Battery health status detection methods, systems, terminal devices, and storage media

Abstract

This application relates to the field of battery testing technology, and discloses a method, system, terminal device, and storage medium for detecting battery health status. The method includes: acquiring real-time battery parameters of a power battery; inputting these parameters into a health status prediction model to obtain a health status estimate; calculating the actual usable capacity based on the health status estimate; calculating a first battery health status based on the actual usable capacity; calculating a second battery health status based on the measured equivalent internal resistance of the power battery; determining the comprehensive battery health status based on the first and second battery health statuses; and correcting the comprehensive battery health status to obtain a target battery health status. This battery health status detection method can accurately and efficiently detect battery health status without affecting battery performance.

Description

Technical Field

[0001] This application relates to the field of battery testing technology, and in particular to a method, system, terminal device, and storage medium for detecting battery health status. Background Technology

[0002] The State of Health (SOH) of a power battery is a core parameter for measuring its aging level, remaining lifespan, safety, and residual value. It is a key testing indicator for scenarios such as used car inspection, battery maintenance, and secondary utilization. Currently, the mainstream SOH testing methods in the industry require full charge and discharge operations on the battery, which has limited testing efficiency. Frequent full charge and discharge operations accelerate the aging of the power battery, and the methods are complex to operate, making it difficult to balance accuracy and efficiency, and are not suitable for power batteries of different brands and aging levels. Summary of the Invention

[0003] In view of this, embodiments of this application provide a battery health status detection method, system, terminal device, and storage medium, which can effectively improve the problems of low detection efficiency, high operation complexity, and difficulty in balancing accuracy and efficiency in traditional battery health status detection.

[0004] In a first aspect, embodiments of this application provide a battery health status detection method, including: The real-time battery parameters of the power battery are obtained, and the real-time battery parameters are input into the health status prediction model to predict the estimated health status of the power battery; the health status prediction model is trained based on unlabeled datasets and labeled datasets. The estimated actual usable capacity of the power battery is obtained by multiplying the estimated health status value and the rated battery capacity of the power battery. The estimated actual usable capacity is corrected based on the mapping error correction coefficient of the mapping relationship between the static open-circuit voltage and the state of charge of the power battery, so as to obtain the actual usable capacity of the power battery. The first battery health state of the power battery is calculated based on the actual usable capacity. The second battery health status of the power battery is calculated based on the measured equivalent internal resistance of the power battery. The overall battery health status of the power battery is determined based on the first battery health status and the second battery health status. The error influence factors of the overall battery health status are weighted and summed to obtain the error correction coefficient. The overall battery health status is then corrected based on the error correction coefficient to obtain the target battery health status.

[0005] Secondly, embodiments of this application provide a battery health status detection system, including: The data acquisition module is used to obtain real-time battery parameters of the power battery; The health status prediction module is used to input the real-time battery parameters into the health status prediction model to predict the estimated health status of the power battery; the health status prediction model is trained based on unlabeled datasets and labeled datasets. The first health status calculation module is used to calculate the product of the health status estimate and the rated battery capacity of the power battery to obtain the actual usable capacity estimate of the power battery; the actual usable capacity estimate is corrected based on the mapping error correction coefficient of the mapping relationship between the static open circuit voltage and the state of charge of the power battery to obtain the actual usable capacity of the power battery; and the first battery health status of the power battery is calculated based on the actual usable capacity. The second health status calculation module is used to calculate the second battery health status of the power battery based on the measured equivalent internal resistance of the power battery. A comprehensive health status calculation module is used to determine the comprehensive battery health status of the power battery based on the first battery health status and the second battery health status. The health status correction module is used to perform weighted summation of error influence factors of the overall battery health status to obtain error correction coefficients, and to correct the overall battery health status based on the error correction coefficients to obtain the target battery health status.

[0006] Thirdly, embodiments of this application provide a terminal device, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the above-described battery health status detection method.

[0007] Fourthly, embodiments of this application provide a readable storage medium storing a computer program that executes the above-described battery health status detection method when run on a processor.

[0008] The embodiments of this application have the following beneficial effects: This embodiment of a battery health status detection method includes: acquiring real-time battery parameters of a power battery, predicting an estimated health status of the power battery, and obtaining an estimated health status output by a health status prediction model; calculating the actual usable capacity of the power battery based on the estimated health status, and calculating a first battery health status based on the actual usable capacity; calculating a second battery health status based on the measured equivalent internal resistance of the power battery; determining a comprehensive battery health status based on the first and second battery health status; and correcting the comprehensive battery health status based on an error correction coefficient to obtain a target battery health status. Based on the above scheme, this battery health status detection method does not require the battery to be fully charged or fully discharged. It only requires inputting the collected real-time battery parameters into the health status prediction model to obtain an estimated health status, which in turn allows for the determination of the battery's actual usable capacity. This ensures that battery performance is not affected, is simple to operate, and improves detection efficiency. It supports the detection of the actual usable capacity of power batteries of different brands, types, and aging levels. Furthermore, it can comprehensively calculate the comprehensive battery health status of the power battery based on both capacity and internal resistance, and correct the comprehensive battery health status to obtain a final accurate target battery health status, ensuring the accuracy of battery health status detection. Attached Figure Description

[0009] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 This paper illustrates a first flowchart of a battery health status detection method according to an embodiment of this application. Figure 2 A schematic diagram of voltage differential feature extraction according to an embodiment of this application is shown; Figure 3 This paper illustrates a second flowchart of the battery health status detection method according to an embodiment of this application. Figure 4 This paper illustrates a schematic diagram of incremental capacity feature extraction according to an embodiment of this application. Figure 5 This diagram illustrates the hierarchical structure of the health status prediction model according to an embodiment of this application. Figure 6 A schematic diagram of a battery health status detection system according to an embodiment of this application is shown.

