - parameter type - aging stage - correction time" and stored in the platform's threshold database. At the same time, the deviation characteristics are monitored in real time. When the deviation parameter output by the subsequent deviation characteristic calculation stage changes by more than 10% (such as the average temperature deviation changing from 2.5℃ to 3℃), or the aging stage changes (such as entering the aging stage from the stable operation stage), the threshold re-correction process is automatically triggered to update the corresponding threshold in the threshold database and push the updated threshold to the fault warning stage to ensure the timeliness and accuracy of the threshold.

[0111] As a specific implementation method, fault early warning includes:

[0112] Cross-validation using multi-dimensional operational data is used to determine the fault point and fault type. Among them, abnormal temperature accompanied by voltage drop is identified as a heat fault, voltage drop accompanied by sudden increase in internal resistance is identified as a liquid leakage fault, and temperature rise accompanied by shell deformation data is identified as a bulging fault.

[0113] A three-level early warning mechanism is adopted for graded response based on the type and severity of the fault. The first-level early warning sends an SMS notification to the maintenance personnel and is applicable to minor anomalies in the early stage. The second-level early warning sends a pop-up warning on the platform and automatically generates a maintenance work order, which is applicable to confirmed medium risks. The third-level early warning triggers an on-site audible and visual warning and links the energy storage system to cut off the circuit of the faulty battery pack, which is applicable to serious faults.

[0114] As one specific implementation method, the method further includes:

[0115] The corrected thresholds are indexed in a multidimensional way according to the battery cell number, parameter type, aging stage and correction time, and stored centrally to form the corresponding threshold files.

[0116] Based on the stored threshold files, the dynamic changes of deviation characteristics are continuously tracked. When the change of deviation characteristics exceeds the preset ratio or the current aging stage changes, the threshold re-correction process is automatically triggered. The recalculated threshold is used to overwrite and store the threshold files and push them synchronously.

[0117] In practical implementation, this embodiment includes two key stages: fault location and alarm. The fault location stage compares the real-time data uploaded from the acquisition stage with the corrected threshold in real time. If the temperature of a single battery cell exceeds the corrected temperature threshold for 5 consecutive minutes (e.g., the temperature exceeds 40°C during aging), or the internal resistance exceeds the corrected internal resistance threshold for 3 consecutive acquisitions, the system combines voltage and current data for cross-verification (e.g., voltage drop accompanied by temperature anomalies) to accurately locate the corresponding faulty cell (physical location determined by sensor number) and fault type (e.g., abnormal temperature associated with "overheating fault," sudden voltage drop + sudden increase in internal resistance associated with "leakage fault," and temperature rise + casing deformation data associated with "bulging fault"). The alarm stage triggers warnings through a "three-level alarm mechanism": Level 1 warning (potential fault) pushes an SMS to the maintenance personnel's mobile phone; Level 2 warning (minor fault) sends a platform pop-up alarm and automatically generates a maintenance work order; Level 3 warning (serious fault) triggers on-site audible and visual alarms, and simultaneously links the energy storage equipment to cut off the faulty battery pack circuit.

[0118] In specific implementation, this embodiment first initiates the data acquisition phase. Sensors collect real-time data on the battery's voltage, current, temperature, and internal resistance, which is then uploaded to the intelligent operation and management platform via data transmission. The platform performs noise reduction processing on the data (removing abnormal fluctuation values, such as instantaneous current spikes) to ensure data validity. Next, the performance curve analysis phase calls upon historical data to fit a baseline curve according to the life cycle stage, while simultaneously receiving real-time data to generate a dynamic performance curve, and displaying a superimposed comparison chart of the two curves on the platform interface. Subsequently, the threshold dynamic correction phase begins. First, the aging degree determination phase calculates the rate of change of internal resistance and determines the current aging stage based on the operating time. Then, the deviation characteristic calculation phase calculates the threshold dynamic correction according to a 12-step process. The hourly time window calculates the average deviation, maximum deviation, and duration of deviation between the dynamic curve and the baseline curve. Next, the threshold correction algorithm retrieves the initial baseline threshold for the corresponding stage, calculates the correction coefficient based on the deviation parameters, and derives the corrected threshold using the formula, passing boundary limit verification. Finally, the threshold storage and update stage stores the corrected threshold in the database and pushes it synchronously to the fault warning stage, while monitoring deviation and aging stage changes to trigger threshold updates. Finally, in the fault judgment and warning stage, the fault warning stage compares real-time data with the corrected threshold in real time. If the data exceeds the threshold range, cross-validation is used to determine the fault point and fault type, triggering the corresponding level of alarm and recording fault information (time, location, type) in the operation and maintenance log.

