A battery cell voltage monitoring method and system

Abstract

The present application relates to the technical field of battery management, and more particularly to a battery pack cell voltage monitoring method and system, the monitoring method comprising: collecting the voltage of each cell in the battery pack during charging; calculating the rising abnormal score and the falling abnormal score of the battery pack at any time during charging, and performing abnormal monitoring on the battery pack cell voltage according to the size of the rising abnormal score and the falling abnormal score, wherein the rising abnormal score of the battery pack at any time reflects the difference between the rising value of each cell voltage and the average of the rising values of the remaining cell voltages. By introducing two independent indicators, i.e., the voltage rising abnormal score and the voltage falling abnormal score, the rising amplitude abnormality and the falling behavior abnormality of the cell voltage are quantitatively analyzed, which not only can identify the abnormal points deviating from the voltage trend during charging, but also can effectively capture serious risk signals such as sudden cell voltage drop, thereby improving the accuracy of abnormal monitoring.

Description

Technical Field

[0001] This invention relates to the field of battery management technology. More specifically, this invention relates to a method and system for monitoring the voltage of battery pack cells. Background Technology

[0002] A battery pack is a power system composed of multiple battery cells connected in series, parallel, or a hybrid configuration. It is commonly used in electric vehicles, energy storage power stations, and portable electronic devices as their power source. Connecting batteries in series increases voltage, while parallel connections increase capacity, thus meeting different power and energy demands. The performance of a battery pack depends not only on the quality of individual battery cells but also on the overall structural design and management strategy. To ensure the safety, reliability, and lifespan of the battery pack, a Battery Management System (BMS) is typically installed. The BMS enables real-time monitoring of parameters such as voltage, temperature, and current for each battery cell and provides multiple protection functions, including overvoltage, undervoltage, overtemperature, and short-circuit protection. Furthermore, the BMS supports state estimation (such as State of Charge (SOC) and State of Health (SOH), thermal management control, and equalization management, effectively preventing thermal runaway, extending battery pack lifespan, and improving the system's intelligence and fault response capabilities.

[0003] When monitoring the voltage of a battery pack, since the pack contains multiple cells and the voltage of each cell can be monitored, current methods for detecting voltage anomalies during charging mostly rely on fixed thresholds. These methods struggle to address differences in cell characteristics and the non-linear trends in voltage changes during charging. Furthermore, temporary voltage drops during passive equalization are easily misinterpreted as anomalies in existing systems, leading to a high false alarm rate and impacting the reliability and accuracy of battery pack voltage monitoring. Summary of the Invention

[0004] This invention provides a method and system for monitoring cell voltage in battery packs, aiming to solve the problem that current technologies mostly rely on fixed threshold judgment methods for monitoring cell voltage anomalies during battery pack charging, which are insufficient to handle differences in cell characteristics and the nonlinear trends of voltage changes during charging. Furthermore, the temporary drop in cell voltage during passive equalization is easily misidentified as an anomaly in existing systems, leading to a high false alarm rate.

[0005] In a first aspect, the present invention provides a method for monitoring the cell voltage of a battery pack. The method includes: collecting the voltage of each cell in the battery pack during charging; calculating the rise anomaly score and fall anomaly score of the battery pack at any given moment during charging; and monitoring the cell voltage anomalies based on the magnitude of the rise and fall anomaly scores. The rise anomaly score at any given moment reflects the difference between the voltage rise value of each cell and the average voltage rise value of the remaining cells at that moment. The fall anomaly score is related to the state of the cell at that moment. If the cell is in a passive equilibrium state, the fall anomaly score of the battery pack is positively correlated with the number of consecutive voltage drop moments exceeding the target drop stage and the voltage drop value at each voltage drop moment. Conversely, the fall anomaly score of the battery pack is positively correlated with the voltage drop value at that moment. By constructing a cell voltage monitoring method based on a combination of trend consistency analysis and anomaly score evaluation, accurate identification and graded early warning of abnormal voltage fluctuations in the battery pack during charging are achieved.

