A battery warning system

Abstract

The application relates to a battery early warning system, comprising a data module, an algorithm module and an index module. The data module is used to provide a data set and a feature set. The data in the data set is obtained by detecting the running state of the battery. The features in the feature set are calculated based on the data in the data set. The algorithm module is used to provide a plurality of early warning algorithms. The input data of each early warning algorithm is the data in the data set and / or the features in the feature set. The index module is used to provide a plurality of early warning indexes. The early warning indexes are used to evaluate the output data of the corresponding early warning algorithms to obtain the early warning result of the battery. By dividing the early warning algorithm of the battery into the data and feature layer, the algorithm layer and the index layer, when the early warning algorithm needs to be updated or optimized, only the algorithm layer and / or the index layer need to be adjusted locally, without the need to redevelop all the contents related to the early warning algorithm, so that the development workload is significantly reduced, the development cycle is shortened, and the development cost is reduced.

Description

Technical Field

[0001] This application relates to the field of vehicle technology, and more particularly to a battery warning system. Background Technology

[0002] As battery technology continues to advance, the requirements for battery performance, safety, and lifespan management are also constantly increasing. Meanwhile, new failure modes or abnormal behaviors may emerge in actual operation, such as thermal runaway, abnormal internal resistance, and abnormal self-discharge. These necessitate continuous updates and optimization of battery warning algorithms to improve the accuracy, timeliness, and reliability of battery warnings.

[0003] Battery warning algorithms are a crucial component of the Battery Management System (BMS), used for real-time monitoring of battery status, anomaly identification, and risk warnings. They are a key technology for ensuring safe battery operation. Because battery systems in different vehicle models can vary significantly in battery type, battery pack structure, sensor configuration, and control strategies, each model's development team typically develops its own battery warning algorithm for its specific vehicle.

[0004] In addition, different BMS may use different communication protocols, such as CAN bus, LIN bus, Ethernet, etc. These protocols differ in terms of data format, transmission rate, and data parsing method. Therefore, it is necessary to develop a separate battery warning algorithm for each communication protocol to ensure the correct parsing and processing of data.

[0005] Therefore, the development of current battery early warning algorithms involves a large workload, long cycle, high resource consumption, and high development cost, which has become a major challenge in the research and development of battery management systems. Summary of the Invention

[0006] This application provides a battery warning system that can reduce development workload, shorten development cycle, reduce resource consumption and development costs.

[0007] To achieve the above objectives, according to a first aspect of this application, a battery warning system is provided, comprising a data module, an algorithm module, and an indicator module. The data module provides a dataset and a feature set. The data in the dataset is obtained by detecting the battery's operating state. The features in the feature set are calculated based on the data in the dataset. The algorithm module provides multiple warning algorithms. The input data for each warning algorithm is the data in the dataset and / or the features in the feature set. The indicator module provides multiple warning indicators. The warning indicators are used to evaluate the output data of the corresponding warning algorithm to obtain a battery warning result.

[0008] Optionally, the multiple early warning algorithms include a zeroth early warning algorithm. The indicator module is further configured to modify the early warning indicators corresponding to the zeroth early warning algorithm and the first use case when the first use case of the zeroth early warning algorithm is changed. And / or, the indicator module is further configured to delete the early warning indicators corresponding to the zeroth early warning algorithm and the second use case when the second use case of the zeroth early warning algorithm is reduced. And / or, the indicator module is further configured to provide early warning indicators corresponding to the zeroth early warning algorithm and the third use case when a third use case of the zeroth early warning algorithm is added.

[0009] Optionally, the algorithm module is further configured to provide a new early warning algorithm. The algorithm module is also configured to establish the input relationship between the data in the dataset and / or the features in the feature set and the new early warning algorithm. The indicator module is further configured to provide early warning indicators corresponding to the new early warning algorithm.

[0010] Optionally, the data module is deployed on the vehicle, while the algorithm module and the indicator module are deployed on a cloud platform.

[0011] Optionally, the algorithm module includes a first algorithm module and a second algorithm module. The first algorithm module provides at least one first warning algorithm, the processing time of which is in the millisecond range. The second algorithm module provides at least one second warning algorithm, the processing time of which is longer than milliseconds. The indicator module includes a first indicator module and a second indicator module. The first indicator module provides at least one first warning indicator, which is used to evaluate the output data of the corresponding warning algorithm in the first algorithm module to obtain a warning result for the battery. The second indicator module provides at least one second warning indicator, which is used to evaluate the output data of the corresponding warning algorithm in the second algorithm module to obtain a warning result for the battery. The data module, the first algorithm module, and the first indicator module are deployed on the vehicle, while the second algorithm module and the second indicator module are deployed on a cloud platform.

[0012] Optionally, the first indicator module is further configured to upload the first warning result to the cloud platform after obtaining the first warning result based on the first warning indicator.

[0013] Optionally, the dataset includes first data. The data module is further configured to provide a correspondence between the first data and detection data of battery operating status under different vehicle models. And / or, the data module is further configured to provide a correspondence between the first data and detection data of battery operating status under different communication protocol versions.

[0014] Optionally, the feature set includes one or more of the following: charging segment features, discharging segment features, and resting segment features. The charging segment features are the behavioral characteristics of the battery during constant current or constant voltage charging. The discharging segment features are the behavioral characteristics of the battery during constant current or constant voltage discharging. The resting segment features are the behavioral characteristics of the battery in a current-free state after charging and discharging are completed.

