A method and apparatus for determining alarm thresholds in a battery management system (BMS).

Abstract

This application provides a method and apparatus for determining alarm thresholds in a battery management system (BMS). By collecting battery operation data and determining user behavior characteristics, normalization and smoothing are performed to obtain standard smoothed user behavior characteristics. These standard smoothed user behavior characteristics and battery operation data are input into a first model to obtain alarm threshold correction parameters and a user behavior intensity score. Based on the user behavior intensity score, the user's type is determined, and the sign of the alarm threshold correction parameters is also determined. A preset alarm threshold is obtained, and the corrected alarm threshold is obtained by multiplying the sum of the numerical value and the signed alarm threshold correction parameters. The historical stable alarm threshold and the corrected alarm threshold are then weighted and fused to obtain a stable alarm threshold, which is then determined as the target alarm threshold. In summary, differentiated alarm thresholds can be dynamically generated based on the driver's actual driving habits, adapting to various scenarios.

Description

Technical Field

[0001] This application relates to the field of battery management systems (BMS) for new energy vehicles, and in particular to a method and apparatus for determining alarm thresholds in a battery management system (BMS). Background Technology

[0002] Currently, existing BMS systems use a uniform static alarm threshold, which does not differentiate between driver behavior differences, making them prone to frequent false alarms and missed alarms. Furthermore, threshold adjustments are highly dependent on the cloud, and the vehicle's local computing power is insufficient to achieve real-time personalized adaptation, making it difficult to adapt to various application scenarios. Summary of the Invention

[0003] This application provides a method and apparatus for determining alarm thresholds in a battery management system (BMS), which to some extent solves the problems of existing BMS using uniform static alarm thresholds, failing to distinguish driver behavior differences, being prone to frequent false alarms and missed alarms, and having threshold adjustments that are highly dependent on the cloud, with insufficient local computing power on the vehicle to achieve real-time personalized adaptation, making it difficult to adapt to various application scenarios.

[0004] According to one aspect of this application, a method for determining alarm thresholds in a battery management system (BMS) is provided. The method includes: collecting battery operating data and determining user behavior characteristics based on the battery operating data; the user behavior characteristics include: acceleration aggression, braking frequency, fast charging usage rate, deep discharge frequency, high temperature exposure, and load fluctuation rate; normalizing and smoothing the user behavior characteristics to obtain corresponding standard smoothed user behavior characteristics; inputting the standard smoothed user behavior characteristics and battery operating data into a first model to obtain alarm threshold correction parameters and user behavior intensity scores; the alarm threshold correction parameters include: differential pressure alarm threshold correction parameters, temperature alarm threshold correction parameters, and charging current correction parameters. The system includes alarm threshold correction parameters and discharge current alarm threshold correction parameters; the first model is a lightweight decision model; based on user behavior intensity scores, the user type is determined, and based on the user type, the sign of the alarm threshold correction parameters is determined; the user types include: aggressive users, normal users, and conservative users; the factory-preset alarm threshold is obtained, and the corrected alarm threshold is obtained by multiplying the sum of the value and the alarm threshold correction parameters with signs by the factory-preset alarm threshold; historical stable alarm thresholds are obtained, and the historical stable alarm thresholds and the corrected alarm thresholds are weighted and fused to obtain the stable alarm threshold, and the stable alarm threshold is determined as the target alarm threshold.

[0005] Furthermore, according to one aspect of the method of this application, the method further includes: performing a security verification process on the stabilized alarm threshold, and determining the stabilized alarm threshold after the verification process as the target alarm threshold; the security verification process includes: hard constraints and rate of change limits.

[0006] Furthermore, according to one aspect of the method of this application, the method further includes: obtaining a factory-preset warning threshold; multiplying the sum of a value and a positive or negative alarm threshold correction parameter by the preset warning threshold to obtain a corrected warning threshold; if the warning threshold is less than the alarm threshold, obtaining a historical stable warning threshold; weighting and fusing the historical stable warning threshold and the corrected warning threshold to obtain a stable warning threshold; and determining the stable warning threshold as the target warning threshold.

[0007] Furthermore, according to one aspect of the method of this application, battery operating data includes: voltage, temperature, charging current, discharging current, and real-time power; collecting battery operating data and determining user behavior characteristics based on the battery operating data, including: determining single acceleration value, number of accelerations, and maximum permissible acceleration value based on discharging current and real-time power, and calculating the ratio of each single acceleration value to the maximum permissible acceleration value within the number of accelerations, and taking the arithmetic mean of all ratios to obtain the acceleration aggression level; determining the total number of brakings per unit cycle based on charging current, and obtaining the braking frequency by dividing the total number of brakings by the unit cycle; and determining the braking frequency based on charging current. The fast charging time and total charging time are determined separately. The fast charging time is divided by the total charging time to obtain the fast charging utilization rate. Based on the charging current and discharging current, the total number of charge-discharge cycles and the number of charge-discharge cycles with a state of charge below 20% are determined. The deep discharge frequency is obtained by dividing the number of charge-discharge cycles by the total number of charge-discharge cycles. Based on temperature, the operating time when the temperature is above 45 degrees Celsius and the total driving operating time are determined. The high temperature exposure is obtained by dividing the operating time by the total driving operating time. Based on real-time power, the power mean is determined, and then the root mean square value of the difference between the real-time power and the power mean is determined to obtain the load fluctuation rate.

[0008] Furthermore, according to one aspect of the method of this application, user behavior features are normalized and smoothed to obtain corresponding standard smoothed user behavior features, including: for any user behavior feature, performing numerical normalization on user behavior features within at least two driving cycles to obtain normalized user behavior features; and performing weighted fusion smoothing on the normalized user behavior features of the current driving cycle and the normalized user behavior features of the previous driving cycle to obtain standard smoothed user behavior features.

[0009] Furthermore, according to one aspect of the method of this application, based on a user behavior intensity score, the user's type is determined, and based on the user's type, the sign of the alarm threshold correction parameter is determined, including: obtaining a first threshold and a second threshold; the first threshold being greater than the second threshold; when the user behavior intensity score is greater than or equal to the first threshold, the user's type is determined to be an aggressive user, and the discharge current alarm threshold correction parameter is determined to be negative, while the differential pressure alarm threshold correction parameter, temperature alarm threshold correction parameter, and charging current upper limit alarm threshold correction parameter are determined to be positive; when the user behavior intensity score is less than the first threshold but greater than or equal to the second threshold, the user's type is determined to be a normal user, and the discharge current alarm threshold correction parameter, differential pressure alarm threshold correction parameter, temperature alarm threshold correction parameter, and charging current upper limit alarm threshold correction parameter are all determined to be positive; when the user behavior intensity score is less than the second threshold, the user's type is determined to be a conservative user, and the discharge current alarm threshold correction parameter, differential pressure alarm threshold correction parameter, temperature alarm threshold correction parameter, and charging current upper limit alarm threshold correction parameter are all determined to be negative.

[0010] Furthermore, according to one aspect of the method of this application, a factory-preset alarm threshold is obtained, and the corrected alarm threshold is obtained by multiplying the sum of a value and an alarm threshold correction parameter with a positive or negative sign by the preset alarm threshold. This includes: obtaining the preset alarm threshold from the local storage area of ​​the BMS; adding the value and the alarm threshold correction parameter with a corresponding positive or negative sign to obtain a correction ratio, and multiplying the correction ratio by the preset alarm threshold to obtain the corrected alarm threshold.

