Battery health state determination method and device, equipment, medium and program product

Abstract

The embodiment of the invention provides a battery health state determination method and device, equipment, a medium and a program product, and relates to the technical field of battery measurement and calculation. The method comprises the following steps: acquiring state monitoring data of a to-be-detected battery in a charging process; determining a key feature set based on the state monitoring data; wherein the key feature set comprises a plurality of key features; inputting the key feature set into a trained battery health state prediction model to obtain a battery health state prediction result output by the battery health state prediction model; and determining the battery health state of the to-be-tested battery based on the battery health state prediction result. According to the embodiment of the invention, the state monitoring data in the charging process of the battery is collected, the key features are extracted, and the internal relation between the key features and the health state of the battery is learned through the training model, so that the prediction accuracy of the health state of the battery can be improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of battery estimation, in particular to a battery health state determination method, device, equipment, medium and program product. BACKGROUND

[0002] Battery SOH (State of Health) is usually defined as the percentage of the current capacity to the initial capacity. The Ah integration method is a commonly used basic method in battery state estimation (including SOC and SOH), and its core principle is to calculate the battery state change by accumulating the integral of current and time of charging and discharging.

[0003] However, the scheme of estimating battery state by Ah integration method has the following disadvantages: 1. Single parameter, unable to capture multi-factor coupling effect, low estimation accuracy; 2. Due to the error of sensor measurement, small current measurement error will be amplified to significant capacity error through the integration process; 3. Under dynamic current working conditions (such as vehicle start-stop, load mutation, etc.), rapid change of current will cause integration delay or sampling error, further reducing the estimation accuracy.

[0004] In summary, there is an urgent need for a scheme that can improve the accuracy of battery capacity estimation. SUMMARY

[0005] The purpose of the embodiments of the present application is to provide a battery health state determination method, device, equipment, medium and program product to improve the accuracy of battery capacity estimation.

[0006] In a first aspect, the embodiments of the present application provide a battery health state determination method, comprising: obtaining state monitoring data of a battery to be measured in a charging process; determining a key feature set based on the state monitoring data; wherein the key feature set includes a plurality of key features; inputting the key feature set into a trained battery health state prediction model to obtain a battery health state prediction result output by the battery health state prediction model; determining the battery health state of the battery to be measured based on the battery health state prediction result.

[0007] In the embodiments of the present application, by collecting the state monitoring data in the battery charging process and extracting a plurality of key features, the internal relationship between the plurality of key features and the battery health state is learned through model training, thereby improving the accuracy of battery health state prediction.

[0008] In some embodiments, the determining the battery health state of the battery to be measured based on the battery health state prediction result comprises: Based on the state monitoring data, the current battery capacity of the battery under test is calculated by using the ampere-hour integration method; Based on the current battery capacity and the initial battery capacity of the battery under test, the reference battery health state of the battery under test is determined; The battery health state of the battery under test is determined in combination with the battery health state prediction result and the reference battery health state.

[0009] In the embodiments of the present application, the final battery health state is determined in combination with the model prediction result and the ampere-hour integration estimation result, which further improves the accuracy and reliability of battery health state prediction.

[0010] In some embodiments, the inputting of the key feature set into the trained battery health state prediction model to obtain the battery health state prediction result output by the battery health state prediction model comprises: A plurality of key feature sets are inputted into the trained battery health state prediction model respectively to obtain a plurality of corresponding battery health state candidate prediction results; wherein the plurality of key feature sets are determined based on a plurality of groups of state monitoring data respectively, and the plurality of groups of state monitoring data are respectively acquired based on the battery under test in a plurality of charging processes; The battery health state prediction result is determined based on the plurality of battery health state candidate prediction results.

[0011] In the embodiments of the present application, a plurality of groups of data are acquired by repeating a plurality of charging processes and extracting key features, and then a plurality of corresponding candidate prediction results are predicted respectively, and finally the final battery health state prediction result is determined by comprehensively determining these candidate prediction results, which further improves the accuracy and reliability of battery health state prediction.

[0012] In some embodiments, the determination of the battery health state of the battery under test based on the battery health state prediction result comprises: The credibility of the battery health state prediction result is determined based on the standard deviation of the plurality of battery health state candidate prediction results; In the case that the credibility of the battery health state prediction result is lower than a preset credibility threshold, the current battery capacity of the battery under test is calculated by using the ampere-hour integration method based on the state monitoring data; The battery health state of the battery under test is determined based on the current battery capacity and the initial battery capacity of the battery under test.

[0013] In the embodiments of the present application, the confidence of the prediction result is determined according to the standard deviation of the plurality of candidate prediction results, and when the confidence is low, the model prediction result is discarded, and the ampere-hour integration method is used to estimate the battery health state, thereby further improving the accuracy and reliability of the battery health state prediction.

[0014] In some embodiments, the state monitoring data includes voltage state data, current state data, and temperature state data. The plurality of key features include at least two of constant current charging time, voltage platform slope, temperature change amount, and charging efficiency.

