Intelligent overload protection and fault prediction system suitable for mobile power supply

Abstract

The invention relates to the technical field of mobile power supply monitoring, in particular to an intelligent overload protection and fault prediction system suitable for a mobile power supply, and the system comprises a real-time impedance monitoring module which continuously monitors the internal impedance of a battery in the use process of the mobile power supply, and generates a real-time impedance analysis result; and according to the real-time impedance analysis result, adjusting the charging current, matching the real-time state of the battery, and generating an impedance adjustment decision. According to the invention, the internal impedance of the mobile power supply is continuously monitored and analyzed, and the charging current can be dynamically adjusted, so that the accurate adaptation of the real-time state of the battery is realized, the safety of the charging process is improved, and the service life of the battery is prolonged; on the basis of continuous comparison between the voltage and current data and the preset normal range and real-time detection of abnormal fluctuation, the response speed to the abnormal state of the battery is improved, and the safety risk in the battery operation process is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of mobile power supply monitoring, and particularly to an intelligent overload protection and fault prediction system applicable to mobile power supplies. Background Art

[0002] The intelligent overload protection and fault prediction system is for the mobile power supply used in new energy vehicles, which can automatically identify and intelligently protect against overload during operation, and predict and diagnose possible faults in advance to reduce the occurrence of safety accidents and improve the operation safety and reliability of the mobile power supply of new energy vehicles.

[0003] In the actual operation of the prior art, the judgment of overload protection and fault prediction only depends on the abnormal detection of voltage and current after the fault occurs, and it is difficult to realize the real-time tracking of the internal state change of the battery, resulting in a single monitoring method and obvious lag. Due to the lack of attention to the real-time impedance state of the battery, abnormal signs cannot be detected in time during the gradual deterioration of the battery, resulting in a passive warning mechanism and insufficient accuracy of the prediction results. Therefore, improvements are needed. Summary of the Invention

[0004] The purpose of the present invention is to solve the disadvantages existing in the prior art, and to propose an intelligent overload protection and fault prediction system applicable to mobile power supplies.

[0005] To achieve the above purpose, the present invention adopts the following technical solutions: An intelligent overload protection and fault prediction system applicable to mobile power supplies includes:

[0006] A real-time impedance monitoring module, which continuously monitors the internal impedance of the battery during the use of the mobile power supply, generates a real-time impedance analysis result; according to the real-time impedance analysis result, adjusts the charging current, matches the real-time state of the battery, and generates an impedance adjustment decision;

[0007] A voltage and current abnormal detection module, which continuously monitors the voltage and current data of the battery, compares with the preset normal working range, and generates a voltage and current monitoring result; based on the voltage and current monitoring result, determines whether abnormal fluctuations in voltage or current are detected, and if detected, immediately issues an abnormal warning and generates an abnormal state alarm;

[0008] A battery maintenance scheduling module, which calculates the charge and discharge efficiency of the battery according to the abnormal state alarm and the charge and discharge cycle of the battery, generates a battery health assessment result; according to the battery health assessment result, obtains the maintenance and replacement schedule of the battery and generates a maintenance scheduling plan;

[0009] A fault prediction and alarm module, which combines the impedance adjustment decision and the maintenance scheduling plan, and uses a statistical model to predict and estimate the occurring battery faults, and generates a fault prediction analysis result.

[0010] Preferably, the step of obtaining the real-time impedance analysis result is as follows: continuously collect the internal impedance changes of the battery under various usage states, record the impedance values of each collection, and obtain the original battery impedance data;

[0011] Based on the original battery impedance data, apply moving window averaging, calculate the average impedance value in each window, identify and remove the data points that exceed twice the average impedance value, and generate the real-time impedance analysis result.

[0012] Preferably, the step of obtaining the impedance adjustment decision is as follows: based on the real-time impedance analysis result, calculate the charging current, and the calculation formula is:

[0013]

[0014] where, I new represents the adjusted charging current, I base represents the initial charging current, α represents the adjustment sensitivity parameter, R current represents the current impedance value, represents the impedance change rate, T cell represents the battery temperature, β and γ are adjustment parameters to reflect the influence of the impedance change rate and temperature on the charging current, R base represents the reference impedance value;

[0015] Based on the charging current, cooperate with the state of the battery to adjust the charging strategy, and form and implement the impedance adjustment decision.

[0016] Preferably, the step of obtaining the voltage and current monitoring result is as follows: record the voltage and current data of the mobile power supply in various working states in real time, and form the original data set of the battery voltage and current;

[0017] Based on the original data set of the battery voltage and current, by comparing each data point with the preset normal working voltage and current ranges, identify and mark all the abnormal points that exceed the normal working voltage and current ranges, and form the voltage and current monitoring result.

[0018] Preferably, the step of obtaining the abnormal state alarm is as follows: receive the voltage and current monitoring result, analyze the voltage and current data of all time points, extract the measured values of each time point, and calculate the voltage change amount, current change amount, battery temperature, battery load state, remaining battery capacity ratio, and ambient temperature at each time point, and establish a voltage and current fluctuation data set after sorting;

[0019] Based on the voltage and current fluctuation data set, calculate the abnormal fluctuation determination value, and the calculation formula is:

[0020]

[0021] Among them, A is the abnormal fluctuation determination value, Q i is the voltage value measured at the i-th time, J j is the current value measured at the j-th time, is the average value of the current measurement values, m and n are the measurement times of voltage and current respectively, T cell is the battery temperature, T e is the ambient temperature, L is the current load state of the battery, C r is the remaining capacity ratio of the battery, Q max is the maximum voltage value within the sampling period, Q min is the minimum voltage value within the sampling period, Q i-1 is the voltage value measured at the (i - 1)-th time;

[0022] Based on the abnormal fluctuation determination value, it is detected whether it exceeds the set threshold, and if it exceeds, an abnormal state alarm is triggered.

[0023] Preferably, the steps for obtaining the battery health assessment result are: based on the abnormal state alarm and the charge and discharge cycle data of the battery, sort out the charge amount, discharge amount, cycle duration, and the associated number of abnormal state alarms in each cycle to obtain a battery charge and discharge cycle data set;

[0024] Based on the battery charge and discharge cycle data set, calculate the battery charge and discharge efficiency, and the calculation formula is:

[0025]

[0026] Among them, E is the battery charge and discharge efficiency, D k is the discharge amount of the k-th cycle, C k is the charge amount of the k-th cycle, p is the number of observation cycles, N a is the number of abnormal alarms in the k-th cycle, T avg is the average charge and discharge cycle duration, T max is the longest single charge and discharge cycle duration during the observation period;

[0027] Based on the battery charge and discharge efficiency, evaluate the health state of the battery and generate a battery health assessment result.

[0028] Preferably, the steps for obtaining the maintenance scheduling plan are: based on the battery health assessment result, calculate the next maintenance replacement time, and the calculation formula is:

[0029]

[0030] Among them, T next is the next maintenance replacement time, H max is the designed life of the battery, H current is the remaining life evaluated currently, Cinitial is the initial capacity of the battery, C current is the current capacity, E initial is the initial charging efficiency of the battery, E current is the current charging efficiency, N cycles is the cumulative charge-discharge cycle number of the battery;

[0031] Based on the next maintenance replacement time, combining the maintenance cycle and the replacement strategy, formulate a maintenance and replacement schedule to generate a maintenance scheduling plan.