[0015] According to a second aspect of this application, a battery warning method is provided, which is applied to the battery warning system described in the first aspect. The battery warning system includes a data module, an algorithm module, and an indicator module. The method includes: the data module acquiring target data and calculating target features based on the target data; the algorithm module inputting the target data and / or the target features into a target warning algorithm to obtain output data of the target warning algorithm; and the indicator module evaluating the output data using a target warning indicator corresponding to the target warning algorithm to obtain a target warning result.

[0016] According to a third aspect of this application, a computer program product is provided, comprising a computer program that, when executed by a processor, implements the aforementioned battery warning method.

[0017] The battery warning system of this application divides the battery warning algorithm into a data and feature layer, an algorithm layer, and an indicator layer. The data and feature layer is responsible for collecting and processing detection data during battery operation (such as voltage, current, temperature, SOC, etc.) and converting it into physically meaningful features (such as voltage fluctuation rate, temperature gradient, capacity decay rate, etc.), providing unified and standardized input data for the warning algorithm. The algorithm layer uses corresponding warning algorithms for risk identification and warning judgment. The indicator layer defines the judgment criteria for warnings (such as thresholds, weights, confidence levels, etc.) to evaluate and classify the output data of the warning algorithm, ultimately generating warning results. When the warning algorithm needs to be updated or optimized, only local adjustments need to be made to the algorithm layer and / or the indicator layer, without redeveloping all content related to the warning algorithm. For example, if it is necessary to improve the accuracy of the warning, the warning algorithm can be optimized or the warning indicator threshold can be adjusted; if it is necessary to adapt to new battery types or communication protocols, only adaptation processing needs to be performed at the data and feature layer, without affecting the structure and logic of the algorithm layer and the indicator layer. This layered design significantly reduces the development workload, shortens the development cycle, and reduces resource consumption, thereby effectively reducing development costs.

[0018] Other features and advantages of this application will be described in detail in the following detailed description section. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a schematic diagram of the structure of a battery warning system provided in an embodiment of this application;

[0021] Figure 2 This is a schematic diagram illustrating an application scenario of a battery warning system provided in an embodiment of this application;

[0022] Figure 3 This is a schematic diagram of a deployment scheme for a battery warning system provided in an embodiment of this application;

[0023] Figure 4 This is a schematic diagram of another deployment scheme for a battery warning system provided in an embodiment of this application. Detailed Implementation

[0024] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the protection scope of this application.

[0025] To address the issues of high workload, long development cycle, high resource consumption, and high development cost in current battery warning algorithms, this application provides a battery warning system. By hierarchically and modularly configuring the battery warning algorithm, the updating, iteration, and optimization of the warning algorithm become more flexible, while also facilitating the development of new algorithms.

[0026] See Figure 1 , Figure 1 This is a schematic diagram of a battery warning system provided in an embodiment of this application. Figure 1 As shown, the battery warning system 10 includes a data module 11, an algorithm module 12, and an indicator module 13.

[0027] Data module 11 is used to provide datasets and feature sets.

[0028] In some embodiments, the data in the dataset can be voltage, current, temperature, etc., during battery operation. This data is obtained by detecting the battery's operating status. This data is typically acquired by sensors and acquisition modules in the Battery Management System (BMS) to reflect the real-time status of the battery during different operating stages such as charging, discharging, and resting. It serves as the basis for evaluating battery performance, health status, and safety.

[0029] In some embodiments, the features in the feature set are calculated based on data in the dataset. Taking data in the dataset, including battery voltage, current, and temperature during operation, as an example, one or more of the following can be extracted: charging segment features, discharging segment features, and resting segment features. Charging segment features are the behavioral characteristics of the battery during constant current or constant voltage charging. Discharging segment features are the behavioral characteristics of the battery during constant current or constant voltage discharging. Resting segment features are the behavioral characteristics of the battery in a current-free state after charging and discharging are completed. These features include, but are not limited to, voltage fluctuation rate, temperature gradient, and charging / discharging rate. Specifically:

[0030] (1) Characteristics of the charging section

[0031] Voltage fluctuation rate: During the charging process, the rate of change of voltage (such as ΔV / Δt) reflects the electrochemical reaction rate inside the battery and can be used to determine whether the battery is in a normal charging state or whether there is an abnormality (such as lithium plating, internal short circuit, etc.).

[0032] Charge-discharge rate (C-rate): This is the ratio of charging current to the battery's rated capacity, used to describe charging speed. Different C-rates can affect the battery's health and lifespan.

[0033] Temperature gradient: During the charging process, the temperature of different parts of the battery may be different due to electrochemical reactions and internal resistance heating. An excessively large temperature gradient may indicate that there is non-uniformity or thermal runaway risk inside the battery.

[0034] Charging efficiency: By comparing the input electrical energy with the actual electrical energy stored in the battery, the energy loss during the charging process can be assessed.

[0035] The above-mentioned charging segment characteristics can be used to determine whether the battery is in a normal charging state and to identify abnormal charging behaviors (such as overcharging, lithium plating, internal short circuit, etc.).

[0036] (2) Characteristics of the discharge segment

[0037] Voltage fluctuation rate: The rate of change of voltage during discharge (such as ΔV / Δt) reflects the change of internal resistance and electrochemical reaction state of the battery, and can be used to determine whether the battery is aging or has abnormalities.

[0038] Discharge rate (C-rate): Similar to charging rate, it describes the rate of discharge and affects battery performance and lifespan.