[0011] Explanation of key component symbols: 200 - Battery health status detection system; 210 - Data acquisition module; 220 - Health status prediction module; 230 - First health status calculation module; 240 - Second health status calculation module; 250 - Comprehensive health status calculation module; 260 - Health status correction module. Detailed Implementation

[0012] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0013] The components of the embodiments of this application described and illustrated in the accompanying drawings can be arranged and designed in a variety of different configurations. Therefore, the following detailed description of the embodiments of this application provided in the drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0014] In the following text, the terms "comprising," "having," and their cognates, which may be used in various embodiments of this application, are intended only to indicate a particular feature, number, step, operation, element, component, or combination thereof, and should not be construed as primarily excluding the presence of one or more other features, numbers, steps, operations, elements, components, or combinations thereof, or adding the possibility of one or more combinations thereof. Furthermore, the terms "first," "second," "third," etc., are used only for distinguishing descriptions and should not be construed as indicating or implying relative importance.

[0015] Unless otherwise specified, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which the various embodiments of this application pertain. Terms (such as those defined in commonly used dictionaries) shall be interpreted as having the same meaning as in their contextual meaning in the relevant technical field and shall not be construed as having an idealized or overly formal meaning, unless clearly defined in the various embodiments of this application.

[0016] The following detailed description of some embodiments of this application is provided in conjunction with the accompanying drawings. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0017] Currently, the mainstream SOH (State of Harm) testing method in the industry is the "full charge / discharge capacity method." This method requires charging the power battery from a fully discharged state to a full charge, and then fully discharging it. The actual maximum capacity is obtained by accumulating the discharge amount, and then the SOH is calculated. In addition, there are simplified methods based on internal resistance comparison, but the accuracy is lower and can only roughly determine the battery aging trend.

[0018] Existing technologies have the following prominent problems in practical applications, and are particularly unsuitable for rapid inspection of used cars: The testing time is too long: It takes 2-8 hours to complete a test using the full charge and discharge capacity method. In the scenario of used car inspection, users need to obtain test results quickly. The long testing time leads to extremely low efficiency and cannot meet the needs of batch testing or instant testing. Battery damage during testing: Frequent full charge and discharge cycles accelerate the aging of power batteries, especially for batteries in used cars that already have some degradation, which will further damage battery performance and affect the objectivity of vehicle residual value assessment. High operational complexity: Full charge and discharge testing requires professional personnel to operate specialized charging and discharging equipment, and strict control of parameters such as charge and discharge rate and ambient temperature is required. The operation threshold is high and it is not suitable for rapid on-site testing. Accuracy and efficiency are difficult to balance: The simplified internal resistance method is fast, but it can only roughly estimate SOH through the internal resistance ratio, and the error is usually more than ±8%, which cannot meet the needs of accurate valuation and fault diagnosis of used cars. High data dependence: Traditional data-driven SOH estimation methods require a large amount of complete charge and discharge labeled data, which is costly and time-consuming to collect, and is difficult to adapt to the application needs of a large amount of incomplete charging data in real-world scenarios.

[0019] In existing technologies, some improved solutions attempt to shorten the detection time, but still require partial charge-discharge operations, resulting in limited improvement in detection efficiency and failing to resolve the contradiction between accuracy and efficiency. Furthermore, traditional methods do not fully integrate knowledge of battery electrochemical mechanisms, leading to poor model generalization ability and difficulty in adapting to power batteries of different brands and aging levels. Therefore, there is an urgent need for a rapid SOH detection method that does not require full charge-discharge, has high detection speed, high accuracy, is easy to operate, and is suitable for on-site inspection scenarios of used cars.

[0020] To address the aforementioned issues, this application provides a battery health status detection method, system, terminal device, and storage medium. It eliminates the need for full charge / discharge cycles, is easy to operate, and rapidly infers the battery's current usable capacity by collecting real-time battery parameters and combining them with a knowledge-guided health status prediction model. Furthermore, it calculates and corrects the overall battery health status from multiple perspectives to obtain the target battery health status. This method is suitable for rapid on-site testing of used cars, while simultaneously reducing data collection costs and improving the accuracy of battery health status detection.

[0021] The following describes the battery health status detection method with reference to some specific embodiments.

[0022] Figure 1 A flowchart of a battery health status detection method according to an embodiment of this application is shown. Exemplarily, the battery health status detection method includes the following steps: S110: Obtain the real-time battery parameters of the power battery, input the real-time battery parameters into the health status prediction model, and predict the estimated health status of the power battery.

[0023] For example, real-time battery parameters include one or more of the following: the static open-circuit voltage of the power battery, short-time load parameters, measured equivalent internal resistance, charging segment characteristics, and auxiliary correction characteristics. Short-time load parameters include load voltage and load current; charging segment characteristics include, but are not limited to, the voltage-capacity charging curve, voltage differential characteristics, and incremental capacity characteristics of the charging segment; auxiliary correction characteristics include, but are not limited to, ambient temperature, temperature correction coefficient, and the number of charge-discharge cycles of the power battery.

