WO2022161002A1 - 电池健康状态预测方法、装置、电子设备和可读存储介质 - Google Patents

电池健康状态预测方法、装置、电子设备和可读存储介质 Download PDF

Info

Publication number
WO2022161002A1
WO2022161002A1 PCT/CN2021/138740 CN2021138740W WO2022161002A1 WO 2022161002 A1 WO2022161002 A1 WO 2022161002A1 CN 2021138740 W CN2021138740 W CN 2021138740W WO 2022161002 A1 WO2022161002 A1 WO 2022161002A1
Authority
WO
WIPO (PCT)
Prior art keywords
battery
health
sample
real
current
Prior art date
Application number
PCT/CN2021/138740
Other languages
English (en)
French (fr)
Inventor
杨静
刘峰
黄智信
杨磊
周文璨
张兴
李丹
Original Assignee
北京嘀嘀无限科技发展有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 北京嘀嘀无限科技发展有限公司 filed Critical 北京嘀嘀无限科技发展有限公司
Publication of WO2022161002A1 publication Critical patent/WO2022161002A1/zh

Links

Images

Classifications

    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60L—PROPULSION OF ELECTRICALLY-PROPELLED VEHICLES; SUPPLYING ELECTRIC POWER FOR AUXILIARY EQUIPMENT OF ELECTRICALLY-PROPELLED VEHICLES; ELECTRODYNAMIC BRAKE SYSTEMS FOR VEHICLES IN GENERAL; MAGNETIC SUSPENSION OR LEVITATION FOR VEHICLES; MONITORING OPERATING VARIABLES OF ELECTRICALLY-PROPELLED VEHICLES; ELECTRIC SAFETY DEVICES FOR ELECTRICALLY-PROPELLED VEHICLES
    • B60L58/00—Methods or circuit arrangements for monitoring or controlling batteries or fuel cells, specially adapted for electric vehicles
    • B60L58/10—Methods or circuit arrangements for monitoring or controlling batteries or fuel cells, specially adapted for electric vehicles for monitoring or controlling batteries
    • B60L58/12—Methods or circuit arrangements for monitoring or controlling batteries or fuel cells, specially adapted for electric vehicles for monitoring or controlling batteries responding to state of charge [SoC]
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60L—PROPULSION OF ELECTRICALLY-PROPELLED VEHICLES; SUPPLYING ELECTRIC POWER FOR AUXILIARY EQUIPMENT OF ELECTRICALLY-PROPELLED VEHICLES; ELECTRODYNAMIC BRAKE SYSTEMS FOR VEHICLES IN GENERAL; MAGNETIC SUSPENSION OR LEVITATION FOR VEHICLES; MONITORING OPERATING VARIABLES OF ELECTRICALLY-PROPELLED VEHICLES; ELECTRIC SAFETY DEVICES FOR ELECTRICALLY-PROPELLED VEHICLES
    • B60L58/00—Methods or circuit arrangements for monitoring or controlling batteries or fuel cells, specially adapted for electric vehicles
    • B60L58/10—Methods or circuit arrangements for monitoring or controlling batteries or fuel cells, specially adapted for electric vehicles for monitoring or controlling batteries
    • B60L58/16—Methods or circuit arrangements for monitoring or controlling batteries or fuel cells, specially adapted for electric vehicles for monitoring or controlling batteries responding to battery ageing, e.g. to the number of charging cycles or the state of health [SoH]
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • G01R31/36—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
    • G01R31/367—Software therefor, e.g. for battery testing using modelling or look-up tables
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • G01R31/36—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
    • G01R31/382—Arrangements for monitoring battery or accumulator variables, e.g. SoC
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • G01R31/36—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
    • G01R31/392—Determining battery ageing or deterioration, e.g. state of health
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T10/00—Road transport of goods or passengers
    • Y02T10/60—Other road transportation technologies with climate change mitigation effect
    • Y02T10/70—Energy storage systems for electromobility, e.g. batteries