[0006] Furthermore, the system monitors battery cell voltage anomalies, including: tiered early warning based on the magnitude of rising and falling anomaly scores, with tiers divided into Level 1 and Level 2 warnings. Tiered warning thresholds are set based on the distribution characteristics of the voltage anomaly scores to achieve Level 1 or Level 2 warning responses to cell voltage status. This mechanism enables the system to not only respond quickly when an anomaly first appears, but also to issue higher-level alarms (Level 2 warning is higher than Level 1 warning) when multiple anomaly indicators are simultaneously met, thus improving sensitivity to potential faults.

[0007] Furthermore, the graded early warning includes: in response to the battery pack's abnormal score at the current moment being greater than the rising threshold, or the abnormal score at the current moment being greater than the falling threshold, if it is determined that the current battery pack is abnormal, then a first-level early warning is issued.

[0008] Furthermore, the graded early warning includes: in response to the battery pack's abnormal score at the current moment being greater than the rising threshold and its abnormal score at the current moment being greater than the falling threshold during the charging process, if it is determined that the current battery pack is abnormal, a second-level early warning is issued, and the empirical value of the rising threshold is 0.8.

[0009] Furthermore, the method for calculating the voltage drop anomaly score of the battery pack at any given time includes: if the battery cell is not in a passive balancing state, calculating the sum of the voltage drop values ​​of all cells in the battery pack at that time, normalizing the sum of voltage drop values, and using the normalized value as the voltage drop anomaly score of the battery pack at that time. By combining the cell's current state (e.g., whether it is in the balancing phase) and historical behavior, the actual magnitude of the cell voltage drop is quantified, reflecting the severity of the drop; a larger value indicates a more significant voltage drop and a higher potential risk.

[0010] Furthermore, the method for calculating the battery pack's voltage drop anomaly score at any given moment includes: if the cell is in a passive equalization state, calculating the product of the number of consecutive voltage drop moments exceeding the target drop stage for that cell before that moment and the voltage drop value at each voltage drop moment, and normalizing the sum of the products of all cell voltages, using the normalized value as the battery pack's voltage drop anomaly score at that moment. This process calculates whether the voltage drop is abnormal, distinguishing between voltage drops caused by normal equalization and genuine anomalies. It ensures that both the magnitude of the voltage drop and the probability of an anomaly are considered, avoiding false alarms due to large voltage drops or missed alarms due to low probability.

[0011] Furthermore, the calculation method for the battery pack's voltage rise anomaly score at any given time includes: calculating the average of the voltage rise values ​​of each cell and the remaining cells at that time; normalizing the sum of all means; and using the normalized value as the battery pack's voltage rise anomaly score. By comparing the voltage rise amplitude of a single cell with the average rise amplitude of other cells, it is possible to effectively identify which cells' voltage rise behavior significantly deviates from the overall trend. The greater the difference, the more abnormal the cell's voltage change is, potentially indicating an internal fault or connection problem.

[0012] Furthermore, the calculation method for the abnormal voltage drop score of each cell at this moment includes: summing the voltage drops of all cells at this moment compared to the previous moment, and using this sum as the abnormal voltage drop score for each cell at this moment. Quantifying the actual magnitude of the voltage drop of the cells can reflect the severity of the drop; the larger the value, the more significant the voltage drop and the higher the potential risk.

[0013] Furthermore, the voltage rise at any given moment is the difference between the voltage at that moment and the voltage at the previous moment.

[0014] In a second aspect, the present invention also provides a battery pack cell voltage monitoring system, including a processor and a memory, wherein the memory stores a computer program, and the processor executes the computer program to implement the battery pack cell voltage monitoring method described in any of the above claims.

[0015] Beneficial effects:

[0016] (i) By introducing two independent indicators, “voltage rise anomaly score” and “voltage drop anomaly score”, the abnormal rise and fall behavior of the cell voltage can be quantitatively analyzed. This can not only identify abnormal points of voltage deviation during charging, but also effectively capture serious risk signals such as sudden drop in cell voltage, thereby improving the accuracy of anomaly monitoring.