[0015] In the embodiments of the present application, by collecting voltage, current, and temperature state data, and generating constant current charging time, voltage platform slope, temperature change amount, and charging efficiency, the accuracy of the battery health state prediction is further improved.

[0016] In some embodiments, the determination method of the constant current charging time includes determining the constant current charging phase of the battery under test based on the current state data, and determining the constant current charging time based on the duration of the constant current charging phase. The determination method of the voltage platform slope includes determining the voltage change amount of the battery under test from the stable voltage to the highest voltage based on the voltage state data, and determining the voltage platform slope based on the voltage change amount. The determination method of the temperature change amount includes determining the temperature change amount based on the maximum temperature value and the minimum temperature value of the temperature state data. The determination method of the charging efficiency includes obtaining the electric energy consumed by the battery under test during the charging process, and obtaining the electric quantity increase value of the battery under test during the charging process, and determining the charging efficiency based on the ratio of the electric quantity increase value to the electric energy.

[0017] In the embodiments of the present application, by calculating the corresponding key features based on the voltage, current, and temperature state data in a specific manner, the accuracy of the battery health state prediction is further improved.

[0018] In some embodiments, the determination of the key feature set based on the state monitoring data includes: Obtaining a plurality of groups of state monitoring data corresponding to the battery under test in a plurality of charging processes; Determining a plurality of reference key feature sets corresponding to the plurality of groups of state monitoring data; Determining the key feature set based on the plurality of reference key feature sets using an average algorithm.

[0019] In the embodiments of the present application, by acquiring multiple sets of state monitoring data and extracting features respectively, and then using an average algorithm to determine the final key feature set from multiple reference key feature sets, the accuracy of battery health state prediction is further improved.

[0020] In some embodiments, the inputting the key feature set into the trained battery health state prediction model to obtain a battery health state prediction result output by the battery health state prediction model comprises: acquiring a first key feature set and a second key feature set; wherein the first key feature set is determined based on state monitoring data of the battery to be tested in a fast charging scenario charging process, and the second key feature set is determined based on state monitoring data of the battery to be tested in a slow charging scenario charging process, the charging power of the fast charging scenario being greater than that of the slow charging scenario; inputting the first key feature set into a trained first battery health state prediction model to obtain a first battery health state prediction result output by the first battery health state prediction model; wherein the first battery health state prediction model is trained based on pre-collected fast charging scenario sample data; inputting the second key feature set into a trained second battery health state prediction model to obtain a second battery health state prediction result output by the second battery health state prediction model; wherein the second battery health state prediction model is trained based on pre-collected slow charging scenario sample data; combining the first battery health state prediction result and the second battery health state prediction result to determine the battery health state prediction result.

[0021] In the embodiments of the present application, by collecting data for fast charging scenarios and slow charging scenarios respectively, and using models trained for different scenarios for prediction respectively, and finally determining the final prediction result by combining the prediction results of the two, the accuracy of battery health state prediction is further improved.

[0022] In a second aspect, the embodiments of the present application provide a battery health state determination apparatus, comprising: a data acquisition module configured to acquire state monitoring data of a battery to be tested in a charging process; a feature determination module configured to determine a key feature set based on the state monitoring data; wherein the key feature set comprises multiple key features; a result prediction module configured to input the key feature set into a trained battery health state prediction model to obtain a battery health state prediction result output by the battery health state prediction model; A state determination module is configured to determine the battery health state of the battery under test based on the battery health state prediction result.

[0023] In a third aspect, an electronic device is provided, which includes a memory, a processor, and a computer program stored in the memory and capable of running on the processor, wherein the processor implements the method of any of the embodiments of the first aspect when running the program.

[0024] In a fourth aspect, a computer readable storage medium is provided, which stores a computer program, and the computer program, when run by a processor, implements the method of any of the embodiments of the first aspect.

[0025] In a fifth aspect, a computer program product is provided, which includes a computer program, and the computer program, when run by a processor, implements the method of any of the embodiments of the first aspect. BRIEF DESCRIPTION OF DRAWINGS

[0026] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments of the present application. It should be understood that the following drawings only show some embodiments of the present application, and therefore should not be regarded as a limitation on the scope, and for those skilled in the art, other related drawings can also be obtained without creative labor on the basis of these drawings.

[0027] Figure 1 A flowchart of a battery health state determination method provided by the embodiments of the present application; Figure 2 A structural diagram of a battery health state determination device provided by the embodiments of the present application; Figure 3 A structural diagram of an electronic device provided by the embodiments of the present application. DETAILED DESCRIPTION

[0028] The technical solutions of the embodiments of the present application will be described below in conjunction with the drawings in the embodiments of the present application.

[0029] It should be noted that: similar reference numerals and letters represent similar items in the following drawings, and therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. Meanwhile, in the description of the present application, the terms "first", "second", etc. are only used for distinguishing description, and cannot be understood as indicating or implying relative importance.