[0024] In one embodiment, the detection device includes positive and negative electrode detection clips and a BMS (Battery Management System) interface adapter. The positive and negative electrode detection clips are used to connect to the positive and negative terminals of the power battery, and the BMS interface adapter is used to interface with the BMS of the power battery to ensure a stable connection. The power battery is allowed to stand still for a period of time, such as 5 to 10 seconds, to allow the battery to reach a stable state, and the open-circuit voltage and ambient temperature are collected during this period. The open-circuit voltage during this period is the terminal voltage of the power battery in its resting state, and the ambient temperature is the surface temperature of the power battery.

[0025] A constant load current is applied to the power battery for a short period (e.g., 1-5 seconds). The load voltage and load current at the instant the load current is applied, as well as the voltage recovery curve after the load current is removed, are collected. The voltage-capacity charging curve of the battery during the current charging stage is collected to extract incremental capacity characteristics.

[0026] In one embodiment, the static open-circuit voltage and short-time load parameters of the power battery are collected, including the load voltage and load current; the voltage difference between the static open-circuit voltage and the load voltage is calculated, and the ratio of the voltage difference to the load current is calculated to obtain the measured equivalent internal resistance.

[0027] In this embodiment, the measured equivalent internal resistance is the DC equivalent internal resistance of the power battery. Based on Ohm's law, and combined with the static open-circuit voltage, load voltage, and load current, the measured equivalent internal resistance is calculated using the following formula: ; In the formula, This represents the measured equivalent internal resistance. This indicates the open-circuit voltage at rest. Indicates the load voltage. This indicates the load current.

[0028] In one embodiment, the voltage differential characteristic is calculated by differentiating the acquired voltage-capacity charging curve using the following formula:

[0029] In the formula, This represents the voltage change corresponding to a unit change in capacity. Indicates voltage. Indicates capacity, extracts The voltage differential characteristics of the curve, such as peak height, peak voltage position, and peak region area, are used to reflect the degradation of electrochemical reaction activity within the battery. For example... Figure 2 The diagram shown illustrates voltage differential feature extraction. The horizontal axis represents battery capacity, and the vertical axis represents the voltage differential with respect to capacity. The curve is labeled with three core voltage differential features: peak height, peak voltage location, and peak region area.

[0030] In another implementation, incremental capacity feature extraction involves calculating the incremental capacity value based on the voltage-capacity charging curve of a charging segment. This extracts the incremental capacity (IC) feature, which is strongly correlated with battery aging, and serves as a core indicator for knowledge guidance. This process can be completed without requiring a complete charging curve. The incremental capacity value can be calculated using theoretical formulas or practical engineering formulas. An example of a theoretical formula is as follows: ; In the formula, incremental capacity Essentially, it is the derivative of battery capacity with respect to voltage, reflecting the rate of change in the amount of electricity a battery can hold when the voltage changes slightly.

[0031] The calculation formula for practical engineering formulas is as follows: ; In the formula, This represents the change in capacity corresponding to two adjacent voltage sampling points in the voltage-capacity charging curve. This represents the voltage change between two adjacent voltage sampling points in the voltage-capacity charging curve. Since the voltage-capacity charging curve consists of discretely collected voltage and capacity data, precise differentials cannot be directly calculated. However, when the voltage sampling interval is sufficiently small, the approximate differential value approaches the theoretical differential value infinitely, meeting the accuracy requirements for IC feature extraction.

[0032] In one embodiment, the temperature correction factor is calculated as follows: The temperature correction factor is calculated based on the ambient temperature, and the formula for calculating the temperature correction factor is as follows: Where T is the ambient temperature. The standard testing temperature is, for example, 25°C. This is a temperature correction factor used to compensate for the effects of temperature on voltage and internal resistance.

[0033] In one embodiment, such as Figure 3 As shown, the health status prediction model is trained based on unlabeled and labeled datasets. The training process of the health status prediction model includes the following steps: S111, construct an unlabeled dataset, which includes charging segment features of different power batteries.

[0034] S112, a regression model is trained based on an unlabeled dataset to obtain an initial health status prediction model.

[0035] In one embodiment, to address the dependence of traditional data-driven models on large amounts of labeled data, reduce data acquisition costs, and simultaneously improve model generalization ability and interpretability, this application employs a two-stage transfer learning strategy to train a knowledge-guided deep learning model. The first stage is an unsupervised pre-training stage based on unlabeled datasets. The core objective is to enable the model's feature extractor to implicitly grasp the degradation patterns of the power battery without relying on labeled data, thereby reducing the need for labeled data in subsequent fine-tuning stages.

[0036] First, a large number of incomplete charging curves of power batteries of different brands, aging levels, and types were collected. Without labeling the State of Harm (SOH) values, only the charging segment features were retained. The voltage-capacity charging curve from the charging segment features was selected. Each incomplete charging curve was preprocessed to remove abnormal noise data. The 3σ criterion was used to eliminate abnormal voltage and capacity fluctuations. Simultaneously, the incremental capacity features of the charging segments were selected to obtain the final unlabeled dataset.

[0037] Then, the unlabeled dataset is input into the model's feature extractor and regressor. The feature extractor and regressor are trained using the fitting accuracy of the incremental capacity features as the loss function (e.g., mean squared error, MSE). The training iterations are 50-100 epochs, with a learning rate of 1e-4 and a batch size of 32, until the loss function converges (e.g., a convergence threshold of 1e-5). Finally, after training, the feature extractor can automatically extract incremental capacity features strongly correlated with battery aging, enabling preliminary identification of battery degradation states without labeled data, laying the foundation for subsequent SOH estimation. Figure 4The diagram shown is a schematic diagram of incremental capacity feature extraction, with the aging-sensitive peak area marked, indicating that the incremental capacity feature is strongly positively correlated with the degree of aging.