[0017] (ii) During the charging process, the short-term drop in cell voltage may be caused by passive equalization, which is a normal phenomenon. This invention calculates the probability that the drop is abnormal by statistically analyzing historical data and establishing a target drop stage model, and combining the differences between the cell and the minimum voltage value and the equalization threshold. This effectively distinguishes between abnormalities caused by equalization and abnormalities caused by non-equalization, significantly reduces the misjudgment rate, and improves the reliability and accuracy of the battery pack voltage monitoring process. Attached Figure Description

[0018] Figure 1 This is a schematic diagram illustrating a process for monitoring the cell voltage of a battery pack according to an embodiment of the present invention. Detailed Implementation

[0019] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0020] like Figure 1 As shown, S101: Collects the voltage of each cell in the battery pack during the charging process.

[0021] In one embodiment, the battery pack consists of multiple cells, and the voltage of each cell can be monitored. Therefore, the voltage of all cells can be monitored. Specifically, each cell is equipped with an independent high-precision voltage sensor (such as a 12-bit or 16-bit ADC) to ensure minimal voltage measurement error. The sensor is connected to the BMS via a multiplexer or parallel acquisition circuit, reducing hardware complexity and monitoring the voltage of each cell.

[0022] It's important to note that during battery pack charging, we expect the voltage of each cell to change smoothly and in a coordinated manner according to the anticipated trend. If the voltage change trend of some cells deviates from the overall trend, then the voltage of those cells may be an outlier. Transforming this "trend consistency" into a quantifiable indicator allows for effective monitoring of battery pack voltage. Specifically, during charging, we expect the voltage of all normal cells to rise. If the voltage of individual cells drops, or the rise is significantly less than that of other cells, an anomaly may exist. Therefore, we need to obtain the rise anomaly score and fall anomaly score of the cell voltage at any given moment in the battery pack, and monitor the cell voltage for anomalies based on the magnitude of these scores.

[0023] Step S102: Calculate the abnormal score of the battery pack at any given time.

[0024] In one embodiment, the abnormal voltage rise score of the battery pack at any given time reflects the difference between the voltage rise of each cell at that time and the average voltage rise of the remaining cells. The voltage rise at any given time is the difference between the voltage at that time and the voltage at the previous time. That is, for any cell, the difference between its voltage rise and the average voltage rise of the remaining cells is calculated (the difference is the absolute value of the difference between its voltage rise and the average voltage rise of the remaining cells). A larger difference indicates that the voltage rise of that cell deviates significantly from the average voltage rise of the remaining cells, suggesting a greater likelihood of an abnormal voltage rise in that cell. Then, the mean of the sum of the voltage differences of all cells is normalized, and the normalized value is used as the abnormal voltage rise score of the battery pack at that time. The normalization process uses the sigmoid function.

[0025] For example, a battery pack contains three cells (cells A, B, and C). At a certain moment, their voltage rise values ​​are: cell A: 2 mV, cell B: 2 mV, and cell C: 10 mV, respectively. Then the difference in voltage rise of cell A is... 4. The difference in cell B is as follows: The difference in cell C is The abnormal rise score of the battery pack at that moment is then represented by the sigmoid function. .

[0026] S103: Calculate the abnormal score of the battery pack at any given time.

[0027] In one embodiment, the abnormal voltage drop score is related to the state of the cell at that moment. If the cell is in a passive equilibrium state, the abnormal voltage drop score of the battery pack is positively correlated with the number of consecutive voltage drop moments of the cell that exceed the target voltage drop stage before that moment and the voltage drop value of each voltage drop moment. Conversely, the abnormal voltage drop score of the battery pack is positively correlated with the voltage drop value at that moment.

[0028] In one embodiment, if the cells are not in an balanced state at a given moment, a formula is provided for calculating the abnormal voltage drop score of each cell at any given moment. The formula is as follows: ; Indicates the first The abnormal score of the battery pack at any given time. Indicates the first Time of the first The voltage change of the first cell (the voltage change is the first...) Time of the first The voltage of the first cell is the same as that of the second cell. Time of the first (the difference in voltage between individual cells) This indicates the number of cells in the battery pack. This represents the maximum value function.