[0039] Temperature gradient: During discharge, the temperature change inside the battery may be related to electrochemical reactions and internal resistance. An excessively large temperature gradient may indicate that there is non-uniformity inside the battery.

[0040] Voltage-capacity curve (V-Curve): By observing the relationship between voltage and capacity during discharge, the health status and capacity decay of the battery can be determined.

[0041] The above-mentioned discharge segment characteristics can be used to determine whether the battery is in a normal discharge state and to identify battery aging, capacity decay, or abnormal discharge behavior (such as over-discharge, internal short circuit, etc.).

[0042] (3) Characteristics of the static section

[0043] Voltage fluctuation rate: During the resting process, the slow change in battery voltage may reflect self-discharge behavior or the continuous occurrence of internal electrochemical reactions.

[0044] Temperature change rate: During the resting period, a slow rise or fall in battery temperature may be related to self-discharge, internal chemical reactions, or thermal runaway.

[0045] Voltage recovery characteristics: During the resting process, the battery voltage may tend to stabilize, and its recovery speed and stability can be used to judge the health status of the battery.

[0046] Self-discharge rate: The self-discharge rate of a battery can be estimated by the change in voltage during resting period, and is used to assess the battery's storage performance and aging condition.

[0047] The above-mentioned resting period characteristics can be used to evaluate the battery's performance in a non-operating state, such as self-discharge behavior, aging trend, and health status.

[0048] In summary, by monitoring data such as voltage, current, and temperature during battery operation, various characteristics of the charging, discharging, and resting phases can be extracted, including voltage fluctuation rate, temperature gradient, and charge / discharge rate. These characteristics are crucial for battery state assessment, anomaly detection, and health prediction. They can serve as important inputs for tasks such as battery state of health (SOH) estimation, fault diagnosis, and early warning detection in a battery management system (BMS).

[0049] Algorithm module 12 provides various early warning algorithms. The input data for each early warning algorithm consists of data from the dataset and / or features from the feature set. Each early warning algorithm targets different characteristics and potential risks during battery operation, used to identify abnormal states, predict performance degradation, detect faults or potential failures, thereby improving the safety, reliability, and lifespan of the battery system. The following are some common early warning algorithms:

[0050] (1) Cutoff Point Detection Algorithm: The input data of the cutoff point detection algorithm includes the battery's voltage, capacity, charging segment characteristics, and discharging segment characteristics during operation. The output data includes the inflection point voltage and inflection point capacity of the voltage-capacity curve during charging and discharging (such as the charging platform inflection point and the discharging platform inflection point). The cutoff point detection algorithm can accurately estimate the battery's capacity change and is suitable for tasks such as battery capacity estimation, SOH assessment, and voltage-capacity curve analysis during charging and discharging.

[0051] (2) Capacity Estimation Algorithm: The input data of the capacity estimation algorithm includes the battery's voltage, current, temperature, charging characteristics, discharging characteristics, and resting characteristics during operation. The output data includes the battery's current capacity and capacity change rate. The capacity estimation algorithm can determine whether the battery is in normal working condition and is suitable for tasks such as SOH assessment, battery life prediction, and capacity decay monitoring.

[0052] (3) Voltage Difference Algorithm: The input data of the voltage difference algorithm includes the battery's voltage, voltage difference, charging segment characteristics, discharging segment characteristics, and resting segment characteristics during operation. The output data includes the voltage difference between each cell in the battery, the voltage difference distribution characteristics, and the voltage difference fluctuation range. By analyzing the voltage changes (voltage difference) at different time points during the charge and discharge cycle, the voltage difference algorithm can identify whether the voltage fluctuation is abnormal, thereby determining whether the battery has internal short circuits, lithium plating, thermal runaway, or other faults. It is suitable for tasks such as battery anomaly detection, cell consistency monitoring, and thermal runaway early warning.

[0053] (4) Insulation Detection Algorithm: The input data of the insulation algorithm includes the voltage, current, and temperature of the battery during operation, and the output data includes the insulation resistance value and the insulation status judgment result. The insulation algorithm can detect the insulation impedance between the battery and the casing and the battery cells, and determine whether there are safety hazards such as insulation failure and leakage. It is suitable for tasks such as battery system insulation status monitoring and safety early warning.

[0054] (5) Thermal Runaway Detection Algorithm: The input data of the thermal runaway detection algorithm includes the battery's voltage, current, temperature, charging segment characteristics, discharging segment characteristics, and resting segment characteristics during operation. The output data includes the thermal runaway risk level, thermal runaway feature identification results (such as the number of primary features (e.g., abnormal temperature, abnormal voltage), the number of secondary features (e.g., sensor disconnection, voltage fluctuation), and the sum of primary and secondary features). The thermal runaway detection algorithm can identify whether the battery is in the early stage of thermal runaway, issue timely warnings, and prevent thermal runaway accidents from occurring. It is suitable for tasks such as battery thermal management and safety warnings.

[0055] (6) Cell Consistency Detection Algorithm: The input data for the cell consistency detection algorithm includes the voltage, current, capacity, charging segment characteristics, discharging segment characteristics, and resting segment characteristics of the battery during operation. The output data includes cell consistency score and cell voltage consistency. The cell consistency detection algorithm can detect the consistency of parameters such as voltage, capacity, and internal resistance among the cells in the battery pack, determine whether there are problems such as cell aging or imbalance, and thus perform cell balancing or replacement. It is suitable for tasks such as battery pack management and cell balancing control.