[0119] The following example uses the operation and maintenance of a communication base station battery. A base station uses a 24-cell 12V lead-acid battery pack (newly commissioned for 1 year, initially in the "stable operation phase"). The data acquisition process collects the voltage (normal range 12.0-13.5V), current (charging 0-10A, discharging 0-20A), temperature (normal 15-30℃), and internal resistance (normal 30-45mΩ) of each cell every minute and uploads them to the intelligent operation and management platform. The performance curve analysis process calls the historical data of this battery model to fit the baseline curve of the stable operation phase (such as voltage 12.2-13.3V, temperature 18-28℃), and generates the current dynamic curve, showing that the temperature of the 15th cell has been maintained at 32℃ for 3 consecutive hours (higher than the upper limit of the baseline curve 28℃). The process then proceeds to the threshold dynamic correction stage: The aging determination stage calculates the battery's initial internal resistance as 35mΩ, real-time internal resistance as 37mΩ, and internal resistance change rate as 5.7% (≤10%). Combined with a runtime of 1 year (≤2 years), the previously misjudged "stable operation stage" is automatically corrected to "new commissioning stage." The deviation characteristic calculation stage calculates the average deviation of the battery's dynamic temperature curve from the new commissioning stage baseline curve (temperature 15-32℃) as 3℃, the maximum deviation as 4℃, and the deviation duration as 3 hours, based on a 12-hour window. The threshold correction algorithm stage retrieves the initial baseline threshold of 32℃ for the new commissioning stage, calculates a correction coefficient of 0.03 using a mapping model, and the corrected threshold is 32℃ × (1 + 0.03) = 32.96℃, rounded to 33℃ (not exceeding the 50℃ boundary limit). The threshold storage and update stage stores the corrected threshold of 33℃ corresponding to "15th battery - temperature - new commissioning stage - 2024XX month XX day" in the database and pushes it to the fault warning stage. Within the next hour, the temperature of the 15th battery rose to 34°C (exceeding the corrected threshold of 33°C), and the voltage dropped from 12.8V to 12.3V. The fault warning system identified it as a "potential overheating fault," triggering a level one warning and sending an SMS to the maintenance personnel. Two hours later, the temperature rose to 36°C, triggering a level two warning and generating a maintenance work order. The maintenance personnel inspected the battery and found that the battery casing was slightly bulging. After timely replacement, the fault was prevented from escalating.

[0120] This embodiment integrates historical records of the battery from its manufacturing process to its entire lifecycle with a real-time sensor network, achieving multi-dimensional data fusion and acquisition of four core parameters: voltage, current, temperature, and internal resistance. Compared to traditional single-dimensional monitoring methods, this constructs a more comprehensive data support system. Based on this, using cleaned, high-quality historical data, the system fits the fluctuation ranges of voltage, current, temperature, and internal resistance for different aging stages (new commissioning, stable operation, and aging), constructing a full-lifecycle performance benchmark curve. This effectively overcomes the limitations of traditional "one-size-fits-all" parameter judgment logic. During operation, the solution accurately determines the current aging stage based on the real-time internal resistance change rate and operating time. By calculating the average deviation, maximum deviation, and duration of deviation between the dynamic performance curve and the benchmark curve within a specific time window, and using a built-in mapping model to automatically calculate correction coefficients, a deviation-driven threshold dynamic correction mechanism is established. This achieves adaptive matching between the threshold and the battery's aging degree, solving the false alarm and missed alarm problems caused by static thresholds. Ultimately, when the real-time monitoring value exceeds the dynamically corrected threshold range, cross-validation of multi-dimensional data (such as abnormal temperature accompanied by voltage drop indicating overheating, and sudden voltage drop accompanied by sudden increase in internal resistance indicating leakage) enables precise location of the fault type and fault point, and triggers graded early warning response based on severity, thereby significantly improving the pertinence and safety of battery operation and maintenance.

[0121] Example 2:

[0122] like Figure 2 As shown, this embodiment provides a battery fault early warning device, which includes:

[0123] Acquisition unit 10 is used to acquire the historical operating dataset and real-time operating data of the target battery;

[0124] The processing unit 20, connected to the acquisition unit 10, is used to construct performance benchmark curves corresponding to different aging stages based on historical operating datasets; and to determine the current aging stage of the target battery based on real-time operating data, and generate dynamic performance curves characterizing the current working state.

[0125] The calculation unit 30, connected to the processing unit 20, is used to calculate the deviation characteristics between the dynamic performance curve and the performance reference curve based on the aging stage.