[0030] It should be noted that the prior art generally uses the ampere-hour integration method (Ah integration method) to estimate the battery capacity (or state of charge). For example, for the state of health (SOH) of the battery, it is generally defined as the percentage of the current capacity (C current) to the initial capacity (C initial), that is .

[0031] The specific calculation steps are: 1. Accumulate the charge and discharge electric quantity: by integrating the change of current with time, the actual available capacity of the battery from full to empty (or from empty to full) is counted. 2. Compare the initial capacity: compare the actual available capacity with the nominal initial capacity of the battery to obtain the SOH.

[0032] Although this traditional method is simple in principle and easy to implement, the method has the following significant disadvantages: 1. Only rely on the current integration result, without fusing voltage, temperature, internal resistance and other multi-source data, cannot identify abnormalities through cross-validation, and cannot capture multi-factor coupling effect, and the precision is low; 2. The current sensor (such as shunt, Hall effect sensor) has inherent error (usually ±1%~±3%), and small current measurement error will be amplified to significant capacity error through the integration process; 3. If the current sensor fails (such as wire breakage or signal anomaly), the ampere-hour integration method will be completely invalid, and there is no other data chain to support state estimation; 4. Under dynamic current conditions (such as vehicle start-stop, load mutation), the rapid change of current will cause integration delay or sampling error, further reducing the estimation accuracy.

[0033] In view of the problems existing in the prior art, the battery health state determination method provided by the embodiments of the present application learns the internal relationship between a plurality of key features and the battery health state through training a model, and predicts the health state of the battery by using the internal relationship, thereby avoiding the case that the measurement error of a single parameter seriously affects the estimation result, and effectively improving the accuracy and reliability of the battery health state estimation.

[0034] As shown in Figure 1 , the battery health state determination method provided by the embodiments of the present application can include the following steps: S1, obtaining state monitoring data of a battery to be measured in a charging process.

[0035] Specifically, the battery to be measured is a battery whose health state needs to be estimated at present.

[0036] Exemplarily, the battery to be measured can be discharged to zero, and then charging is started, and the state monitoring data in the charging process is collected in real time (according to a preset collection frequency).

[0037] Exemplarily, the state monitoring data can include multiple data, such as battery voltage (V), current (I), temperature (T) and time (t) data, etc., and the collection frequency of each data can be different.

[0038] S2, determining a key feature set based on the state monitoring data; wherein the key feature set includes multiple key features.

[0039] Exemplarily, based on the collected state monitoring data, multiple key features can be generated, such as voltage time sequence features, current time sequence features, temperature change features, etc.

[0040] It can be understood that the set composed of multiple key features is the key feature set.

[0041] S3, inputting the key feature set into the trained battery health state prediction model to obtain a battery health state prediction result output by the battery health state prediction model.

[0042] Exemplarily, the battery health state prediction model can adopt a suitable machine learning model according to requirements, and the battery health state prediction model can be trained based on pre-collected key feature set sample data and corresponding battery health state label results.

[0043] Exemplarily, the battery health state prediction model can adopt a support vector regression (SVR) model. It can be understood that the insensitive loss function allows the model not to update the parameters within a certain error range, and has natural robustness to outliers and noise.

[0044] S4, determining the battery health state of the to-be-tested battery based on the battery health state prediction result.

[0045] After obtaining the battery health state prediction result output by the prediction model, the final battery health state of the to-be-tested battery can be determined according to a preset post-processing step or corresponding condition judgment.

[0046] Exemplarily, an estimation method (such as ampere-hour integration method) different from the above model prediction can be used to determine the battery health state prediction range of the to-be-tested battery, if the battery health state prediction result does not exceed the prediction range, the battery health state prediction result is used as the battery health state of the to-be-tested battery, otherwise the battery health state prediction result is discarded, and the battery health state of the to-be-tested battery is recalculated and determined based on the battery health state prediction range.

[0047] Based on this, by collecting state monitoring data in the battery charging process and extracting a plurality of key features, the internal relationship between the plurality of key features and the battery health state is learned through model training, thereby improving the accuracy of battery health state prediction.

[0048] In some embodiments, step S4, determining the battery health state of the battery under test based on the battery health state prediction result, can include: S401, based on the state monitoring data, calculating the current battery capacity of the battery under test by using the ampere-hour integration method; S402, determining the reference battery health state of the battery under test based on the current battery capacity and the initial battery capacity of the battery under test; S403, determining the battery health state of the battery under test in combination with the battery health state prediction result and the reference battery health state.

[0049] Exemplarily, the state monitoring data can include the current-time curve of the battery under test in the charging process. By integrating the current-time curve, the actual available capacity of the battery under test from no electricity to full can be calculated, and the current battery health state of the battery under test (reference battery health state) can be estimated.

[0050] Then, in combination with the battery health state prediction result and the reference battery health state, the final output battery health state of the battery under test is determined.

[0051] Exemplarily, the battery health state prediction result and the reference battery health state can be respectively pre-set with a weight parameter, and weighted average operation is performed to obtain the final battery health state of the battery under test.