[0038] S113, construct a labeled dataset, which includes real-time battery parameters and corresponding real battery health status for different power batteries.

[0039] In this embodiment, the labeled dataset corresponds to the actual health status values ​​of the batteries, providing a calibration benchmark for model fine-tuning. The actual health status values ​​range from 50.0% to 100.0%, covering all battery aging levels, including excellent, good, average, and requiring replacement. The actual health status values ​​can be measured using the traditional full charge-discharge method, ensuring the accuracy of the labeled dataset and avoiding model fine-tuning failure due to label bias.

[0040] Taking the labeled data for ternary lithium batteries as an example, the real-time battery parameters of the complete labeled data include: static open-circuit voltage of 3.65V, load voltage of 3.58V, load current of 10A, measured equivalent internal resistance of 7mΩ, ambient temperature of 25℃, temperature correction factor of 1.0, peak height of 0.8V / Ah, peak voltage of 3.6V, and peak area of ​​1.2V. The voltage-capacity charging curve for the Ah charging segment includes 10 sampling points. The complete labeled data also includes the actual battery health status. Optionally, the complete labeled data may also include auxiliary labels: rated capacity 50Ah, standard internal resistance 8mΩ, battery type is ternary lithium, service life is 1 year, and charge-discharge cycle count is 320 times.

[0041] In one implementation, the number of labeled datasets is 50 to 300, which is sufficient to meet the fine-tuning requirements without the need for a large number of samples. This solves the data dependency problem of traditional models, and the samples need to cover different aging levels, temperatures, and battery types to avoid model overfitting.

[0042] S114, train the initial health status prediction model based on the labeled dataset to obtain the health status prediction model.

[0043] In one implementation, the second stage is a fine-tuning stage based on a small amount of labeled data: the core objective is to adapt the initial health status prediction model to the SOH estimation task, and further improve the SOH estimation accuracy of the model by fine-tuning it with a small amount of labeled data.

[0044] Specific steps: First, replace the top-level regressor of the model with a Somatic Health (SOH) regressor. The SOH regressor can use two fully connected layers, with 64 neurons in the hidden layer and 1 neuron in the output layer. The activation function is the Sigmoid function, mapping the output to the 0%-100% range. Second, freeze most parameters of the feature extractor, such as unfreezing the parameters of the last three convolutional layers, to prevent the destruction of effective features obtained during pre-training. Then, input the labeled dataset into the model, using the root mean square error (RMSE) between the model's output SOH estimate and the true SOH value as the loss function for fine-tuning training. Fine-tuning iterations are performed for 20-50 rounds, with a learning rate set to 1e-5 (lower than the pre-training stage to avoid parameter oscillations). The batch size is set to 16 until the loss function converges (e.g., a convergence threshold of 5e-6). Finally, after fine-tuning, the model can quickly process real-time battery parameters, outputting high-precision health state estimates, and requires only a small amount of labeled data to achieve the preset accuracy requirements, significantly reducing the cost and cycle time of data collection.

[0045] In another embodiment, such as Figure 5 The diagram shows the hierarchical structure of the health status prediction model, which consists of four parts from top to bottom: input layer, feature extractor, regressor, and output layer. The input layer contains real-time battery parameters; the feature extractor consists of 13 stacked convolutional modules, each labeled as containing a convolutional layer, a batch normalization layer, and a ReLU activation layer, used to automatically extract aging-related features of the charging segment; the regressor consists of a global average pooling layer and a fully connected layer, used to receive the features output by the feature extractor and output a health status estimate; the output layer outputs the health status estimate.

[0046] S120 calculates the product of the health status estimate and the rated battery capacity of the power battery to obtain the estimated actual usable capacity of the power battery.

[0047] S130, the estimated actual usable capacity is corrected based on the mapping error correction coefficient of the mapping relationship between the static open-circuit voltage and the state of charge of the power battery, so as to obtain the actual usable capacity of the power battery, and the first battery health state of the power battery is calculated based on the actual usable capacity.

[0048] Demonstratively, a unique and stable mapping relationship exists between the open-circuit voltage and the state of charge (SOC) of a power battery. This relationship is an inherent characteristic of power batteries, and the mapping curve differs for different battery types. It can be pre-calibrated experimentally; that is, the corresponding SOC can be accurately determined from the current open-circuit voltage. For example, for a ternary lithium battery, an open-circuit voltage of 3.65V corresponds to an 80% SOC, 3.70V corresponds to a 90% SOC, and 3.58V corresponds to a 70% SOC, forming a complete mapping table. The SOC is the ratio of the current remaining usable capacity to the current actual usable capacity, i.e.: ; In the formula, Indicates the state of charge. Indicates the current remaining available capacity. This indicates the current actual available capacity.

[0049] In one embodiment, by combining the health status estimate and the rated battery capacity, a preliminary estimate of the actual usable capacity is first calculated. The formula for calculating the actual usable capacity estimate is as follows: ; In the formula, This represents the estimated actual available capacity. This represents an estimate of health status. This indicates the rated battery capacity.

