WO2023221587A1 - 电动车辆的动力电池的健康状态确定方法及服务器 - Google Patents

电动车辆的动力电池的健康状态确定方法及服务器 Download PDF

Info

Publication number
WO2023221587A1
WO2023221587A1 PCT/CN2023/077917 CN2023077917W WO2023221587A1 WO 2023221587 A1 WO2023221587 A1 WO 2023221587A1 CN 2023077917 W CN2023077917 W CN 2023077917W WO 2023221587 A1 WO2023221587 A1 WO 2023221587A1
Authority
WO
WIPO (PCT)
Prior art keywords
battery
fault
vehicle
historical
health
Prior art date
Application number
PCT/CN2023/077917
Other languages
English (en)
French (fr)
Inventor
沈小杰
廖增成
Original Assignee
深圳市道通合创数字能源有限公司
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 深圳市道通合创数字能源有限公司 filed Critical 深圳市道通合创数字能源有限公司
Publication of WO2023221587A1 publication Critical patent/WO2023221587A1/zh

Links

Classifications

    • 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
    • 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]
    • G01R31/378—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC] specially adapted for the type of battery or accumulator
    • 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]
    • G01R31/396—Acquisition or processing of data for testing or for monitoring individual cells or groups of cells within a battery

Definitions

  • This application relates to the field of new energy technology, and specifically to a method and server for determining the health status of a power battery of an electric vehicle.
  • the existing methods for evaluating the SOH value of power batteries mainly include electrochemical mechanism methods and big data artificial intelligence evaluation methods.
  • the electrochemical mechanism method is to obtain the SOH value of the power battery by analyzing the voltage, current and other data of the power battery through the charge and discharge process under certain conditions. Since the charging and discharging process of power batteries requires certain charging and discharging time and charging and discharging conditions, the efficiency of evaluating the SOH value of the above method is relatively low.
  • the big data artificial intelligence evaluation method uses long-term data analysis algorithms to obtain the SOH value of the power battery from a large amount of historical data. Since the SOH decay of a power battery is a slow and irreversible process, when conducting the first SOH evaluation of a power battery of an electric vehicle that has been used for a period of time, this method will require the use of long-term data analysis algorithms for subsequent evaluation due to the lack of initial values. When , the convergence of the algorithm is relatively slow and the SOH value cannot be output quickly and reliably.
  • One purpose of the embodiments of the present application is to provide a method and server for determining the health status of a power battery of an electric vehicle, so as to improve the technical problem of low efficiency in the existing technology when evaluating the health status of a power battery.
  • embodiments of the present application provide a method for determining the health status of a power battery of an electric vehicle, including:
  • battery data of the power battery and vehicle characteristics of the electric vehicle where the battery data includes real-time battery-related parameters and real-time fault type parameters;
  • the comprehensive health evaluation value of the power battery is determined.
  • the real-time battery-related parameter is one of driving mileage, cumulative charging capacity or cumulative discharge capacity of the power battery.
  • determining the comprehensive health evaluation value of the power battery based on the normal attenuation value and the fault attenuation value includes:
  • SOH t Std - ⁇ t - ⁇ t , determine the comprehensive health evaluation value of the power battery.
  • Std is the standard battery health status value
  • SOH t is the comprehensive health evaluation value
  • ⁇ t is the normal health status attenuation value
  • ⁇ t is the fault health state attenuation value.
  • the normal attenuation model is trained based on first training data of power batteries of a plurality of first historical vehicles that have the same characteristics as the vehicle;
  • the fault attenuation model is trained based on the second training data of power batteries of a plurality of second historical vehicles with the same characteristics as the vehicle and the normal attenuation model.
  • the first historical vehicle is a vehicle that has not experienced a specified battery failure
  • the second historical vehicle is a vehicle that has experienced a specified battery failure
  • the specified battery fault is any one of overvoltage fault, undervoltage fault, charging overcurrent fault, discharge overcurrent fault, high temperature fault, low temperature fault or specified serious fault.
  • the normal attenuation model is: Among them, ⁇ t is the normal health state attenuation value, is the normal decay rate, and P t is the real-time battery-related parameter.