[0032] Preferably, the steps for obtaining the fault prediction analysis result are: Based on the impedance adjustment decision and the maintenance scheduling plan, estimate the predicted fault probability of the battery. The calculation formula is:

[0033]

[0034] where F represents the fault probability, λ is the baseline failure rate fitted according to the fault data, Da is the number of days since the last maintenance, Ca is the number of charging times since the last maintenance, and Na is the number of discharging times since the last maintenance;

[0035] Based on the fault probability, conduct fault prediction analysis to guide the maintenance and replacement strategy to obtain the fault prediction analysis result.

[0036] Compared with the prior art, the advantages and positive effects of the present invention are:

[0037] In the present invention, by continuously monitoring and analyzing the internal impedance of the mobile power supply, the charging current can be dynamically adjusted to achieve precise adaptation to the real-time state of the battery, improve the safety of the charging process and the service life of the battery; based on the voltage and current data and the preset normal range for continuous comparison and real-time detection of abnormal fluctuations, the response speed to the abnormal state of the battery is improved, and the safety risk during the operation of the battery is reduced; combining the battery charge-discharge cycle to calculate the health assessment result, formulating a maintenance and replacement schedule, making the maintenance of the battery more targeted, and reducing the waste of resources caused by untimely or excessive maintenance; integrating the impedance adjustment decision and the maintenance plan, predicting possible faults through statistical analysis, realizing the active identification and early disposal of potential safety hazards of the battery, and improving the operation reliability and overall safety level of the mobile power supply. Brief Description of the Drawings

[0038] Figure 1 is the system flow chart of the present invention. Detailed Embodiments

[0039] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0040] Please refer to Figure 1 , the present invention provides a technical solution: an intelligent overcurrent protection and fault prediction system applicable to a mobile power supply includes:

[0041] A real-time impedance monitoring module, during the use of the mobile power supply, continuously monitors the internal impedance of the battery, generates a real-time impedance analysis result; according to the real-time impedance analysis result, adjusts the charging current, matches the real-time state of the battery, and generates an impedance adjustment decision;

[0042] A voltage and current anomaly detection module, continuously monitors the voltage and current data of the battery, compares with the preset normal operating range, and generates a voltage and current monitoring result; based on the voltage and current monitoring result, determines whether voltage or current abnormal fluctuations are detected, and if detected, immediately issues an anomaly warning and generates an abnormal state alarm;

[0043] A battery maintenance scheduling module, according to the abnormal state alarm and the charge and discharge cycle of the battery, calculates the charge and discharge efficiency of the battery, generates a battery health assessment result; according to the battery health assessment result, obtains the maintenance and replacement schedule of the battery, and generates a maintenance scheduling plan;

[0044] A fault prediction and alarm module, combines the impedance adjustment decision and the maintenance scheduling plan, and uses a statistical model to predict the estimated battery faults to be occurred, and generates a fault prediction analysis result.

[0045] The steps for obtaining the real-time impedance analysis result are as follows: continuously collect the internal impedance changes of the battery under various usage states, record the impedance values collected each time, and obtain the original battery impedance data;

[0046] Based on the original battery impedance data, apply a sliding window average to calculate the average impedance value in each window, identify and remove the data points that exceed twice the average impedance value, and generate a real-time impedance analysis result.

[0047] Specifically, continuously observe the change in the internal impedance of the battery under different temperature ranges (e.g., 0°C to 90°C), different voltage ranges (e.g., 0V to 24V), and different load current ranges (e.g., 0A to 5A). Identify the acquisition time and environmental conditions according to the ranges preset for these usage states. When recording each acquired impedance value, simultaneously mark the measurement time, working voltage, current magnitude, and temperature information. To avoid missing any extreme states, an additional sampling frequency of 10 times per second is set and continuous observation is maintained. Directly compare the observed impedance values with possible actual usage scenarios to confirm the integrity and consistency of the data. For example, when the temperature sampling value falls between 0°C and 90°C, the voltage value is in the range of 0V to 24V, and the current value is in the range of 0A to 5A, it is regarded as a valid record; otherwise, the corresponding acquisition result will be separately marked for further confirmation or troubleshooting in subsequent processes. This marking method is achieved by maintaining a judgment flag field in the database. The threshold and range are both derived from a comprehensive analysis of the technical parameters provided by the battery supplier and the results of multiple experimental measurements, and the reliability is verified with actual examples. For example, in several tests, impedance measurements of the same type of battery in an environment of 15°C, 25°C, and 40°C obtain similar resistance ranges, thus determining that the above temperature and voltage-current ranges can cover the main usage conditions. Each acquisition record is real-time recorded into a data table to form serialized data, providing an index for subsequent processing and maintaining a smooth time sequence to obtain the original battery impedance data.

[0048] Based on the original battery impedance data obtained previously, perform a moving window average process on each acquired impedance value in combination with the known observation sequence length. To clarify the window size, a configuration with a window width of 10 is selected from actual examples and can be fine-tuned according to different battery capacities and test requirements. First, add up all the impedance values within each window and divide by 10 to obtain the average impedance value, denoted as Then use this as the benchmark to calculate the threshold where the coefficient of 2 is selected based on the statistical evaluation of the impedance distribution of the same model battery in repeated experiments and combined with the actual measurement accuracy and error range. For example, if the average impedance value within a certain window is 0.05Ω, the threshold is 0.10Ω. When a data point exceeds 0.10Ω, it will be identified as abnormal and marked as data to be excluded. To explain the process of identifying this abnormal value, it is also necessary to comprehensively compare the overall distribution of the data in the same batch in other window segments and confirm that it is not caused by a measurement equipment failure. This judgment process also refers to the previously set temperature, voltage, and current ranges and combines the repeated comparison results of impedance measurements. Once it is confirmed that these data points exceeding the threshold do indeed deviate from the normal level, they will be excluded. After scanning all the data in the windows, summarize the effective impedance mean sequence of each window to generate the real-time impedance analysis result.

[0049] The steps for obtaining the impedance adjustment decision are as follows: Based on the real-time impedance analysis results, calculate the charging current, and the calculation formula is:

[0050]

[0051] Where, I new represents the adjusted charging current, I base represents the initial charging current, α represents the adjustment sensitivity parameter, R current represents the current impedance value, represents the change rate of impedance, T cell represents the battery temperature, β and γ are adjustment parameters to reflect the influence of the impedance change rate and temperature on the charging current, R base represents the reference impedance value;

[0052] Based on the charging current, cooperate with the state of the battery to adjust the charging strategy, and form and implement the impedance adjustment decision.

[0053] Specifically, the advantage of the formula is that by introducing multiple parameters such as α, β, and γ into the denominator and comprehensively considering multiple monitoring results such as the current impedance value, the impedance change rate, and the temperature factor, the charging current can be dynamically adjusted more flexibly in actual operations. Especially when the impedance deviates or the temperature is in different ranges, a more refined charging strategy can be formed, so as to obtain a charging process that takes into account both safety and efficiency in the overall system.

[0054] I base The steps for obtaining the parameter are as follows: When initially establishing the system, to obtain the reference charging current, it is necessary to conduct charging experiments on several batteries of the same model or series and record the current values in the initial charging stage of the batteries, and continuously track the performance of each battery for a period of time. Decompose and statistically analyze the recorded initial charging current data according to different temperature ranges. For example, compare the temperature in the range of 0°C to 90°C. If the initial charging current of most batteries is measured to fall within the range of 2.0A to 2.3A under the condition of 25°C for multiple times, then 2.0A can be established as the initial reference value for this model. For other temperature ranges, collect and summarize the reference currents corresponding to each temperature in the same way. Finally, take the average value of the appropriate range to determine I base . In a set of data from 5 repeated experiments, if 2.0A, 2.1A, 2.1A, 2.2A, 2.0A are obtained, the average value 2.08A can be comprehensively evaluated, and I base is rounded to 2.1A.