Definitions

  • the present application relates to the field of computer technology, and in particular, to a battery health state prediction method, apparatus, electronic device and readable storage medium.
  • new energy vehicles use electric energy as the main energy.
  • the battery life of energy vehicles is strongly correlated with the performance of new energy vehicles.
  • the life of the battery can be represented by the state of health (SOH) of the battery.
  • SOH is a parameter that cannot be directly obtained and needs to be determined by a prediction algorithm.
  • embodiments of the present invention provide a battery state of health prediction method, apparatus, electronic device, and readable storage medium, so as to improve the accuracy of battery state of health prediction.
  • a battery state of health prediction method is provided, the method is applied to an electronic device, and the method includes:
  • the current state of health of the battery corresponding to the current mileage is determined.
  • the current battery state of health, the current mileage and the current battery parameters are input into the pre-trained machine learning model to determine the target battery state of health corresponding to the target trolley in the target mileage.
  • a battery health state prediction apparatus is provided, the apparatus is applied to electronic equipment, and the apparatus includes:
  • the first determination module is used to determine the current mileage and current battery parameters of the target tram.
  • the second determination module is configured to determine the current state of health of the battery corresponding to the current mileage based on the preset first correspondence between the mileage and the state of health of the battery.
  • the prediction module is configured to input the current battery state of health, the current mileage and the current battery parameters into a pre-trained machine learning model to determine the target battery state of health corresponding to the target mileage of the target tram.
  • an embodiment of the present invention provides an electronic device, including a memory and a processor, where the memory is configured to store one or more computer program instructions, wherein the one or more computer program instructions are processed by the The controller executes to implement the method as described in the first aspect.
  • an embodiment of the present invention provides a computer-readable storage medium on which computer program instructions are stored, the computer program instructions implementing the method according to the first aspect when executed by a processor.
  • an embodiment of the present invention provides a computer program product, including a computer program/instruction, which implements the method according to the first aspect when the computer program/instruction is executed by a processor.
  • the current state of health corresponding to the current mileage of the target tram can be determined through the preset first correspondence, and then the current state of health of the battery, the current mileage and the current battery parameters can be synthesized based on the pre-trained machine learning model, To predict the target battery health state corresponding to the target mileage of the target tram, compared with the related art, the embodiment of the present invention also combines the current battery health state in the prediction process, so that the prediction result is more accurate.
  • FIG. 1 is a schematic diagram of a method for predicting a battery state of health according to an embodiment of the present invention
  • FIG. 2 is a flowchart of a method for predicting a state of health of a battery provided by an embodiment of the present invention
  • FIG. 3 is a schematic diagram of a mileage segment in a first correspondence provided by an embodiment of the present invention.
  • FIG. 4 is a flowchart of another battery state of health prediction method provided by an embodiment of the present invention.
  • FIG. 5 is a schematic diagram of a battery capacity curve at 25 degrees Celsius according to an embodiment of the present invention.
  • FIG. 6 is a schematic diagram of a second correspondence provided by an embodiment of the present invention.
  • FIG. 7 is a schematic diagram of a first correspondence provided by an embodiment of the present invention.
  • Fig. 8 is a kind of experimental result comparison diagram provided by the embodiment of the present invention.
  • FIG. 9 is a schematic structural diagram of an apparatus for predicting a battery state of health according to an embodiment of the present invention.
  • FIG. 10 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
  • new energy vehicles are applied to various scenarios (such as new energy vehicles). Private cars, new energy buses and shared trams, etc.), new energy vehicles use electricity as the main energy source, that is, new energy vehicles obtain electricity through batteries. Therefore, the battery life of new energy vehicles is strongly related to the performance of new energy vehicles.
  • the life of the battery can be represented by the state of health (SOH) of the battery.
  • SOH is a parameter that cannot be directly obtained and needs to be determined by a prediction algorithm.
  • the SOH of the battery can be estimated from two dimensions. First, it can be based on the dimension of the electrochemical mechanism, by observing the loss of active lithium ions, the collapse of the material lattice, the positive and negative electrode materials inside the battery, the size of the separator, However, due to confidentiality reasons, the above reaction mechanisms and parameters are not easy to obtain, so it is difficult to estimate the battery SOH in this way. Second, the SOH of the battery can be estimated by the external characteristic parameters of the battery (such as current, voltage, temperature, etc.), but this method cannot accurately represent the attenuation of the battery, that is, the future SOH of the battery cannot be accurately predicted by this method.
  • the external characteristic parameters of the battery such as current, voltage, temperature, etc.
  • an embodiment of the present invention provides a battery state of health prediction method, which can be applied to an electronic device, and the electronic device can predict the battery SOH of a target tram in a certain target mileage in the future
  • the electronic device may be a terminal or a server
  • the terminal may be a smart phone, a tablet computer, or a personal computer (Personal Computer, PC), etc.
  • the server may be a single server, a server cluster configured in a distributed manner, or a Cloud Server.
  • FIG. 1 is a schematic diagram of a battery state of health prediction method provided by an embodiment of the present invention.
  • the schematic diagram includes: a target tram 11 , a first correspondence 12 , and a machine learning model 13 .
  • the electronic device can obtain the current mileage and current battery parameters of the target tram 11, where the target tram 11 may be the electric vehicle shown in FIG. This is not limited, the current mileage is used to represent the accumulated mileage of the target tram 11 , and the current battery parameter is the relevant parameter of the battery in the target tram 11 .
  • the electronic device can determine the current state of health of the battery corresponding to the current mileage according to the preset first correspondence 12, wherein the first correspondence 12 is the correspondence between the mileage and the state of health of the battery, when the mileage is known , the battery health state corresponding to the mileage can be determined according to the first correspondence 12 .
  • the electronic device may input the current mileage, current battery state of health and current battery parameters of the target tram 11 into the pre-trained machine learning model 13, so that the machine learning model 13 outputs the target battery state of health of the target tram 11, wherein the target
  • the battery health state is used to represent the battery health state of the target tram 11 at a certain distance in the future, that is, the machine learning model 13 can predict the target tram 11 according to the current mileage of the target tram 11, the current battery health state and the current battery parameters. Battery health status for any future mileage.
  • the current mileage and current battery parameters of the target tram are determined.
  • the current mileage is used to represent the accumulated mileage of the target tram.
  • the life of a tram can usually be characterized by the accumulated mileage.
  • the target tram can also be other types of trams, such as electric two-wheeled vehicles, electric tricycles, etc. Not limited.
  • the current battery parameters may include battery temperature, discharge current, and discharge voltage, wherein the battery temperature refers to the chemical, electrochemical changes, electron migration, and substances that occur due to the internal structure of the battery during use.
  • the phenomenon of heating on the surface of the battery caused by transmission and other reasons can generally be measured by the battery temperature sensor.
  • the above-mentioned battery temperature, discharge current, and discharge voltage are only an optional implementation of the current battery parameters, and the current battery parameters may also include other data, which are not limited in this embodiment of the present invention.
  • step 22 based on the preset first correspondence between the mileage and the state of health of the battery, the current state of health of the battery corresponding to the current mileage is determined.
  • the first correspondence may be a pre-established correspondence, which may represent the correspondence between the mileage of a certain type of electric car and the state of health of the battery of the electric car under normal circumstances.
  • FIG. 3 is a schematic diagram of a mileage segment in a first correspondence provided by an embodiment of the present invention, and the schematic diagram includes: a number axis including a plurality of consecutive mileage segments.
  • the mileage life of an electric vehicle is about 120,000 kilometers, that is to say, when an electric vehicle travels to about 120,000 kilometers, its lifespan will be exhausted.
  • 120,000 kilometers can be used as the upper limit of mileage, and multiple mileage segments are preset.
  • the number axis in FIG. 3 is used to represent the mileage range of 0-120,000 kilometers.
  • 50,000 kilometers is used as the length of one mileage segment, and 120,000 kilometers is divided into 24 mileage segments.
  • a mileage segment with an excessively large range may result in too many or too few sample trams corresponding to a certain mileage segment, for example, for 10,000 kilometers-2
  • the mileage of 10,000 kilometers corresponds to 15 sample trams, and 20,000-30,000 kilometers corresponds to 1 sample tram, which will make the sample distribution uneven, resulting in a low prediction accuracy.
  • the range included in the mileage segment is too small (for example, 10,000 kilometers is used as the length of a mileage segment), the difference in data corresponding to adjacent mileage segments will be small, resulting in a waste of computing resources.
  • the first correspondence may be determined based on the following steps:
  • step 41 the real-time battery parameters of a plurality of sample trams and the attenuation coefficients of the corresponding models of the sample trams are acquired.
  • the model of the sample tram is the same as that of the target tram, and the real-time mileage of each sample tram corresponds to a preset mileage segment.
  • the real-time battery health state of the sample trolley is determined based on the real-time battery parameters and the attenuation coefficient.
  • the attenuation coefficient can correct the real-time battery parameters when determining the real-time battery health state of the sample tram.
  • the attenuation coefficient may be determined based on the battery capacity curve of the corresponding model of the sample electric vehicle, and the battery capacity curve is a curve obtained by fitting at a specific temperature.
  • FIG. 5 is a schematic diagram of a battery capacity curve at 25 degrees Celsius provided by an embodiment of the present invention.
  • the schematic diagram includes: a time axis (unit is hours), a battery voltage axis (unit is volts), Current/battery capacity axis (units are A/Ah), battery voltage curve, battery capacity curve, and charge current curve.
  • the schematic diagram of the battery capacity curve shown in FIG. 5 is obtained by fitting based on the battery data of the same type of tram during charging. During the fitting process, the battery charge of each tram of the same type can be obtained. The data of current, voltage, battery temperature and time during the state of charge (SOC) charging from 20% to 100%, and these data are fitted to obtain the battery capacity curve shown in Figure 5.
  • SOC state of charge
  • the attenuation coefficient can be determined through the difference. Specifically, the formula of the attenuation coefficient can be as follows:
  • P is used to characterize the attenuation coefficient, which can correct the battery data when the battery is charging to a level of 25 degrees Celsius
  • Charge Energy is used to characterize the charge energy, that is, the current when the battery is fully charged
  • soc end is used to characterize the end of charging.
  • the remaining power of the battery, soc begin is used to characterize the remaining power of the battery at the beginning of charging
  • tem_min end is used to characterize the minimum temperature of the cell when the battery is charged.
  • the real-time battery health state of each sample tram can be determined based on the real-time battery parameters and the attenuation coefficient.
  • the real-time battery parameters may include the charging current and the accumulated charging time
  • step 42 may be specifically performed as: determining the accumulated charging capacity of the sample electric vehicle based on the charging current, the accumulated charging time and the attenuation coefficient, And the real-time battery health status of the sample tram is determined based on the accumulated charging capacity and the nominal capacity of the sample tram.
  • the cumulative charging power of the sample tram can be calculated based on the following formula:
  • W t is used to characterize the cumulative charging capacity of the sample electric car
  • P t is used to characterize the attenuation coefficient
  • I t is used to characterize the charging current
  • t is used to characterize the time.
  • the accumulative charging power calculated based on the charging current has a high accuracy.
  • an accurate real-time battery health state of the sample tram can be determined.
  • the process of determining the real-time battery state of health of the sample trolley based on the accumulated charging power and the nominal capacity of the sample trolley may be performed as follows: based on the accumulated charging power and the nominal capacity of the sample trolley, determine the sample The real-time cycle times of the tram, and the real-time battery health state corresponding to the real-time cycle times is determined based on the real-time cycle times and the preset second correspondence.
  • the number of real-time cycles is used to represent the number of charge/discharge cycles of the sample tram
  • the second correspondence is used to represent the corresponding relationship between the number of cycles and the state of health of the battery
  • the nominal capacity refers to the battery capacity measured after the battery is fully discharged, namely The nominal capacity of a battery can be used to characterize the amount of power that the battery can store.