[0052] Based on this, in combination with the model prediction result and the estimation result based on the ampere-hour integration method, the final battery health state is determined comprehensively, further improving the accuracy and reliability of battery health state prediction.

[0053] In some embodiments, step S3, inputting the key feature set into the trained battery health state prediction model to obtain the battery health state prediction result output by the battery health state prediction model, can include: S301, respectively inputting a plurality of key feature sets into the trained battery health state prediction model to obtain a plurality of battery health state candidate prediction results corresponding thereto; wherein the plurality of key feature sets are determined based on a plurality of groups of state monitoring data respectively, and the plurality of groups of state monitoring data are respectively collected based on the battery under test in a plurality of charging processes; S302, determining the battery health state prediction result based on the plurality of battery health state candidate prediction results.

[0054] It should be noted that by repeatedly charging and discharging the battery under test, a plurality of sets of state monitoring data of the battery under test in a plurality of charging processes can be collected, each charging process corresponding to a set of state monitoring data, and each set of state monitoring data can extract a corresponding set of key features.

[0055] Based on this, the battery health state prediction model can be used to predict each set of key features respectively to obtain a plurality of candidate battery health state prediction results corresponding thereto. Each time the battery health state prediction model is used to predict a set of key features, a candidate battery health state prediction result corresponding thereto is obtained.

[0056] Finally, the battery health state prediction result can be determined based on the plurality of candidate battery health state prediction results obtained by prediction.

[0057] Exemplarily, after removing the highest value and the lowest value from the plurality of candidate battery health state prediction results, the remaining several candidate battery health state prediction results can be averaged to obtain the final battery health state prediction result.

[0058] Based on this, by repeating the charging multiple times to obtain multiple sets of data and extract key features, and then respectively predicting a plurality of candidate prediction results corresponding thereto, and finally determining the final battery health state prediction result by synthesizing these candidate prediction results, the accuracy and reliability of the battery health state prediction are further improved.

[0059] In some embodiments, step S4 of determining the battery health state of the battery under test based on the battery health state prediction result can include: S411, determining the credibility of the battery health state prediction result based on the standard deviation of the plurality of candidate battery health state prediction results; S412, in the case where the credibility of the battery health state prediction result is lower than a preset credibility threshold, calculating the current battery capacity of the battery under test based on the state monitoring data and using the ampere-hour integral method; S413, determining the battery health state of the battery under test based on the current battery capacity and the initial battery capacity of the battery under test.

[0060] It should be noted that after the plurality of candidate battery health state prediction results are obtained in step S301, the standard deviation of the candidate battery health state prediction results can be calculated.

[0061] It can be understood that the standard deviation can be used to describe the distribution of data, which reflects the deviation degree between the data points of a group of data and the mean value thereof. Therefore, when the standard deviation is too large, it indicates that the deviation degree of a plurality of battery state of health candidate prediction results from the average value thereof is large, which is unstable, and at this time, it can be considered that the battery state of health prediction result predicted and calculated by the model has low reliability.

[0062] Exemplarily, according to the preset corresponding relationship between the standard deviation and the reliability, the corresponding reliability can be determined based on the standard deviation of the plurality of battery state of health candidate prediction results.

[0063] Exemplarily, when it is determined that the reliability of the battery state of health prediction result is lower than the preset reliability threshold, the battery state of health prediction result is discarded, and the ampere-hour integral method is used to calculate the current battery capacity of the battery under test, and then the battery state of health of the battery under test is determined based on the current battery capacity and the initial battery capacity of the battery under test. When it is determined that the reliability of the battery state of health prediction result is not lower than the preset reliability threshold, the battery state of health prediction result can be used to determine the final battery state of health of the battery under test.

[0064] Therefore, by determining the reliability of the prediction result according to the standard deviation of a plurality of candidate prediction results, and discarding the model prediction result when the reliability is low, and using the ampere-hour integral method to estimate the battery state of health, the accuracy and reliability of the battery state of health prediction are further improved.

[0065] In some embodiments, the state monitoring data includes voltage state data, current state data, and temperature state data. The plurality of key features include at least two of constant current charging time, voltage platform slope, temperature change amount, and charging efficiency.

[0066] It should be noted that the voltage state data, the current state data, and the temperature state data can all be state change curves over time.

[0067] Exemplarily, the voltage state data refers to a voltage value change curve over time, wherein the voltage value can be collected at a preset collection frequency (such as once per second). The current state data and the temperature state data are the same, and the collection frequencies of the state data can be different.

[0068] In some embodiments, the determination manner of the constant current charging time includes determining a constant current charging phase of the battery under test based on the current state data, and determining the constant current charging time based on the duration of the constant current charging phase. The determination manner of the voltage platform slope includes determining a voltage change amount of the battery under test from a stable voltage to a highest voltage based on the voltage state data, and determining the voltage platform slope based on the voltage change amount. The determination manner of the temperature variation amount comprises: determining the temperature variation amount based on the maximum temperature value and the minimum temperature value of the temperature state data. The determination manner of the charging efficiency comprises: obtaining the electric energy consumed by the battery under test in the charging process, and obtaining the electric quantity increase value of the battery under test in the charging process, and determining the charging efficiency based on the ratio of the electric quantity increase value to the electric energy.