[0011] Furthermore, according to one aspect of the method of this application, a historical stable alarm threshold is obtained, and a weighted fusion of the historical stable alarm threshold and the corrected alarm threshold is performed to obtain a stable alarm threshold, and the stable alarm threshold is determined as the target alarm threshold, including: obtaining the historical stable alarm threshold from the BMS local storage area; obtaining a preset trend coefficient and determining it as the first weight of the corrected alarm threshold, and subtracting the preset trend coefficient from the value to determine the second weight of the historical stable alarm threshold; multiplying the first weight by the corrected alarm threshold, and adding the second weight multiplied by the historical stable alarm threshold to obtain the stable alarm threshold, which is then determined as the target alarm threshold.

[0012] Furthermore, according to one aspect of the method of this application, a safety verification process is performed on the stabilized alarm threshold, and the stabilized alarm threshold after the verification process is determined as the target alarm threshold. This includes: when the safety verification process is a hard constraint, determining whether the stabilized alarm threshold is greater than a preset battery safety limit value, and when the stabilized alarm threshold is greater than the preset battery safety limit value, updating the stabilized alarm threshold to the battery safety limit value; when the safety verification process is a rate of change limit, determining whether the change range of the current stabilized alarm threshold relative to the previous stabilized alarm threshold is greater than 5%, and when the change range is greater than 5%, adjusting the stabilized alarm threshold to a boundary value allowed by 5%.

[0013] According to another aspect of this application, a battery management system (BMS) alarm threshold determination device is provided. The device includes: a data acquisition unit for acquiring battery operating data and determining user behavior characteristics based on the battery operating data; the user behavior characteristics include: acceleration aggression, braking frequency, fast charging usage rate, deep discharge frequency, high temperature exposure, and load fluctuation rate; a processing unit for normalizing and smoothing the user behavior characteristics to obtain corresponding standard smoothed user behavior characteristics; and a first determination unit for inputting the standard smoothed user behavior characteristics and battery operating data into a first model to obtain alarm threshold correction parameters and user behavior intensity scores; the alarm threshold correction parameters include: differential pressure alarm threshold correction parameters, temperature alarm threshold correction parameters, and charging current correction parameters. The system includes alarm threshold correction parameters and discharge current alarm threshold correction parameters; the first model is a lightweight decision model; the second determination unit is used to determine the user's type based on the user behavior intensity score, and to determine the sign of the alarm threshold correction parameters based on the user's type; the user types include: aggressive user, normal user, and conservative user; the correction unit is used to obtain the factory-preset alarm threshold, and to obtain the corrected alarm threshold by multiplying the sum of the value and the alarm threshold correction parameters with signs by the preset alarm threshold; the stabilization unit is used to obtain the historical stable alarm threshold, to perform weighted fusion of the historical stable alarm threshold and the corrected alarm threshold to obtain the stabilized alarm threshold, and to determine the stabilized alarm threshold as the target alarm threshold.

[0014] This application provides a method and apparatus for determining alarm thresholds in a battery management system (BMS). The application collects battery operating data and determines user behavior characteristics based on this data. These user behavior characteristics include: acceleration aggression, braking frequency, fast charging usage, deep discharge frequency, high-temperature exposure, and load fluctuation. The user behavior characteristics are then normalized and smoothed to obtain corresponding standard smoothed user behavior characteristics. These standard smoothed user behavior characteristics and battery operating data are input into a first model to obtain alarm threshold correction parameters and a user behavior intensity score. The alarm threshold correction parameters include: differential pressure alarm threshold correction parameters, temperature alarm threshold correction parameters, charging current upper limit alarm threshold correction parameters, and discharge current alarm threshold correction parameters. The system employs a lightweight decision-making model to correct user behavior intensity scores. It determines the user's type and, based on that type, the sign of the alarm threshold correction parameter. User types include aggressive, average, and conservative users. The system obtains the factory-preset alarm threshold and multiplies it by the sum of the numerical value and the signed alarm threshold correction parameter to obtain the corrected alarm threshold. It also obtains historical stable alarm thresholds and weightedly fuses these with the corrected threshold to obtain a stable alarm threshold, which is then used as the target alarm threshold. Compared to existing fixed, uniform thresholds and cloud-based offline calculation adjustment schemes, this application allows for local computation on the BMS vehicle-mounted embedded hardware. It leverages a lightweight decision-making model to adapt to the limited computing power of the vehicle, quantitatively distinguishes users with different driving styles based on six-dimensional battery usage behavior characteristics, automatically matches the threshold correction direction and magnitude, and combines historical threshold weighting and smoothing to suppress frequent threshold jumps caused by short-term operating condition fluctuations, thus balancing personalized adaptation with battery operation safety. In summary, the technical solution provided in this application can dynamically generate differentiated alarm thresholds based on the driver's actual driving habits, significantly reducing the probability of false alarms and missed alarms in the battery monitoring process. It does not rely on the network cloud to achieve real-time local calculation and can adapt to a variety of application scenarios.

[0015] It should be understood that both the foregoing general description and the following detailed description are exemplary and intended to provide further illustration of the claimed technology. Attached Figure Description

[0016] The above and other objects, features, and advantages of this application will become more apparent from the more detailed description of the embodiments of this application in conjunction with the accompanying drawings. The accompanying drawings are used to provide a further understanding of the embodiments of this application and form part of the specification. They are used together with the embodiments of this application to explain this application and do not constitute a limitation thereof. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0017] Figure 1A flowchart illustrating a method for determining alarm thresholds in a battery management system (BMS) according to an embodiment of this application; Figure 2 This is a structural block diagram of a battery management system (BMS) alarm threshold determination device provided in an embodiment of this application. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this application more apparent, exemplary embodiments according to this application will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments of this application. It should be understood that this application is not limited to the exemplary embodiments described herein.

[0019] Currently, existing BMS systems use a uniform static alarm threshold, which does not differentiate between driver behavior differences, making them prone to frequent false alarms and missed alarms. Furthermore, threshold adjustments are highly dependent on the cloud, and the vehicle's local computing power is insufficient to achieve real-time personalized adaptation, making it difficult to adapt to various application scenarios.

[0020] Therefore, to address the aforementioned problems, this application provides a method for determining alarm thresholds in a battery management system (BMS). Compared to existing fixed, uniform thresholds and cloud-based offline calculation and adjustment schemes, this application allows all calculations to be completed locally on the embedded hardware of the BMS in the vehicle. It utilizes a lightweight decision model adapted to the limited computing power of the vehicle, quantitatively distinguishes users with different driving styles based on six-dimensional battery usage behavior characteristics, automatically matches the direction and magnitude of threshold correction, and combines historical threshold weighted smoothing to suppress frequent threshold jumps caused by short-term operating condition fluctuations, thus balancing personalized adaptation with battery operation safety. In summary, the technical solution provided by this application can dynamically generate differentiated alarm thresholds based on the driver's actual driving habits, significantly reducing the probability of false alarms and missed alarms during battery monitoring. It does not rely on cloud-based real-time local calculations and can adapt to various application scenarios.

[0021] This application provides a method for determining alarm thresholds in a battery management system (BMS). Please refer to... Figure 1 , Figure 1 This is a flowchart illustrating a method for determining alarm thresholds in a battery management system (BMS) according to an embodiment of this application. Figure 1 As shown, the method includes: In step S101, battery operation data is collected, and user behavior characteristics are determined based on the battery operation data. User behavior characteristics include: acceleration aggression, braking frequency, fast charging usage rate, deep discharge frequency, high temperature exposure, and load fluctuation rate. In this application, battery operating data can be understood as raw monitoring data collected in real time by the BMS to characterize the real-time operating state of the power battery, which may include five basic parameters: voltage, temperature, charging current, discharging current, and real-time power. Specific battery operating data will be detailed later.