[0055] The steps to obtain the α parameter are as follows: α is used to characterize the adjustment sensitivity of the charging current with respect to impedance and temperature changes. It is necessary to quantify the impact of impedance fluctuations and temperature differences on the actual charging current after integrating multi-day sampling data. The specific method is to first record the current deviation caused by impedance fluctuations over several days, and then compare it in combination with the temperature range. For example, during 30 consecutive days of monitoring, the reduction in current caused by the impedance changing from 0.05 Ω to 0.06 Ω is statistically calculated. For example, if the average current drops from 2.1 A to 1.95 A, then the adjustment ratio can be obtained approximately Then cross-compare with the temperature fluctuation range (such as from 25 °C to 40 °C), and calculate cumulatively according to the time series data, so as to generate multiple pairs of corresponding values of the adjustment ratio and the temperature difference. Then obtain α according to the regression method. The regression process is through: where ΔI i is the current change amount at the i-th acquisition, I base,i is the reference current, ΔT i is the difference between the temperature and the reference temperature, and n is the number of observations. If α≈0.28 is obtained in a set of actual calculations, then take its average value of 0.3 according to the relatively stable stage of the monitoring data.

[0056] R current The steps to obtain the R parameter are as follows: Obtain the R current value by measuring the impedance of the battery at the current moment. Synchronize the original battery impedance data obtained previously in time and extract the most recent valid impedance measurement value. Assume that this measurement value is collected and registered at 9:00 am in a day, then directly incorporate it into the current impedance information. If the measured impedance is 0.06 Ω at this time, then record it as R current = 0.06 Ω.

[0057] The steps to obtain the parameter are as follows: This parameter represents the change rate of impedance. It is necessary to continuously monitor several recent impedance measurement values, and then obtain it through differential or slope calculation methods. The specific steps can be referred to as follows: First, obtain several impedance values at intervals of 1 hour within a day Use Divide the difference between two adjacent measurements by the measurement interval. For example, if the impedance measured at 10:00 am is 0.055 Ω and the impedance measured at 11:00 am is 0.06 Ω, then

[0058] T cell The steps to obtain the T parameter are as follows: T cellis the current temperature of the battery. The temperature is recorded by placing a temperature measurer at a position close to the inside or the surface of the device. The position is selected at the vicinity of the positive and negative connection terminals so as to reflect the internal temperature rise. By continuously collecting 24-hour data and differentiating the monitoring results of the high-temperature section (greater than 40 °C), the normal-temperature section (20 °C to 40 °C), and the low-temperature section (0 °C to 20 °C), the most matching temperature value of the battery at the current moment is found. For example, in actual operation, if the temperature probe measures 25.5 °C at this time period, then T cell can be taken as 25.5.

[0059] The steps for obtaining the β parameter are as follows: β is used to reflect the weight of the impedance change rate on the current regulation part. When establishing the correlation with the actual current adjustment, the current decrease amount caused by increase can be calculated first for several measurement periods, and statistical analysis is performed. Then, the linear or non-linear fitting is performed on the difference between this decrease amplitude and itself to obtain the β value. In the example, the and the corresponding current decrease ratio within different time periods in two weeks may be collected first. By: M is the number of samplings. This ratio is determined according to the distribution range of the relatively stable stage of the monitoring environment. If β = 1.2 is obtained in a certain calculation, it can be finally selected between 1.1 and 1.2 after multiple comparisons.

[0060] The steps for obtaining the γ parameter are as follows: γ is used to reflect the correction amount of the temperature to the current change. Similar to the way of obtaining β, it is necessary to conduct multiple analyses on the relationship between the temperature range and the actual charging current adjustment during the system operation. First, the current data in each temperature range is extracted, and the difference statistics are made with the current data at the reference temperature. Then, these differences are related to the temperature deviation to obtain the mapping relationship. Specifically, it can be adopted: P represents the number of monitoring days, and I temp,p represents the current value measured within the corresponding temperature section on the p-th day. T base can take the temperature at which this type of battery usually works (such as 25 °C). After weighting the data of multiple days, a relatively stable average value is selected as γ. If 1.05 is obtained after multiple operations, it can be set within the range of 1.0 to 1.1.

[0061] R base The steps for obtaining the R parameter are as follows: R baseis the reference impedance value for comparison. It is obtained by centrally collecting the impedance data measured for multiple batteries of the same model at room temperature (25°C ± 5°C), and after excluding extreme cases, taking the average of the remaining impedance values as the final reference impedance value. In a centralized measurement of 20 batteries, if most of the measured impedance values are between 0.04Ω and 0.06Ω, the intermediate value of 0.05Ω can be selected as R base 。

[0062] Calculation process:

[0063] When the values of the previous 9 parameters have been obtained, the operation can be carried out. Assume that the observation conditions at the current moment are as follows:

[0064] I base =2.1A, α = 0.3, R current =0.06Ω, T cell =

[0065] 25.5, β = 1.1, γ = 1.05, R base =0.05Ω;

[0066] First, write in the numerator part:

[0067] I base =2.1A;

[0068] Next, calculate the internal term in the denominator. Let:

[0069]

[0070] Then divide by R base :

[0071]

[0072] Take the natural logarithm of it and take the absolute value:

[0073] |ln(536.81)| ≈ 6.28;

[0074] Subsequently, multiply by α:

[0075]

[0076] Add 1 to it:

[0077] 1 + 1.884 = 2.884;

[0078] Finally, divide the numerator by this result: This result indicates that when the system real-time monitors that the current impedance, temperature and other parameters are in the above state, the calculated I newIt is about 0.73A, which corresponds to the more conservative charging current intensity at this moment. If the impedance or temperature conditions change subsequently, the same formula can be used to calculate again and obtain a new charging current value, thereby completing the subsequent adjustment of the charging strategy.

[0079] Based on the previously obtained charging current value and the actual charging and discharging records that have been registered, the current charging current is compared with the activity state of the battery at different temperatures. First, set the voltage, current and temperature data to be read every 15 minutes in a day and mark the timestamp. In order to ensure the accuracy of the data, all the read values ​​must be determined in intervals, such as comparing the temperature with the 0℃ to 90℃ interval, the voltage with the 0V to 24V interval, and the current with the 0A to 5A interval. The records that are out of range are marked as data that need to be reviewed. Then, the confirmed valid data is selected to calculate the real-time charging rate and analyze the duration of the charging stage. Each set of data is paired with the impedance measurement record and focus is placed on whether the impedance changes with the linkage of temperature and current. The impedance of the same period should also be compared. The impedance mean value is used as a parameter to compare with the impedance analysis results obtained previously. If the impedance exceeds the specified value for multiple consecutive times, it is necessary to check the source of the value and confirm whether it reaches the pre-set warning threshold. These warning thresholds come from a comprehensive analysis of multiple sample test experiences and multi-scenario monitoring results. If the impedance is found to be higher than 0.06Ω or the temperature is higher than 45°C during the judgment process, it will be listed as a higher risk range, and combined with the actual temperature rise rate to determine whether it is necessary to terminate charging in time or enter the next safety review link. In order to complete the corresponding processing, it is also necessary to establish more detailed grading standards within this range to determine the number of consecutive abnormalities or the degree of abnormality. When all comparisons are completed, the charging strategy adjustment behavior at the current moment can be obtained, and the impedance adjustment decision can be formed and implemented.