[0029] It should be noted that, through The voltage drops of all cells in the battery are summed to detect serious, direct anomalies. During charging, any voltage drop in a cell is a strong abnormal signal, potentially indicating an internal problem (such as an internal short circuit) or a connection failure to the BMS. A larger value indicates more cells with voltage drops, and a larger voltage drop value indicates a more severe anomaly. The higher the abnormal score of the cell voltage drop at any given time, the higher the probability of an abnormality in the battery pack.

[0030] In one embodiment, BMS balancing is a key function of the battery management system, designed to address the inherent inconsistencies between individual cells in a battery pack. The aim is to prevent overcharging and over-discharging: In a series-connected battery pack, if cells differ in capacity, during charging, the smaller cell will reach full charge first, leading to overcharging if charging continues. Conversely, during discharging, the smaller cell will deplete its charge first, leading to over-discharging if discharging continues. Both overcharging and over-discharging severely damage cells, shorten their lifespan, and can even cause safety issues (such as thermal runaway). Many battery packs on the market currently employ passive balancing because the circuit design is simple, low-cost, and easy to implement. This method dissipates excess energy as heat by connecting a resistor in parallel across the higher-voltage cell until its voltage drops to a level similar to other cells. During passive balancing, the voltage of the balanced cell may briefly drop, which is normal as energy is being dissipated. Therefore, during the charging process of the battery pack, the voltage drop of the high-voltage cells under passive balancing is normal. Thus, it is necessary to consider the probability that the voltage drop of each cell is abnormal.

[0031] In one embodiment, a formula is provided to calculate the probability that the voltage drop of each cell at any given time is abnormal. The formula is as follows: In the formula, Indicates the first The cell voltage at the first The probability that a voltage drop at any given moment is abnormal. No. The first cell in the first Voltage at time, This indicates that all cells in the battery pack are in the [number]th [period]. The minimum voltage value at time [time]. To balance the threshold, Indicates the first The number of consecutive voltage drops prior to a given moment. Indicates the target descent phase. This represents the maximum value function.

[0032] In the formula, When it is greater than 0, it means that the first Voltage of each cell Minimum of all cell voltages Greater than the equilibrium threshold The equalization threshold The empirical value is 30mv, indicating that a passive balancing system needs to intervene at this point. The voltage of each cell will drop within the target time period. During this period, the voltage of the first cell will decrease. The voltage drop of the individual cells within the target time period is normal. However, during the equalization process, the voltage of the first cell... The voltage of a single cell will drop in the short term; this drop caused by balancing is usually short-lived and not long-term. If the duration of the continuous drop exceeds the target drop period, it indicates that the voltage drop is abnormal. The cell voltage at the first The higher the degree of abnormality at any given moment, the better. In other words, if the cell voltage continues to drop within the target drop phase, it indicates that the battery cell is in an equilibrium state, and the short-term drop in cell voltage is normal, so the probability of the cell voltage drop being abnormal is 0. If the cell voltage continues to drop beyond the target drop phase, there may be signs of abnormality. The probability of the cell voltage drop being abnormal is positively correlated with the time exceeding the target drop phase; the longer the time exceeds the target drop phase, the higher the probability of abnormality.

[0033] In one embodiment, determining the target voltage decline phase includes: acquiring historical voltage data of each cell in multiple battery packs during charging; assembling the voltage data of each cell into a sequence to obtain a historical voltage sequence for each cell; clustering all historical voltage sequences to obtain two clusters; selecting the cluster containing the most sequences as the cluster under normal conditions; counting the moments of continuous decline in the passive equilibrium state among all sequences in the normal state cluster; using these moments of continuous decline in the passive equilibrium state as sample segments to obtain sample segments for all sequences in the normal state cluster; then counting the proportion of sample segments of different lengths to obtain the proportion of all sample segments of different lengths; accumulating the proportions of each sample segment (accumulating sequentially from low to high according to the length of the sample segment); comparing the accumulated value with a threshold; and determining the time length corresponding to the sample segment whose accumulated value is greater than the threshold as the target voltage decline phase, where the empirical value of the threshold is 80%. By extracting the normal state decline period through historical data clustering and using the sample segment cumulative proportion method to determine the target voltage decline phase, the subjectivity of manually setting a fixed time threshold is avoided, making the anomaly judgment more adaptive and universal. The empirical value of setting the threshold to 80% can cover most samples with normal descent behavior, reducing misjudgments caused by excessive convergence.