[0050] Then, based on the definition formula of the state of charge obtained from the mapping, the current remaining available capacity is deduced. The formula for calculating the current remaining available capacity is: ; In the formula, This represents the state of charge (SCC) obtained by mapping the open-circuit voltage of the power battery to its actual state of charge (SCC). Based on the obtained current remaining usable capacity and the battery's true SCC, the actual usable capacity is derived in reverse. The formula for calculating the actual usable capacity is as follows:

[0051] ; Simplified accurate formula:

[0052] ; Further simplification: ; This formula is the one used to correct the estimated actual available capacity based on the mapping error correction coefficient. The actual available capacity is obtained by multiplying the mapping error correction coefficient by the estimated actual available capacity. Indicates the battery's true state of charge. The mapping error correction coefficient is the mapping relationship between the open-circuit voltage and the state of charge of a power battery. It is a small deviation generated when the state of charge and the open-circuit voltage are mapped to each other. It is used to eliminate the error between the actual state of charge and the mapped state of charge during the mapping process. , The value ranges from 0.99 to 1.01 and is used to correct minor errors in the mapping process.

[0053] In this embodiment, the estimated value of actual available capacity is calibrated by the mapping error correction coefficient between open-circuit voltage and state of charge, eliminating the small error in the mapping between open-circuit voltage and state of charge, ensuring the back-calculation accuracy of actual available capacity, and meeting the accuracy requirements for rapid detection of actual available capacity.

[0054] Exemplarily, the first battery health state is a quantitative indicator characterizing the degree to which the current actual usable capacity of a power battery is maintained relative to its factory rated capacity. In one embodiment, the percentage of actual usable capacity to the rated capacity of the power battery is calculated to obtain the state characterization value of the first battery health state. The formula for calculating the state characterization value of the first battery health state is as follows: ; In the formula, The state characterization value represents the health status of the first battery. This indicates the rated capacity, which is the capacity of the battery when it leaves the factory.

[0055] S140 calculates the second battery health status of the power battery based on the measured equivalent internal resistance of the power battery.

[0056] Exemplary, the second battery health state is an indicator for assessing the degree of battery aging based on changes in internal resistance. In one embodiment, the percentage of the standard internal resistance and the measured equivalent internal resistance of the power battery is calculated to obtain the state characterization value of the second battery health state. The formula for calculating the state characterization value of the second battery health state is: ; In the formula, The state characterization value represents the health status of the second battery. This represents the measured equivalent internal resistance. This indicates the standard internal resistance, which is the standard DC internal resistance of the power battery and can be obtained from the parameters provided by the manufacturer.

[0057] S150, determine the overall battery health status of the power battery based on the health status of the first battery and the health status of the second battery.

[0058] In one embodiment, the product of the state representation value of the first battery health state and the first preset weight is calculated to obtain the first weighted state representation value; the product of the state representation value of the second battery health state and the second preset weight is calculated to obtain the second weighted state representation value, and the sum of the first preset weight and the second preset weight is one; the sum of the first weighted state representation value and the second weighted state representation value is used as the state representation value of the comprehensive battery health state.

[0059] In this embodiment, a weighted fusion formula is used to combine the first battery health state and the second battery health state to obtain the comprehensive battery health state. The calculation formula for the state characterization value of the comprehensive battery health state is as follows: ; In the formula, This represents the overall battery health status value. Indicates the first preset weight. This represents the second preset weight, where α ranges from 0.6 to 0.7, and β ranges from 0.3 to 0.4. The first and second preset weights can be adaptively adjusted according to different brands and types of batteries.

[0060] The preset weights reflect the relative contribution of different health characterization dimensions to the overall battery degradation state. The battery health state is calculated using capacity-type and impedance-type indicators, achieving error cancellation and dynamic compensation, and improving the accuracy of the overall battery health state calculation.

[0061] S160, the error influence factors of the overall battery health status are weighted and summed to obtain the error correction coefficient. The overall battery health status is corrected based on the error correction coefficient to obtain the target battery health status.

[0062] In one embodiment, the error influencing factors include one or more of the following: static open-circuit voltage error, voltage differential characteristic error, ambient temperature error, battery aging degree, battery type identification, and load detection error.

[0063] For example, the static open-circuit voltage error is the difference between the actual sampled static open-circuit voltage and the true static open-circuit voltage, determined by the accuracy of the static open-circuit voltage acquisition equipment. The voltage differential characteristic error is the deviation between the calculated value and the true value of the voltage differential characteristic, determined by the voltage sampling interval. The ambient temperature error is the difference between the current detection ambient temperature and the standard temperature, used to correct for the influence of temperature on battery characteristics. The battery aging degree can be set as a preliminary health status estimate to adapt to the error correction rules of different aging stages; the more severe the aging, the greater the battery aging degree. The battery type identifier is set as a binary identifier to match the error correction parameters for different battery types; for example, the battery type identifier for lithium iron phosphate batteries is 0, and the battery type label for ternary lithium batteries is 1. The load detection error is the error between the actual short-time load parameters and the true short-time load parameters, used to correct the influence of load parameters on the calculation of battery health status.

[0064] In one embodiment, during the weighted summation process, the error impact factors are first normalized. This can be achieved by minimizing and maximizing each error impact factor to the 0-1 range. Then, the six types of error impact factors are weighted and fused to calculate the error correction coefficient. The formula for calculating the error correction coefficient is as follows: ; In the formula, This represents the error correction factor. This represents the normalized resting open-circuit voltage error. This represents the weighting coefficient corresponding to the static open-circuit voltage error; This represents the normalized voltage differential characteristic error. This represents the weighting coefficient corresponding to the voltage differential characteristic error; This represents the normalized ambient temperature error. This represents the weighting coefficient corresponding to the ambient temperature error. This indicates the normalized degree of battery aging. This indicates the weighting coefficient corresponding to the degree of battery aging. This indicates the normalized battery type identifier. This indicates the weighting coefficient corresponding to the battery type identifier; This represents the normalized load detection error. This represents the weighting coefficient corresponding to the load detection error. Wherein, .