[0022] In this application, user behavior characteristics can be understood as quantitative indicators derived from raw battery operating data, capable of quantifying the driver's charging and driving habits, and used to reflect the differences in battery usage load among different users. The user behavior characteristics in this application include: acceleration aggression, braking frequency, fast charging usage rate, deep discharge frequency, high-temperature exposure, and load fluctuation rate. Specifically, these six characteristics comprehensively characterize user battery usage behavior from six dimensions: acceleration operation, braking frequency, charging method, low-charge usage, high-temperature conditions, and power fluctuation. The specific determinations will be elaborated upon later.

[0023] In step S102, the user behavior features are normalized and smoothed to obtain the corresponding standard smoothed user behavior features. In this application, normalization smoothing can be understood as first mapping each type of feature to the same numerical range to eliminate dimensional differences, and then filtering the fluctuations caused by a single driving condition through a multi-driving-cycle sliding window weighted operation.

[0024] In this application, the standard smoothed user behavior features can be understood as standardized feature data that has been uniformly formatted and stable after numerical unification and temporal smoothing and noise reduction, and can be directly input into the model for calculation.

[0025] In step S103, standard smoothed user behavior characteristics and battery operation data are input into the first model to obtain alarm threshold correction parameters and user behavior intensity scores; the alarm threshold correction parameters include: differential pressure alarm threshold correction parameters, temperature alarm threshold correction parameters, charging current upper limit alarm threshold correction parameters, and discharge current alarm threshold correction parameters; the first model is a lightweight decision model; In this application, the first model can be understood as a small decision model that is adapted to BMS embedded low computing power hardware, has a simple structure, and fast inference speed. It does not require cloud computing power support and can complete real-time calculations locally on the vehicle.

[0026] In this application, alarm threshold correction parameters can be understood as numerical coefficients output by the model, used to characterize the magnitude of adjustment required for the corresponding alarm threshold. The alarm threshold correction parameters in this application include: differential pressure alarm threshold correction parameters, temperature alarm threshold correction parameters, charging current upper limit alarm threshold correction parameters, and discharge current alarm threshold correction parameters. Specifically, these four types of correction parameters correspond to the four core monitoring alarm items: battery differential pressure, cell temperature, maximum charging current, and maximum discharging current.

[0027] In this application, the user behavior intensity score can be understood as a quantitative score calculated by comprehensively considering six dimensions of behavioral characteristics, used to evaluate the aggressiveness of the user's driving and electricity consumption behavior as a whole.

[0028] In step S104, the user's type is determined based on the user behavior intensity score, and the sign of the alarm threshold correction parameter is determined based on the user's type; the user's type includes: aggressive user, normal user, and conservative user; In this application, the user type can be understood as a classification of driver electricity usage style based on a behavioral intensity score range. The user types in this application include: aggressive users, average users, and conservative users. Specifically, high scores correspond to aggressive users, medium scores to average users, and low scores to conservative users.

[0029] In this application, the positive or negative sign of the alarm threshold correction parameter can be understood as an indicator of the direction of threshold adjustment. A positive sign means to relax the corresponding alarm threshold, and a negative sign means to tighten the corresponding alarm threshold.

[0030] In step S105, the factory-preset alarm threshold is obtained, and the corrected alarm threshold is obtained by multiplying the sum of the value one and the alarm threshold correction parameter with positive and negative signs by the preset alarm threshold. In this application, the preset alarm threshold can be understood as the fixed basic alarm threshold set by the manufacturer at the time the battery leaves the factory and stored locally in the BMS.

[0031] In this application, the revised alarm threshold can be understood as a temporary alarm threshold after personalized adjustment based on user behavior style.

[0032] In step S106, the historical stable alarm threshold is obtained, and the historical stable alarm threshold and the corrected alarm threshold are weighted and fused to obtain the stable alarm threshold, and the stable alarm threshold is determined as the target alarm threshold.

[0033] In this application, the historical stable alarm threshold can be understood as the alarm threshold that has been smoothed and stored after the calculation of the previous complete driving cycle.

[0034] In this application, the stabilized alarm threshold can be understood as a stable alarm threshold obtained by integrating the current temporary correction threshold with the historical threshold and suppressing short-term fluctuations.

[0035] Specifically, in this embodiment of the application, the following steps may be included when determining the BMS alarm threshold: The system collects real-time data on the voltage, temperature, charging current, discharging current, and real-time power of the vehicle's power battery; calculates six-dimensional user behavior characteristics based on five types of raw data; performs normalization and multi-cycle smoothing and noise reduction on all features; inputs the standardized features and real-time battery data into a local lightweight decision model, outputting four types of threshold correction coefficients and user behavior intensity scores; classifies user types according to the score range and sets the positive and negative adjustment directions for each correction parameter; calculates personalized temporary thresholds using factory-set basic thresholds combined with signed correction coefficients; retrieves historical stable thresholds from the previous cycle and obtains the target alarm threshold through fixed-weighted fusion.

[0036] In addition to the methods described above, the methods of this application also include: After stabilization, the alarm threshold is subjected to a safety verification process, and the stable alarm threshold that passes the verification process is determined as the target alarm threshold. The safety verification process includes hard constraints and rate of change limits.

[0037] In this application, the safety verification process can be understood as providing a safety fallback control over the stable alarm threshold obtained after weighted fusion, preventing the dynamically adjusted threshold from exceeding the battery's safe operating range or causing frequent alarm jumps due to excessively large single adjustment amplitudes, thus ensuring battery operating safety and alarm logic stability. The safety verification process in this application includes: hard constraints and rate of change limits. Hard constraints are used to limit the threshold from exceeding the extreme safety parameters specified by the battery manufacturer, while rate of change limits are used to control the magnitude of single-round threshold adjustments, avoiding threshold abrupt changes that could interfere with BMS monitoring and judgment.

[0038] Specifically, when determining the target alarm threshold, the following can be included: First, the corrected alarm threshold is calculated using user behavior characteristics and a lightweight model. Then, it is weighted and fused with historical stable alarm thresholds to obtain a stable alarm threshold. Hard constraint verification and rate of change limit verification are performed sequentially. If both verifications pass, the stable alarm threshold is used as the final usable target alarm threshold. If either verification fails, the stable alarm threshold is corrected according to the corresponding verification rules, and then determined as the target alarm threshold after correction.

[0039] Furthermore, this application also provides a specific security verification process, including: When the safety verification process is a hard constraint, determine whether the alarm threshold after stabilization is greater than the preset battery safety limit. If the alarm threshold after stabilization is greater than the preset battery safety limit, update the alarm threshold after stabilization to the battery safety limit. When the security verification process is limited by the rate of change, it is determined whether the change of the current stable alarm threshold relative to the previous stable alarm threshold is greater than 5%. If the change is greater than 5%, the stable alarm threshold is adjusted to the boundary value allowed by 5%.

[0040] In one embodiment of this application, when performing hard constraint verification, the stable alarm thresholds corresponding to differential pressure, temperature, upper limit of charging current, and upper limit of discharging current can be compared one by one with their respective battery safety limit values. If any stable alarm threshold exceeds the corresponding safety limit value, the threshold is directly replaced with the matching battery safety limit value. After completing the verification of all parameters, the next step of verification is performed.