[0034] For example, under normal conditions, a cluster contains sequences A, B, C, D, E, and F, where the sample segment of sequence A is... The sample segment of sequence B is The sample segment of sequence C is The sample segment of sequence D is The sample segment of sequence E is The sample segment of sequence F is ,like and The time lengths are equal. , and The time lengths are equal. The time lengths of all of them are not equal, and and The time length is less than , and The length of time, , and The time length is less than The length of the time segment; at this point, the proportion of sample segments of different lengths is calculated, if and The time length accounts for 2 / 6. , and The time length accounts for 3 / 6. The time length accounts for 1 / 6, at which point... and The proportion of time length and , and The percentages of time duration are accumulated. If the sum is greater than 80%, then... , and The time length is used as the target descent phase.

[0035] In another embodiment, if the cells are in an equal state at that moment, based on the above description, another formula is provided for calculating the voltage drop anomaly score at any given moment, specifically: . Indicates the first The abnormal score of the battery pack at any given time. Indicates the first The cell voltage at the first The probability that a voltage drop at any given moment is abnormal. The score indicates an abnormal drop in cell voltage. This indicates the number of cells in the battery pack. Let represent the hyperbolic tangent function. Where, Indicates the first The cell voltage at the first The voltage drop at time and the first The cell voltage at the first The voltage drop at time point is the product of the probability that it is abnormal; the larger the drop, the higher the probability that it is abnormal at time point. The greater the voltage anomaly at time 1, the more likely it is that the voltage anomaly at time 2 is to occur. The cell voltage at the first The higher the probability that a voltage drop at a given moment is abnormal, the more likely it is to be. The higher the voltage drop score of the battery pack at any given moment, the better. This formula takes into account the normal phenomenon of a brief drop in cell voltage during passive balancing, thus solving the problem of existing systems easily misidentifying this as an anomaly, leading to a high false alarm rate.

[0036] S104: Monitor the voltage of the battery pack cells for abnormalities.

[0037] In one embodiment, the rising and falling anomaly scores of the battery pack at any given moment during charging are calculated, and the cell voltage of the battery pack is monitored for anomalies based on these scores. Specifically, a graded early warning system is implemented based on the satisfaction of certain conditions regarding the magnitude of the rising and falling anomaly scores, with the early warning system divided into Level 1 and Level 2 warnings. For example, if the rising anomaly score of the battery pack at the current moment is greater than a rising threshold, or if the rising anomaly score at the current moment is greater than a falling threshold, then if either condition is met, the current battery pack is determined to be abnormal, and a Level 1 warning is issued, where the empirical value of the rising threshold is 0.8. If the rising anomaly score of the battery pack at the current moment is greater than both the rising and falling thresholds, i.e., both conditions are met, then the current battery pack is determined to be abnormal, and a Level 2 warning is issued, where the empirical value of the falling threshold is 0.8. This satisfies the requirements for monitoring the cell voltage of the battery pack. It should be noted that in other embodiments, the rising and falling thresholds can be adjusted according to the implementation situation.

[0038] The present invention also provides a battery pack cell voltage monitoring system. The system includes a processor and a memory, the memory storing computer program instructions. When the processor executes the computer program instructions, it implements a battery pack cell voltage monitoring method according to the first aspect of the present invention.

[0039] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and therefore will not be described in detail here.