  • the first training data includes first historical battery-related parameters and first battery health status values of a plurality of first historical vehicles that are the same as the vehicle characteristics;
  • the normal attenuation rate is: is the normal decay rate, ⁇ i is the undetermined decay rate of the i-th first historical vehicle, and n is the total number of first historical vehicles participating in training the normal decay model;
  • the undetermined attenuation rate of the i-th first historical vehicle is calculated based on the first historical battery-related parameters and the first battery health state value of the i-th first historical vehicle.
  • the to-be-determined attenuation rate of the i-th first historical vehicle is:
  • the unit of eta i is %/10000km
  • M wi is the mileage of the i-th first historical vehicle
  • SOH wi is the first battery health status value of the i-th first historical vehicle.
  • the fault attenuation model is: Among them, ⁇ t is the fault health state attenuation value, x j is the jth real-time fault type parameter, and ⁇ j is the fault attenuation rate of the jth real-time fault type parameter.
  • the real-time fault type parameter is one of the number of overvoltage faults, the number of undervoltage faults, the number of charging overcurrent faults, the number of discharge overcurrent faults, the number of high temperature faults, the number of low temperature faults or the number of designated serious faults.
  • the second training data includes second battery health status values, second historical battery-related parameters and historical fault type parameters of a plurality of second historical vehicles that are the same as the vehicle characteristics;
  • the fault attenuation rate of the j-th second historical vehicle is calculated based on the linear regression algorithm by performing determinant calculation on the health difference values and historical fault type parameters of multiple second historical vehicles;
  • the health difference value of the j-th second historical vehicle is the difference between the expected health value of the j-th second historical vehicle and the j-th second battery health value;
  • the expected health value of the j-th second historical vehicle is calculated based on the j-th second historical battery-related parameters and the normal attenuation model.
  • the fault decay rate of the jth real-time fault type parameter is:
  • ⁇ yj is the health difference value of the j-th second historical vehicle, is the expected health value of the j-th second historical vehicle, SOH yj is the second battery health value of the j-th second historical vehicle, Std is the standard battery health state value, is the normal decay rate, P yj is the fault history battery-related parameter of the j-th second historical vehicle, x sj is the j-th historical fault type parameter of the s-th second historical vehicle, ⁇ j is the j-th historical fault type The parameter's fault decay rate.
  • an embodiment of the present application provides a server, including:
  • the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor, so that the at least one processor can perform the above-mentioned determination of the health status of the power battery of the electric vehicle. method.
  • embodiments of the present application provide a computer-readable storage medium that stores computer-executable instructions, and the computer-executable instructions are used to cause the computer to execute the above-mentioned power battery of an electric vehicle. How to determine the health status of Law.
  • inventions of the present application provide a computer program product.
  • the computer program product includes a computer program stored on a non-volatile computer-readable storage medium.
  • the computer program includes program instructions. When the program When the instruction is executed by the electronic device, the electronic device is caused to execute the above-mentioned method for determining the health status of the power battery of the electric vehicle.
  • this application has at least the following beneficial effects:
  • the battery data of the power battery and the vehicle characteristics of the electric vehicle are obtained, and the battery data includes real-time battery-related parameters.
  • the normal health state attenuation value of the power battery is calculated.
  • the power battery is calculated.
  • the fault health state attenuation value determines the comprehensive health evaluation value of the power battery based on the normal health state attenuation value and the fault health state attenuation value.
  • the health status of the power battery can be quickly assessed without the need for the first health status value of the power battery, thus improving the technical problem of inefficiency in the existing technology when assessing the health status of the power battery.
  • Figure 1 is a schematic structural diagram of a power battery health status determination system provided by an embodiment of the present application
  • Figure 2 is a schematic flowchart of a method for determining the health status of a power battery of an electric vehicle provided by an embodiment of the present application;
  • Figure 3 is a schematic diagram of a scenario for training a normal attenuation model and a fault attenuation model provided by an embodiment of the present application;
  • Figure 4 is a schematic structural diagram of a device for determining the health state of a power battery of an electric vehicle provided by an embodiment of the present application
  • Figure 5 is a schematic circuit structure diagram of a server provided by an embodiment of the present application.
  • the embodiment of the present application provides a power battery health status determination system.
  • the health status determination system 100 includes a vehicle communication device 11 (Vehicle Communication Interface, VCI) and a server 12.