[0056] In summary, there are many types of early warning algorithms, each with its specific application scenarios and detection targets. In practical applications, multiple early warning algorithms are usually combined to achieve comprehensive monitoring and early warning of battery status.

[0057] The indicator module 13 provides various early warning indicators. These indicators are used to evaluate the output data of the corresponding early warning algorithm to obtain the battery's early warning result. Specifically, early warning indicators are key parameters used to evaluate the output data of the corresponding early warning algorithm to determine whether the battery is in an abnormal state or has a potential risk of failure.

[0058] Using the aforementioned inflection point detection algorithm, capacity detection algorithm, differential pressure algorithm, insulation algorithm, thermal runaway early warning algorithm, and cell consistency detection algorithm as examples of various early warning algorithms in algorithm module 12, we will introduce various early warning indicators in indicator module 13 and their application process.

[0059] For the inflection point detection algorithm, the output data includes the inflection point voltage threshold (specific value) and the inflection point charge level (specific value). The corresponding warning indicators can be the inflection point voltage deviation threshold, the inflection point charge level deviation rate threshold, and the inflection point validity judgment (valid or invalid). The inflection point voltage deviation refers to the deviation between the inflection point voltage threshold and the calibrated inflection point voltage. The inflection point charge level deviation rate refers to the proportion of the deviation between the inflection point charge level and the calibrated inflection point charge level. The inflection point validity judgment refers to whether it is a valid inflection point (e.g., whether the voltage change rate meets preset conditions). If the inflection point voltage deviation calculated based on the inflection point voltage threshold (specific value) exceeds the inflection point voltage deviation threshold, or the inflection point charge level deviation rate calculated based on the inflection point charge level (specific value) exceeds the inflection point charge level deviation rate threshold, or the inflection point validity judgment is invalid, a warning is triggered, indicating that the battery state is abnormal or aging. The warning result for the battery is: the battery may be in an aging or performance degradation state.

[0060] For capacity detection algorithms, the output data includes the battery's current capacity (specific value) and capacity change rate (specific value). Corresponding warning indicators can be capacity decay rate thresholds and capacity change rate thresholds. The capacity decay rate refers to the ratio of the battery's current capacity to its initial capacity. The capacity change rate refers to the amount of capacity change per unit time. If the capacity decay rate calculated based on the battery's current capacity (specific value) exceeds the capacity decay rate threshold, or the capacity change rate (specific value) exceeds the capacity change rate threshold, a warning is triggered. The warning result for the battery is: battery capacity is decreasing, which may affect battery system performance or lifespan.

[0061] For the differential voltage algorithm, the output data includes the voltage difference between the cells in the battery, the voltage difference distribution characteristics, and the voltage difference fluctuation range. Corresponding warning indicators can be the maximum voltage difference threshold and the voltage difference fluctuation rate threshold. The maximum voltage difference refers to the voltage difference between the highest-voltage and lowest-voltage cells in the battery. The voltage difference fluctuation rate refers to the rate of change of the voltage difference over time. If the maximum voltage difference calculated based on the voltage difference between the cells in the battery exceeds the maximum voltage difference threshold, or if the voltage difference fluctuation rate calculated based on the voltage difference between the cells in the battery exceeds the voltage difference fluctuation rate threshold, a warning is triggered. The warning result for the battery is: inconsistency among the cells in the battery pack may lead to performance degradation or thermal runaway risk.

[0062] For the insulation algorithm, the output data includes insulation resistance value (specific numerical value), insulation status judgment result (normal or abnormal), etc. The corresponding warning indicators can be preset safety values ​​(such as 100kΩ), insulation status judgment results, etc. If the insulation resistance value is lower than the preset safety value, or the insulation status judgment result is abnormal, a warning is triggered. The warning result for the battery is: the battery system may have leakage or insulation failure, and immediate inspection is required.

[0063] For thermal runaway early warning algorithms, the output data includes the thermal runaway risk level (specific level) and thermal runaway feature identification results (such as the number of primary features, the number of secondary features, and the sum of primary and secondary features). The corresponding early warning indicators can be preset risk levels, thresholds for the number of primary features, thresholds for the number of secondary features, and thresholds for the sum of primary and secondary features. If the thermal runaway risk level (specific level) is higher than the preset risk level, or the number of primary features exceeds the primary feature threshold, or the number of secondary features exceeds the secondary feature threshold, or the sum of primary and secondary features exceeds the primary and secondary feature sum threshold, an early warning is triggered. The battery's warning result is: the battery may experience thermal runaway, and immediate safety measures are required.

[0064] The cell consistency detection algorithm outputs data including cell consistency score (specific value) and cell voltage consistency (specific value). Corresponding warning indicators can be preset score values ​​or preset voltage difference thresholds (e.g., 50mV). If the cell consistency score (specific value) is lower than the preset score value, or the cell voltage consistency (specific value) exceeds the preset voltage difference threshold, a warning is triggered. The warning result for the battery is: inconsistency among cells in the battery pack may affect battery pack life and safety, requiring equalization processing.

[0065] In summary, there are many types of early warning indicators, and each early warning algorithm has its corresponding early warning indicators.

[0066] In the example scenario where the battery early warning system 10 is applied to the aforementioned data set containing battery operating voltage, current, and temperature, and using charging, discharging, and resting characteristics as features in the feature set, and employing multiple early warning algorithms such as inflection point detection, capacity detection, differential pressure detection, insulation detection, thermal runaway early warning, and cell consistency detection, and using the corresponding early warning indicators for each of these algorithms as multiple early warning indicators, the specific content of the battery early warning system 10 is as follows: Figure 2 As shown.