[0065] In another embodiment, the product of the error correction coefficient and the state characterization value of the overall battery health state is calculated to obtain the battery health state correction amount; the sum of the state characterization value of the overall battery health state and the battery health state correction amount is calculated to obtain the state characterization value of the initial battery health state; the state characterization value of the initial battery health state is limited to a preset calibration range to obtain the state characterization value of the target battery health state.

[0066] As an example, the formula for calculating the initial battery health state characterization value is as follows: ; In the formula, The state characterization value represents the initial battery health state.

[0067] In one embodiment, the preset calibration range is a reasonable numerical range set to ensure that the battery health status output conforms to physical reality. The initial battery health status characterization value needs to be controlled within the preset calibration range, such as 50.0% to 100.0%. If it exceeds or falls below the range, the boundary value is taken to ensure that the output result conforms to the actual aging law of the power battery and avoid invalid values.

[0068] In another embodiment, the final target battery health status can be output, along with auxiliary parameters such as detection time, real-time battery parameters, and battery aging level. A target battery health status of 100%~90% is considered excellent, 90%-80% is considered good, 80%-70% is considered average, and a target battery health status of less than or equal to 70% indicates that it needs to be replaced, for user reference.

[0069] Example 1: Taking a certain brand of ternary lithium battery (rated capacity of 50Ah, standard internal resistance of new battery is 8mΩ) as an example, the SOH test was performed using the battery health status detection method of this application. The specific steps are as follows: Test preparation: Connect the positive and negative test clips of the test device to the positive and negative terminals of the battery, connect the BMS interface adapter to the BMS interface of the battery, let the battery stand for 8 seconds, and the ambient temperature T=25℃. Real-time battery parameter acquisition: The static open-circuit voltage is 3.65V. A load current of 0.2C (10A) is applied for 3 seconds. The load voltage is 3.58V and the load current is 10A. At the same time, the voltage-capacity data of the charging segment are acquired. Key parameter calculations: DC internal resistance is calculated as (3.65-3.58) / 10=7mΩ; temperature correction coefficient is calculated as 1+0.01×(25-25)=1.0; voltage differential characteristics of the charging segment are extracted, with a peak height of 0.8V / Ah, a peak voltage of 3.6V, and a peak area of ​​1.2V·Ah. Model prediction and feature fusion: Real-time battery parameters are input into the trained health status prediction model, and the estimated health status value is 92.5%. Based on the overall State of Health (SOH) calculation: the current actual usable capacity is estimated to be 46.25 Ah, the first battery's health status is 92.5%, and the second battery's health status is 114.3% (the upper limit of 100% is used because the internal resistance is lower than the standard value); the overall battery health status is 95.1%. Error calibration: After error correction, the final target battery health status is 95.0%; Output results: The target battery health status is output. The detection time is 1 minute and 20 seconds, the ambient temperature is 25℃, the measured equivalent internal resistance is 7mΩ, and the aging level is excellent, which is consistent with the actual aging status of the battery (1 year of use, slight degradation).

[0070] Example 2, taking a certain brand of lithium iron phosphate battery (rated capacity of 60Ah, standard internal resistance of 10mΩ for new batteries) as an example, the specific steps are as follows: Test preparation: Connect the test device to the battery, let the battery stand for 10 seconds, and the ambient temperature T=15℃; Parameter acquisition: The static open circuit voltage was 3.2V. A 0.15C (9A) load was applied for 4 seconds. The load voltage was 3.12V and the load current was 9A. The voltage-capacity data of the charging segment were also acquired. Key parameter calculations: DC internal resistance is calculated as (3.65-3.58) / 10 = 7mΩ; temperature correction coefficient is calculated as 1 + 0.01 × (25-25) = 1.0; the voltage differential characteristics of the charging segment are extracted, with a peak height of 0.7V / Ah, a peak voltage of 3.2V, and a peak area of ​​1.0V·Ah. Model prediction and feature fusion: Real-time battery parameters are input into the trained health status prediction model, and the estimated health status value is 88.2%. Based on the overall State of Health (SOH) calculation: the current actual usable capacity is estimated to be 52.92 Ah, the first battery's health status is 88.2%, the second battery's health status is 112.5%, and we take the upper limit of 100%; the overall battery health status is 93.3%. Error calibration: After calibration, the final target battery health status is 93.2%; Output results: The target battery health status is output. The detection time is 1 minute and 40 seconds, the ambient temperature is 15℃, the measured equivalent internal resistance is 8.89mΩ, and the aging level is good, which is consistent with the actual aging status of the battery (2 years of use, moderate degradation).

[0071] Example 3: Rapid SOH Detection of Aging Batteries (Comparative Verification) A ternary lithium battery (rated capacity 50Ah, actual full charge / discharge test battery health status 82.0%) that has been used for 3 years was selected. The battery health status detection method of this application was compared with the traditional full charge / discharge method and the traditional internal resistance method. The results are as follows: Table 1: Comparison of the battery health status detection method of this application with the full charge / discharge method and the traditional internal resistance method.

[0072] As shown in Table 1, the battery health status detection method of this application has a detection speed much faster than the traditional full charge and discharge method, and the detection accuracy is basically the same as that of the traditional full charge and discharge method, without causing battery damage; compared with the traditional internal resistance method, the detection accuracy is greatly improved, which fully demonstrates the advantages of this application.