[0080] The steps for obtaining the voltage and current monitoring results are as follows: real-time recording of the voltage and current data of the mobile power supply under various working conditions to form a battery voltage and current original data set;

[0081] Based on the battery voltage and current raw data set, by comparing each data point with the preset normal operating voltage and current range, all abnormal points that exceed the normal operating voltage and current range are identified and marked to form the voltage and current monitoring results.

[0082] Specifically, according to the operation data of the mobile power supply prepared previously, record all the voltage and current data under different load modes, continuously monitor the power supply in various states such as standby, light load, and full load, and obtain real-time samples every 10 seconds. To ensure the accuracy of the data, check whether the temperature falls between 0°C and 90°C, the voltage falls between 0V and 24V, and the current falls between 0A and 5A during sampling, and record the current timestamp and corresponding working condition information at the same time. These judgment ranges are obtained from repeated measurements and statistical inductions of multiple mobile power supplies, and after repeatedly testing the device in normal temperature, low temperature, and high temperature environments, a threshold system suitable for this model of mobile power supply is formed. For example, for a power supply with a capacity of 10,000 mAh, if the voltage generally remains in the range of 5V to 12V and the current remains in the range of 1A to 3A at normal temperature, it is considered normal. When encountering changes in ambient temperature or extreme conditions, the threshold range can be increased or decreased accordingly. Perform item-by-item comparison in the obtained reference interval. If the voltage or current value in a certain sampling is not within the range of 0V to 24V or 0A to 5A, give a mark through a specially set judgment field. If the judgment field indicates that it exceeds the interval, list this type of data as an observation value that requires additional investigation. If the judgment field does not indicate an excess, it is regarded as preliminary valid data. Subsequently, summarize all the preliminary valid data to obtain the original dataset of battery voltage and current.

[0083] After obtaining the original dataset of battery voltage and current summarized previously, select each record in it and compare it with the previously set normal voltage range and normal current range. Compare the voltage with the range of 0V to 24V and the current with the range of 0A to 5A, and mark each record that does not meet the conditions one by one. These ranges are obtained from the collation and statistics of the measured values of multiple charge and discharge cycles. Classify the reasons for the abnormally marked values into specific types such as too high voltage, too low voltage, too large current, or too small current. If the voltage in a single sampling exceeds 24V, it is determined as a voltage anomaly point and marked. If the current in a single sampling is higher than 5A or lower than 0A, the same marking is also carried out. If the temperature information in the record exceeds 90°C, it also needs to be separately marked as over-temperature data. Considering that in some extreme cases, the values may have short-term jumps, so multiple over-limit observations in multiple consecutive sampled data are regarded as long-term anomalies, otherwise only short-term offset processing is performed. All the marked data that are still determined to be over-limit after investigation will be summarized and stored in the anomaly list for subsequent investigation and detection. When the comparison and marking of all records are completed, the voltage and current monitoring results can be generated.

[0084] The steps to obtain the abnormal state alarm are as follows: receive the voltage and current monitoring results, parse the voltage and current data at all time points, extract the measured values at each time point, and calculate the voltage change amount, current change amount, battery temperature, battery load status, remaining battery capacity ratio, and ambient temperature at each time point. After collation, establish a voltage and current fluctuation dataset;

[0085] Based on the voltage and current fluctuation dataset, calculate the abnormal fluctuation determination value. The calculation formula is as follows:

[0086]

[0087] where A is the abnormal fluctuation determination value, Q i is the voltage value of the i-th measurement, J j is the current value of the j-th measurement, is the average value of the current measurement values, m and n are the number of voltage and current measurements respectively, T cell is the battery temperature, T e is the ambient temperature, L is the current load state of the battery, C r is the remaining capacity ratio of the battery, Q max is the maximum voltage value within the sampling period, Q min is the minimum voltage value within the sampling period, Q i-1 is the voltage value of the (i - 1)-th measurement;

[0088] Based on the abnormal fluctuation determination value, detect whether it exceeds the set threshold. If it exceeds, trigger an abnormal state alarm.

[0089] Specifically, receive the voltage and current monitoring results obtained previously. Extract the corresponding voltage and current measurement values for each time point and archive them on the same time axis. After statistically obtaining the basic status of all time points, record the temperature parameters, the current load state of the battery, and the remaining capacity ratio, etc. of this time point. To ensure the comparability of the data, pack all the parameters of the same time point for comparison and judge the data integrity. When there are missing measurement values or broken measurement value periods, divide them into the abnormal record list and arrange for subsequent review. The time points here can be collected at fixed time intervals, for example, input once every 30 seconds and mark the change of the ambient temperature in the range of 0°C to 90°C. Subsequently, calculate the voltage change amount at each time point based on the difference between the voltage value at this time point and the voltage value at the previous time point and summarize it. Process the data of all periods in this way to form a voltage change sequence. Apply the same method to the current dimension to obtain a current change amount sequence. After uniformly organizing them, add fields such as battery temperature, battery load state, remaining battery capacity ratio, and ambient temperature. During this process, if it is detected that the voltage amplitude is greater than 24V or the current amplitude is greater than 5A, include this record in a separate high-order record queue and conduct subsequent analysis in combination with the known power demand range. When the induction of all period records is completed, the corresponding voltage and current fluctuation dataset can be obtained.

[0090] The benefit of the formula is that by comprehensively superimposing factors such as voltage fluctuation, current deviation, temperature difference, load status and remaining capacity ratio, the impact of multiple aspects on abnormal fluctuations can be reflected in the same expression, and different correction elements are loaded by setting the denominator, so that different parameters can form a relatively reasonable balance during calculation. The entire formula helps to obtain a more intuitive abnormal judgment value in the monitoring link and take into account multiple types of factors in the calculation.

[0091] Q i The steps to obtain the parameters are: Q i Refers to the voltage value measured at the ith time. The specific value is obtained by periodically sampling the voltage measuring device during the monitoring process. For example, in the 12th record at the sampling time of a certain day, the voltage measurement value is 12.3V, then Q 12 Recorded as 12.3V.

[0092] Q i-1 The steps to obtain the parameters are: Q i-1 Indicates the voltage value measured at the i-1th time, which is located next to Q in the data table. i Previously, it was used to calculate the difference between two measurements. In order to accurately obtain Q i-1 , it is necessary to ensure that the time sequence of the records is continuous and the acquisition frequency is constant when generating the voltage sequence. A cumulative count can be added to each record. When calculating the voltage change, it is only necessary to directly index the i-1th record. For example, in the above text, the previous record of the 12th record is the 11th record. If the voltage in the 11th record is 11.5V, then Q 11 It is equal to 11.5V.

[0093] J j The steps to obtain the parameters are: j The jth measured current value needs to be recorded under the same periodic sampling mechanism. The range can be set between 0A and 5A, and the sampling time and load conditions should be noted in each record. In order to rigorously reflect the change of current under different conditions, it is necessary to list the corresponding current sequence at the same time as the voltage sequence. The two correspond one to one according to the same time step. If the current measurement value is 2.3A at a certain time, the jth record can be located and J j Written as 2.3A.