[0067] In another embodiment, a machine learning model can also be used to characterize the warning indicators corresponding to the warning algorithm. Specifically, the output data of the warning algorithm is input into the machine learning model, which then calculates the battery warning result based on the algorithm's output data. This machine learning model is trained using the correspondence between the historical output data of the warning algorithm and the historical warning results of the battery as training data. As can be seen, the machine learning model can also act as a warning indicator, and it can also evaluate the output data of the warning algorithm to obtain the battery warning result.

[0068] In some embodiments, the multiple early warning algorithms in algorithm module 12 include a zero-early warning algorithm.

[0069] When the first use case of the zero-warning algorithm is changed, the indicator module 13 maintains the usability of the original zero-warning algorithm by modifying the warning indicators corresponding to the zero-warning algorithm and the first use case.

[0070] For example, the zeroth warning algorithm is an insulation algorithm. Changing the primary application scenario of the zeroth warning algorithm could be, for instance, the introduction of new battery usage specifications in location A. In this case, indicator module 13 can continue using the insulation algorithm by modifying its preset safety value in location A (e.g., from 100kΩ to 120kΩ). Compared to related solutions where a new battery usage specification is introduced in location A, requiring each vehicle's R&D team to redevelop an insulation algorithm corresponding to the new specification for their respective models, the battery warning system 10 can generate a new insulation algorithm simply by modifying the warning indicator in location A. At most, it modifies the specific values ​​of the warning indicator for each vehicle model in location A, without needing to redevelop the specific code for the insulation algorithm or re-establish the input relationship between the insulation algorithm and certain battery data or features. This results in less development workload, shorter development cycle, less resource consumption, and lower development costs.

[0071] In reducing the second use case of the zero-warning algorithm, the indicator module 13 maintains the usability of the original zero-warning algorithm in other scenarios by deleting the warning indicators corresponding to the zero-warning algorithm and the second use case.

[0072] For example, the zeroth warning algorithm is an insulation algorithm. Reducing the second use case of the zeroth warning algorithm could be, for instance, the failure of some battery usage specifications in location B. In this case, indicator module 13 can maintain the use of the insulation algorithm in locations A and C by deleting the preset safety value of the insulation algorithm in location B. Compared to related solutions where some battery usage specifications fail in location B, each vehicle's R&D team needs to delete the insulation algorithm corresponding to the failed battery usage specifications for their respective models. This includes deleting the input relationship between the insulation algorithm and certain battery data or features, deleting the specific code of the insulation algorithm, and deleting the warning indicators of the insulation algorithm in location B. The battery warning system 10 only needs to delete the preset safety value of the insulation algorithm in location B, or at most, the specific values ​​of the warning indicators of the insulation algorithm on each vehicle model in location B. It does not need to delete the specific code of the insulation algorithm or the input relationship between the insulation algorithm and certain battery data or features. This results in less development workload, shorter cycle, less resource consumption, and lower development cost.

[0073] In the case of adding a third use case for the zero-warning algorithm, the indicator module 13 increases the usability of the original zero-warning algorithm in the new scenario by providing warning indicators corresponding to the zero-warning algorithm and the third use case.

[0074] For example, if the zeroth warning algorithm is an insulation algorithm, a third application scenario for the newly added zeroth warning algorithm could be the introduction of various vehicle models in location D. In this case, the indicator module 13 can add a preset safety value (e.g., 80kΩ) for the insulation algorithm in location D based on the battery usage specifications of location D, thus promoting the use of the existing insulation algorithm in location D. Compared to related solutions where various vehicle models are introduced in location D, requiring each model's R&D team to redevelop insulation algorithms corresponding to the battery usage specifications of location D for their respective models, the battery warning system 10 can generate new application scenarios for the insulation algorithm simply by adding a warning indicator for the insulation algorithm in location D. At most, it adds the specific values ​​of the warning indicator for each vehicle model in location D, without needing to redevelop the specific code of the insulation algorithm or re-establish the input relationship between the insulation algorithm and certain battery data or features. This results in less development workload, shorter cycle, less resource consumption, and lower development cost.

[0075] In some embodiments, the algorithm module 12 is further used to provide a new warning algorithm and establish the input relationship between the data in the dataset and / or the features in the feature set and the new warning algorithm. The indicator module 13 is further used to provide a warning indicator corresponding to the new warning algorithm. It can be seen that when a new warning algorithm is needed, compared to the relevant solutions where the R&D teams of each vehicle model need to develop a set of code for a new warning algorithm and its corresponding warning indicators for their respective vehicle models, and establish the input relationship between their respective developed warning algorithms and the battery data or features in the corresponding vehicle models, the battery warning system 10 can keep the data and features in the data module 11 unchanged, and achieve the applicability of the new warning algorithm in various scenarios by adding a set of code for the warning algorithm in the algorithm module 12, establishing the input relationship between the warning algorithm and the data in the dataset and / or the features in the feature set, and adding the warning indicators of the warning algorithm in each vehicle model and each region in the indicator module 13. This reduces the development workload, shortens the cycle, consumes fewer resources, and lowers the development cost.