[0069] Specifically, for the constant current charging time, it is necessary to first determine the constant current charging phase of the battery under test in the current charging process. Exemplarily, the constant current charging phase refers to a charging phase in which the charging current is constant within a preset range, or the constant current charging phase refers to a charging phase in which the fluctuation range of the charging current is less than a preset threshold.

[0070] After determining the constant current charging phase, the duration of the constant current charging phase can be counted as the constant current charging time.

[0071] For the voltage platform slope, the time when the battery has a stable voltage is first determined. It can be understood that the battery has an unstable voltage at the beginning of charging (starting from zero electric quantity), and thus needs to reach a certain electric quantity to have a stable voltage. By detecting the rate of change of the voltage, when the rate of change is lower than a preset threshold, it is considered as the time when the battery has a stable voltage.

[0072] Then, the voltage variation amount between the highest voltage of the battery from the time when the battery has a stable voltage to the end of charging is obtained, and then the voltage platform slope is determined based on the ratio of the voltage variation amount to the time.

[0073] For the temperature variation amount, the maximum temperature value and the minimum temperature value in the charging process can be obtained based on the temperature state data respectively, and the temperature variation amount is determined based on the difference between the two.

[0074] For the charging efficiency, it refers to the ratio between the actual increase of the electric quantity and the actual consumption of the electric energy of the battery under test in the charging process. Exemplarily, the electric energy consumed by the battery under test in the charging process is first obtained, which can be estimated by ampere-hour integration method or can be determined by reading a metering electric meter; then the electric quantity increase value of the battery under test in the charging process is obtained, which can also be estimated by the ampere-hour integration method for the actual electric quantity of the battery under test from full charge to empty. Finally, the charging efficiency is determined based on the ratio of the electric quantity increase value to the electric energy.

[0075] Based on this, by calculating the corresponding key features based on the voltage, current and temperature state data in a specific manner, the accuracy of the battery health state prediction is further improved.

[0076] In some embodiments, the step S2 of determining the key feature set based on the state monitoring data can comprise: S201, acquire a plurality of groups of state monitoring data corresponding to the to-be-tested battery in a plurality of charging processes; S202, determine a plurality of reference key feature sets corresponding to the plurality of groups of state monitoring data; S203, determine a key feature set based on the plurality of reference key feature sets by using an averaging algorithm.

[0077] It should be noted that the to-be-tested battery can be repeatedly charged and discharged, and the state monitoring data corresponding to each charging process can be acquired. Each group of state monitoring data can correspond to the determination of a reference key feature set. Then, each key feature in the reference key feature set is subjected to an averaging operation to obtain a corresponding final key feature. Finally, all key features obtained by the averaging operation are combined to determine the final key feature set.

[0078] Based on this, by acquiring a plurality of groups of state monitoring data and extracting features respectively, and then using an averaging algorithm to determine a final key feature set from a plurality of reference key feature sets, the accuracy of battery health state prediction is further improved.

[0079] In some embodiments, step S3, inputting the key feature set into the trained battery health state prediction model to obtain the battery health state prediction result output by the battery health state prediction model, can include: S311, acquire a first key feature set and a second key feature set; wherein the first key feature set is determined based on state monitoring data of the to-be-tested battery in a fast charging scenario charging process, and the second key feature set is determined based on state monitoring data of the to-be-tested battery in a slow charging scenario charging process; the charging power of the fast charging scenario is greater than that of the slow charging scenario; S312, input the first key feature set into a trained first battery health state prediction model to obtain a first battery health state prediction result output by the first battery health state prediction model; wherein the first battery health state prediction model is trained based on pre-collected fast charging scenario sample data; S313, input the second key feature set into a trained second battery health state prediction model to obtain a second battery health state prediction result output by the second battery health state prediction model; wherein the second battery health state prediction model is trained based on pre-collected slow charging scenario sample data; S314, determine the battery health state prediction result in combination with the first battery health state prediction result and the second battery health state prediction result.

[0080] It should be noted that one battery health state prediction model can be trained based on the fast charging scenario and the slow charging scenario respectively. Among them, exemplarily, the charging power of the fast charging scenario is generally 60kW-240kW, and part of the super charging can reach 350kW or more; the charging power of the slow charging scenario is generally 3.5kW-14kW (mainly 7kW).

[0081] For the battery to be tested, by repeatedly charging and discharging, charging can be carried out in the fast charging scenario and in the slow charging scenario respectively.

[0082] Exemplarily, in the fast charging scenario, the battery to be tested can also be charged and discharged multiple times, and the state monitoring data in each charging process can be collected. After determining the corresponding key feature set, the key feature set can be input into the first battery health state prediction model trained based on the fast charging scenario for prediction to obtain the corresponding first battery health state prediction result. By the same token, by collecting the state monitoring data of the slow charging scenario and extracting the key feature set, the second battery health state prediction result can be obtained by using the second battery health state prediction model for prediction.