[0094] The steps to obtain the parameters are: is the mean of the current measurement values. It is necessary to sum up all the current records in the observation interval and divide them by the number of measurements. The number of measurements here only corresponds to the records included in the effective monitoring range during the period. It can be achieved through an addition and division, for example: Where n effis the number of current valid data points. Assuming that 180 current values are recorded during a 30 - minute monitoring period and all fall within the range of 0A to 5A, then add up these 180 records and divide by 180 to obtain For example, if the sum of all current records is 350A, then

[0095] The steps to obtain the m parameter are as follows: m refers to the number of voltage measurements, which is the total number of valid voltage records within a certain time period. For example, if measurements are taken once every 10 seconds for 6 hours, then ideally m = 6 hours ×

[0096] (3600 seconds ÷ 10 seconds) = 2160.

[0097] The steps to obtain the n parameter are as follows: n represents the number of current measurements. Similar to m, it is also the cumulative count of current sampling times in the data table.

[0098] T cell The steps to obtain the T parameter are as follows: T cell represents the battery temperature. A temperature acquisition device needs to be set at the battery shell or internal fitting to monitor, and the temperature value is synchronously collected at each time point. The numerical range is usually based on between 0℃ and 90℃. If the temperature is monitored at 32.5℃ at a certain moment, then T cell is established as 32.5.

[0099] T e The steps to obtain the T parameter are as follows: T e represents the ambient temperature. A special temperature measurement device needs to be set at the external environment location for recording, generally in an area close to the space where the battery is located. It is collected through the same time stamp and corresponds to T cell If the current ambient temperature is 26℃, then T e = 26.

[0100] The steps to obtain the L parameter are as follows: L represents the current load status of the battery, which is a numerical field used to describe the load level. It can be obtained by monitoring the real - time output power and dividing it according to power intervals, or linearly quantifying the output power within a certain range. For example, when the power is between 0W and 10W, L = 0.2; when the power is between 10W and 20W, L = 0.4, etc. After dividing all power intervals into several segments, L is determined by which interval the actually measured power falls into. Here is an example of the value - taking process. For example, when the output power is 12W, find the interval where this power is located and get L = 0.4.

[0101] C r The steps to obtain the C parameter are as follows: C rRefers to the ratio of the remaining battery capacity, which needs to be obtained through capacity statistics at the battery management level. Capacity statistics can be comprehensively calculated by means of the rated capacity at the time of factory and the current discharge measurement. After determining the remaining capacity value, divide the remaining capacity by the rated capacity to obtain the ratio. For example, for a 20000 mAh battery, the remaining capacity at the current moment is 14000 mAh.

[0102] Q max The steps to obtain the parameter are: Q max is the maximum voltage value within the sampling period. First, all valid Qs in this period i need to be extracted and traversed within the same sequence to find the maximum value, and then record it as Q max , for example, 300 voltage values are collected in a segment. After comparison, it is found that the highest voltage value reaches 19.8 V, then Q max = 19.8 V.

[0103] Q min The steps to obtain the parameter are: Q min is the minimum voltage value within the sampling period. For example, among the 300 voltage values in the same segment, the lowest voltage is found to be 5.5 V, then record Q min = 5.5 V.

[0104] Calculation process:

[0105] After all parameters are obtained, substitute them into the above expression for step-by-step calculation. The following is an example:

[0106] m = 10, n = 10, Q 1..10 =

[0107] {12.2, 12.7, 13.0, 12.8, 12.9, 13.2, 13.3, 12.6, 12.4, 13.1};

[0108] Q max = 13.3, Q min = 12.2, T cell = 32.5, T e = 26, L = 0.4, C r = 0.7;

[0109] J 1..10 = {1.8, 1.9, 2.0, 2.1, 1.6, 1.7, 2.2, 1.9, 2.0, 2.1};

[0110] First, calculate the voltage change part:

[0111]

[0112] Since the first Q0 can be the same as the first one to avoid having no previous value, the first item is 0. Here, the absolute values are taken after subtracting term by term and then accumulated, resulting in:

[0113] 0 + 0.5 + 0.3 + 0.2 + 0.1 + 0.3 + 0.1 + 0.7 + 0.2 + 0.7 = 2.9;

[0114]

[0115] Then calculate the part of the current fluctuation. First, find

[0116]

[0117] Calculate each item in turn and accumulate, getting a total of approximately 0.51;

[0118]

[0119] Then the temperature difference term:

[0120]

[0121] Finally, the remaining capacity ratio term:

[0122]

[0123] Sum up the above four parts: A = 0.29 + 0.714 + 4.64 + 0.333 ≈ 5.977. This result indicates that the comprehensive abnormal fluctuation determination value for this time period is approximately 5.977. This value can be used in a pre-established threshold system to determine whether the current fluctuation crosses the warning line. When the threshold set during monitoring is, for example, 5, then A > 5 means it is higher than the threshold, thus triggering an abnormal state alarm. When A ≤ 5, it is considered that the abnormal condition has not been reached, and thus stable monitoring is executed.

[0124] The steps to obtain the battery health assessment result are as follows: Based on the abnormal state alarm and the charge-discharge cycle data of the battery, organize the charge amount, discharge amount, cycle duration, and the associated number of abnormal state alarms in each cycle to obtain the battery charge-discharge cycle dataset;

[0125] Based on the battery charge-discharge cycle dataset, calculate the battery charge-discharge efficiency. The calculation formula is:

[0126]

[0127] Among them, E is the battery charge-discharge efficiency, D k is the discharge amount in the k-th cycle, C k is the charge amount in the k-th cycle, p is the number of observation cycles, N a is the number of abnormal alarms in the k-th cycle, Tavg is the average charge-discharge cycle duration, T max is the longest single charge-discharge cycle duration during the observation period;

[0128] Based on the battery charge-discharge efficiency, evaluate the health status of the battery and generate a battery health assessment result.

[0129] Specifically, according to the obtained abnormal status alarms and the charge-discharge cycle data of the battery, first extract the charge-discharge time periods matching the current device for each cycle and record the corresponding start and end times, then summarize the charge amount data and discharge amount data occurring in each cycle from these records, and at the same time register the start time and end time of each cycle within the same indexing system and calculate the cycle duration. If there are multiple abnormal status alarms in some cycles, add an alarm count field at the corresponding cycle record. Each cycle record must undergo a complete data comparison to confirm whether there are deviations or omissions and be audited against a pre-established valid range. For example, compare the charging current within the range of 0A to 5A, the voltage within the range of 0V to 24V, and combine whether the temperature is within the range of 0°C to 90°C to confirm whether the cycle monitoring data is available. When any observed value exceeds the range, add a corresponding mark in the cycle record and sort it in chronological order so that all cycles obtain information such as the charge amount, discharge amount, and associated abnormal alarm times. After all cycle data has been checked, summarize the sorted information in a tokenized manner and arrange it in cycle sequence for convenient subsequent retrieval. When all required entries have been verified, a battery charge-discharge cycle dataset can be obtained.

[0130] The advantage of the formula is that it calculates a value that can comprehensively evaluate the battery charge-discharge efficiency by integrating multiple dimensions of elements such as the discharge amount, charge amount, abnormal alarm times, and cycle duration distribution of each cycle, and introduces the correction factor of N a and to make the high-frequency anomalies or long-term charge-discharge states have a more intuitive impact on the efficiency evaluation.