[0076] Considering the significant differences in battery systems across different vehicle models in terms of battery type, battery pack structure, sensor configuration, and control strategies, and the fact that different BMSs may employ different communication protocols such as CAN bus, LIN bus, and Ethernet, which differ considerably in data format, transmission rate, and data parsing methods, the resulting battery detection data formats across different vehicle models are diverse and complex to adapt. To address this issue, data module 11 establishes a correspondence between the data in the dataset (such as the first data) and the detection data of the battery's operating status under different vehicle models, and / or establishes a correspondence between the data in the dataset (such as the first data) and the detection data of the battery's operating status under different communication protocol versions. This ensures that the data in the dataset is uniform and standardized, representing detection data with the same physical meaning under different scenarios and usage conditions. In this way, the features in the feature set calculated based on the standard data in the dataset also possess uniformity and standardization, thereby fixing the input relationship between the early warning algorithm and the data in the dataset, as well as the input relationship between the early warning algorithm and the features in the feature set. This design allows the warning algorithm to adapt flexibly to changes such as new data versions or removed protocol fields when faced with battery detection data from different vehicle models or different communication protocol versions, without requiring redevelopment or adjustment at the algorithm level. It only requires establishing or updating the correspondence between the dataset and the actual detection data, without affecting the input relationship between the warning algorithm and the data in the dataset and / or the features in the feature set. This modification method is highly practical and flexible, effectively addressing dynamic changes in battery system and vehicle detection data, avoiding the need for redevelopment of the warning algorithm due to changes in underlying battery or vehicle detection data, thus significantly improving the system's scalability, compatibility, and maintenance efficiency. Furthermore, this method enhances the robustness of the warning system, enabling it to operate stably in various vehicle models and communication protocol environments, meeting the needs of complex and ever-changing real-world application scenarios.

[0077] In summary, the battery early warning system 10 divides the battery early warning algorithm into a data and feature layer, an algorithm layer, and an indicator layer, achieving modular design and flexible system upgrades. In the data and feature layer, it is responsible for collecting and processing detection data during battery operation (such as voltage, current, temperature, SOC, etc.) and converting it into physically meaningful features (such as voltage fluctuation rate, temperature gradient, capacity decay rate, etc.), providing unified and standardized input data for the early warning algorithm. In the algorithm layer, corresponding early warning algorithms are used for risk identification and early warning judgment. In the indicator layer, early warning judgment criteria (such as threshold, weight, confidence level, etc.) are defined to evaluate and classify the output data of the early warning algorithm, ultimately generating early warning information.

[0078] When the early warning algorithm needs updating or optimization, only local adjustments are required at the algorithm layer and / or indicator layer, without redeveloping all content related to the algorithm. For example, to improve the accuracy of early warnings, the algorithm can be optimized or the threshold values ​​for early warning indicators can be adjusted; to adapt to new battery types or communication protocols, adaptation processing is only needed at the data and feature layers, without affecting the structure and logic of the algorithm and indicator layers. This layered design significantly reduces development workload, shortens the development cycle, and reduces resource consumption, thereby effectively reducing development costs.

[0079] In addition, this layered architecture enhances the maintainability and scalability of the system, enabling the battery warning system 10 to quickly respond to the evolution of battery technology, updates to communication protocols, and adaptation needs of different vehicle models, providing a more flexible and efficient solution for battery safety management and intelligent warning.

[0080] In some embodiments, see Figure 3 , Figure 3 This is a schematic diagram of a deployment scheme for a battery warning system provided in an embodiment of this application. Figure 3 As shown, the data module 11 of the battery warning system 10 is deployed on the vehicle, and the algorithm module 12 and indicator module 13 of the battery warning system 10 are deployed on the cloud platform.

[0081] In existing solutions, the cloud centrally manages early warning algorithms through a .so dynamic library. If an algorithm needs updating or optimization, the entire dynamic library must be redeployed. This not only requires updating the algorithm modules on the cloud server but may also involve adaptation and verification on the vehicle side, resulting in high development and deployment costs and a long development cycle. In this technical solution, however, the data and feature layers are deployed on the vehicle side. The vehicle side is responsible for collecting and processing raw data, converting it into features with unified physical meaning, thus shielding the differences between different vehicle models and communication protocols. The cloud platform only needs to receive data and features in a unified format for algorithm processing and early warning judgment, significantly reducing the adaptation costs for different vehicle models and communication protocols. Furthermore, this modular design gives the system excellent scalability. When adding a new vehicle model or communication protocol, only the data and feature extraction logic needs to be adapted on the vehicle side, while the cloud platform remains unchanged, greatly improving the system's maintainability and scalability. Moreover, the vehicle side only uploads data and features in a unified format to the cloud, reducing data transmission volume and communication bandwidth consumption. Simultaneously, the cloud only needs to process a small amount of feature data, improving cloud computing efficiency and optimizing system resource utilization. Furthermore, this technical solution decouples the algorithm from the data, allowing the development, testing, and maintenance of the algorithm in the cloud to be carried out independently of the data acquisition and processing logic on the vehicle side, thereby reducing the complexity of development and maintenance, and reducing the development cycle and resource consumption.