[0083] Exemplarily, the first battery health state prediction result and the second battery health state prediction result can be determined based on a plurality of candidate prediction results, wherein each candidate prediction result corresponds to the state monitoring data of one charging process (same charging scenario).

[0084] Finally, the battery health state prediction result is determined based on the first battery health state prediction result and the second battery health state prediction result.

[0085] Exemplarily, the first battery health state prediction result and the second battery health state prediction result can be weighted and averaged based on a preset weight to obtain the corresponding battery health state prediction result.

[0086] Based on this, by collecting data for the fast charging scenario and the slow charging scenario respectively, and using the models trained for different scenarios for prediction, and finally combining the prediction results of the two to determine the final prediction result, the accuracy of battery health state prediction is further improved.

[0087] The method flow of the embodiments of the present application is exemplarily illustrated as follows: Step 1: Charging curve data collection In the charging process, the voltage (V), current (I), temperature (T) and time (t) data of the battery to be tested are collected.

[0088] Step 2: Feature extraction 2.1, Constant current charging time (tcc): from SOC10% to 80%, the constant current charging stage is the constant current charging time; 2.2, Voltage platform slope (k): the slope of the voltage-time curve (from the beginning of the stable voltage until the highest voltage at the end of the charging); 2.3, Temperature change rate (ΔT): the difference between the highest and lowest temperatures of the battery during the charging process.

[0089] 2.4, Charging efficiency (η): the actual charging efficiency calculated according to the law of conservation of energy.

[0090] Step 3: Feature pre-processing De-noising: remove abnormal charging data (such as curve mutations caused by charging interruption, faults, etc.); Data standardization: normalize the extracted features to the interval [0, 1]; Step 4: SOH prediction model Use the support vector regression (SVR) model, input the above feature parameters, and output the battery SOH prediction value: SOH = f(tCC, k, ΔT, η) where the model parameters are obtained by training historical data.

[0091] Step 5: Model updating A sliding window mechanism can be used to periodically update the model parameters using newly collected sample data, or to retrain the model to maintain prediction accuracy.

[0092] It should be noted that the core of the present application is to explore the internal relationship between the voltage, current, temperature, and other parameter curves during the battery charging process and the battery state of health (SOH). By extracting features such as constant current charging time, voltage platform slope, and temperature rise rate from the charging process data, these features are highly related to the degree of battery aging, for example, after the battery ages, the constant current charging time will be shortened, and the voltage platform slope will be larger.

[0093] Then, these key features are input into a machine learning model (such as support vector regression SVR) for training. After the model learns the non-linear relationship between the key features and SOH, it can predict the current SOH of the battery based on the features extracted from the real-time charging curve. Compared with traditional methods, this technology does not need to disassemble the battery, realizes high-precision prediction, and can effectively adapt to factors such as battery aging and temperature changes, providing key data support for electric vehicle battery management systems, helping to optimize battery usage strategies, and prolonging battery life.

[0094] The following specific examples are used to illustrate the present application: Suppose a battery testing agency receives a batch of new energy vehicle battery testing orders from a ride-hailing company. These batteries have been used in ride-hailing for 2-3 years, and the ride-hailing company wants to know the battery health status in order to decide whether to replace the battery and ensure the stability of the operation.

[0095] The detection agency can use the battery state of health determination method proposed in this application. First, the battery is connected to a professional charging device to simulate the fast charging and slow charging scenarios in actual use, and collect voltage, current and temperature data during charging. In a simulated fast charging process, the device charges a battery at 2C rate (assuming the battery capacity is 100Ah, and the current is 200A, then the charging rate is 2C), from 30% to 80% capacity. During this period, the high-precision sensor collects data every second, recording that the voltage gradually rises from 330V to 380V, and the current gradually decreases from the initial 100A to 80A as the battery capacity increases, and the battery temperature rises from 25℃ to 30℃.

[0096] Then the key features are extracted. The detection personnel extract a plurality of key features from various charging curves to form a key feature set. For example, the length of the constant current charging stage, when the battery is originally new, the constant current charging time is about 30 minutes, and this time the detection finds that the constant current charging time is shortened to 25 minutes, which is a sign of battery aging. At the same time, the voltage rise slope is calculated, and the voltage rise slope of a new battery under this charging rate is about 0.15V / min, and now the voltage rise slope of this battery increases to 0.2V / min.

[0097] Then, the detection agency uses the trained prediction model to predict the battery state of health. These prediction models are trained using a large amount of battery charging curve data of different brands, models and service life, and the prediction process requires the use of prediction models trained with data of the same brand and model as the battery to be tested.

[0098] The input of the prediction model is the key feature set extracted, and the output of the prediction model is the predicted value of the battery state of health (SOH). After inputting the key feature set of this battery into the model, the SOH prediction value is 75%. According to industry standards, when the battery SOH is lower than 80%, its performance and endurance will decrease significantly. The detection agency judges that the battery has aged to some extent, and suggests that the online car company closely monitor the subsequent use of the battery, and if conditions permit, consider replacing the severely aged battery to avoid vehicle breakdown due to battery problems and affect operation.