  • the server 12 is communicatively connected to the vehicle communication device 11.
  • communication connections include wired communication connections or wireless communication connections
  • wired communication connections include various communication connections that use tangible media such as metal wires and optical fibers to transmit information.
  • Wireless communication connections include 5G communication, 4G communication, 3G communication, 2G communication, CDMA, Bluetooth, wireless broadband, ultra-wideband communication, near field communication, CDMA2000, GSM, ISM, RFID, UMTS/3GPPw/HSDPA, WiMAX, Wi-Fi Or ZigBee etc.
  • the vehicle communication device 11 is used to plug into the OBD interface (On Board Diagnostics, OBD) of the electric vehicle 13.
  • the vehicle communication device 11 communicates with the electric vehicle 13 based on the OBD interface to obtain vehicle data of the electric vehicle 13.
  • the vehicle data includes faults. code, battery data or vehicle characteristics. Fault codes are used to indicate fault types of electric vehicles.
  • the battery data is data associated with the power battery of the electric vehicle.
  • the vehicle characteristics are vehicle type characteristics and/or location characteristics used to represent the electric vehicle.
  • the vehicle communication device 11 obtains the battery data and vehicle characteristics of the electric vehicle 13, and then packages the battery data and vehicle characteristics and sends them to the server 12.
  • the server 12 selects the corresponding normal attenuation model 14 and fault attenuation model according to the vehicle characteristics. 15, and input the battery data into the normal attenuation model 14 and the fault attenuation model 15 respectively, and determine the comprehensive health evaluation value of the power battery based on the output results of the normal attenuation model 14 and the output results of the fault attenuation model 15.
  • the server here may be a physical server or a logical server virtualized by multiple physical servers.
  • the server can also be a server group composed of multiple servers that can communicate with each other, and each functional module can be distributed on each server in the server group.
  • the embodiments of the present application provide a method for determining the health status of the power battery of an electric vehicle.
  • the method provided by this embodiment is applied in automobile aftermarket diagnosis, battery maintenance, public charging, insurance loss assessment, Multiple application scenarios such as residual value evaluation. Please refer to Figure 2.
  • the methods for determining the health status of the power battery of electric vehicles include:
  • the battery data includes real-time battery-related parameters and real-time fault type parameters.
  • the vehicle communication device communicates with the electric vehicle.
  • the electric vehicle sends battery data and vehicle characteristics of the power battery to the vehicle communication device.
  • the vehicle communication device packages the battery data and vehicle characteristics and sends them to the server.
  • the battery data includes real-time battery-related parameters.
  • the real-time battery-related parameters are parameters used to represent the health status of the power battery.
  • the real-time fault type parameters are used to represent the number of times the corresponding fault type occurs in the power battery.
  • the real-time battery-related parameter is one of driving mileage, cumulative charging capacity, or cumulative discharge capacity of the power battery.
  • the mileage is the number of kilometers traveled by an electric vehicle, and the unit of mileage can be per 10,000 kilometers (10,000 kilometers).
  • driving mileage There is a correlation between driving mileage and the health status of the power battery. The greater the driving mileage, the stronger the attenuation of the power battery. The smaller the driving mileage, the weaker the attenuation of the power battery. Therefore, the driving mileage can reflect the health status of the power battery. .
  • the cumulative charging capacity is the sum of the charging capacity of the power battery in multiple charges
  • the cumulative discharge capacity is the sum of the discharge capacity of the power battery in multiple discharges.
  • the charging capacity is C1.
  • the charging capacity is C2
  • the cumulative charging capacity C1 + C2.
  • the battery data includes real-time fault type parameters.
  • the real-time fault type parameters are the number of overvoltage faults, the number of undervoltage faults, the number of charging overcurrent faults, the number of discharge overcurrent faults, and the number of high temperature faults. One of the number of low temperature failures or other serious failures.
  • the number of overvoltage faults is the total number of overvoltage faults, where the overvoltage fault e1 includes an overvoltage fault of each single battery or a total system overvoltage fault, etc.
  • the number of undervoltage faults is the total number of undervoltage faults, where the undervoltage fault e2 includes an undervoltage fault of each single cell, a short circuit fault within a single battery, or a total system voltage undervoltage fault, etc.