[0082] In some embodiments, see Figure 4 , Figure 4 This is a schematic diagram of another deployment scheme for a battery warning system provided in an embodiment of this application. For example... Figure 4 As shown, the algorithm module 12 in the battery warning system 10 includes a first algorithm module 121 and a second algorithm module 122. The first algorithm module 121 provides at least one first warning algorithm, with a processing time in the millisecond range, for example, less than 1000 milliseconds. The second algorithm module 122 provides at least one second warning algorithm, with a processing time exceeding the millisecond range, for example, exceeding 1000 milliseconds. The indicator module 13 in the battery warning system 10 includes a first indicator module 131 and a second indicator module 132. The first indicator module 131 provides at least one first warning indicator, which is used to evaluate the output data of the corresponding warning algorithm in the first algorithm module 121 to obtain a battery warning result. The second indicator module 132 provides at least one second warning indicator, which is used to evaluate the output data of the corresponding warning algorithm in the second algorithm module 122 to obtain a battery warning result. The data module 11, the first algorithm module 121, and the first indicator module 131 are deployed on the vehicle, while the second algorithm module 122 and the second indicator module 132 are deployed on a cloud platform.

[0083] In related solutions, the cloud centrally manages the early warning algorithm by deploying a .so dynamic library, while the vehicle only acts as a data production terminal, responsible for collecting battery operating status data and uploading it to the cloud for processing. This architecture has significant shortcomings when facing sudden and dynamic operating scenarios. For example, in cases of transient current changes or abnormal battery fluctuations during fast charging, data must first go through multiple stages such as uploading, queuing, calculation, and feedback, resulting in a significant delay in the overall response. Since the early warning algorithm cannot complete the judgment and response on the vehicle in the first instance, it is easy to miss the early warning window, reducing the system's active protection capability and even potentially causing battery safety hazards. In this technical solution, a reasonable layered deployment is adopted to address the processing time and response requirements of the early warning algorithm: early warning algorithms with short processing times and requiring rapid responses (such as SOH assessment and fast charging identification) are directly deployed on the vehicle. Utilizing the vehicle's real-time computing capabilities, the initial judgment and early warning response of the battery status are completed within milliseconds, thereby ensuring that the early warning mechanism can be triggered immediately in the event of sudden anomalies, improving the system's active protection capability and safety. Meanwhile, for early warning algorithms that require long processing times and accurate calculations (such as battery life prediction, thermal runaway risk assessment, and multi-dimensional state fusion analysis), they are deployed in the cloud. This fully leverages the powerful computing resources and data processing capabilities of the cloud to perform precise calculations on complex models, thereby ensuring the accuracy and reliability of the early warning results. This layered deployment strategy not only improves the real-time performance and response speed of the early warning system but also ensures the accuracy and stability of the early warning results, achieving a balance between ensuring battery safety and balancing system performance and computational efficiency.

[0084] In one specific embodiment, in an application scenario where the data module 11, the first algorithm module 121, and the first indicator module 131 are deployed on a vehicle, and the second algorithm module 122 and the second indicator module 132 are deployed on a cloud platform, the first indicator module 131 is also used to upload the first warning result to the cloud platform after obtaining the first warning result based on the first warning indicator.

[0085] In this technical solution, the warning results uploaded from the vehicle provide the cloud with a complete record of warning events, including key information such as warning time, warning type, warning level, and vehicle status. This data can be used for subsequent fault analysis, performance evaluation, and user behavior research, providing data support for battery system optimization, warning algorithm improvement, and vehicle safety strategy formulation. Furthermore, the warning results uploaded from the vehicle can also serve as a real-time data stream, providing the cloud model with the latest operational status information, improving the model's timeliness and adaptability. Especially when facing new battery technologies, new usage scenarios, or sudden operating conditions, it can quickly adjust model parameters and improve prediction accuracy.

[0086] According to one aspect of this application, a battery warning method is provided, applicable to the aforementioned battery warning system 10. The method includes the following steps:

[0087] During the operation of the target vehicle, the data module 11 in the battery warning system 10 detects the operating status of the target battery in the target vehicle and obtains target data as data in the dataset. Furthermore, the data module 11 calculates target features based on the target data as features in the feature set. The data module 11 also sends the target data and / or target features to the algorithm module 12 in the battery warning system 10.

[0088] Subsequently, the algorithm module 12 selects a target early warning algorithm from its stored pool of early warning algorithms based on the early warning task. For example, if the early warning task is thermal runaway risk assessment, the target early warning algorithm is a thermal runaway early warning algorithm. Alternatively, if the early warning task is SOH assessment, the target early warning algorithm is a capacity detection algorithm. The algorithm module 12 inputs the target data and / or target features into the target early warning algorithm to obtain the output data of the target early warning algorithm. The algorithm module 12 also sends the output data of the target early warning algorithm to the indicator module 13 in the battery early warning system 10.

[0089] Next, the indicator module 13 selects the target early warning indicator corresponding to the target early warning algorithm from its stored collection of early warning indicators. For example, if the target early warning algorithm is a thermal runaway early warning algorithm, the target early warning indicator would be a preset risk level, a threshold for the number of occurrences of primary features, etc. Alternatively, if the target early warning algorithm is a capacity detection algorithm, the target early warning indicator would be a capacity decay rate threshold, a capacity change rate threshold, etc. The indicator module 13 uses the target early warning indicator to evaluate the output data of the target early warning algorithm and obtain the target early warning result.

[0090] exist Figure 3 In the deployment scheme of the battery warning system shown, the data module 11 is deployed on the vehicle side, and the operation it performs is actually completed by the microcontroller unit (MCU) in the target vehicle; while the algorithm module 12 and the indicator module 13 are deployed on the cloud platform, and the operation they perform is completed by the server on the cloud platform.