[0126] The correction unit 40 is connected to the calculation unit 30 and is used to dynamically correct the preset initial benchmark threshold based on the deviation characteristics to obtain a dynamic target threshold range that adapts to the current operating state.

[0127] The comparison unit 50, connected to the correction unit 40, is used to compare the real-time monitoring value in the real-time operation data with the dynamic target threshold range. If the real-time monitoring value exceeds the dynamic target threshold range, it is determined that the target battery has a fault and a fault warning is triggered.

[0128] The apparatus in this embodiment is capable of performing the method in Embodiment 1.

[0129] Example 3:

[0130] like Figure 3 As shown, the present invention provides an electronic device, which includes a memory 200 and a processor 100. The memory 200 stores a computer program. When the processor 100 runs the computer program stored in the memory 200, the processor 100 executes the battery fault warning method according to Embodiment 1.

[0131] It is understood that the above embodiments are merely exemplary implementations used to illustrate the principles of the present invention, and the present invention is not limited thereto. For those skilled in the art, various modifications and improvements can be made without departing from the spirit and essence of the present invention, and these modifications and improvements are also considered to be within the scope of protection of the present invention.

Claims

1. A method for early warning of battery failure, characterized in that, The method includes: Obtain the historical and real-time operational data of the target battery; Based on the historical operating dataset, a performance benchmark curve corresponding to different aging stages is constructed; and based on the real-time operating data, the current aging stage of the target battery is determined, and a dynamic performance curve characterizing the current working state is generated. Based on the aging stage, the deviation characteristics between the dynamic performance curve and the performance benchmark curve are calculated, and the preset initial benchmark threshold is dynamically corrected based on the deviation characteristics to obtain a dynamic target threshold range that adapts to the current operating state. The real-time monitoring value in the real-time operating data is compared with the dynamic target threshold range. If the real-time monitoring value exceeds the dynamic target threshold range, it is determined that the target battery has a fault and a fault warning is triggered.

2. The method for early warning of battery faults according to claim 1, characterized in that, Determining the current aging stage of the target battery based on the real-time operating data specifically includes: The rate of change of internal resistance of the target battery is calculated based on the real-time operating data, where the rate of change of internal resistance = (real-time internal resistance - initial internal resistance) / initial internal resistance × 100%; The current aging stage is determined based on the internal resistance change rate and the running time. The determination of the current aging stage based on the internal resistance change rate and operating time specifically includes: When the rate of change of internal resistance is within the first rate of change range and the running time is within the first duration range, it is determined to be a new commissioning phase; When the rate of change of internal resistance is in the second rate of change range and the running time is in the second time range, it is determined to be in the stable operation phase. When the rate of change of internal resistance is in the third rate of change range or the running time is in the third duration range, it is determined to be in the aging stage. Wherein, the upper limit of the first rate of change interval is less than the lower limit of the second rate of change interval, and the upper limit of the second rate of change interval is less than the lower limit of the third rate of change interval; the upper limit of the first duration interval is less than the lower limit of the second duration interval, and the upper limit of the second duration interval is less than the lower limit of the third duration interval.

3. The method for early warning of battery faults according to claim 1, characterized in that, The step of calculating the deviation characteristics between the dynamic performance curve and the performance benchmark curve based on the aging stage includes: The dynamic performance curve and the performance benchmark curve corresponding to the aging stage are synchronously segmented according to a preset time window. Within each time window, the average deviation, maximum deviation, and duration of deviation of the dynamic performance curve relative to the performance baseline curve corresponding to the aging stage are calculated respectively. Wherein, the average deviation value is the difference between the average value of the dynamic performance curve within the time window and the average value of the performance benchmark curve corresponding to the aging stage; the maximum deviation value is the difference between the peak value of the dynamic performance curve within the time window and the peak value of the performance benchmark curve corresponding to the aging stage; the deviation duration is the cumulative duration for which the value of the dynamic performance curve exceeds the upper limit of the fluctuation range threshold or falls below the lower limit of the fluctuation range threshold of the performance benchmark curve corresponding to the aging stage. The average deviation value, the maximum deviation value, and the duration of the deviation are determined as the deviation characteristics between the dynamic performance curve and the performance benchmark curve corresponding to the aging stage.