  • the number of charging overcurrent faults is the total number of charging overcurrent fault types, where the charging overcurrent fault e3 includes fast and slow charging overcurrent faults or abnormal charging current faults.
  • the number of discharge overcurrent faults is the total number of discharge overcurrent faults, where the discharge overcurrent fault e4 includes discharge overcurrent fault, abnormal discharge current fault or short circuit fault, etc.
  • High-temperature faults e5 include battery charging high-temperature faults, battery discharge high-temperature faults, environmental high-temperature faults, thermal management high-temperature faults, or controller high-temperature faults.
  • the number of low temperature faults is the total number of low temperature faults.
  • the low temperature fault e6 includes battery charging low temperature fault, battery discharge low temperature fault, environmental low temperature fault or thermal management low temperature fault, etc.
  • the number of specified serious faults is the total number of specified serious faults, in which the specified serious fault e7 is a fault other than the above 6 fault types.
  • the vehicle characteristics include the model characteristics of the electric vehicle, and the model characteristics are used for the model of the electric vehicle.
  • the model characteristics can be represented by MMYB information, and the MMYB information is the manufacturer (Make), model ( The collective name for Model, Year, Battery and other information.
  • the vehicle characteristics include territorial characteristics of the electric vehicle.
  • the territorial characteristics are used to represent the area where the electric vehicle is driven.
  • the territorial characteristics can be represented by geographical location information, and the geographical location information can be generated by positioning the electric vehicle's positioning system. , for example, the geographical location information is Nanshan District, Shenzhen, Guangdong.
  • the normal attenuation model is the model used to calculate the attenuation value of the normal health state.
  • the attenuation value of the normal health state is the driven force.
  • the attenuation value of the power battery is evaluated.
  • the normal attenuation model can output the normal health state attenuation value of the power battery.
  • this embodiment can construct a normal attenuation model based on vehicle characteristics. Different vehicle characteristics correspond to different normal attenuation models, which is beneficial to improving the accuracy of calculating the normal health state attenuation value.
  • this embodiment obtains vehicle characteristics, and searches the normal model library for a normal attenuation model that matches the vehicle characteristics according to the vehicle characteristics and normal feature labels, where the normal model library includes multiple normal feature labels and normal feature labels.
  • the normal decay model corresponding to the feature label is a normal attenuation model that matches the vehicle characteristics according to the vehicle characteristics and normal feature labels.
  • the fault attenuation model is a model used to calculate the attenuation value of the fault health state.
  • the fault health state attenuation value is used to evaluate the attenuation value of the power battery from the fault dimension of the power battery.
  • the fault attenuation model can output the fault health state attenuation value of the power battery.
  • this embodiment can construct a fault attenuation model based on vehicle characteristics. Different vehicle characteristics correspond to different fault attenuation models. , which is helpful to improve the accuracy of calculating the fault health state attenuation value.
  • this embodiment obtains vehicle characteristics, searches for a fault attenuation model that matches the vehicle characteristics in the fault model library according to the vehicle characteristics and fault characteristic labels, where the fault model library includes multiple fault characteristic labels and fault characteristic labels.
  • the fault attenuation model corresponding to the feature label.
  • the vehicle characteristics are car model characteristics. This embodiment searches for normal attenuation models and fault attenuation models that match the car model characteristics. That is, a single car model feature can correspond to a normal attenuation model and a fault attenuation model respectively.
  • the vehicle features are local features. This embodiment searches for normal attenuation models and fault attenuation models that match the local features respectively. That is, a single local feature can correspond to a normal attenuation model and a fault attenuation model respectively.
  • the vehicle characteristics are vehicle type characteristics and location characteristics.
  • This embodiment searches for a normal attenuation model and a fault attenuation model that match the vehicle type characteristics and the location characteristics respectively. That is, the normal attenuation model and the fault attenuation model are both represented by The characteristics of the vehicle model and the characteristics of the territory are jointly determined.
  • the comprehensive health evaluation value is a value used to evaluate the health of the power battery, where the comprehensive health evaluation value includes the battery health status value.
  • the battery health status value can be expressed by the SOH value as mentioned above.
  • the standard battery health status value Std can be customized by engineers according to business requirements. For example, the standard battery health status value Std is 100%.