[0073] Figure 6 A schematic diagram of a battery health status detection system 200 according to an embodiment of this application is shown. Exemplarily, the battery health status detection system 200 includes: The data acquisition module 210 is used to acquire real-time battery parameters of the power battery.

[0074] In one embodiment, the data acquisition module 210 is also used to acquire the static open-circuit voltage and short-time load parameters of the power battery, the short-time load parameters including the load voltage and the load current; calculate the voltage difference between the static open-circuit voltage and the load voltage, calculate the ratio of the voltage difference to the load current, and obtain the measured equivalent internal resistance. The health status prediction module 220 is used to input real-time battery parameters into the health status prediction model to obtain the health status estimate output by the health status prediction model; the health status prediction model is trained based on unlabeled datasets and labeled datasets.

[0075] The first health state calculation module 230 is used to calculate the product of the health state estimate and the rated battery capacity of the power battery to obtain the estimated actual usable capacity of the power battery; the estimated actual usable capacity is corrected based on the mapping error correction coefficient of the mapping relationship between the static open circuit voltage and the state of charge of the power battery to obtain the actual usable capacity of the power battery; and the first battery health state of the power battery is calculated based on the actual usable capacity.

[0076] In one embodiment, the first health status calculation module 230 is further used to calculate the product of the mapping error correction coefficient and the estimated actual available capacity to obtain the actual available capacity.

[0077] In one embodiment, the first health state calculation module 230 is further used to calculate the percentage of the actual available capacity and the rated capacity of the power battery to obtain the state characterization value of the first battery health state.

[0078] The second health state calculation module 240 is used to calculate the second battery health state of the power battery based on the measured equivalent internal resistance of the power battery.

[0079] In one embodiment, the second health state calculation module 240 is further used to calculate the percentage of the standard internal resistance of the power battery to the measured equivalent internal resistance, so as to obtain the state characterization value of the second battery health state.

[0080] The comprehensive health status calculation module 250 is used to determine the comprehensive battery health status of the power battery based on the health status of the first battery and the health status of the second battery.

[0081] In one embodiment, the comprehensive health status calculation module 250 is further configured to calculate the product of the state representation value of the first battery health status and the first preset weight to obtain a first weighted state representation value; calculate the product of the state representation value of the second battery health status and the second preset weight to obtain a second weighted state representation value, wherein the sum of the first preset weight and the second preset weight is one; and use the sum of the first weighted state representation value and the second weighted state representation value as the state representation value of the comprehensive battery health status.

[0082] The health status correction module 260 is used to perform weighted summation of error influence factors of the overall battery health status to obtain error correction coefficients, and to correct the overall battery health status based on the error correction coefficients to obtain the target battery health status.

[0083] In one embodiment, the health status correction module 260 is further configured to perform a weighted summation of the error influence factors of the overall battery health status to obtain an error correction coefficient; wherein, the error influence factors include one or more of the following: static open circuit voltage error, voltage differential characteristic error, ambient temperature error, battery aging degree, battery type identification, and load detection error.

[0084] In another embodiment, the health status correction module 260 is further configured to calculate the product of the error correction coefficient and the state characterization value of the overall battery health status to obtain the battery health status correction amount; calculate the sum of the state characterization value of the overall battery health status and the battery health status correction amount to obtain the state characterization value of the initial battery health status; and limit the state characterization value of the initial battery health status within a preset calibration range to obtain the state characterization value of the target battery health status.

[0085] It is understood that the system in this embodiment corresponds to the battery health status detection method in the above embodiment, and the options in the above embodiment are also applicable to this embodiment, so they will not be described again here.

[0086] This application also provides a terminal device, exemplary of which includes a processor and a memory, wherein the memory stores a computer program, and the processor executes the computer program to enable the terminal device to perform the functions of the various modules in the above-described battery health status detection method or battery health status detection system.

[0087] The processor can be an integrated circuit chip with signal processing capabilities. The processor can be a general-purpose processor, including at least one of a Central Processing Unit (CPU), Graphics Processing Unit (GPU), Network Processor (NP), Digital Signal Processor (DSP), Application-Specific Integrated Circuit (ASIC), Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. The general-purpose processor can be a microprocessor or any conventional processor, capable of implementing or executing the methods, steps, and logic block diagrams disclosed in the embodiments of this application.

[0088] Memory can be, but is not limited to, Random Access Memory (RAM), Read Only Memory (ROM), Programmable Read-Only Memory (PROM), Erasable Programmable Read-Only Memory (EPROM), and Electrically Erasable Programmable Read-Only Memory (EEPROM). Memory is used to store computer programs, and the processor can execute these programs upon receiving execution instructions.

[0089] This application also provides a computer-readable storage medium for storing computer programs used in the aforementioned terminal devices. For example, the computer-readable storage medium may include, but is not limited to, various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0090] In the several embodiments provided in this application, it should be understood that the disclosed systems and methods can also be implemented in other ways. The apparatus embodiments described above are merely illustrative; for example, the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that, in alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0091] In addition, the functional modules or units in the various embodiments of this application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0092] If a function is implemented as a software module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a smartphone, personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application.