[0099] Please refer to Figure 2 , Figure 2 The composition block diagram of the battery state of health determination device provided by some embodiments of the application is shown. It should be understood that the battery state of health determination device is the same as the above Figure 1The method embodiments correspond to each other, and can perform each step involved in the above method embodiments. The specific functions of the battery health state determination apparatus can be referred to the description above. To avoid repetition, the detailed description is appropriately omitted here.

[0100] Figure 2 The battery health state determination apparatus comprises at least one software function module stored in the memory in the form of software or firmware or solidified in the battery health state determination apparatus, and the battery health state determination apparatus comprises: The data acquisition module 210 is configured to acquire state monitoring data of the battery to be tested in a charging process. The feature determination module 220 is configured to determine a key feature set based on the state monitoring data, wherein the key feature set comprises a plurality of key features. The result prediction module 230 is configured to input the key feature set into the trained battery health state prediction model to obtain a battery health state prediction result output by the battery health state prediction model. The state determination module 240 is configured to determine the battery health state of the battery to be tested based on the battery health state prediction result.

[0101] It can be understood that the above-described device item embodiments correspond to the method item embodiments of the present application. The battery health state determination apparatus provided by the embodiments of the present application can realize the battery health state determination method provided by any one of the method item embodiments of the present application.

[0102] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the above-described device can refer to the corresponding process in the foregoing method, and will not be described in more detail here.

[0103] As shown in Figure 3 some embodiments of the present application provide an electronic device 300, which comprises a memory 310, a processor 320, and a computer program stored in the memory 310 and executable on the processor 320, wherein the processor 320 reads the program from the memory 310 through the bus 330 and implements the method of any embodiment of the battery health state determination method as described above when executing the program.

[0104] The processor 320 can process digital signals and can include various computing structures. For example, a complex instruction set computer structure, a reduced instruction set computer structure, or a structure implementing a combination of multiple instruction sets. In some examples, the processor 320 can be a microprocessor.

[0105] The memory 310 can be used to store instructions executed by the processor 320 or data related to the instructions during execution. The instructions and / or data can include code for implementing some or all of the functions of one or more modules described in the embodiments of the present application. The processor 320 of the embodiments of the present disclosure can be used to execute the instructions in the memory 310 to implement the methods shown above. The memory 310 includes a dynamic random access memory, a static random access memory, a flash memory, an optical memory, or other memory well known to those skilled in the art.

[0106] Some embodiments of the present application also provide a computer readable storage medium having stored thereon a computer program, which, when executed by a processor, performs the method of the method embodiments.

[0107] Some embodiments of the present application also provide a computer program product, which, when executed on a computer, causes the computer to perform the method of the method embodiments.

[0108] It should be noted that each of the embodiments in the present specification is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same parts of each embodiment can be referred to each other. For device embodiments, since they are basically similar to method embodiments, the description is relatively simple, and the relevant parts can be referred to the part of the method embodiments.

[0109] In several embodiments provided in the present application, it should be understood that the disclosed devices and methods can also be implemented in other ways. The device embodiments described above are only schematic. For example, the flowcharts and block diagrams in the drawings show the possible implementation architectures, functions and operations of the devices, methods and computer program products according to the embodiments of the present application. In this regard, each block in the flowcharts or block diagrams can represent a module, a program segment or a part of code, which contains one or more executable instructions for implementing the specified logical functions. It should also be noted that in some alternative implementations, the functions noted in the blocks can occur in different orders than that shown in the figure. For example, two consecutive blocks can actually be executed substantially in parallel, and sometimes they can be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and the combination of blocks in the block diagrams and / or flowcharts, can be implemented by a dedicated hardware-based system for executing the specified functions or actions, or can be implemented by a combination of special-purpose hardware and computer instructions.

[0110] In addition, each functional module in each embodiment of the present 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.

[0111] If the functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various program code storage media.

[0112] The above is only an embodiment of the present application and does not limit the protection scope of the present application. For those skilled in the art, the present application can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application. It should be noted that similar reference numbers and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings.

[0113] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited thereto. Any skilled person in the art can easily think of changes or replacements within the technical range disclosed by the present application, which shall be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

[0114] It is to be noted that, as used in this specification and the appended claims, the singular forms "a," "an," and "the" include plural referents unless the context clearly dictates otherwise. Thus, for example, reference to "a component" can include a combination of two or more components. Additionally, the terms "comprise," "comprises," and "comprising," or any variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements is not necessarily limited to those elements, but can include other elements not expressly listed or inherent to such process, method, article, or apparatus. Furthermore, unless otherwise indicated herein, the terms "first," "second," "third," etc., are used herein merely as labels, and are not intended to impose ordinal import.