[0040] In this invention, the aforementioned memory can be any tangible medium containing or storing a program that can be used or combined with an instruction execution system, apparatus, or device. For example, a computer-readable storage medium can be any suitable magnetic or magneto-optical storage medium, such as Resistive Random Access Memory (RRAM), Dynamic Random Access Memory (DRAM), Static Random Access Memory (SRAM), Enhanced Dynamic Random Access Memory (EDRAM), High-Bandwidth Memory (HBM), Hybrid Memory Cube (HMC), etc., or any other medium that can be used to store desired information and can be accessed by an application, module, or both. Any such computer storage medium can be part of a device or accessible to or connected to a device. Any application or module described in this invention can be implemented using computer-readable / executable instructions stored or otherwise maintained on such a computer-readable medium.

[0041] The embodiments described above are merely examples of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention.

Claims

1. A battery cell voltage monitoring method, characterized by, The monitoring method comprises: collecting voltages of each battery cell in the battery pack during charging; calculating an abnormal score of the battery pack at any time during charging and an abnormal score of the battery pack at any time during charging, and performing abnormal monitoring on the battery cell voltage of the battery pack according to the abnormal score and the abnormal score, wherein the abnormal score of the battery pack at any time reflects the difference between the voltage rising value of each battery cell at the time and the average of the voltage rising values of the remaining battery cells; the abnormal score is related to the state of the battery cell at the time, if the battery cell is in a passive balancing state, the abnormal score of the battery pack is positively correlated with the number of times that the voltage of the battery cell exceeds the target falling stage at the time before the continuous voltage falling time and the voltage falling value at each voltage falling time, otherwise, the abnormal score of the battery pack is positively correlated with the voltage falling value at the time.

2. The battery cell voltage monitoring method of claim 1, wherein, The abnormal monitoring on the battery cell voltage of the battery pack comprises: grading early warning according to the abnormal score and the abnormal score, wherein the grading early warning is divided into first-level early warning and second-level early warning.

3. The battery cell voltage monitoring method of claim 2, wherein, The grading early warning comprises: in response to the abnormal score of the battery pack at the current time during charging being greater than the rising threshold, or the abnormal score of the battery pack at the current time during charging being greater than the falling threshold, it is determined that the current battery pack is abnormal, and first-level early warning is performed.

4. The battery cell voltage monitoring method of claim 2, wherein, The grading early warning comprises: in response to the abnormal score of the battery pack at the current time during charging being greater than the rising threshold, and the abnormal score of the battery pack at the current time during charging being greater than the falling threshold, it is determined that the current battery pack is abnormal, and second-level early warning is performed.

5. The battery cell voltage monitoring method of claim 1, wherein, The calculation method of the abnormal score of the battery pack at any time comprises: if the battery cell is not in a passive balancing state, the sum of the voltage falling values of all battery cells in the battery pack at the time is calculated, and the sum of the voltage falling values is normalized, and the normalized value is taken as the abnormal score of the battery pack at the time.

6. The battery cell voltage monitoring method of claim 1, wherein, The calculation method of the abnormal score of the battery pack at any time comprises: if the battery cell is in a passive balancing state, the product of the number of times that the voltage of the battery cell exceeds the target falling stage at the time before the continuous voltage falling time and the voltage falling value at each voltage falling time is calculated, and the sum of the products of all battery cell voltages is normalized, and the normalized value is taken as the abnormal score of the battery pack at the time.

7. The battery cell voltage monitoring method of claim 1, wherein, The calculation method of the abnormal score of the battery pack at any time comprises: the difference between the voltage rising value of each battery cell at the time and the average of the voltage rising values of the remaining battery cells is calculated, and the average of the sum of all differences is normalized, and the normalized value is taken as the abnormal score of the battery pack.

8. The battery cell voltage monitoring method of claim 1, wherein, The voltage rising value at any time is the difference between the voltage at the time and the voltage at the previous time.

9. The battery cell voltage monitoring method of claim 3, wherein, The experience value of the rising threshold is 0.

8.

10. A battery cell voltage monitoring system comprising a processor and a memory, wherein, The memory stores a computer program, and the processor executes the computer program to realize the battery cell voltage monitoring method according to any one of claims 1-9. The memory stores a computer program, and the processor executes the computer program to realize the battery cell voltage monitoring method according to any one of claims 1-9.

Images