[0041] In another embodiment of this application, when performing the rate of change limit verification, the stable alarm threshold saved in the previous round of calculation can be retrieved, and the ratio of the difference between the current threshold and the previous threshold to the previous threshold can be calculated item by item. If the absolute value of the ratio exceeds five percent, the current stable alarm threshold is corrected according to the boundary value of increasing or decreasing by five percent, so as to ensure that the threshold adjustment range in each round does not exceed the limit range.

[0042] In addition to the methods described above, the methods of this application also include: Obtain the factory-preset warning threshold, and multiply the sum of the value 1 and the alarm threshold correction parameter (with positive or negative sign) by the preset warning threshold to obtain the corrected warning threshold; if the warning threshold is less than the alarm threshold... Obtain historical stable early warning thresholds, weight and fuse the historical stable early warning thresholds and the corrected early warning thresholds to obtain stable early warning thresholds, and determine the stable early warning thresholds as the target early warning thresholds.

[0043] In this application, the preset warning threshold can be understood as a basic warning benchmark value set by the manufacturer at the time of battery shipment and pre-stored in the local storage of the BMS, used to provide early warning of abnormal battery parameters. The warning threshold has a larger safety margin than the alarm threshold. The preset warning threshold in this application is a risk advance warning threshold, and its value is always less than the preset alarm threshold with the same parameters; the warning only provides an indication and does not limit power, while the alarm triggers mandatory battery safety protections such as current limiting and power cut-off.

[0044] In this application, the revised warning threshold can be understood as a temporary warning threshold obtained by using the same positive and negative correction parameters as the alarm threshold to perform personalized adjustment of the factory basic warning threshold.

[0045] In this application, the historical stable warning threshold can be understood as the warning threshold that has been calculated and stored after the previous complete vehicle driving cycle and has undergone smoothing processing.

[0046] In this application, the target early warning threshold can be understood as the final early warning threshold obtained by merging the current temporary modified early warning threshold and the historical stable early warning threshold, and smoothing short-term operating condition fluctuations, which can be used for real-time early warning in BMS.

[0047] Specifically, in the embodiments of this application, determining the target early warning threshold may include: The system reads the factory-preset warning threshold from the BMS local storage, calculates the correction factor by combining it with the alarm threshold correction parameters whose signs have been determined, and multiplies the preset warning threshold by the correction factor to obtain the corrected warning threshold. It also retrieves the historical stable warning threshold saved in the previous cycle, assigns weights through preset trend coefficients to complete weighted fusion, obtains the stable warning threshold after weighted fusion, and directly uses the stable warning threshold as the target warning threshold for early monitoring of battery status.

[0048] The following will specifically explain how this application utilizes battery operation data to determine user behavior characteristics, including: Battery operating data includes: voltage, temperature, charging current, discharging current, and real-time power; based on the collected battery operating data, user behavior characteristics are determined, including: Based on the discharge current and real-time power, the single acceleration value, the number of accelerations and the maximum allowable acceleration value are determined. The ratio of each single acceleration value to the maximum allowable acceleration value within the number of accelerations is calculated, and the arithmetic mean of all ratios is taken to obtain the degree of acceleration aggression. Based on the charging current, the total number of braking events per unit cycle is determined, and the braking frequency is obtained by dividing the total number of braking events by the unit cycle. Based on the charging current, the fast charging duration and the total charging duration are determined respectively. The fast charging utilization rate is obtained by dividing the fast charging duration by the total charging duration. Based on the charging current and discharging current, the total number of charge-discharge cycles and the number of charge-discharge cycles when the state of charge is less than 20% are determined. The deep discharge frequency is obtained by dividing the number of charge-discharge cycles by the total number of charge-discharge cycles. Based on temperature, the operating time and total driving operating time when the temperature is above 45 degrees Celsius are determined. The high temperature exposure is obtained by dividing the operating time by the total driving operating time. Based on real-time power, the average power is determined, and then the root mean square value of the difference between real-time power and the average power is determined to obtain the load fluctuation rate.

[0049] In one embodiment of this application, determining the degree of acceleration aggression includes: during vehicle acceleration, the battery continuously outputs discharge power; the instantaneous power output of the entire vehicle can be calculated by using real-time power and discharge current; further, the acceleration value corresponding to a single acceleration is derived; each acceleration operation is recorded within a single complete driving cycle, the total number of accelerations is counted, the actual acceleration of each acceleration is compared with the maximum permissible acceleration specified by the vehicle manufacturer, and the arithmetic mean of all ratios is calculated. A higher average result indicates that the user presses the accelerator pedal more aggressively and the driving behavior is more aggressive, thereby quantifying the degree of acceleration aggression. For example, this application also provides a specific formula for calculating the degree of acceleration aggression, satisfying the following:

[0050] in, To accelerate the degree of radicalization; The number of accelerations within the statistical period; This represents the acceleration value for a single acceleration. This is the maximum permissible acceleration of the vehicle.

[0051] In one embodiment of this application, determining braking frequency includes: when the vehicle is coasting or braking, the motor performs energy recovery, generating a regenerative charging current on the battery side; each braking operation is distinguished by identifying the charging current pulse signal, the total number of braking triggers within a fixed driving cycle is counted, and the total number of braking triggers is divided by the total driving time of that cycle to obtain the braking frequency per unit time. A higher value indicates a more prominent driving habit of frequent starts and stops and frequent braking. For example, this application also provides a specific formula for calculating braking frequency, satisfying the following:

[0052] in, For braking frequency; The number of braking actions within a statistical period T is defined as follows: T is a standardized statistical duration, which can be one hour. This application selects one hour as the statistical period to cover the user's complete commute, avoiding distortion of indicators due to short-term driving data, while also preventing the smoothing out of short-term driving characteristics by an excessively long period, ensuring stable indicators and comparability across different scenarios. In one embodiment of this application, determining the fast charging utilization rate includes: pre-setting a fast charging current threshold; when the charging current exceeds this threshold, the current condition is determined to be fast charging; calculating the cumulative charging time that meets the fast charging current condition within a single statistical period, and simultaneously recording the total charging time of all charging behaviors within that period; dividing the cumulative fast charging time by the total charging time to obtain the fast charging utilization rate, which intuitively reflects whether the user prefers fast charging or slow charging in daily life. For example, this application also provides a specific formula for calculating the fast charging utilization rate, satisfying the following:

[0053] in, To increase the usage rate of fast charging; The fast charging time is the time within the statistical period. Total charging time.

[0054] In one embodiment of this application, determining the deep discharge frequency includes: calculating the battery's state of charge (SOC) in real time based on continuously collected charge and discharge currents, and fully recording each charge and discharge cycle; distinguishing between depleted cycles where the SOC drops below 20% during a single discharge, counting the total number of such deep discharge cycles, and dividing by the total number of all complete charge and discharge cycles within the cycle to obtain the deep discharge frequency. A higher value indicates that the user frequently uses the battery in a low-charge range, resulting in greater battery life loss. For example, this application also provides a specific formula for calculating the deep discharge frequency, satisfying the following:

[0055] in, This is the deep discharge frequency; The number of cycles for which the SOC is below 20%; This represents the total number of loops.