[0093] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A battery state of health detection method, characterized by, include: The real-time battery parameters of the power battery are obtained, and the real-time battery parameters are input into the health status prediction model to predict the estimated health status of the power battery. The health status prediction model was trained based on unlabeled and labeled datasets; The estimated actual usable capacity of the power battery is obtained by multiplying the estimated health status value and the rated battery capacity of the power battery. The estimated actual usable capacity is corrected based on the mapping error correction coefficient of the mapping relationship between the static open-circuit voltage and the state of charge of the power battery, so as to obtain the actual usable capacity of the power battery. The first battery health state of the power battery is calculated based on the actual usable capacity. The second battery health status of the power battery is calculated based on the measured equivalent internal resistance of the power battery. The overall battery health status of the power battery is determined based on the first battery health status and the second battery health status. The error influence factors of the overall battery health status are weighted and summed to obtain the error correction coefficient. The overall battery health status is then corrected based on the error correction coefficient to obtain the target battery health status. The error influencing factors include one or more of the following: static open-circuit voltage error, voltage differential characteristic error, ambient temperature error, battery aging degree, battery type identification, and load detection error. The correction of the overall battery health status based on the error correction coefficient includes: The battery health state correction amount is obtained by multiplying the error correction coefficient and the state characterization value of the comprehensive battery health state. The sum of the state characterization value of the overall battery health state and the battery health state correction amount is calculated to obtain the state characterization value of the initial battery health state. The initial battery health status is limited to a preset calibration range to obtain the target battery health status.

2. The battery state of health detection method of claim 1, wherein The mapping error correction coefficient based on the mapping relationship between the static open-circuit voltage and the state of charge of the power battery corrects the estimated actual usable capacity, including: The actual available capacity is obtained by multiplying the mapping error correction coefficient by the estimated actual available capacity.

3. The battery state of health detection method of claim 1, wherein, The calculation of the first battery health status of the power battery based on the actual available capacity includes: The percentage of the actual usable capacity to the rated capacity of the power battery is calculated to obtain the state characterization value of the first battery's health status.

4. The battery health status detection method according to claim 1, characterized in that, The real-time battery parameters include the measured equivalent internal resistance. The acquisition of the real-time battery parameters of the power battery includes: Collect the static open-circuit voltage and short-time load parameters of the power battery, wherein the short-time load parameters include load voltage and load current; Calculate the voltage difference between the static open-circuit voltage and the load voltage, and calculate the ratio of the voltage difference to the load current to obtain the measured equivalent internal resistance; The calculation of the second battery health state of the power battery based on the measured equivalent internal resistance of the power battery includes: The percentage of the standard internal resistance of the power battery to the measured equivalent internal resistance is calculated to obtain the state characterization value of the second battery's health status.

5. The battery health status detection method according to claim 1, characterized in that, Determining the overall battery health status of the power battery based on the first battery health status and the second battery health status includes: Calculate the product of the state representation value of the first battery health state and the first preset weight to obtain the first weighted state representation value; The second weighted state characterization value is obtained by multiplying the state characterization value of the second battery health state with the second preset weight, and the sum of the first preset weight and the second preset weight is one. The sum of the first weighted state representation value and the second weighted state representation value is used as the state representation value of the comprehensive battery health state.

6. The battery health status detection method according to claim 1, characterized in that, The training process of the health status prediction model includes: Construct the unlabeled dataset, which includes charging segment features of different power batteries; A regression model is trained based on the unlabeled dataset to obtain an initial health status prediction model. Construct the labeled dataset, which includes real-time battery parameters and corresponding real battery health states for different power batteries; The initial health status prediction model is trained based on the labeled dataset to obtain the health status prediction model; The real-time battery parameters include one or more of the following: static open-circuit voltage, short-time load parameters, measured equivalent internal resistance, charging segment characteristics, and auxiliary correction characteristics of the power battery.

7. A battery health status detection system, characterized in that, include: The data acquisition module is used to obtain real-time battery parameters of the power battery; The health status prediction module is used to input the real-time battery parameters into the health status prediction model and predict the estimated health status of the power battery. The health status prediction model was trained based on unlabeled and labeled datasets; The first health status calculation module is used to calculate the product of the health status estimate and the rated battery capacity of the power battery to obtain the estimated actual usable capacity of the power battery. The estimated actual usable capacity is corrected based on the mapping error correction coefficient of the mapping relationship between the static open-circuit voltage and the state of charge of the power battery, so as to obtain the actual usable capacity of the power battery. The first battery health state of the power battery is calculated based on the actual usable capacity. The second health status calculation module is used to calculate the second battery health status of the power battery based on the measured equivalent internal resistance of the power battery. A comprehensive health status calculation module is used to determine the comprehensive battery health status of the power battery based on the first battery health status and the second battery health status. The health status correction module is used to perform weighted summation of error influence factors of the overall battery health status to obtain error correction coefficients, and to correct the overall battery health status based on the error correction coefficients to obtain the target battery health status. The error influencing factors include one or more of the following: static open circuit voltage error, voltage differential characteristic error, ambient temperature error, battery aging degree, battery type identification, and load detection error. The health status correction module is also used to calculate the product of the error correction coefficient and the state characterization value of the comprehensive battery health status to obtain the battery health status correction amount. The sum of the state characterization value of the overall battery health state and the battery health state correction amount is calculated to obtain the state characterization value of the initial battery health state. The initial battery health status is limited to a preset calibration range to obtain the target battery health status.

8. A terminal device, characterized in that, The device includes a memory and a processor, the memory storing a computer program and the processor executing the computer program to implement the battery health status detection method according to any one of claims 1-6.

9. A readable storage medium, characterized in that, It stores a computer program that, when run on a processor, executes the battery health status detection method according to any one of claims 1 to 6.

Images