4. The method for early warning of battery faults according to claim 1, characterized in that, The step of dynamically correcting the preset initial benchmark threshold based on the deviation characteristics to obtain a dynamic target threshold range that adapts to the current operating state specifically includes: Retrieve the initial baseline threshold corresponding to the current aging stage; Based on the preset deviation and threshold mapping model, the threshold correction coefficient is calculated according to the average deviation value, maximum deviation value and deviation duration in the deviation characteristics; The initial reference threshold is calculated using the corresponding correction formula selected according to the parameter type; where, for temperature parameters, the corrected threshold = initial reference threshold × (1 + threshold correction coefficient); for voltage parameters, the corrected threshold = initial reference threshold × (1 - threshold correction coefficient). The calculated corrected threshold is subject to boundary constraints to ensure it does not exceed the physical tolerance limit of the target battery, and the corrected threshold after boundary constraint correction is used as the range value of the dynamic target threshold.

5. The method for early warning of battery faults according to claim 1, characterized in that, The method further includes: The corrected thresholds are indexed in a multidimensional way according to the battery cell number, parameter type, aging stage and correction time, and stored centrally to form the corresponding threshold files. Based on the stored threshold files, the dynamic changes of deviation characteristics are continuously tracked. When the change of deviation characteristics exceeds the preset ratio or the current aging stage changes, the threshold re-correction process is automatically triggered. The recalculated threshold is used to overwrite and store the threshold files and push them synchronously.

6. The method for early warning of battery faults according to claim 1, characterized in that, The fault warning specifically includes: Cross-validation using multi-dimensional operational data is used to determine the fault point and fault type. Among them, abnormal temperature accompanied by voltage drop is identified as a heat fault, voltage drop accompanied by sudden increase in internal resistance is identified as a leakage fault, and temperature rise accompanied by shell deformation data is identified as a bulging fault. A three-level early warning mechanism is adopted for graded response based on the fault type and its severity. The first-level early warning sends an SMS notification to the maintenance personnel and is applicable to minor anomalies in the early stage. The second-level early warning sends a pop-up warning on the platform and automatically generates a maintenance work order, which is applicable to confirmed medium risks. The third-level early warning triggers an on-site audible and visual warning and links the energy storage system to cut off the circuit of the faulty battery pack, which is applicable to serious faults.

7. The method for early warning of battery faults according to any one of claims 1 to 6, characterized in that, The construction of performance benchmark curves corresponding to different aging stages based on the historical running dataset specifically includes: Based on the historical operation dataset, the life cycle of the target battery is divided into the new commissioning stage, the stable operation stage, and the aging stage. The historical operation dataset includes the factory test data and annual operation and maintenance data of the target battery. The timestamp of the factory test data is the factory time of the target battery, and the time span of the annual operation and maintenance data covers the entire life cycle from the new commissioning of the target battery to its aging and scrapping. For the new commissioning stage, stable operation stage, and aging stage, respectively, the fluctuation range curves of the voltage, current, temperature, and internal resistance parameters of the target battery are fitted to form the performance benchmark curve of the target battery throughout its entire life cycle.

8. The method for early warning of battery faults according to claim 7, characterized in that, The process of fitting fluctuation range curves of the target battery's voltage, current, temperature, and internal resistance parameters for the new commissioning phase, stable operation phase, and aging phase respectively includes: For the new commissioning stage, stable operation stage, and aging stage, voltage, current, temperature, and internal resistance data of the target battery are extracted at different time points or under the corresponding state of charge within the corresponding stage. For each type of parameter data in each stage, calculate the mean and standard deviation of the data at the same time node or under the same state of charge; and use the time node or state of charge as the horizontal axis, the corresponding mean of the data as the median, and the standard deviation of the data at a preset multiple as the fluctuation boundary to fit and generate the corresponding fluctuation range curve for each type of parameter in each stage.

9. A battery fault early warning device, characterized in that, include: The acquisition unit is used to acquire the historical operating dataset and real-time operating data of the target battery. A processing unit, connected to the acquisition unit, is used to construct performance benchmark curves corresponding to different aging stages based on the historical running dataset. Furthermore, based on the real-time operating data, the current aging stage of the target battery is determined, and a dynamic performance curve characterizing the current operating state is generated; A calculation unit, connected to the processing unit, is used to calculate the deviation characteristics between the dynamic performance curve and the performance reference curve based on the aging stage. The correction unit, connected to the calculation unit, is used to dynamically correct the preset initial benchmark threshold based on the deviation characteristics to obtain a dynamic target threshold range that adapts to the current operating state. The comparison unit, connected to the correction unit, is used to compare the real-time monitoring value in the real-time operating data with the dynamic target threshold range. If the real-time monitoring value exceeds the dynamic target threshold range, it is determined that the target battery has a fault and a fault warning is triggered.

10. An electronic device comprising a memory and a processor, characterized in that, The memory stores a computer program, and when the processor runs the computer program stored in the memory, the processor executes the battery fault early warning method according to any one of claims 1 to 8.

Images