  • the cumulative charging capacity of the sample tram can be divided by the nominal capacity of the battery of the sample tram to determine the number of real-time cycles .
  • the formula for determining the real-time cycle times of the sample tram can be as follows:
  • N t is used to characterize the real-time cycle times, that is, the current cycle times of the sample trolley battery
  • W t is used to characterize the cumulative charging capacity of the sample trolley
  • Capacity BOL is used to characterize the nominal capacity of the sample trolley battery.
  • the real-time cycle times corresponding to the sample tram can represent the exact battery cycle times of the sample tram.
  • the real-time battery health state corresponding to the real-time cycle number can be determined.
  • FIG. 6 is a schematic diagram of a second correspondence provided by an embodiment of the present invention, and the schematic diagram includes: a battery Cycle number axis, SOH/% axis and line segment used to characterize SOH/% as a function of battery cycle number.
  • Figure 6 is the corresponding relationship between the number of battery cycles determined with reference to the 1C current rate and the battery capacity decay (ie SOH), where C is used to represent the battery charge and discharge capacity rate, and 1C represents the battery when the battery is fully discharged in one hour. current intensity. For example, a battery with a nominal capacity of 2200mAh is discharged for 1 hour at 1C intensity, and the discharge current is 2200mA at this time.
  • SOH battery capacity decay
  • the SOH/% of the battery is 100, that is, when the number of battery cycles is 0, the battery life consumption is 0 (that is, the SOH of the battery is 100% at this time) ), as the number of battery cycles increases, the life of the battery is gradually consumed, and the SOH/% of the battery is getting lower and lower.
  • the SOH/% corresponding to the real-time cycle times can be determined, that is, the SOH corresponding to the real-time cycle times can be determined.
  • the second correspondence may also be in the form of a table, as shown in Table 1 below.
  • Table 1 is another table of the second correspondence provided in this embodiment of the present invention, and the table is as follows :
  • step 43 for each mileage segment, a first correspondence is established based on the real-time battery health status of each sample tram.
  • step 43 may be performed as: determining the average value of the real-time battery state of health of each sample tram corresponding to the mileage segment, and establishing the average value of the real-time battery state of health and the first value of the mileage segment Correspondence.
  • FIG. 7 is a schematic diagram of a first correspondence provided by an embodiment of the present invention.
  • the schematic diagram includes: a number axis including a plurality of consecutive mileage segments and an SOH value corresponding to each mileage segment.
  • the SOH corresponding to the current mileage can be determined according to the first correspondence shown in FIG. 7 , and further, the target can be determined based on the SOH corresponding to the current mileage.
  • the target battery health status corresponding to the target mileage of the tram can be determined according to the first correspondence shown in FIG. 7 , and further, the target can be determined based on the SOH corresponding to the current mileage.
  • step 23 the current battery state of health, the current mileage and the current battery parameters are input into the pre-trained machine learning model to determine the target battery state of health corresponding to the target trolley in the target mileage.
  • the machine learning model is a prediction model, which can output the battery health state corresponding to the target tram at a certain mileage in the future.
  • the machine learning model can be based on XGBoost (Extreme Gradient Boosting ), support vector machine (Support Vector Machine, SVM) or linear regression algorithm (LinearRegression) construction.
  • XGBoost is an optimized distributed gradient boosting library. It is an improvement to the gradient boosting algorithm. Newton's method is used to solve the extreme value of the loss function, and the loss function Taylor is expanded to the second order. In addition, the regularization term is added to the loss function, so , XGBoost is essentially an improvement on the gradient boosting decision tree (Gradient Boosting Decision Tree, GBDT) algorithm, which is more efficient, flexible and portable.
  • Gradient Boosting Decision Tree GBDT
  • SVM is a kind of generalized linear classifier that performs binary classification on data according to supervised learning. Its decision boundary is the maximum-margin hyperplane for learning samples. SVM uses hinge loss The function (hinge loss) calculates the empirical risk (empirical risk) and adds a regularization term to the solution system to optimize the structural risk (structural risk), which is a classifier with sparsity and robustness. SVM can perform nonlinear classification through the kernel method, which is one of the common kernel learning methods.
  • Linear regression algorithm is a statistical analysis method that uses regression analysis in mathematical statistics to determine the quantitative relationship between two or more variables. It is widely used.
  • a regression analysis in which the least squares function models the relationship between one or more independent variables and the dependent variable.
  • This function is a linear combination of one or more model parameters called regression coefficients, in the case of only one independent variable. It is called simple regression, and the case of more than one independent variable is called multiple regression.
  • the current mileage of the target tram A is 50,000 kilometers
  • the current mileage of the target tram A (50,000 kilometers) can be determined through the first correspondence in the embodiment of the present invention (for example, the correspondence shown in FIG. 7 ). km) corresponding to the SOH (that is, the current battery state of health), then, the current battery health state, current mileage and current battery parameters can be input into the pre-trained machine learning model to determine the target tram A when the battery is 100,000 kilometers away.
  • the state of health that is, the target battery state of health).
  • the current state of health corresponding to the current mileage of the target tram can be determined through the preset first correspondence, and then the current state of health of the battery, the current mileage and the current battery parameters can be synthesized based on the pre-trained machine learning model, To predict the target battery health state corresponding to the target mileage of the target tram, compared with the related art, the embodiment of the present invention also combines the current battery health state in the prediction process, so that the prediction result is more accurate.
  • the embodiment of the present invention randomly selects 4 types of target trams (target tram A, target tram B, target tram C, and target tram D) to conduct experiments, and the battery health state prediction method provided by the embodiment of the present invention is used. Predict the target trams of these four types respectively.
  • the predicted capacity can be calculated based on the predicted SOH.
  • the first data set and the second data set respectively represent the mileage parameters and battery parameters of each target tram in two different time periods.
  • the embodiment of the present invention selects the data (mileage parameters and battery parameters) of the target tram E in the first time period, and predicts that the target tram E is 120,000 kilometers (target mileage) based on the data in the first time period ), and then, the embodiment of the present invention also selects the data (mileage parameters and battery parameters) of the target tram E in the second time period, and predicts the target tram E at 120,000 kilometers (target driving distance) based on the data in the second time period. mileage) SOH.
  • FIG. 8 is a comparison diagram of experimental results provided by an embodiment of the present invention, and the comparison diagram includes: a mileage axis (unit is 10,000 kilometers), an SOH (%) axis, based on the first time The SOH change trend line segment of the target tram E predicted by the data of the segment, the SOH change trend line segment of the target tram E predicted based on the data of the second time period, and the measured data.
  • an embodiment of the present invention also provides a battery state of health prediction device.
  • the device includes: a first determination module 91 , a second determination module 92 and a prediction module 93 .
  • the first determination module 91 is used to determine the current mileage and current battery parameters of the target tram.
  • the second determination module 92 is configured to determine the current state of health of the battery corresponding to the current mileage based on the preset first correspondence between the mileage and the state of health of the battery.
  • the prediction module 93 is configured to input the current battery state of health, the current mileage and the current battery parameters into the pre-trained machine learning model, so as to determine the target battery state of health corresponding to the target mileage of the target tram.
  • the first correspondence is determined based on the following modules:
  • the acquisition module is used to acquire the real-time battery parameters of multiple sample trams and the attenuation coefficients of the corresponding models of the sample trams.
  • the sample trams are of the same model as the target tram, and the real-time mileage of each sample tram corresponds to a preset mileage segment.
  • the third determination module is configured to, for each sample tram, determine the real-time battery health state of the sample tram based on the real-time battery parameters and the attenuation coefficient.
  • An establishment module is used to establish a first correspondence relationship for each mileage segment based on the real-time battery health status of each sample tram.
  • the real-time battery parameters include charging current and accumulated charging time.
  • the third determination module is specifically used for:
  • the accumulated charging capacity of the sample electric vehicle is determined.
  • the real-time battery health status of the sample tram is determined.
  • the third determination module is specifically used for:
  • the real-time cycle number of the sample trolley is determined, and the real-time cycle number is used to characterize the number of charge/discharge cycles of the sample trolley.
  • the real-time battery state of health corresponding to the real-time cycle number is determined based on the real-time cycle number and a preset second correspondence relationship, and the second correspondence relationship is used to represent the corresponding relationship between the cycle number and the battery state of health.
  • build a module specifically for:
  • a first correspondence between the average real-time battery state of health and the mileage segment is established.
  • the attenuation coefficient is determined based on the battery capacity curve of the corresponding model of the sample tram, and the battery capacity curve is a curve obtained by fitting at a specific temperature.
  • the current battery parameters include battery temperature, discharge current and discharge voltage.
  • the machine learning model is constructed based on XGBoost, support vector machine or linear regression algorithm.
  • the current state of health corresponding to the current mileage of the target tram can be determined through the preset first correspondence, and then the current state of health of the battery, the current mileage and the current battery parameters can be synthesized based on the pre-trained machine learning model, To predict the target battery health state corresponding to the target mileage of the target tram, compared with the related art, the embodiment of the present invention also combines the current battery health state in the prediction process, so that the prediction result is more accurate.
  • FIG. 10 is a schematic diagram of an electronic device according to an embodiment of the present invention.
  • the electronic device shown in FIG. 10 is a general address query apparatus, which includes a general computer hardware structure, which at least includes a processor 101 and a memory 102 .
  • the processor 101 and the memory 102 are connected by a bus 103.
  • the memory 102 is adapted to store instructions or programs executable by the processor 101 .
  • the processor 101 may be an independent microprocessor, or may be a set of one or more microprocessors.
  • the processor 101 executes the instructions stored in the memory 102 to execute the above-described method flow of the embodiments of the present invention to process data and control other devices.
  • the bus 103 connects the above-mentioned various components together, while connecting the above-mentioned components to the display controller 104 and the display device and the input/output (I/O) device 105 .
  • the input/output (I/O) device 105 may be a mouse, a keyboard, a modem, a network interface, a touch input device, a somatosensory input device, a printer, and other devices known in the art.
  • input/output devices 105 are connected to the system through input/output (I/O) controllers 106 .
  • embodiments of the present invention may be provided as a method, an apparatus (apparatus) or a computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may employ a computer program product embodied on one or more computer-readable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.
  • computer-readable storage media including, but not limited to, disk storage, CD-ROM, optical storage, etc.
  • These computer program instructions may be stored in a computer readable memory capable of directing a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer readable memory result in an article of manufacture comprising instruction means, the instruction means Implements the function specified in a flow chart or flows.
  • These computer program instructions may also be provided to the processor of a general purpose computer, special purpose computer, embedded processor or other programmable data processing device to produce a machine such that the instructions executed by the processor of the computer or other programmable data processing device produce Means for implementing the functions specified in a flow or flows of a flowchart.
  • Another embodiment of the present invention relates to a non-volatile storage medium for storing a computer-readable program, the computer-readable program being used for a computer to execute some or all of the above method embodiments.
  • the aforementioned storage medium includes: U disk, mobile hard disk, Read-Only Memory (ROM, Read-Only Memory), Random Access Memory (RAM, Random Access Memory), magnetic disk or optical disk and other media that can store program codes .
  • Another embodiment of the present invention relates to a computer program product, including computer programs/instructions, and when the computer program/instructions are executed by a processor, some or all of the above method embodiments can be implemented.
  • a computer program product (computer program/instruction) may be executed by a processor to specify relevant hardware (including the processor itself), thereby implementing all or part of the methods in the foregoing embodiments. step.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Sustainable Development (AREA)
  • Sustainable Energy (AREA)
  • Power Engineering (AREA)
  • Transportation (AREA)
  • Mechanical Engineering (AREA)
  • Secondary Cells (AREA)
  • Electric Propulsion And Braking For Vehicles (AREA)
  • Charge And Discharge Circuits For Batteries Or The Like (AREA)