[0056] In one embodiment of this application, determining the high-temperature exposure includes: continuously collecting the real-time temperature of the battery cells, accumulating the continuous operating time during which the battery cell temperature is consistently above 45 degrees Celsius within a single driving cycle, and simultaneously calculating the total driving time of the entire vehicle during that cycle. The high-temperature operating time is divided by the total driving time to obtain the high-temperature exposure, which is used to characterize the proportion of users who drive the vehicle for extended periods in high-temperature environments and congested low-speed conditions. For example, this application also provides a specific formula for calculating the high-temperature exposure, satisfying the following:

[0058] in, High temperature exposure; Operating time when the battery temperature is above 45°C; Total driving time.

[0059] In one embodiment of this application, determining the load volatility includes: collecting real-time battery power data at all sampling times within a period; first calculating the average power mean of this set of power data; then calculating the difference between each real-time power sample value and the mean value point by point; and taking the root mean square result of all differences as the load volatility. This indicator reflects the severity of fluctuations in power demand during driving; aggressive driving and frequent acceleration and deceleration will significantly increase the load volatility value. For example, this application also provides a specific formula for calculating the load volatility, satisfying the following:

[0060] in, For load volatility; Number of samples; Real-time power; This represents the average power.

[0061] The following will explain in detail how to obtain standard smoothed user behavior characteristics, including: For any user behavior feature, the user behavior features within at least two driving cycles are numerically normalized to obtain the normalized user behavior features. We use the normalized user behavior features of the current driving cycle and the normalized user behavior features of the previous driving cycle to perform weighted fusion and smoothing to obtain standard smoothed user behavior features.

[0062] Specifically, in the embodiments of this application, determining the standard smoothed user behavior characteristics may include: First, six types of user behavior characteristics—acceleration aggression, braking frequency, fast charging usage rate, deep discharge frequency, high temperature exposure, and load fluctuation rate—are processed separately. The original calculated values ​​of each type of characteristic are collected within at least two complete vehicle start-stop driving cycles. The maximum and minimum value normalization method is used to uniformly map the original characteristics of different dimensions and different numerical ranges to the standard numerical range of 0~1, eliminating the model calculation bias caused by the differences in the numerical range of various characteristics, and outputting the normalized user behavior characteristics. Secondly, a fixed weight allocation rule is selected to complete the time-series smooth weighted fusion. The normalized features of the current driving cycle are assigned a weight in the range of 0.1 to 0.3, and the normalized features of the previous driving cycle are assigned the remaining weight. The normalized features corresponding to the two cycles are multiplied by their respective weights and then added together to achieve the filtering and suppression of short-term single driving condition fluctuations. Finally, the stable value output after weighted fusion is the standard smooth user behavior feature. This feature has a uniform value range and smooth temporal fluctuations, and can be directly input into the lightweight decision model to complete the subsequent calculation of threshold correction parameters.

[0063] For example, this application also provides a specific smoothing calculation formula, which satisfies the following:

[0064] in, The normalized user behavior feature is the j-th dimension feature of the t-th period; This is the smoothing coefficient (values ​​range from 0.6 to 0.8). The normalized user behavior features are the j-th dimension features of the (t-1)-th period.

[0065] The following will explain in detail how to determine the sign of the alarm threshold parameter, including: Obtain the first threshold and the second threshold; the first threshold is greater than the second threshold; When the user's behavior intensity score is greater than or equal to the first threshold, the user is determined to be an aggressive user. The discharge current alarm threshold correction parameter is set to a negative sign, while the differential pressure alarm threshold correction parameter, temperature alarm threshold correction parameter, and charging current upper limit alarm threshold correction parameter are set to a positive sign. When the user's behavior intensity score is less than the first threshold and greater than or equal to the second threshold, the user is determined to be a normal user, and the discharge current alarm threshold correction parameter, differential pressure alarm threshold correction parameter, temperature alarm threshold correction parameter and charging current upper limit alarm threshold correction parameter are all set to positive signs. When the user's behavior intensity score is less than the second threshold, the user is determined to be a conservative user, and the correction parameters for the discharge current alarm threshold, differential pressure alarm threshold, temperature alarm threshold, and charging current upper limit alarm threshold are all negative.

[0066] In this application, the first threshold can be understood as a behavioral intensity dividing line between aggressive and normal users, used to determine whether a driver's driving behavior is excessively aggressive. The specific calculation process for determining the first threshold may include: performing statistical clustering based on a large number of behavioral intensity score samples from users with different driving styles, extracting the lower limit of the score for the aggressive user group as the dividing line standard, and optimizing and correcting it using battery failure samples to finally obtain a critical score that stably distinguishes aggressive driving behavior. A preferred first threshold in this application is 0.7.

[0067] In this application, the second threshold can be understood as a behavioral intensity dividing line between ordinary users and conservative users, used to determine whether the driver's driving behavior is too mild and restrained. The specific calculation process for determining the second threshold may include: performing stratified statistical analysis on massive user behavior intensity scores, extracting the upper limit of the conservative user group's score as the dividing line standard, and calibrating and adjusting it in conjunction with battery underreporting risk samples to obtain a critical score that stably distinguishes conservative driving behavior. A preferred second threshold in this application is 0.3.

[0068] Specifically, determining the sign of the alarm threshold parameter can include the following steps: The first step is to read the user behavior intensity score output by the lightweight decision model and retrieve the first threshold of 0.7 and the second threshold of 0.3 pre-stored locally in the BMS. The second step involves comparing the behavior intensity score with two thresholds to classify users: if the score is ≥0.7, the user is classified as aggressive, and the discharge current correction parameter is negative to relax the threshold, while the differential pressure, temperature, and charging current upper limit correction parameters are positive to tighten the threshold; if the score is 0.3≤score<0.7, the user is classified as normal, and all four correction parameters are positive to slightly tighten the threshold to balance safety and user experience; if the score is <0.3, the user is classified as conservative, and all four correction parameters are negative to further tighten all alarm thresholds and improve safety redundancy. The third step is to record the positive or negative sign of each type of alarm threshold correction parameter, and then input it into the subsequent basic threshold correction ratio calculation process to complete the personalized threshold adjustment direction locking.

[0069] The following will explain in detail how to determine the corrected alarm threshold, including: Retrieve preset alarm thresholds from the BMS local storage area; Add the value 1 to the alarm threshold correction parameter with the corresponding positive or negative sign to obtain the correction ratio, and then multiply the correction ratio by the preset alarm threshold to obtain the corrected alarm threshold.

[0070] In this application, the preset alarm threshold can be understood as a fixed basic alarm threshold defined by the manufacturer and pre-stored in the BMS local storage chip during the power battery manufacturing stage. It is a unified benchmark threshold that does not differentiate between user driving behaviors and includes safety alarm benchmark values ​​corresponding to four types of parameters: differential pressure, temperature, charging current limit, and discharging current limit. For example, this application also provides a specific calculation formula for determining the preset alarm threshold, satisfying the following:

[0071] in, For the first The factory preset alarm thresholds corresponding to the class parameters; For the first Historical statistical mean of class parameters; Standard deviation; This is a safety factor (valued between 1.2 and 1.5, adjusted according to the parameter type).