Claims

1. A method of determining a state of health of a battery, the method comprising: The method comprises: acquiring state monitoring data of a battery under test during a charging process; determining a set of key features based on the state monitoring data, wherein the set of key features comprises a plurality of key features; inputting the set of key features into a trained battery health state prediction model to obtain a battery health state prediction result output by the battery health state prediction model; determining a battery health state of the battery under test based on the battery health state prediction result.

2. The battery state of health determination method of claim 1, wherein, The determination of the battery health state of the battery under test based on the battery health state prediction result comprises: calculating a current battery capacity of the battery under test by using an ampere-hour integration method based on the state monitoring data; determining a reference battery health state of the battery under test based on the current battery capacity and an initial battery capacity of the battery under test; determining the battery health state of the battery under test in combination with the battery health state prediction result and the reference battery health state.

3. The battery state of health determination method of claim 1, wherein, The inputting of the set of key features into the trained battery health state prediction model to obtain the battery health state prediction result output by the battery health state prediction model comprises: inputting a plurality of sets of key features into the trained battery health state prediction model respectively to obtain a plurality of battery health state candidate prediction results respectively, wherein the plurality of sets of key features are determined based on a plurality of groups of state monitoring data respectively, and the plurality of groups of state monitoring data are respectively acquired based on the battery under test in a plurality of charging processes; determining the battery health state prediction result based on the plurality of battery health state candidate prediction results.

4. The battery state of health determination method of claim 3, wherein, The determination of the battery health state of the battery under test based on the battery health state prediction result comprises: determining a credibility of the battery health state prediction result based on a standard deviation of the plurality of battery health state candidate prediction results; in a case where the credibility of the battery health state prediction result is lower than a preset credibility threshold, calculating a current battery capacity of the battery under test by using an ampere-hour integration method based on the state monitoring data; determining a battery health state of the battery under test based on the current battery capacity and an initial battery capacity of the battery under test.

5. The battery state of health determination method of claim 1, wherein, The state monitoring data comprises voltage state data, current state data and temperature state data. The plurality of key features comprises at least two of a constant-current charging time, a voltage platform slope, a temperature variation and a charging efficiency.

6. The battery state of health determination method of claim 5, wherein, The determination of the constant-current charging time comprises determining a constant-current charging phase of the battery under test based on the current state data, and determining the constant-current charging time based on a duration of the constant-current charging phase; The determination of the voltage platform slope comprises determining a voltage variation of the battery under test from a stable voltage to a highest voltage based on the voltage state data, and determining the voltage platform slope based on the voltage variation; The determination of the temperature variation comprises determining the temperature variation based on a highest temperature value and a lowest temperature value of the temperature state data. The determination manner of the charging efficiency comprises: acquiring the electric energy consumed by the battery under test in the charging process, and acquiring the electric quantity increase value of the battery under test in the charging process; and determining the charging efficiency based on the ratio of the electric quantity increase value to the electric energy.

7. The battery state of health determination method of claim 1, wherein, The determining the key feature set based on the state monitoring data comprises: acquiring a plurality of groups of state monitoring data corresponding to the battery under test in a plurality of charging processes; determining a plurality of reference key feature sets corresponding to the plurality of groups of state monitoring data; determining the key feature set based on the plurality of reference key feature sets by using an average algorithm.

8. The battery state of health determination method of claim 1, wherein, The inputting the key feature set into the trained battery health state prediction model to obtain the battery health state prediction result output by the battery health state prediction model comprises: acquiring a first key feature set and a second key feature set; wherein the first key feature set is determined based on the state monitoring data of the battery under test in a fast charging scenario charging process, and the second key feature set is determined based on the state monitoring data of the battery under test in a slow charging scenario charging process; the charging power of the fast charging scenario is greater than that of the slow charging scenario; inputting the first key feature set into a trained first battery health state prediction model to obtain a first battery health state prediction result output by the first battery health state prediction model; wherein the first battery health state prediction model is trained based on pre-collected fast charging scenario sample data; inputting the second key feature set into a trained second battery health state prediction model to obtain a second battery health state prediction result output by the second battery health state prediction model; wherein the second battery health state prediction model is trained based on pre-collected slow charging scenario sample data; combining the first battery health state prediction result and the second battery health state prediction result to determine the battery health state prediction result.

9. A battery state of health determination apparatus characterized by comprising: The method comprises: a data acquisition module configured to acquire state monitoring data of a battery under test in a charging process; a feature determination module configured to determine a key feature set based on the state monitoring data; wherein the key feature set comprises a plurality of key features; a result prediction module configured to input the key feature set into a trained battery health state prediction model to obtain a battery health state prediction result output by the battery health state prediction model; a state determination module configured to determine a battery health state of the battery under test based on the battery health state prediction result.

10. An electronic device, comprising: The computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the battery health state determination method of any one of claims 1-8.

11. A computer readable storage medium, characterized in that, The computer program is stored on the computer readable storage medium and executable on the processor to execute the battery health state determination method of any one of claims 1-8.

12. A computer program product, characterised in that, The computer program product comprises a computer program which, when executed by a processor, implements the method of determining the state of health of a battery according to any one of claims 1 to 8.

Images