[0131] D k The acquisition steps of the D parameter are: D k represents the discharge amount of the k-th cycle, and the actual discharge value corresponding to this cycle needs to be queried in the previously obtained battery charge-discharge cycle dataset. For example, in a battery with a capacity of 10000 mAh, if a complete discharge process consumes 2800 mAh in a certain instance, then D k = 2800. After continuously collecting several cycles,

[0132] C k The acquisition steps of the C parameter are: C k represents the charge amount of the k-th cycle, similar to D kThe same is to read specific values from the previously summarized periodic dataset. For example, under the same battery, if the measurement in this period is 3000 mAh, then C k = 3000. By accumulating the data of several periods, we can obtain

[0133] The steps to obtain the p parameter are as follows: p represents the number of observation periods, which is obtained by counting all the complete charge-discharge processes experienced by the battery within a certain time range. The statistical method is to number each valid period in the period table and associate it with time. When the same battery may accumulate several periods after multiple cycles, generally p valid periods can be recorded within one week or a longer time. Each period has information such as the starting charging time, the ending discharging time, and the number of abnormal alarms. Finally, the total number obtained by statistics is recorded as p.

[0134] N a The steps to obtain the parameter are as follows: N a represents the number of abnormal alarms in the k-th period. It is necessary to count all the alarms triggered within this period when viewing the period data. The alarm information can be associated with the abnormal status alarms obtained previously. If a total of 3 alarms are recorded in the k-th period, then N a = 3.

[0135] T avg The steps to obtain the parameter are as follows: T avg represents the average charge-discharge cycle duration. First, it is necessary to record the start and end times of each period and calculate the length of this period, and then add up the length values of all periods and divide by the number of periods p. For example, if p = 5 periods, and the period durations are 3.5 hours, 4 hours, 3.8 hours, 3.9 hours, and 4.1 hours respectively, then

[0136] T max The steps to obtain the parameter are as follows: T max represents the longest single charge-discharge cycle duration during the observation period. It is to find the one with the longest duration among the same batch of observation periods and use its duration as T max , which can be obtained by traversing the durations of all periods and comparing the maximum value. Still taking the case of p = 5 periods as an example, the maximum value among the above 5 period durations is 4.1 hours, then T max = 4.1.

[0137] Calculation process:

[0138] The following gives an actual calculation example. Let p = 3 and obtain from the periodic dataset:

[0139] D1 = 2500 mAh, D2 = 2800 mAh, D3 = 2700 mAh;

[0140] C1=3000mAh, C2=3100mAh, C3=3050mAh;

[0141] N a =4,T avg = 3.5 hours, T max =4 hours;

[0142] First calculate the ratio of discharge to charge:

[0143]

[0144] Next process

[0145]

[0146] Final treatment Multiplying these three parts together, we get: E = 0.874 × 20 × 0.875 ≈ 15.313; the result shows that the charge and discharge efficiency of the battery in this example during these three cycles is 15.313. By definition, it is not a pure efficiency value in the sense of a percentage, but a comprehensive evaluation value defined by this formula. If the same calculation is performed for subsequent monitoring or other batteries, a horizontal comparison can be made. If the value is found to be less than 10, it means that the efficiency is low or there are more abnormal alarms in this stage. If it exceeds 20, it may indicate that the overall performance between cycles is better or the charge and discharge match is higher. Other data can be combined to further evaluate the battery health status.

[0147] Based on the battery charge and discharge efficiency calculated previously, the efficiency sequence that has been sorted out is compared one by one under the same observation environment, and the data rows with larger deviations are selected for key inspection. Then the remaining efficiency values ​​are summarized in sections and correlated with the corresponding operating temperature, current, voltage and load ranges for statistics. If low efficiency or sudden high abnormal alarms occur in multiple observations, whether the temperature range recorded in each cycle falls between 0℃ and 90℃, whether the voltage range falls between 0V and 24V, and whether the current range falls between 0A and 5A is confirmed whether there is an out-of-range operation. Then, the possible charging rate instability points are traced. If no significant abnormality is found, the sample is merged into the final health status score sequence. Here, the score calculation method needs to be separately counted and verified with the previous efficiency results. Then, the corresponding score label is added to each cycle in the data table and the health level is divided. Subsequently, this overall evaluation data is referenced with the historical operation data to obtain the final battery health evaluation result.

[0148] The steps to obtain the maintenance scheduling plan are as follows: Based on the battery health assessment results, calculate the next maintenance replacement time. The calculation formula is:

[0149]

[0150] Among them, T next is the replacement time for the next maintenance, H max is the designed battery life, H current is the remaining life currently evaluated, C initial is the initial capacity of the battery, C current is the current capacity, E initial is the initial charging efficiency of the battery, E current is the current charging efficiency, N cycles is the cumulative charge-discharge cycle number of the battery;

[0151] Based on the replacement time for the next maintenance, combined with the maintenance cycle and replacement strategy, formulate a maintenance and replacement schedule to generate a maintenance scheduling plan.

[0152] Specifically, the benefit of the formula is that by introducing the logarithmic relationship of the remaining life, capacity ratio, efficiency ratio, and cumulative charge-discharge cycle number in the product, battery state information in different dimensions can be integrated within the same calculation framework, and the specific time for the next maintenance replacement can be further quantified according to the battery health assessment result.

[0153] H max The acquisition steps of the parameter are as follows: H max represents the designed battery life, which needs to be read from the technical specifications or factory standards of this type of battery and corrected by combining actual measurement data to obtain a benchmark value that can be used within the monitoring cycle. If the designed life is calibrated on the nameplate as 800 charge-discharge cycles or 2-year service life, for example, the designed life of a certain mobile power supply is about 1.5 years and is converted to 13140 hours, then H max = 13140.

[0154] H current The acquisition steps of the parameter are as follows: H current is the remaining life currently evaluated, and the battery needs to be subjected to multiple health detections. For example, it is estimated that the remaining life is about 0.9 years and is recorded in H current For example, 0.9 years corresponds to 7884 hours, then H current can be taken as 7884 here.

[0155] C initial The acquisition steps of the parameter are as follows: C initial is the initial capacity of the battery, which is usually given by the manufacturer or the quality inspection link when it is just out of the factory. For example, a battery nominally rated at 20000 mAh can actually reach about 19500 mAh during the initial factory inspection.

[0156] C currentThe steps to obtain parameter C are as follows: current It represents the current capacity and needs to measure or estimate the available capacity of the battery after the device has been used for a period of time. The measurement methods include recording the discharge amount under specific constant current or constant power discharge conditions and comparing it with the rated capacity. If it is detected that the available capacity is only 17000 mAh after a period of use, then C current = 17000.

[0157] E initial The steps to obtain parameter E are as follows: initial It represents the initial charging efficiency of the battery and requires combining actual test data or multiple charge-discharge experiments on this type of battery performed previously to obtain a relatively stable efficiency value. Usually, the measurement results of multiple complete charge-discharge cycles in the initial stage can be averaged to obtain E initial , for example, if the average charge-discharge efficiency is calculated to be approximately 92% after continuously measuring 5 complete charge-discharge cycles at the factory stage and comparing the ratio of the actual input and output capacities, then E initial = 0.92.

[0158] E current The steps to obtain parameter E are as follows: current It is the current charging efficiency, similar to the previously obtained E initial , and also needs to be obtained through real-time monitoring or the previous evaluation and calculation process. When the efficiency of the battery decreases after several cycles of use, for example, if the average efficiency of the recent 10 charge-discharge cycles is only 0.87, then E current = 0.87 and record it.