[0072] Specifically, determining the revised alarm threshold may include the following steps: Read the four preset alarm thresholds stored in the BMS local storage area: differential pressure, temperature, charging current limit, and discharging current limit. Retrieve the four types of alarm threshold correction parameters that have been determined in the previous process and have positive and negative indicators; The correction factor is calculated separately for each type of alarm parameter. The calculation method is to add the corresponding alarm threshold correction parameter with positive or negative sign to the value 1. Multiply the preset alarm threshold corresponding to each type of parameter by the correction factor calculated by itself to complete the personalized magnitude adjustment based on the user's behavior style; The value obtained after multiplication is the alarm threshold after the corresponding parameter correction. After all four types of parameters have been calculated in sequence, the process enters the next stage of historical threshold weighted fusion.

[0073] For example, this application also provides a specific formula for calculating the modified alarm threshold, satisfying the following:

[0074] in, For the first The corrected alarm thresholds corresponding to the monitoring parameters; For the first Alarm threshold correction parameters with positive and negative signs.

[0075] The following will explain in detail how to determine the alarm threshold after stabilization, including: Obtain historical stable alarm thresholds from the BMS local storage area; The preset trend coefficient is obtained and determined as the first weight of the corrected alarm threshold, and the value minus the preset trend coefficient is determined as the second weight of the historical stable alarm threshold. The target alarm threshold is determined by multiplying the first weight by the corrected alarm threshold and adding the second weight by the historical stable alarm threshold.

[0076] In this application, the historical stable alarm threshold can be understood as the previous stable alarm threshold that has been calculated, weighted and smoothed, and stored in the local storage of the BMS after the previous complete vehicle start-stop driving cycle has been completed. It includes the stable threshold values ​​corresponding to the four types of parameters: differential pressure, temperature, charging current limit, and discharging current limit, which are used to suppress the large jump in threshold caused by the fluctuation of single driving conditions.

[0077] Specifically, the alarm threshold after stabilization can include: Retrieve the historical stable alarm thresholds obtained from the previous round of complete calculation stored in the BMS local storage, and synchronously read the corrected alarm thresholds calculated in the current process. The pre-stored preset trend coefficient is read as the first weight. The trend coefficient ranges from 0.1 to 0.3, representing the proportion of the current temporary correction threshold in the fusion calculation. The second weight is obtained by subtracting the trend coefficient from the value 1, which represents the fusion proportion of the historical stable alarm threshold. The four parameters—differential pressure, temperature, upper limit of charging current, and upper limit of discharging current—are calculated independently. The calculation formula is: stable alarm threshold = first weight × corrected alarm threshold + second weight × historical stable alarm threshold. After each type of parameter is weighted and summed, the corresponding stable alarm threshold is output. This threshold integrates the current user's real-time behavior characteristics and long-term historical usage habits to avoid sudden changes in the threshold caused by a single aggressive driving. The stable alarm thresholds corresponding to all four types of parameters are determined as the final target alarm thresholds that can be used for BMS monitoring.

[0078] For example, this application also provides a specific formula for calculating the alarm threshold after stabilization, satisfying the following:

[0079] in, For the current number The first driving cycle Alarm threshold after class parameters stabilize; The preset trend coefficient (first weight) has a value range of 0.1 to 0.3 and is used to control the fusion ratio of the current corrected threshold. For the previous driving cycle The historical stable alarm threshold corresponding to the class parameter.

[0080] This application also provides a device for determining alarm thresholds in a battery management system (BMS). Figure 2 A structural block diagram of a battery management system (BMS) alarm threshold determination device provided in this application embodiment is shown below. Figure 2 As shown, the battery management system (BMS) alarm threshold determination device 200 includes: The data acquisition unit 201 is used to collect battery operation data and determine user behavior characteristics based on the battery operation data. The user behavior characteristics include: acceleration aggression, braking frequency, fast charging usage rate, deep discharge frequency, high temperature exposure, and load fluctuation rate. Processing unit 202 is used to perform normalization and smoothing processing on user behavior features to obtain corresponding standard smoothed user behavior features; The first determining unit 203 is used to input standard smoothed user behavior characteristics and battery operation data into the first model to obtain alarm threshold correction parameters and user behavior intensity scores; the alarm threshold correction parameters include: differential pressure alarm threshold correction parameters, temperature alarm threshold correction parameters, charging current upper limit alarm threshold correction parameters and discharge current alarm threshold correction parameters; the first model is a lightweight decision model; The second determining unit 204 is used to determine the type of a user based on the user behavior intensity score, and to determine the sign of the alarm threshold correction parameter based on the type of the user; the types of users include: aggressive users, normal users, and conservative users; The correction unit 205 is used to obtain the factory-preset alarm threshold and multiply the sum of the value one and the alarm threshold correction parameter with positive and negative signs by the preset alarm threshold to obtain the corrected alarm threshold. The stabilization unit 206 is used to obtain the historical stable alarm threshold, perform weighted fusion of the historical stable alarm threshold and the corrected alarm threshold to obtain the stable alarm threshold, and determine the stable alarm threshold as the target alarm threshold.

[0081] In one exemplary embodiment, the stabilization unit 206 is further configured to: perform a security verification process on the stabilized alarm threshold, and determine the stabilized alarm threshold after the verification process as the target alarm threshold; the security verification process includes: hard constraints and rate of change limits.

[0082] In one exemplary embodiment, the correction unit 205 is further configured to: obtain a factory-preset warning threshold, multiply the sum of a value and a sign-positive-negative alarm threshold correction parameter by the preset warning threshold to obtain a corrected warning threshold; if the warning threshold is less than the alarm threshold, obtain a historical stable warning threshold, perform weighted fusion of the historical stable warning threshold and the corrected warning threshold to obtain a stable warning threshold, and determine the stable warning threshold as the target warning threshold.

[0083] In one exemplary embodiment, battery operating data includes: voltage, temperature, charging current, discharging current, and real-time power; the acquisition unit 201 is specifically used to: determine the single acceleration value, the number of accelerations, and the maximum allowable acceleration value based on the discharging current and real-time power, and calculate the ratio of each single acceleration value to the maximum allowable acceleration value within the number of accelerations, and take the arithmetic mean of all ratios to obtain the acceleration aggression level; determine the total number of brakings per unit cycle based on the charging current, and obtain the braking frequency by dividing the total number of brakings by the unit cycle; and determine the fast charging time and total charging time based on the charging current. The fast charging utilization rate is obtained by dividing the fast charging time by the total charging time. Based on the charging current and discharging current, the total number of charge-discharge cycles and the number of charge-discharge cycles with a state of charge below 20% are determined. The deep discharge frequency is obtained by dividing the number of charge-discharge cycles by the total number of charge-discharge cycles. Based on the temperature, the operating time at temperatures above 45 degrees Celsius and the total driving operating time are determined. The high temperature exposure is obtained by dividing the operating time by the total driving operating time. Based on the real-time power, the average power is determined, and then the root mean square value of the difference between the real-time power and the average power is determined to obtain the load fluctuation rate.

[0084] In one exemplary embodiment, the processing unit 202 is specifically configured to: perform numerical normalization processing on user behavior features within at least two driving cycles for any user behavior feature, to obtain normalized user behavior features; and perform weighted fusion smoothing processing on the normalized user behavior features of the current driving cycle and the normalized user behavior features of the previous driving cycle to obtain standard smoothed user behavior features.