[0091] exist Figure 4In the battery warning system deployment scheme shown, the data module 11, the first algorithm module 121, and the first indicator module 131 are deployed on the vehicle side, while the second algorithm module 122 and the second indicator module 132 are deployed on the cloud platform. Therefore, if the target warning algorithm is stored in the first algorithm module 121 (such as a capacity detection algorithm), the operations performed by the data module 11, algorithm module 12, and indicator module 13 in the above battery warning method are actually all completed by the MCU in the target vehicle. If the target warning algorithm is stored in the second algorithm module 122 (such as a thermal runaway warning algorithm), the operations performed by the data module 11 in the above battery warning method are actually completed by the MCU in the target vehicle, while the operations performed by the algorithm module 12 and indicator module 13 are completed by the server on the cloud platform.

[0092] It should be noted that the battery warning method provided in this application has all the beneficial effects of the battery warning system 10 described above, and will not be repeated here.

[0093] According to one aspect of this application, a computer program product or computer program is provided, comprising computer instructions stored in a computer-readable storage medium. A processor of an electronic device reads the computer instructions from the computer-readable storage medium, and executes the computer instructions, causing the electronic device to perform the operations performed by the data module 11 in the battery warning method provided in the above embodiments, and / or, perform the operations performed by the algorithm module 12 in the battery warning method provided in the above embodiments, and / or, perform the operations performed by the indicator module 13 in the battery warning method provided in the above embodiments.

[0094] In the description of this application, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more features. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0095] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0096] The embodiments, implementation methods, and related technical features of this application can be combined and substituted for each other without conflict.

[0097] The above are merely preferred embodiments of this application and are not intended to limit this application in any way. Any simple modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of this application without departing from the scope of the technical solution of this application shall still fall within the scope of the technical solution of this application.

Claims

1. A battery early warning system, characterized in that, include: Data module, algorithm module, and indicator module. The data module is used to provide a dataset and a feature set, wherein the data in the dataset is obtained by detecting the operating status of the battery, and the features in the feature set are calculated based on the data in the dataset; The algorithm module is used to provide multiple early warning algorithms, wherein the input data for each early warning algorithm is data in the dataset and / or features in the feature set; The indicator module is used to provide a variety of early warning indicators, wherein the early warning indicators are used to evaluate the output data of the corresponding early warning algorithm to obtain the early warning result of the battery.

2. The battery warning system according to claim 1, characterized in that, The various early warning algorithms include the zeroth early warning algorithm. The indicator module is further configured to modify the warning indicator corresponding to the zeroth warning algorithm and the first use case when the first use case of the zeroth warning algorithm is changed; and / or, The indicator module is further configured to, when reducing the second use case of the zeroth early warning algorithm, delete the early warning indicator corresponding to the zeroth early warning algorithm and the second use case; and / or, The indicator module is also used to provide warning indicators corresponding to the zero warning algorithm and the third use scenario when a third use scenario of the zero warning algorithm is added.

3. The battery warning system according to claim 1 or 2, characterized in that, The algorithm module is also used to provide new early warning algorithms; The algorithm module is also used to establish the input relationship between the data in the dataset and / or the features in the feature set and the new early warning algorithm; The indicator module is also used to provide early warning indicators corresponding to the new early warning algorithm.

4. The battery warning system according to any one of claims 1-3, characterized in that, The data module is deployed on the vehicle, while the algorithm module and the indicator module are deployed on the cloud platform.

5. The battery warning system according to any one of claims 1-3, characterized in that, The algorithm module includes a first algorithm module and a second algorithm module. The first algorithm module is used to provide at least one first early warning algorithm, the processing time of which is in the millisecond range. The second algorithm module is used to provide at least one second early warning algorithm, the processing time of which exceeds milliseconds; The indicator module includes a first indicator module and a second indicator module. The first indicator module is used to provide at least one first warning indicator, and the first warning indicator is used to evaluate the output data of the corresponding warning algorithm in the first algorithm module to obtain the warning result of the battery. The second indicator module is used to provide at least one second warning indicator, which is used to evaluate the output data of the corresponding warning algorithm in the second algorithm module to obtain the warning result of the battery. The data module, the first algorithm module, and the first indicator module are deployed on the vehicle, while the second algorithm module and the second indicator module are deployed on the cloud platform.

6. The battery early warning system according to claim 5, characterized in that, The first indicator module is also used to upload the first warning result to the cloud platform after obtaining the first warning result based on the first warning indicator.

7. The battery warning system according to any one of claims 1-6, characterized in that, The dataset includes first data. The data module is further configured to provide a correspondence between the first data and detection data of the battery's operating status under different vehicle models; and / or, The data module is also used to provide the correspondence between the first data and the detection data of the battery's operating status under different communication protocol versions.

8. The battery warning system according to any one of claims 1-7, characterized in that, The feature set includes one or more of the following: charging segment features, discharging segment features, and resting segment features. The charging segment features are the behavioral characteristics of the battery during constant current or constant voltage charging. The discharging segment features are the behavioral characteristics of the battery during constant current or constant voltage discharging. The resting segment features are the behavioral characteristics of the battery in a no-current state after charging and discharging are completed.

9. A battery early warning method, characterized in that, The method is applied to the battery warning system as described in any one of claims 1 to 8, the battery warning system comprising a data module, an algorithm module, and an indicator module, and the method comprising: The data module acquires target data and calculates target features based on the target data. The algorithm module inputs the target data and / or the target features into the target early warning algorithm to obtain the output data of the target early warning algorithm; The indicator module uses the target early warning indicator corresponding to the target early warning algorithm to evaluate the output data and obtain the target early warning result.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the battery warning method of claim 9.

Images