[0159] N cycles The steps to obtain parameter N are as follows: cycles It is used to represent the cumulative number of charge-discharge cycles of the battery and needs to accumulate all the complete charge-discharge cycle times since the start of use. When each discharge is until the battery is exhausted and then recharged to full charge, it is regarded as a complete cycle, and N is obtained by successive accumulation cycles , if multiple rounds of cycles have been registered in the monitoring system and the statistical result is 415 times, then N cycles = 415.

[0160] Calculation process:

[0161] The following gives a practical calculation example. Substitute the relevant indicators monitored after the same battery has been used for about 8 months into the calculation:

[0162] H max = 13140 hours, H current = 7884 hours, C initial = 19500 mAh, C current = 17000 mAh;

[0163] Einitial = 0.92, E current = 0.87, N cycles = 415;

[0164] First, calculate

[0165]

[0166] Then, calculate

[0167]

[0168]

[0169] Subsequently

[0170]

[0171] Then, calculate ln(1 + N cycles ):

[0172] 1 + 415 = 416, ln(416) ≈ 6.03;

[0173] Multiply the above four parts:

[0174] T next = 1.20 × 1.07 × 1.057 × 6.03;

[0175] T next ≈ 1.20 × 1.07 × 1.057 × 6.03 ≈ 8.12;

[0176] If all dimensions here are in hours, then T next ≈ 8.12 hours. This result indicates that the next maintenance replacement time obtained from this formula is approximately 8.12 hours, meaning that under the current monitoring conditions and known attenuation degree, the system can perform further inspection or replacement work according to the preset maintenance rules after about 8 hours. If this value is detected to be greater than 24, it may indicate that it can still maintain a long service time, while if the calculated result is less than 1, it means that maintenance or replacement is required soon, and it can be further split according to the subsequent maintenance cycle.

[0177] Based on the next maintenance replacement time obtained previously, cross-query the maintenance cycle and replacement strategy. First, extract adjacent maintenance periods according to the recorded maintenance cycle setting, and mark key indicators such as the remaining capacity and temperature of the current time period on the same timeline. Then, check item by item whether there is a situation outside the monitoring range where the current is in the range of 0A to 5A, the voltage is in the range of 0V to 24V, and the temperature is in the range of 0°C to 90°C. If it is found that the data falls outside the range multiple times, add an emergency inspection step to the maintenance plan arrangement, and pay attention to whether the cumulative charge and discharge times continue to increase to the set classification threshold when comparing multi-cycle records. These classification thresholds come from the actual operation and maintenance statistics results. For example, in the tracking of the use of the same model power supply for many years, it is confirmed that when the cumulative charge and discharge times are close to 500 times or more, the probability of rapid capacity decay will increase. Therefore, it is set as the preliminary threshold for this model. When the current cumulative times have exceeded the above threshold, a specific encrypted observation plan should be determined. After completing all cross-comparisons, organize the plan containing important time points and inspection contents into a schedule, and arrange the inspection order in the next few days or weeks according to the strategy being executed. Record the time points in the schedule in sequence to connect with the maintenance timeline of the system to avoid omissions or conflicts, thus finally generating a maintenance scheduling plan.

[0178] The steps to obtain the fault prediction analysis result are as follows: Based on the impedance adjustment decision and the maintenance scheduling plan, estimate the predicted fault probability of the battery. The calculation formula is:

[0179]

[0180] where F represents the fault probability, λ is the baseline failure rate fitted according to the fault data, Da is the number of days since the last maintenance, Ca is the number of charge times since the last maintenance, and Na is the number of discharge times since the last maintenance;

[0181] Based on the fault probability, conduct fault prediction analysis to guide the maintenance and replacement strategy, and obtain the fault prediction analysis result.

[0182] Specifically, the benefit of the formula is that by comprehensively considering the influence of days and charge and discharge times in the exponential term, and setting two weight values of 0.7 and 0.3 to distinguish the influence of days and charge and discharge ratio on the failure rate, it can take into account both the dual factors of daily time passage and usage intensity (based on charge and discharge times) in the same function form, and use the baseline failure rate λ as the overall adjustment, enabling the whole method to quantify the fault probability according to historical or current data.

[0183] The steps to obtain the λ parameter are as follows: λ is the baseline failure rate, which is obtained by performing statistical fitting on the previously collected failure data. The specific method is to first group a large number of known failure cases by occurrence time or charge-discharge cycles, and count the frequency of failures in the records. Then, logarithmic linear regression or exponential distribution fitting is used to obtain the failure rate. In the example, the following can be adopted: where f i is the cumulative frequency of failures in the i-th batch, t i is the observation duration or the number of usage cycles of this batch, and M is the number of batches included. After combining these samples, a relatively stable λ can be obtained and denoted as the baseline failure rate. For example, after tracking and statistically analyzing the 2-year usage of 100 batteries and finding that the cumulative annual failure frequency in the population sample is 0.015 / hour, then λ = 0.015 can be taken.

[0184] The steps to obtain the Da parameter are as follows: Da represents the number of days since the last maintenance. It is necessary to find the exact date of the last maintenance in the maintenance record and calculate the difference from the current date. Specifically, a time difference operation can be used to obtain the complete number of days. If the last maintenance date is March 1, 2025, and the current date is March 15, 2025, then Da = 14 days.

[0185] The steps to obtain the Ca parameter are as follows: Ca is the number of charge cycles since the last maintenance. Each time the process from low battery to full charge is completed, it can be regarded as an accumulation of the number of charge cycles. In the example, if a total of 7 complete charge processes are observed in the past 14 days, then Ca = 7.

[0186] The steps to obtain the Na parameter are as follows: Na represents the number of discharge cycles since the last maintenance. Similar to Ca, it is also obtained by accumulating the discharge processes. If 6 complete discharges occur in the past 14 days, then Na = 6.

[0187] Basis for setting the weights of 0.7 and 0.3: In this formula, the pre-factor 0.7 multiplied by Da mainly represents the influence of time factors on the failure rate, and the post-factor 0.3 multiplied by mainly measures the relative intensity of the charge-discharge behavior within the same time. The sum of the two is then multiplied by λ to obtain the exponential term. The weights 0.7 and 0.3 need to perform a sensitivity analysis on "time lapse" and "charge-discharge ratio" after sampling a large number of failure cases. It is determined that the influence of time lapse on the failure ratio is greater, so it is assigned a value of 0.7, and the charge-discharge ratio is assigned a value of 0.3 as a supplement. For different types or models of equipment, regression fitting can also be performed again to obtain slightly different weight values. In the example, 0.7 and 0.3 are still maintained to be compatible with the existing data system.

[0188] Calculation process:

[0189] Take a practical calculation example:

[0190] λ = 0.015, Da = 14, Ca = 7, Na = 6;

[0191] First, calculate the product of 0.7 and Da:

[0192] 0.7 × 14 = 9.8;

[0193] Then, calculate 0.3 and

[0194]

[0195] 0.3 × 1.167 ≈ 0.350;

[0196] Add the two results together:

[0197] 9.8 + 0.350 = 10.15;

[0198] Multiply by λ:

[0199]

[0200] Substitute into the exponent:

[0201] e -0.152 ≈ 0.859;

[0202] F = 1 - 0.859 = 0.141;

[0203] This result indicates that the current failure probability F is approximately 0.141, meaning that the likelihood of the system failing under the current monitoring conditions is about 14.1%. If subsequent detections show an increase or a significant lengthening of the time interval, the calculated F will increase accordingly. When this value exceeds 0.5 or higher, it is often necessary to include it in the high - risk list for timely follow - up and handling.