[0085] In one exemplary embodiment, the second determining unit 204 is specifically configured to: obtain a first threshold and a second threshold; the first threshold is greater than the second threshold; when the user behavior intensity score is greater than or equal to the first threshold, determine that the user belongs to the aggressive user type, determine that the discharge current alarm threshold correction parameter is negative, and the differential pressure alarm threshold correction parameter, temperature alarm threshold correction parameter, and charging current upper limit alarm threshold correction parameter are positive; when the user behavior intensity score is less than the first threshold but greater than or equal to the second threshold, determine that the user belongs to the normal user type, and determine that the discharge current alarm threshold correction parameter, differential pressure alarm threshold correction parameter, temperature alarm threshold correction parameter, and charging current upper limit alarm threshold correction parameter are all positive; when the user behavior intensity score is less than the second threshold, determine that the user belongs to the conservative user type, and determine that the discharge current alarm threshold correction parameter, differential pressure alarm threshold correction parameter, temperature alarm threshold correction parameter, and charging current upper limit alarm threshold correction parameter are all negative.

[0086] In one exemplary embodiment, the correction unit 205 is specifically used to: obtain a preset alarm threshold from the BMS local storage area; add the value one to the alarm threshold correction parameter with corresponding positive and negative signs to obtain a correction ratio, and multiply the correction ratio by the preset alarm threshold to obtain the corrected alarm threshold.

[0087] In one exemplary embodiment, the stabilization unit 206 is specifically configured to: obtain historical stable alarm thresholds from the BMS local storage area; obtain a first weight determined by a preset trend coefficient as the corrected alarm threshold, and subtract the preset trend coefficient from the value to determine a second weight of the historical stable alarm threshold; multiply the first weight by the corrected alarm threshold, and add the second weight multiplied by the historical stable alarm threshold to obtain a stable alarm threshold, which is then determined as the target alarm threshold.

[0088] In one exemplary embodiment, the stabilization unit 206 is specifically configured to: when the safety verification process is a hard constraint, determine whether the stabilized alarm threshold is greater than a preset battery safety limit value, and when the stabilized alarm threshold is greater than the preset battery safety limit value, update the stabilized alarm threshold to the battery safety limit value; when the safety verification process is a rate of change limit, determine whether the change range of the current stabilized alarm threshold relative to the previous stabilized alarm threshold is greater than 5%, and when the change range is greater than 5%, adjust the stabilized alarm threshold to a boundary value allowed by 5%.

[0089] In summary, this application provides a method and apparatus for determining alarm thresholds in a battery management system (BMS). This application collects battery operating data and determines user behavior characteristics based on this data. These user behavior characteristics include: acceleration aggression, braking frequency, fast charging usage rate, deep discharge frequency, high temperature exposure, and load fluctuation rate. The user behavior characteristics are normalized and smoothed to obtain corresponding standard smoothed user behavior characteristics. The standard smoothed user behavior characteristics and battery operating data are input into a first model to obtain alarm threshold correction parameters and user behavior intensity scores. The alarm threshold correction parameters include: differential pressure alarm threshold correction parameters, temperature alarm threshold correction parameters, charging current upper limit alarm threshold correction parameters, and discharge current alarm threshold correction parameters. The system employs a lightweight decision-making model to correct user behavior intensity scores. It determines the user's type and, based on that type, the sign of the alarm threshold correction parameter. User types include aggressive, average, and conservative users. The system obtains the factory-preset alarm threshold and multiplies it by the sum of the numerical value and the signed alarm threshold correction parameter to obtain the corrected alarm threshold. It also obtains historical stable alarm thresholds and weightedly fuses these with the corrected threshold to obtain a stable alarm threshold, which is then used as the target alarm threshold. Compared to existing fixed, uniform thresholds and cloud-based offline calculation adjustment schemes, this application allows for local computation on the BMS vehicle-mounted embedded hardware. It leverages a lightweight decision-making model to adapt to the limited computing power of the vehicle, quantitatively distinguishes users with different driving styles based on six-dimensional battery usage behavior characteristics, automatically matches the threshold correction direction and magnitude, and combines historical threshold weighting and smoothing to suppress frequent threshold jumps caused by short-term operating condition fluctuations, thus balancing personalized adaptation with battery operation safety. In summary, the technical solution provided in this application can dynamically generate differentiated alarm thresholds based on the driver's actual driving habits, significantly reducing the probability of false alarms and missed alarms in the battery monitoring process. It does not rely on the network cloud to achieve real-time local calculation and can adapt to a variety of application scenarios.

[0090] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments claimed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0091] The basic principles of this application have been described above with reference to specific embodiments. However, it should be noted that the advantages, benefits, and effects mentioned in this application are merely examples and not limitations, and should not be considered as essential features of each embodiment of this application. Furthermore, the specific details of the above application are for illustrative and facilitative purposes only, and are not limitations. These details do not limit the application to the necessity of employing the aforementioned specific details for implementation.

[0092] The block diagrams of devices, apparatuses, devices, and systems involved in this application are merely illustrative examples and are not intended to require or imply that they must be connected, arranged, or configured in the manner shown in the block diagrams. As those skilled in the art will recognize, these devices, apparatuses, devices, and systems can be connected, arranged, and configured in any manner. Words such as “comprising,” “including,” “having,” etc., are open-ended terms meaning “including but not limited to,” and are used interchangeably with them. The terms “or” and “and” as used herein refer to the terms “and / or,” and are used interchangeably with them unless the context clearly indicates otherwise. The term “such as” as used herein refers to the phrase “such as but not limited to,” and is used interchangeably with it.

[0093] Additionally, as used herein, the "or" used in a list of items beginning with "at least one" indicates a separate list, such that a list of, for example, "at least one of A, B, or C" means A or B or C, or AB or AC or BC, or ABC (i.e., A and B and C). Furthermore, the word "exemplary" does not imply that the described example is preferred or better than other examples.

[0094] It should also be noted that in the system and method of this application, the components or steps can be decomposed and / or recombined. These decompositions and / or recombinations should be considered as equivalent solutions of this application.

[0095] Various changes, substitutions, and modifications can be made to the technology described herein without departing from the teachings defined by the appended claims. Furthermore, the scope of the claims is not limited to the specific aspects of the processes, machines, manufactures, events, means, methods, and actions described above. Currently existing or later-developed processes, machines, manufactures, events, means, methods, or actions that perform substantially the same function or achieve substantially the same result as the corresponding aspects described herein can be utilized. Therefore, the appended claims include such processes, machines, manufactures, events, means, methods, or actions within their scope.

[0096] The above description of the claimed aspects is provided to enable any person skilled in the art to make or use this application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein can be applied to other aspects without departing from the scope of this application. Therefore, this application is not intended to be limited to the aspects shown herein, but rather to be applied within the widest scope consistent with the principles and novel features of this application.

[0097] The above description has been given for illustrative and descriptive purposes. Furthermore, this description is not intended to limit the embodiments of this application to the forms described herein. Although numerous exemplary aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations thereof.

Claims

1. A method for determining alarm thresholds in a battery management system (BMS), characterized in that, The method includes: Collect battery operation data and determine user behavior characteristics based on the battery operation data; the user behavior characteristics include: acceleration aggression, braking frequency, fast charging usage rate, deep discharge frequency, high temperature exposure, and load fluctuation rate; The user behavior features are normalized and smoothed to obtain the corresponding standard smoothed user behavior features; The user behavior characteristics and battery operation data, which are smoothed according to the standard, are input into the first model to obtain alarm threshold correction parameters and user behavior intensity scores. The alarm threshold correction parameters include: differential pressure alarm threshold correction parameters, temperature alarm threshold correction parameters, charging current upper limit alarm threshold correction parameters, and discharge current alarm threshold correction parameters. The first model is a lightweight decision model. Based on the user behavior intensity score, the user's type is determined, and based on the user's type, the sign of the alarm threshold correction parameter is determined; the user's type includes: aggressive user, normal user, and conservative user; Obtain the factory-preset alarm threshold, and multiply the sum of the value one and the alarm threshold correction parameter with the positive or negative sign by the preset alarm threshold to obtain the corrected alarm threshold. Obtain historical stable alarm thresholds, perform weighted fusion of the historical stable alarm thresholds and the corrected alarm thresholds to obtain stable alarm thresholds, and determine the stable alarm thresholds as target alarm thresholds.