Abstract

一种电池健康状态预测方法、装置、电子设备和可读存储介质，涉及计算机技术领域，可以通过预先设置的第一对应关系确定目标电车当前行驶里程对应的当前健康状态（22），然后可以基于预先训练的机器学习模型综合当前电池健康状态、当前行驶里程和当前电池参数，预测目标电车在目标行驶里程对应的目标电池健康状态（23），在预测的过程中结合了当前电池健康状态，使得预测结果更加准确。

Description

电池健康状态预测方法、装置、电子设备和可读存储介质
交叉引用
本申请要求于2021年1月27日提交的中国专利申请No.202110112655.8的优先权，其全部内容通过引用结合于此。
技术领域
本申请涉及计算机技术领域，特别是涉及一种电池健康状态预测方法、装置、电子设备和可读存储介质。
背景技术
目前，随着人们环保意识的增强，新能源车辆的发展越来越迅速，越来越多的人选择新能源车辆作为出行的交通工具，其中，新能源车辆以电能作为主要能源，因此，新能源车辆的电池寿命与新能源车辆的性能强相关。
电池的寿命可以通过电池的健康状态(State of health，SOH)表示，SOH是一个无法直接获取的参数，需要通过预测算法来确定。
相关技术中，往往会通过电流、电压等外部特性参数来预测未来某一时刻的电池SOH，但是，这样的预测方法很难保证电池SOH的精度。
发明内容
有鉴于此，本发明实施例提供一种电池健康状态预测方法、装置、电子设备和可读存储介质，以提高电池健康状态预测的准确率。
第一方面，提供了一种电池健康状态预测方法，所述方法应用于电子设备，所述方法包括：
确定目标电车的当前行驶里程和当前电池参数。
基于预先设置的行驶里程和电池健康状态的第一对应关系，确定所述当前行驶里程对应的当前电池健康状态。
将所述当前电池健康状态、所述当前行驶里程和所述当前电池参数输入预先训练的机器学习模型，以确定所述目标电车在目标行驶里程对应 的目标电池健康状态。
第二方面，提供了一种电池健康状态预测装置，所述装置应用于电子设备，所述装置包括：
第一确定模块，用于确定目标电车的当前行驶里程和当前电池参数。
第二确定模块，用于基于预先设置的行驶里程和电池健康状态的第一对应关系，确定所述当前行驶里程对应的当前电池健康状态。
预测模块，用于将所述当前电池健康状态、所述当前行驶里程和所述当前电池参数输入预先训练的机器学习模型，以确定所述目标电车在目标行驶里程对应的目标电池健康状态。
第三方面，本发明实施例提供了一种电子设备，包括存储器和处理器，所述存储器用于存储一条或多条计算机程序指令，其中，所述一条或多条计算机程序指令被所述处理器执行以实现如第一方面所述的方法。
第四方面，本发明实施例提供了一种计算机可读存储介质，其上存储计算机程序指令，所述计算机程序指令在被处理器执行时实现如第一方面所述的方法。
第五方面，本发明实施例提供了一种计算机程序产品，包括计算机程序/指令，所述计算机程序/指令被处理器执行时实现如第一方面所述的方法。
通过本发明实施例，可以通过预先设置的第一对应关系确定目标电车当前行驶里程对应的当前健康状态，然后可以基于预先训练的机器学习模型综合当前电池健康状态、当前行驶里程和当前电池参数，预测目标电车在目标行驶里程对应的目标电池健康状态，相较于相关技术，本发明实施例在预测的过程中还结合了当前电池健康状态，使得预测结果更加准确。
附图说明
通过以下参照附图对本发明实施例的描述，本发明实施例的上述以及其它目的、特征和优点将更为清楚，在附图中：
图1为本发明实施例提供的一种电池健康状态预测方法的示意图；
图2为本发明实施例提供的一种电池健康状态预测方法流程图；
图3为本发明实施例提供的一种第一对应关系中行驶里程段的示意 图；
图4为本发明实施例提供的另一种电池健康状态预测方法流程图；
图5为本发明实施例提供的一种25摄氏度下的电池容量曲线的示意图；
图6为本发明实施例提供的一种第二对应关系的示意图；
图7为本发明实施例提供的一种第一对应关系的示意图；
图8为本发明实施例提供的一种实验结果对比图；
图9为本发明实施例提供的一种电池健康状态预测装置的结构示意图；
图10为本发明实施例提供的一种电子设备的结构示意图。
具体实施方式
以下基于实施例对本发明进行描述，但是本发明并不仅仅限于这些实施例。在下文对本发明的细节描述中，详尽描述了一些特定的细节部分。对本领域技术人员来说没有这些细节部分的描述也可以完全理解本发明。为了避免混淆本发明的实质，公知的方法、过程、流程、元件和电路并没有详细叙述。
此外，本领域普通技术人员应当理解，在此提供的附图都是为了说明的目的，并且附图不一定是按比例绘制的。
除非上下文明确要求，否则在说明书的“包括”、“包含”等类似词语应当解释为包含的含义而不是排他或穷举的含义；也就是说，是“包括但不限于”的含义。
在本发明的描述中，需要理解的是，术语“第一”、“第二”等仅用于描述目的，而不能理解为指示或暗示相对重要性。此外，在本发明的描述中，除非另有说明，“多个”的含义是两个或两个以上。
目前，随着新能源车辆的发展越来越迅速，越来越多的人选择新能源车辆作为出行的交通工具，以新能源汽车为例，新能源汽车被应用至各种场景(例如新能源私家车、新能源公交车和共享电车等等)，新能源汽车以电能作为主要能源，即新能源汽车通过电池来获取电能，因此，新能源车辆的电池寿命与新能源车辆的性能强相关。
电池的寿命可以通过电池的健康状态(State of health，SOH)表示，SOH是一个无法直接获取的参数，需要通过预测算法来确定。
相关技术中，可以从两种维度对电池SOH进行估算，其一，可以基于电化学机理的维度，通过观测活性锂离子的损失、材料晶格的塌陷、电池内部正负极材料、隔膜尺寸、电解液离子浓度和扩散系数等电池内部的反应机理和参数来估算电池SOH，但是由于保密等原因，上述反应机理和参数不易获取，所以很难通过此方式估算电池SOH。其二，可以通过电池的外部特性参数(例如电流、电压、温度等)来估算电池的SOH，但是此方式并不能准确表示电池的衰减，也即通过此方式无法准确预测该电池未来的SOH。
为了可以准确的预测电池未来的SOH，本发明实施例提供一种电池健康状态预测方法，该方法可以应用于电子设备，并通过电子设备来预测目标电车在未来某一目标行驶里程的电池SOH，其中，电子设备可以是终端或者服务器，终端可以是智能手机、平板电脑或者个人计算机(Personal Computer，PC)等，服务器可以是单个服务器，也可以是以分布式方式配置的服务器集群，还可以是云服务器。
如图1所示，图1为本发明实施例提供的一种电池健康状态预测方法的示意图，该示意图包括：目标电车11、第一对应关系12和机器学习模型13。
首先，电子设备可以获取目标电车11的当前行驶里程和当前电池参数，其中，目标电车11可以是图1所示的电动汽车，也可以是电动两轮车、电动三轮车等等，本发明实施例对此不做限定，当前行驶里程用于表征目标电车11已经累计行驶的里程，当前电池参数为目标电车11中的电池的相关参数。
然后，电子设备可以根据预先设置的第一对应关系12确定该当前行驶里程对应的当前电池健康状态，其中，第一对应关系12为行驶里程和电池健康状态的对应关系，当已知行驶里程时，可以确定根据第一对应关系12确定该行驶里程对应的电池健康状态。
然后，电子设备可以将目标电车11的当前行驶里程、当前电池健康状态和当前电池参数输入预先训练的机器学习模型13，以使得机器学习模型13输出目标电车11的目标电池健康状态，其中，目标电池健康状态用 于表征目标电车11在未来某一里程时的电池健康状态，也就是说，机器学习模型13可以根据目标电车11的当前行驶里程、当前电池健康状态和当前电池参数预测目标电车11在未来任一里程的电池健康状态。
下面将结合具体实施方式，对本发明实施例提供的一种电池健康状态预测方法进行详细的说明，如图2所示，具体步骤如下：
在步骤21，确定目标电车的当前行驶里程和当前电池参数。
其中，当前行驶里程用于表征目标电车已经累计行驶的里程，在实际应用中，通常可以通过累积行驶的里程来表征一辆电车的寿命，例如，对于一辆电动汽车，该电动汽车行驶到12万公里左右时，该电动汽车会寿命耗尽，并处于报废状态，当然，目标电车也可以是其它类型的电车，例如电动两轮车、电动三轮车等等，本发明实施例对目标电车的类型不做限定。
另外，在一种可选的实施方式中，当前电池参数可以包括电池温度、放电电流和放电电压，其中，电池温度是指电池在使用时由于内部结构发生化学、电化学变化、电子迁移及物质传输等原因而产生的电池表面发热的现象，一般可以通过电池温度传感器来测量电池温度。
当然，上述电池温度、放电电流和放电电压仅为当前电池参数一种可选的实施方式，当前电池参数也可以包括其它数据，本发明实施例对此不作限定。
在步骤22，基于预先设置的行驶里程和电池健康状态的第一对应关系，确定当前行驶里程对应的当前电池健康状态。
其中，第一对应关系可以是一种预先建立的对应关系，其可以表征通常情况下某型号电车的行驶里程与该型号电车电池健康状态的对应关系。