[0204] According to the previously obtained failure probability, verify item by item whether the necessary monitoring and inspections have been completed during the current operation and maintenance period for the impedance adjustment decision and maintenance scheduling plan of the battery. First, extract all timestamps within the past week from the full usage records and check separately whether the current, voltage, and temperature data exceed the pre-established current range of 0A to 5A, voltage range of 0V to 24V, and temperature range of 0°C to 90°C. When any monitoring record exceeds these ranges, a data entry that requires additional investigation is formed. Subsequently, compare all the abnormal entries within this week with the recent trend curve of the failure probability. If the number of abnormal entries increases significantly on the same day and the failure probability shows a large increase, it is recorded as an attention event, and more intensive observation time points are added to the maintenance list, with additional checks at a frequency of every half day or one day. If no obvious out-of-bounds records are found during the comparison process and the failure probability is relatively low, the original maintenance replacement strategy is retained and this period of time is directly regarded as a normal observation cycle. After all the final records have been checked, a failure prediction analysis result can be sorted out and the next work plan can be registered.

Claims

1. An intelligent overload protection and fault prediction system applicable to a mobile power supply, characterized in that, The system includes: A real-time impedance monitoring module that continuously monitors the internal impedance of the battery during the use of the mobile power supply, generates a real-time impedance analysis result; adjusts the charging current according to the real-time impedance analysis result to match the real-time state of the battery, and generates an impedance adjustment decision; A voltage and current anomaly detection module that continuously monitors the voltage and current data of the battery, compares with a preset normal operating range, and generates a voltage and current monitoring result; based on the voltage and current monitoring result, determines whether abnormal fluctuations in voltage or current are detected, and if detected, immediately issues an anomaly warning and generates an abnormal state alarm; A battery maintenance scheduling module that calculates the charge and discharge efficiency of the battery according to the abnormal state alarm and the charge and discharge cycle of the battery, generates a battery health assessment result; obtains a maintenance and replacement schedule for the battery according to the battery health assessment result, and generates a maintenance scheduling plan; A fault prediction and alarm module that combines the impedance adjustment decision and the maintenance scheduling plan, and uses a statistical model to predict and estimate potential battery faults, generating a fault prediction analysis result.

2. The intelligent overload protection and fault prediction system applicable to a mobile power supply according to claim 1, characterized in that, The steps for obtaining the real-time impedance analysis result are as follows: Continuously collect the internal impedance changes of the battery in various usage states, record the impedance values of each collection, and obtain the original battery impedance data; Based on the original battery impedance data, apply a moving window average to calculate the average impedance value in each window, identify and remove data points that exceed twice the average impedance value, and generate a real-time impedance analysis result.

3. The intelligent overload protection and fault prediction system applicable to a mobile power supply according to claim 1, wherein The steps for obtaining the impedance adjustment decision are as follows: Based on the real-time impedance analysis result, calculate the charging current, and the calculation formula is: Among them, I new represents the adjusted charging current, I base represents the initial charging current, α represents the adjustment sensitivity parameter, R current represents the current impedance value, represents the rate of change of impedance, T cell represents the battery temperature, β and γ are adjustment parameters to reflect the influence of the rate of change of impedance and temperature on the charging current, R base represents the reference impedance value; Based on the charging current, adjust the charging strategy in coordination with the state of the battery to form and implement an impedance adjustment decision.

4. The intelligent overload protection and fault prediction system applicable to a mobile power supply according to claim 1, characterized in that, The steps for obtaining the voltage and current monitoring result are as follows: Real-time record the voltage and current data of the mobile power supply in various working states to form an original dataset of battery voltage and current; Based on the original dataset of battery voltage and current, compare each data point with the preset normal operating voltage and current range, identify and mark all abnormal points that exceed the normal operating voltage and current range, and form a voltage and current monitoring result.

5. The intelligent overload protection and fault prediction system applicable to a mobile power supply according to claim 1, characterized in that, The steps for obtaining the abnormal state alarm are as follows: Receive the voltage and current monitoring result, analyze the voltage and current data at all time points, extract the measured value at each time point, and calculate the voltage change amount, current change amount, battery temperature, battery load status, remaining battery capacity ratio, and ambient temperature at each time point. After sorting, establish a voltage and current fluctuation dataset; Based on the voltage and current fluctuation dataset, calculate the abnormal fluctuation determination value, and the calculation formula is: Among them, A is the abnormal fluctuation determination value, Q i is the voltage value measured at the i-th time, J j is the current value measured at the j-th time, is the average value of the current measurement values, m and n are the number of voltage and current measurements respectively, T cell is the battery temperature, T e is the ambient temperature, L is the current load state of the battery, C r is the remaining capacity ratio of the battery, Q max is the maximum voltage value within the sampling period, Q min is the minimum voltage value within the sampling period, Q i-1 is the voltage value measured at the (i - 1)-th time; Based on the abnormal fluctuation determination value, detect whether it exceeds the set threshold. If it exceeds, trigger an abnormal state alarm.

6. The intelligent overload protection and fault prediction system applicable to a mobile power supply according to claim 1, characterized in that, The steps for obtaining the battery health assessment result are as follows: Based on the abnormal state alarm and the charge and discharge cycle data of the battery, sort out the charge amount, discharge amount, cycle duration, and the number of associated abnormal state alarms in each cycle to obtain a battery charge and discharge cycle dataset; Based on the battery charge and discharge cycle dataset, calculate the battery charge and discharge efficiency, and the calculation formula is: Among them, E is the battery charge-discharge efficiency, D k is the discharge amount in the k-th cycle, C k is the charge amount in the k-th cycle, p is the number of observation cycles, N a is the number of abnormal alarms in the k-th cycle, T avg is the average charge-discharge cycle duration, T max is the longest single charge-discharge cycle duration during the observation period; Evaluate the health status of the battery based on the charge and discharge efficiency of the battery, and generate a battery health assessment result.

7. The intelligent overload protection and fault prediction system applicable to a mobile power supply according to claim 1, characterized in that, The steps for obtaining the maintenance scheduling plan are as follows: Based on the battery health assessment result, calculate the next maintenance replacement time, and the calculation formula is: Among them, T next is the replacement time for the next maintenance, H max is the designed battery life, H current is the remaining life evaluated currently, C initial is the initial capacity of the battery, C current is the current capacity, E initial is the initial charging efficiency of the battery, E current is the current charging efficiency, N cycles is the cumulative charge-discharge cycle number of the battery; Based on the next maintenance replacement time, combine the maintenance cycle and the replacement strategy to formulate a maintenance and replacement schedule, and generate a maintenance scheduling plan.

8. The intelligent overload protection and fault prediction system applicable to a mobile power supply according to claim 1, characterized in that, The steps for obtaining the fault prediction analysis result are as follows: Based on the impedance adjustment decision and the maintenance scheduling plan, estimate the predicted fault probability of the battery, and the calculation formula is: Where F represents the fault probability, λ is the baseline failure rate fitted according to the failure data, Da is the number of days since the last maintenance, Ca is the number of charge cycles since the last maintenance, and Na is the number of discharge cycles since the last maintenance; Based on the fault probability, conduct a fault prediction analysis to guide the maintenance and replacement strategy, and obtain a fault prediction analysis result.

Images