2. The method according to claim 1, characterized in that, The method further includes: The stabilized alarm threshold is subjected to a security verification process, and the stabilized alarm threshold that passes the verification process is determined as the target alarm threshold; the security verification process includes: hard constraints and rate of change limits.

3. The method according to claim 1, characterized in that, The method further includes: Obtain the factory-preset warning threshold, and multiply the sum of the value one and the alarm threshold correction parameter with the positive or negative sign by the preset warning threshold to obtain the corrected warning threshold; the warning threshold is less than the alarm threshold. Obtain a historical stable early warning threshold, and then perform a weighted fusion of the historical stable early warning threshold and the corrected early warning threshold to obtain a stable early warning threshold. Finally, determine the stable early warning threshold as the target early warning threshold.

4. The method according to claim 1, wherein the battery operating data includes: Voltage, temperature, charging current, discharging current, and real-time power; The process of collecting battery operating data and determining user behavior characteristics based on that data includes: Based on the discharge current and the real-time power, the single acceleration value, the number of accelerations and the maximum allowable acceleration value are determined, and the ratio of the single acceleration value to the maximum allowable acceleration value within the number of accelerations is calculated. The arithmetic mean of all the ratios is then taken to obtain the degree of acceleration aggression. Based on the charging current, the total number of braking events per unit cycle is determined, and the braking frequency is obtained by dividing the total number of braking events by the unit cycle. Based on the charging current, the fast charging duration and the total charging duration are determined respectively. The fast charging utilization rate is obtained by dividing the fast charging duration by the total charging duration. Based on the charging current and the discharging current, the total number of charge-discharge cycles and the number of charge-discharge cycles in which the state of charge is less than 20% are determined. The deep discharge frequency is obtained by dividing the number of charge-discharge cycles by the total number of charge-discharge cycles. Based on the temperature, the operating time and total driving operating time when the temperature is above 45 degrees Celsius are determined, and the high temperature exposure is obtained by dividing the operating time by the total driving operating time. Based on the real-time power, the average power is determined, and then the root mean square value of the difference between the real-time power and the average power is determined to obtain the load fluctuation rate.

5. The method according to claim 1, characterized in that, The step of normalizing and smoothing the user behavior features to obtain the corresponding standard smoothed user behavior features includes: For any of the user behavior features, the user behavior features within at least two driving cycles are numerically normalized to obtain the normalized user behavior features. The user behavior features are obtained by performing a weighted fusion smoothing process using the normalized user behavior features of the current driving cycle and the normalized user behavior features of the previous driving cycle.

6. The method according to claim 1, characterized in that, The step of determining the user's type based on the user behavior intensity score, and determining the sign of the alarm threshold correction parameter based on the user's type, includes: Obtain a first threshold and a second threshold; the first threshold is greater than the second threshold; When the user behavior intensity score is greater than or equal to the first threshold, the user is determined to be an aggressive user, the discharge current alarm threshold correction parameter is determined to be negative, and the differential pressure alarm threshold correction parameter, the temperature alarm threshold correction parameter and the charging current upper limit alarm threshold correction parameter are determined to be positive. When the user behavior intensity score is less than the first threshold and greater than or equal to the second threshold, the user is determined to be of the ordinary user type, and the discharge current alarm threshold correction parameter, the differential pressure alarm threshold correction parameter, the temperature alarm threshold correction parameter, and the charging current upper limit alarm threshold correction parameter are all positive. When the user behavior intensity score is less than the second threshold, the user is determined to be a conservative user, and the discharge current alarm threshold correction parameter, the differential pressure alarm threshold correction parameter, the temperature alarm threshold correction parameter, and the charging current upper limit alarm threshold correction parameter are all negative.

7. The method according to claim 1, characterized in that, The step of obtaining the factory-preset alarm threshold, and multiplying the sum of a value and the alarm threshold correction parameter with the sign by the preset alarm threshold to obtain the corrected alarm threshold, includes: The preset alarm threshold is obtained from the BMS local storage area; The correction factor is obtained by adding the value 1 to the alarm threshold correction parameter with the corresponding positive or negative sign, and then multiplying the correction factor by the preset alarm threshold to obtain the corrected alarm threshold.

8. The method according to claim 1, characterized in that, The step of obtaining a historical stable alarm threshold, weighting and fusing the historical stable alarm threshold and the corrected alarm threshold to obtain a stable alarm threshold, and determining the stable alarm threshold as the target alarm threshold includes: The historical stable alarm threshold is obtained from the BMS local storage area; A preset trend coefficient is obtained and determined as the first weight of the corrected alarm threshold, and the value is subtracted from the preset trend coefficient and determined as the second weight of the historical stable alarm threshold. The target alarm threshold is determined by multiplying the first weight by the corrected alarm threshold and adding the second weight by the historical stable alarm threshold.

9. The method according to claim 2, characterized in that, The step of performing a security verification process on the stabilized alarm threshold, and determining the stabilized alarm threshold that has passed the verification process as the target alarm threshold, includes: When the safety verification process is the hard constraint, it is determined whether the stabilized alarm threshold is greater than the preset battery safety limit value, and when the stabilized alarm threshold is greater than the preset battery safety limit value, the stabilized alarm threshold is updated to the battery safety limit value. When the security verification process is the change rate limit, it is determined whether the change of the current stable alarm threshold relative to the previous stable alarm threshold is greater than 5%. If the change is greater than 5%, the stable alarm threshold is adjusted to the boundary value allowed by 5%.

10. A battery management system (BMS) alarm threshold determination device, characterized in that, The device includes: The data acquisition unit is used to collect battery operating data and determine user behavior characteristics based on the battery operating data; the user behavior characteristics include: acceleration aggression, braking frequency, fast charging usage rate, deep discharge frequency, high temperature exposure, and load fluctuation rate. The processing unit is used to perform normalization and smoothing processing on the user behavior features to obtain the corresponding standard smoothed user behavior features. The first determining unit is used to input the standard smoothed user behavior characteristics and the battery operation data into the first model to obtain alarm threshold correction parameters and user behavior intensity scores; the alarm threshold correction parameters include: differential pressure alarm threshold correction parameters, temperature alarm threshold correction parameters, charging current upper limit alarm threshold correction parameters, and discharge current alarm threshold correction parameters; the first model is a lightweight decision model; The second determining unit is used to determine the type of the user based on the user behavior intensity score, and to determine the sign of the alarm threshold correction parameter based on the type of the user; the user type includes: aggressive user, normal user, and conservative user; The correction unit is used to obtain the factory-preset alarm threshold, and multiply the sum of the value one and the alarm threshold correction parameter with the positive or negative sign by the preset alarm threshold to obtain the corrected alarm threshold. A stabilization unit is used to obtain historical stable alarm thresholds, perform weighted fusion of the historical stable alarm thresholds and the corrected alarm thresholds to obtain a stable alarm threshold, and determine the stable alarm threshold as the target alarm threshold.

Images