例如，如图3所示，图3为本发明实施例提供的一种第一对应关系中行驶里程段的示意图，该示意图包括：包括多个连续行驶里程段的数轴。
在实际应用中，电动汽车的里程寿命约为12万公里，也就是说，当一辆电动汽车行驶至12万公里左右时，寿命将耗尽。
因此，在本发明实施例中，可以12万公里作为里程上限，预先设置多个行驶里程段，如图3所示，图3中的数轴用于表征0-12万公里的里程范围，其中，本发明实施例又以0.5万公里作为一个行驶里程段长度，将12万公里划分为24个行驶里程段。
在实际应用中，范围过大的行驶里程段(例如以1万公里作为一个行驶里程段长度)可能会导致对应某个行驶里程段的样本电车过多或过少，例如针对1万公里-2万公里的行驶里程段对应15个样本电车，而2万公里-3万公里对应1个样本电车，这样，就会使得样本分布不均匀，进而导致预测的准确率较低。
相应的，若行驶里程段包括的范围过小(例如以0.1万公里作为一个行驶里程段长度)，则会导致相邻行驶里程段对应的数据差异较小，造成计算资源的浪费。
因此，在一种优选的方式中，如图3所示，可以将0.5万公里作为行驶里程段的长度，这样既可以保证样本分布均匀，也可以保证相邻行驶里程段对应的数据具有一定差异。
在一种可选的实施方式中，如图4所示，第一对应关系可以基于如下步骤确定：
在步骤41，获取多个样本电车的实时电池参数以及样本电车对应型号的衰减系数。
其中，样本电车与目标电车的型号相同，每个样本电车的实时行驶里程均对应一个预先设置的行驶里程段。
在步骤42，针对每个样本电车，基于实时电池参数以及衰减系数，确定样本电车的实时电池健康状态。
其中，衰减系数可以在确定样本电车的实时电池健康状态时对实时电池参数进行矫正。
在本发明实施例中，衰减系数可以基于样本电车对应型号的电池容量曲线确定，电池容量曲线为在特定温度下拟合得到的曲线。
具体的如图5所示，图5为本发明实施例提供的一种25摄氏度下的电池容量曲线的示意图，该示意图包括：时间轴(单位为小时)、电池电压轴(单位为伏特)、电流/电池容量轴(单位为安/安时)、电池电压曲线、电池容量曲线和充电电流曲线。
在本发明实施例中，图5所示的电池容量曲线示意图是基于同一型号的电车在充电时的电池数据拟合得到的，在拟合的过程中，可以获取各个同型号电车的电池荷电状态(State of charge，SOC)从20％充电至100％ 过程中的电流、电压、电池温度和时间等数据，并将这些数据进行拟合，得到图5所示的电池容量曲线。
其中，SOC用于表征电池的荷电状态，常用来反映电池的剩余容量，其数值上定义为剩余容量占电池容量的比值，常用百分数表示。SOC的取值范围为0～1(即0％～100％)，当SOC＝0时表示电池放电完全，当SOC＝1时表示电池完全充满。
由于在充电的过程中电池温度往往不能保持恒温，即不能电池温度不能全程维持在25摄氏度，因此，实际充电时的数据与理想状态下充电时的数据(理想状态为图5所示的曲线)之间会存在差异，进而，通过该差异可以确定衰减系数，具体的，衰减系数的公式可以如下：
Figure PCTCN2021138740-appb-000001
其中，P用于表征衰减系数，衰减系数可以将电池充电时的电池数据矫正至25摄氏度的水平，Charge Energy用于表征电荷能量，即电池满电状态下电流，soc end用于表征充电结束时的电池剩余电量，soc begin用于表征充电开始时的电池剩余电量，tem_min end用于表征电池充电时的电芯最低温度。
当确定衰减系数之后，可以基于实时电池参数以及衰减系数，确定各样本电车的实时电池健康状态。
在一种可选的实施方式中，实时电池参数可以包括充电电流和累计充电时间，进而，步骤42具体可以执行为：基于充电电流、累计充电时间以及衰减系数，确定样本电车的累计充电电量，以及基于累计充电电量和样本电车的标称容量，确定样本电车的实时电池健康状态。
具体的，针对样本电车的累计充电电量，可以基于如下公式计算样本电车的累计充电电量：
Figure PCTCN2021138740-appb-000002
其中，W t用于表征样本电车的累计充电电量，P t用于表征衰减系数，I t用于表征充电电流，t用于表征时间。
在实际应用中，由于电池在充电的过程中电流往往为恒流，同时，电池工作过程中放出的电量近乎等于电池充电的电量，因此，基于充电电流 计算的累计充电电量精度较高。
更进一步的，基于准确的累计充电电量和样本电车的标称容量，可以确定样本电车准确的实时电池健康状态。
在一种可选的实施方式中，基于累计充电电量和样本电车的标称容量，确定样本电车的实时电池健康状态的过程可以执行为：基于累计充电电量和样本电车的标称容量，确定样本电车的实时循环次数，以及基于实时循环次数和预先设置的第二对应关系，确定实时循环次数对应的实时电池健康状态。
其中，实时循环次数用于表征样本电车的充/放电循环的次数，第二对应关系用于表征循环次数和电池健康状态的对应关系，标称容量指电池完全放电后测量得到的电池容量，即电池的标称容量可以用于表征该电池可以存储的电量。
具体的，当确定样本电车的累计充电电量(也即样本电车中电池的累计充电电量)后，可以将该样本电车的累计充电电量除以该样本电车电池的标称容量，以确定实时循环次数。
其中，确定样本电车实时循环次数的公式可以如下：
Figure PCTCN2021138740-appb-000003
其中，N t用于表征实时循环次数，也即样本电车电池当前的循环次数，W t用于表征上述样本电车的累计充电电量，Capacity BOL用于表征样本电车电池的标称容量。
在本发明实施例中，由于累计充电电量精度较高，同时标称容量也是一个较为精确的数值，因此，样本电车对应的实时循环次数可以表征该样本电车准确的电池循环次数。
进而，基于该实时循环次数和预先设置的第二对应关系，可以确定该实时循环次数对应的实时电池健康状态。
其中，第二对应关系用于表征循环次数和电池健康状态的对应关系，例如，如图6所示，图6为本发明实施例提供的一种第二对应关系的示意图，该示意图包括：电池循环次数轴、SOH/％轴和用于表征SOH/％随电池循环次数变化的线段。
其中，图6是以1C电流倍率为参照确定的电池循环次数与电池容量衰减(即SOH)之间的对应关系，其中，C用来表示电池充放电能力倍率，1C表示电池一小时完全放电时电流强度。例如，标称容量为2200mAh的电池在1C强度下放电1小时放电完成，此时该放电电流为2200mA。
如图6所示，当电池循环次数为0时，电池的SOH/％为100，也就是说，当电池循环次数为0时，电池的寿命消耗为0(即此时电池的SOH为100％)，随着电池循环次数的增加，电池的寿命逐渐被消耗，电池的SOH/％越来越低。
通过图6所示的第二对应关系，当确定样本电车的实时循环次数后，可以确定该实时循环次数所对应的SOH/％，也即确定该实时循环次数所对应的SOH。
在另一种可选的实施方式中，第二对应关系也可以是表格的形式，如下表一所示，表一为本发明实施例提供的另一种第二对应关系的表格，该表格如下：
表一
电池循环次数 SOH/％
0-199 100
200-399 95
400-599 90
600-799 85
800-899 80
900-1000 75
由表一可知，通过表格的形式同样可以确定电池循环次数与SOH之间的关系，另外，第二对应关系还可以基于其他形式表示，本发明实施例对此不作过多赘述。
在步骤43，针对每个行驶里程段，基于各样本电车的实时电池健康状态，建立第一对应关系。
在一种可选的实施方式中，步骤43可以执行为：确定行驶里程段对应的各样本电车的实时电池健康状态的平均值，以及建立实时电池健康状态的平均值和行驶里程段的第一对应关系。
例如，如图7所示，图7为本发明实施例提供的一种第一对应关系的示意图，该示意图包括：包括多个连续行驶里程段的数轴和每个行驶里程段对应的SOH值。
由图7可知，当确定目标电车的当前行驶里程时，可以依据图7所示的第一对应关系确定该当前行驶里程所对应的SOH，进而，可以基于该当前行驶里程所对应的SOH确定目标电车在目标行驶里程对应的目标电池健康状态。
在步骤23，将当前电池健康状态、当前行驶里程和当前电池参数输入预先训练的机器学习模型，以确定目标电车在目标行驶里程对应的目标电池健康状态。
其中，机器学习模型是一种预测模型，其可以输出目标电车在未来某一行驶里程时所对应的电池健康状态，在一种可选的实施方式中，机器学习模型可以基于XGBoost(Extreme Gradient Boosting)、支持向量机(Support Vector Machine，SVM)或者线性回归算法(LinearRegression)构建。
XGBoost是一个优化的分布式梯度增强库，是对梯度提升算法的改进，求解损失函数极值时使用了牛顿法，将损失函数泰勒展开到二阶，另外损失函数中加入了正则化项，因此，XGBoost本质上是在梯度提升决策树(Gradient Boosting Decision Tree，GBDT)算法的基础上进行了改进，实现了更加高效、灵活以及便携。
SVM是一类按监督学***面(maximum-margin hyperplane)，SVM使用铰链损失函数(hinge loss)计算经验风险(empirical risk)并在求解***中加入了正则化项以优化结构风险(structural risk)，是一个具有稀疏性和稳健性的分类器。SVM可以通过核方法(kernel method)进行非线性分类，是常见的核学习(kernel learning)方法之一。
线性回归算法是利用数理统计中回归分析，来确定两种或两种以上变量间相互依赖的定量关系的一种统计分析方法，运用十分广泛，其中，线性回归算法是利用称为线性回归方程的最小平方函数对一个或多个自变 量和因变量之间关系进行建模的一种回归分析，这种函数是一个或多个称为回归系数的模型参数的线性组合，只有一个自变量的情况称为简单回归，大于一个自变量情况的叫做多元回归。
在一种示例中，目标电车A的当前行驶里程为5万公里，通过本发明实施例的第一对应关系(例如图7所示的对应关系)可以确定目标电车A的当前行驶里程(5万公里)所对应的SOH(即当前电池健康状态)，然后，可以将当前电池健康状态、当前行驶里程和当前电池参数输入预先训练的机器学习模型，以确定目标电车A在10万公里时电池的健康状态(即目标电池健康状态)。
通过本发明实施例，可以通过预先设置的第一对应关系确定目标电车当前行驶里程对应的当前健康状态，然后可以基于预先训练的机器学习模型综合当前电池健康状态、当前行驶里程和当前电池参数，预测目标电车在目标行驶里程对应的目标电池健康状态，相较于相关技术，本发明实施例在预测的过程中还结合了当前电池健康状态，使得预测结果更加准确。
在一种示例中，本发明实施例随机选取4种型号的目标电车(目标电车A、目标电车B、目标电车C和目标电车D)进行实验，通过本发明实施例提供的电池健康状态预测方法对这4种型号的目标电车分别进行预测。
其具体结果如下表所示：
表二
Figure PCTCN2021138740-appb-000004
其中，表二中的容量至电池容量，预测的容量可以基于预测出的SOH计算得出。另外，第一数据集和第二数据集分别表征两个不同时间段内各 目标电车的里程参数和电池参数。
由表二可知，通过本发明实施例，对于各目标电车的电池容量(即目标电车的电池SOH)预测误差均小于5％，具有较高的准确率。
在另一种示例中，本发明实施例选取目标电车E在第一时间段的数据(里程参数和电池参数)，并基于第一时间段的数据预测目标电车E在12万公里(目标行驶里程)时的SOH，然后，本发明实施例还选取目标电车E在第二时间段的数据(里程参数和电池参数)，并基于第二时间段的数据预测目标电车E在12万公里(目标行驶里程)时的SOH。
具体的，如图8所示，图8为本发明实施例提供的一种实验结果对比图，该对比图包括：行驶里程轴(单位为万公里)、SOH(％)轴、基于第一时间段的数据预测的目标电车E的SOH变化趋势线段、基于第二时间段的数据预测的目标电车E的SOH变化趋势线段以及实测数据。
由图8可知，通过本发明实施例，对于目标电车E的SOH预测值和实测值之间误差较小，即本发明实施例的预测方法具有较高的准确率。
基于相同的技术构思，本发明实施例还提供了一种电池健康状态预测装置，如图9所示，该装置包括：第一确定模块91、第二确定模块92和预测模块93。
第一确定模块91，用于确定目标电车的当前行驶里程和当前电池参数。
第二确定模块92，用于基于预先设置的行驶里程和电池健康状态的第一对应关系，确定当前行驶里程对应的当前电池健康状态。
预测模块93，用于将当前电池健康状态、当前行驶里程和当前电池参数输入预先训练的机器学习模型，以确定目标电车在目标行驶里程对应的目标电池健康状态。
可选的，第一对应关系基于如下模块确定：
获取模块，用于获取多个样本电车的实时电池参数以及样本电车对应型号的衰减系数，样本电车与目标电车的型号相同，每个样本电车的实时行驶里程均对应一个预先设置的行驶里程段。
第三确定模块，用于针对每个样本电车，基于实时电池参数以及衰减系数，确定样本电车的实时电池健康状态。
建立模块，用于针对每个行驶里程段，基于各样本电车的实时电池健康状态，建立第一对应关系。
可选的，实时电池参数包括充电电流和累计充电时间。
第三确定模块，具体用于：
基于充电电流、累计充电时间以及衰减系数，确定样本电车的累计充电电量。
基于累计充电电量和样本电车的标称容量，确定样本电车的实时电池健康状态。
可选的，第三确定模块，具体用于：
基于累计充电电量和样本电车的标称容量，确定样本电车的实时循环次数，实时循环次数用于表征样本电车的充/放电循环的次数。
基于实时循环次数和预先设置的第二对应关系，确定实时循环次数对应的实时电池健康状态，第二对应关系用于表征循环次数和电池健康状态的对应关系。
可选的，建立模块，具体用于：
确定行驶里程段对应的各样本电车的实时电池健康状态的平均值。
建立实时电池健康状态的平均值和行驶里程段的第一对应关系。
可选的，衰减系数基于样本电车对应型号的电池容量曲线确定，电池容量曲线为在特定温度下拟合得到的曲线。
可选的，当前电池参数包括电池温度、放电电流和放电电压。
可选的，机器学习模型基于XGBoost、支持向量机或者线性回归算法构建。
通过本发明实施例，可以通过预先设置的第一对应关系确定目标电车当前行驶里程对应的当前健康状态，然后可以基于预先训练的机器学习模型综合当前电池健康状态、当前行驶里程和当前电池参数，预测目标电车在目标行驶里程对应的目标电池健康状态，相较于相关技术，本发明实施例在预测的过程中还结合了当前电池健康状态，使得预测结果更加准确。
图10是本发明实施例的电子设备的示意图。如图10所示，图10所示的电子设备为通用地址查询装置，其包括通用的计算机硬件结构，其至少包括处理器101和存储器102。处理器101和存储器102通过总线103连 接。存储器102适于存储处理器101可执行的指令或程序。处理器101可以是独立的微处理器，也可以是一个或者多个微处理器集合。由此，处理器101通过执行存储器102所存储的指令，从而执行如上所述的本发明实施例的方法流程实现对于数据的处理和对于其它装置的控制。总线103将上述多个组件连接在一起，同时将上述组件连接到显示控制器104和显示装置以及输入/输出(I/O)装置105。输入/输出(I/O)装置105可以是鼠标、键盘、调制解调器、网络接口、触控输入装置、体感输入装置、打印机以及本领域公知的其他装置。典型地，输入/输出装置105通过输入/输出(I/O)控制器106与***相连。
本领域的技术人员应明白，本发明的实施例可提供为方法、装置(设备)或计算机程序产品。因此，本发明可采用完全硬件实施例、完全软件实施例或结合软件和硬件方面的实施例的形式。而且，本发明可采用在一个或多个其中包含有计算机可用程序代码的计算机可读存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品。
本发明是参照根据本发明实施例的方法、装置(设备)和计算机程序产品的流程图来描述的。应理解可由计算机程序指令实现流程图中的每一流程。
这些计算机程序指令可以存储在能引导计算机或其他可编程数据处理设备以特定方式工作的计算机可读存储器中，使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品，该指令装置实现流程图一个流程或多个流程中指定的功能。
也可提供这些计算机程序指令到通用计算机、专用计算机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器，使得通过计算机或其他可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程中指定的功能的装置。
本发明的另一实施例涉及一种非易失性存储介质，用于存储计算机可读程序，所述计算机可读程序用于供计算机执行上述部分或全部的方法实施例。
即，本领域技术人员可以理解，实现上述实施例方法中的全部或部分步骤是可以通过程序来指定相关的硬件来完成，该程序存储在一个存储 介质中，包括若干指令用以使得一个设备(可以是单片机，芯片等)或处理器(processor)执行本发明各实施例所述方法的全部或部分步骤。而前述的存储介质包括：U盘、移动硬盘、只读存储器(ROM，Read-Only Memory)、随机存取存储器(RAM，Random Access Memory)、磁碟或者光盘等各种可以存储程序代码的介质。
本发明的另一实施例涉及一种计算机程序产品，包括计算机程序/指令，计算机程序/指令被处理器执行时可以实现上述部分或全部的方法实施例。
即，本领域技术人员可以理解，本发明实施例可以通过处理器执行计算机程序产品(计算机程序/指令)来指定相关的硬件(包括处理器自身)，进而实现上述实施例方法中的全部或部分步骤。
以上所述仅为本发明的优选实施例，并不用于限制本发明，对于本领域技术人员而言，本发明可以有各种改动和变化。凡在本发明的精神和原理之内所作的任何修改、等同替换、改进等，均应包含在本发明的保护范围之内。

Claims (19)

  1. 一种电池健康状态预测方法，其特征在于，所述方法包括：
    确定目标电车的当前行驶里程和当前电池参数；
    基于预先设置的行驶里程和电池健康状态的第一对应关系，确定所述当前行驶里程对应的当前电池健康状态；以及
    将所述当前电池健康状态、所述当前行驶里程和所述当前电池参数输入预先训练的机器学习模型，以确定所述目标电车在目标行驶里程对应的目标电池健康状态。
  2. 根据权利要求1所述的方法，其特征在于，所述第一对应关系基于如下步骤确定：
    获取多个样本电车的实时电池参数以及所述样本电车对应型号的衰减系数，所述样本电车与所述目标电车的型号相同，每个样本电车的实时行驶里程均对应一个预先设置的行驶里程段；
    针对每个样本电车，基于所述实时电池参数以及所述衰减系数，确定所述样本电车的实时电池健康状态；以及
    针对每个行驶里程段，基于各样本电车的实时电池健康状态，建立所述第一对应关系。
  3. 根据权利要求2所述的方法，其特征在于，所述实时电池参数包括充电电流和累计充电时间；
    所述基于所述实时电池参数以及所述衰减系数，确定所述样本电车的实时电池健康状态，包括：
    基于所述充电电流、所述累计充电时间以及所述衰减系数，确定所述样本电车的累计充电电量；以及
    基于所述累计充电电量和所述样本电车的标称容量，确定所述样本电车的实时电池健康状态。
  4. 根据权利要求3所述的方法，其特征在于，所述基于所述累计充电电量和所述样本电车的标称容量，确定所述样本电车的实时电池健康状态，包括：
    基于所述累计充电电量和所述样本电车的标称容量，确定所述样本电车的实时循环次数，所述实时循环次数用于表征所述样本电车的充/放电循环的次数；以及
    基于所述实时循环次数和预先设置的第二对应关系，确定所述实时循环次数对应的实时电池健康状态，所述第二对应关系用于表征循环次数和电池健康状态的对应关系。
  5. 根据权利要求2所述的方法，其特征在于，所述基于各样本电车的实时电池健康状态，建立所述第一对应关系，包括：
    确定所述行驶里程段对应的各样本电车的实时电池健康状态的平均值；以及
    建立所述实时电池健康状态的平均值和所述行驶里程段的第一对应关系。
  6. 根据权利要求2所述的方法，其特征在于，所述衰减系数基于所述样本电车对应型号的电池容量曲线确定，所述电池容量曲线为在特定温度下拟合得到的曲线。
  7. 根据权利要求1所述的方法，其特征在于，所述当前电池参数包括电池温度、放电电流和放电电压。
  8. 根据权利要求1所述的方法，其特征在于，所述机器学习模型基于XGBoost、支持向量机或者线性回归算法构建。
  9. 一种电池健康状态预测装置，其特征在于，所述装置包括：
    第一确定模块，用于确定目标电车的当前行驶里程和当前电池参数；
    第二确定模块，用于基于预先设置的行驶里程和电池健康状态的第一对应关系，确定所述当前行驶里程对应的当前电池健康状态；以及
    预测模块，用于将所述当前电池健康状态、所述当前行驶里程和所述当前电池参数输入预先训练的机器学习模型，以确定所述目标电车在目标行驶里程对应的目标电池健康状态。
  10. 根据权利要求9所述的装置，其特征在于，所述第一对应关系基于如下模块确定：
    获取模块，用于获取多个样本电车的实时电池参数以及所述样本电车对应型号的衰减系数，所述样本电车与所述目标电车的型号相同，每个样本电车的实时行驶里程均对应一个预先设置的行驶里程段；
    第三确定模块，用于针对每个样本电车，基于所述实时电池参数以及所述衰减系数，确定所述样本电车的实时电池健康状态；以及
    建立模块，用于针对每个行驶里程段，基于各样本电车的实时电池健康状态，建立所述第一对应关系。
  11. 根据权利要求10所述的装置，其特征在于，所述实时电池参数包括充电电流和累计充电时间；
    所述第三确定模块，具体用于：
    基于所述充电电流、所述累计充电时间以及所述衰减系数，确定所述样本电车的累计充电电量；以及
    基于所述累计充电电量和所述样本电车的标称容量，确定所述样本电车的实时电池健康状态。
  12. 根据权利要求11所述的装置，其特征在于，所述第三确定模块，具体用于：
    基于所述累计充电电量和所述样本电车的标称容量，确定所述样本电车的实时循环次数，所述实时循环次数用于表征所述样本电车的充/放电循环的次数；以及
    基于所述实时循环次数和预先设置的第二对应关系，确定所述实时循环次数对应的实时电池健康状态，所述第二对应关系用于表征循环次数和电池健康状态的对应关系。
  13. 根据权利要求10所述的装置，其特征在于，所述建立模块，具体用于：
    确定所述行驶里程段对应的各样本电车的实时电池健康状态的平均值；以及
    建立所述实时电池健康状态的平均值和所述行驶里程段的第一对应关系。
  14. 根据权利要求10所述的装置，其特征在于，所述衰减系数基 于所述样本电车对应型号的电池容量曲线确定，所述电池容量曲线为在特定温度下拟合得到的曲线。
  15. 根据权利要求9所述的装置，其特征在于，所述当前电池参数包括电池温度、放电电流和放电电压。
  16. 根据权利要求9所述的装置，其特征在于，所述机器学习模型基于XGBoost、支持向量机或者线性回归算法构建。
  17. 一种电子设备，包括存储器和处理器，其特征在于，所述存储器用于存储一条或多条计算机程序指令，其中，所述一条或多条计算机程序指令被所述处理器执行以实现如权利要求1-8中任一项所述的方法。
  18. 一种计算机可读存储介质，其特征在于，所述计算机可读存储介质内存储有计算机程序，所述计算机程序被处理器执行时实现权利要求1-8任一项所述的方法。
  19. 一种计算机程序产品，包括计算机程序/指令，其特征在于，所述计算机程序/指令被处理器执行时实现权利要求1-8任一项所述的方法。
PCT/CN2021/138740 2021-01-27 2021-12-16 电池健康状态预测方法、装置、电子设备和可读存储介质 WO2022161002A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202110112655.8A CN112666464B (zh) 2021-01-27 2021-01-27 电池健康状态预测方法、装置、电子设备和可读存储介质
CN202110112655.8 2021-01-27

Publications (1)

Publication Number Publication Date
WO2022161002A1 true WO2022161002A1 (zh) 2022-08-04

Family

ID=75414772

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/138740 WO2022161002A1 (zh) 2021-01-27 2021-12-16 电池健康状态预测方法、装置、电子设备和可读存储介质

Country Status (2)

Country Link
CN (1) CN112666464B (zh)
WO (1) WO2022161002A1 (zh)

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115308629A (zh) * 2022-08-23 2022-11-08 中国第一汽车股份有限公司 电池性能的测试方法、装置、存储介质及电子设备
CN115407211A (zh) * 2022-11-01 2022-11-29 北京航空航天大学 一种电动汽车锂电池健康状态在线预测方法及***
CN115424443A (zh) * 2022-09-20 2022-12-02 安徽江淮汽车集团股份有限公司 一种基于行驶数据的车辆异常监控方法
CN115542186A (zh) * 2022-11-30 2022-12-30 中国电力科学研究院有限公司 一种储能电池状态及一致性评价方法、装置、设备及介质
CN115840145A (zh) * 2022-11-29 2023-03-24 上海玫克生储能科技有限公司 电化学参数辨识方法、装置、设备及存储介质
CN116626526A (zh) * 2023-07-24 2023-08-22 宁德时代新能源科技股份有限公司 电池健康状态的检测方法、装置、终端及存储介质
CN117150275A (zh) * 2023-11-01 2023-12-01 宁德时代新能源科技股份有限公司 机器学习模型构建方法、电池健康度预测方法以及装置
CN117277445A (zh) * 2023-11-21 2023-12-22 国网（北京）新能源汽车服务有限公司 自驾游区域电力调度***
CN117434463A (zh) * 2023-09-21 2024-01-23 中国电子产品可靠性与环境试验研究所（（工业和信息化部电子第五研究所）（中国赛宝实验室）） 动力电池的剩余寿命评估方法、装置、设备和存储介质
WO2024077754A1 (zh) * 2022-10-10 2024-04-18 深圳先进技术研究院 储能电池健康状态预测方法、装置及智能终端

Families Citing this family (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112666464B (zh) * 2021-01-27 2023-11-07 北京嘀嘀无限科技发展有限公司 电池健康状态预测方法、装置、电子设备和可读存储介质
CN113391209B (zh) * 2021-05-26 2022-08-26 江苏小牛电动科技有限公司 电池健康状态的预测方法、装置、***和电池
CN113306449B (zh) * 2021-06-15 2023-11-03 安徽信息工程学院 一种新能源汽车的电池健康提醒方法及***
CN113433458B (zh) * 2021-07-16 2023-06-27 北京现代汽车有限公司 电池健康状态的确定方法和装置
CN114236415A (zh) * 2021-12-16 2022-03-25 湖北亿纬动力有限公司 一种动力电池健康状态检测方法、装置、设备及存储介质
CN115036590A (zh) * 2022-03-29 2022-09-09 东莞新能安科技有限公司 二次电池内阻检测方法、装置及电子设备
CN115219919A (zh) * 2022-07-27 2022-10-21 浙江极氪智能科技有限公司 电池健康状态预测方法、装置、电子设备及可读存储介质
CN115792627B (zh) * 2022-11-14 2023-06-06 上海玫克生储能科技有限公司 锂电池soh分析预测方法、装置、电子设备及存储介质
CN116051081B (zh) * 2023-03-30 2023-08-15 山东智捷专用车制造有限公司 一种新能源电动公铁两用牵引车的运检方法及***
CN116413609B (zh) * 2023-06-08 2023-08-29 江苏正力新能电池技术有限公司 一种电池跳水识别方法、装置、电子设备及存储介质

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103273921A (zh) * 2013-06-14 2013-09-04 清华大学 电动汽车续驶里程估计方法
CN105425167A (zh) * 2015-12-30 2016-03-23 广东顺德中山大学卡内基梅隆大学国际联合研究院 基于行车习惯和电池使用数据的电池重估和电池保险评估***
CN110261790A (zh) * 2019-04-10 2019-09-20 北京海博思创科技有限公司 电池健康状态的预估方法、装置及***
CN111323719A (zh) * 2020-03-18 2020-06-23 北京理工大学 一种电动汽车动力电池组健康状态在线确定方法和***
CN111381170A (zh) * 2020-05-15 2020-07-07 上海工程技术大学 基于大数据的电动汽车电池包健康状态预测方法及***
CN112666464A (zh) * 2021-01-27 2021-04-16 北京嘀嘀无限科技发展有限公司 电池健康状态预测方法、装置、电子设备和可读存储介质

Family Cites Families (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102981123A (zh) * 2012-11-05 2013-03-20 浙江吉利汽车研究院有限公司杭州分公司 动力电池剩余容量估算***及估算方法
KR101792975B1 (ko) * 2017-04-25 2017-11-02 한국기술교육대학교 산학협력단 수치적 시뮬레이션 데이터 기반 배터리의 수명 상태 예측 방법
CN108819722A (zh) * 2018-06-01 2018-11-16 汉腾汽车有限公司 一种电动汽车续航里程预估方法
EP3591414B1 (en) * 2018-07-03 2022-01-19 Electricité de France Method for evaluating an electric battery state of health
CN109934408A (zh) * 2019-03-18 2019-06-25 常伟 一种基于大数据机器学习进行汽车电池rul预测的应用分析方法
CN110221222B (zh) * 2019-04-30 2022-03-29 蜂巢能源科技有限公司 电池安全截止电压预测方法、装置及电池管理***
CN112180280A (zh) * 2020-09-27 2021-01-05 吉林大学 一种考虑电池健康状态的混合动力汽车电池寿命优化方法

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103273921A (zh) * 2013-06-14 2013-09-04 清华大学 电动汽车续驶里程估计方法
CN105425167A (zh) * 2015-12-30 2016-03-23 广东顺德中山大学卡内基梅隆大学国际联合研究院 基于行车习惯和电池使用数据的电池重估和电池保险评估***
CN110261790A (zh) * 2019-04-10 2019-09-20 北京海博思创科技有限公司 电池健康状态的预估方法、装置及***
CN111323719A (zh) * 2020-03-18 2020-06-23 北京理工大学 一种电动汽车动力电池组健康状态在线确定方法和***
CN111381170A (zh) * 2020-05-15 2020-07-07 上海工程技术大学 基于大数据的电动汽车电池包健康状态预测方法及***
CN112666464A (zh) * 2021-01-27 2021-04-16 北京嘀嘀无限科技发展有限公司 电池健康状态预测方法、装置、电子设备和可读存储介质

Cited By (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115308629A (zh) * 2022-08-23 2022-11-08 中国第一汽车股份有限公司 电池性能的测试方法、装置、存储介质及电子设备
CN115424443A (zh) * 2022-09-20 2022-12-02 安徽江淮汽车集团股份有限公司 一种基于行驶数据的车辆异常监控方法
CN115424443B (zh) * 2022-09-20 2024-04-26 安徽江淮汽车集团股份有限公司 一种基于行驶数据的车辆异常监控方法
WO2024077754A1 (zh) * 2022-10-10 2024-04-18 深圳先进技术研究院 储能电池健康状态预测方法、装置及智能终端
CN115407211A (zh) * 2022-11-01 2022-11-29 北京航空航天大学 一种电动汽车锂电池健康状态在线预测方法及***
CN115840145A (zh) * 2022-11-29 2023-03-24 上海玫克生储能科技有限公司 电化学参数辨识方法、装置、设备及存储介质
CN115840145B (zh) * 2022-11-29 2023-07-25 上海玫克生储能科技有限公司 电化学参数辨识方法、装置、设备及存储介质
CN115542186A (zh) * 2022-11-30 2022-12-30 中国电力科学研究院有限公司 一种储能电池状态及一致性评价方法、装置、设备及介质
CN115542186B (zh) * 2022-11-30 2023-03-14 中国电力科学研究院有限公司 一种储能电池状态及一致性评价方法、装置、设备及介质
CN116626526A (zh) * 2023-07-24 2023-08-22 宁德时代新能源科技股份有限公司 电池健康状态的检测方法、装置、终端及存储介质
CN116626526B (zh) * 2023-07-24 2023-12-22 宁德时代新能源科技股份有限公司 电池健康状态的检测方法、装置、终端及存储介质
CN117434463A (zh) * 2023-09-21 2024-01-23 中国电子产品可靠性与环境试验研究所（（工业和信息化部电子第五研究所）（中国赛宝实验室）） 动力电池的剩余寿命评估方法、装置、设备和存储介质
CN117150275A (zh) * 2023-11-01 2023-12-01 宁德时代新能源科技股份有限公司 机器学习模型构建方法、电池健康度预测方法以及装置
CN117150275B (zh) * 2023-11-01 2024-04-09 宁德时代新能源科技股份有限公司 机器学习模型构建方法、电池健康度预测方法以及装置
CN117277445B (zh) * 2023-11-21 2024-03-29 国网（北京）新能源汽车服务有限公司 自驾游区域电力调度***
CN117277445A (zh) * 2023-11-21 2023-12-22 国网（北京）新能源汽车服务有限公司 自驾游区域电力调度***

Also Published As

Publication number Publication date
CN112666464A (zh) 2021-04-16
CN112666464B (zh) 2023-11-07

Similar Documents

Publication Publication Date Title
WO2022161002A1 (zh) 电池健康状态预测方法、装置、电子设备和可读存储介质
CN109716150B (zh) 具有远程参数估计的二次电池管理***
Guo et al. Online estimation of SOH for lithium-ion battery based on SSA-Elman neural network
US11637330B2 (en) Battery charging method and apparatus
US10931128B2 (en) Method and apparatus to predict capacity fade rate of battery
Wei et al. Remaining useful life prediction and state of health diagnosis for lithium-ion batteries using particle filter and support vector regression
US12034319B2 (en) Server-side characterization of rechargeable batteries
Xu et al. An improved CNN-LSTM model-based state-of-health estimation approach for lithium-ion batteries
US11340298B2 (en) Apparatus and method for calculating state of charge of battery by reflecting noise
Liu et al. State of health estimation of power batteries based on multi-feature fusion models using stacking algorithm
Ruan et al. Artificial Intelligence-based health diagnostic of Lithium-ion battery leveraging transient stage of constant current and constant voltage charging
JP6411538B2 (ja) 予測システム、予測プログラム、予測装置
Wang et al. Co-estimation of state of charge and capacity for lithium-ion battery based on recurrent neural network and support vector machine
Shen et al. Toward fast and accurate SOH prediction for lithium-ion batteries
EP3985829B1 (en) Method and apparatus for charging battery
Liu et al. An improved method of state of health prediction for lithium batteries considering different temperature
Che et al. Predictive health assessment for lithium-ion batteries with probabilistic degradation prediction and accelerating aging detection
JP2023139227A (ja) 電池健康状態予測方法、装置、電子機器及び可読記憶媒体
Deng et al. Maximum available capacity and energy estimation based on support vector machine regression for lithium-ion battery
Zhang et al. A pattern-driven stochastic degradation model for the prediction of remaining useful life of rechargeable batteries
Dong et al. Data-driven predictive prognostic model for power batteries based on machine learning
CN114509695A (zh) 电池析锂检测方法、装置、设备、存储介质和计算机程序
Valladolid et al. Lithium-ion SOC optimizer consumption using accelerated particle swarm optimization and temperature criterion
Lin et al. Online estimation of diffusion-induced stress in cathode particles of li-ion batteries
US20240168094A1 (en) Systems, methods, and media for predicting degradation in energy storage devices

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21922592

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21922592

Country of ref document: EP

Kind code of ref document: A1