  • the method for determining the health status of the power battery of the electric vehicle further includes: determining whether the comprehensive health evaluation value is less than or equal to the first preset threshold; if so, generating battery warning information; if not, generating battery warning information based on the comprehensive health evaluation value. Value, generate battery health evaluation information, in which the battery warning information is used to prompt the user to replace or maintain the power battery.
  • the battery warning information can be any form of warning information, such as the battery warning information is text warning information, voice warning information or flash warning information. .
  • the battery health evaluation information is the health level information for evaluating the health status of the power battery. For example, the battery health evaluation information includes "very good", "better” or "average".
  • the first preset threshold can be customized by engineers according to business requirements, for example, the first preset threshold is 70%.
  • the difference from the above embodiments is that the comprehensive health evaluation value includes the battery health attenuation value.
  • This embodiment can add the normal health state attenuation value and the fault health state attenuation value to obtain the battery health attenuation value.
  • the method for determining the health status of the power battery of the electric vehicle further includes: determining whether the comprehensive health evaluation value is greater than or equal to the second preset threshold; if so, generating battery warning information; if not, generating battery warning information based on the comprehensive health evaluation value. value to generate battery health evaluation information.
  • the second preset threshold can be customized by engineers according to business requirements, for example, the second preset threshold is 80%.
  • this embodiment does not need to control the power battery to charge and discharge, which means that the user does not need to set up equipment for controlling the power battery to charge and discharge.
  • the health status of the power battery can be quickly assessed even when the battery is stopped, thereby improving the technical problem of inefficiency in the existing technology when assessing the health status of the power battery, and also expanding the application scope of this method.
  • this embodiment does not require the first health status value of the power battery. That is, for electric vehicles without historical data, the electric vehicle can be detected once and output the electric power value.
  • the comprehensive health evaluation value of the vehicle thus improves the technical problem of inefficiency in the existing technology when evaluating the health status of the power battery.
  • this embodiment can be used to provide the initial SOH value. From this, it can also be seen that The application scope of the method provided by this embodiment is more flexible and wider.
  • electrochemical mechanism methods or big data artificial intelligence evaluation methods often lack initial values, causing the algorithm to converge slowly.
  • this embodiment can provide the initial SOH value, and then use electrochemical mechanism methods or big data later.
  • artificial intelligence evaluation methods it is helpful to improve the convergence speed and accuracy of the above two algorithms.
  • the normal decay model is trained based on first training data of power batteries of a plurality of first historical vehicles that have the same characteristics as the vehicle.
  • the fault attenuation model is trained based on the second training data of the power batteries of multiple second historical vehicles with the same characteristics as the vehicle and the normal attenuation model.
  • the first training data is data used to train and generate a normal attenuation model.
  • the second training data is data used to cooperate with the normal attenuation model to train and generate the fault attenuation model.
  • the first historical vehicle is: the electric vehicle before training to generate a normal attenuation model.
  • the second historical vehicle is: the electric vehicle before the fault attenuation model is generated by training. For example, to train and generate a normal attenuation model at time point t11, this embodiment selects training data of multiple first electric vehicles as the first training data, where the first electric vehicle is the first historical vehicle. At time point t22, the fault attenuation model is trained and generated. In this embodiment, training data of multiple second electric vehicles are selected as the second training data, where the second electric vehicles are second historical vehicles.
  • This embodiment first trains to obtain a normal attenuation model, and then trains on the basis of the normal attenuation model and the second training data to obtain a fault attenuation model. Therefore, this embodiment can use the evolution of the normal attenuation model to generate a fault attenuation model. Compared with Instead of constructing a fault attenuation model separately from the normal attenuation model, the method provided in this embodiment can improve the integration between the normal attenuation model and the fault attenuation model, which is beneficial to generating a more reliable and accurate fault attenuation model.
  • the vehicle characteristics are car model characteristics.
  • the training data of the power battery of the first historical vehicle that is the same as the car model characteristics can be selected as the first training data, and the training data of the power battery of the second historical vehicle that is the same as the car model characteristics can be selected.
  • the training data of the power battery is used as the second training data.
  • the difference from the above embodiments is that the vehicle characteristics are local characteristics.
  • the training data of the power battery of the first historical vehicle that is the same as the local characteristics can be selected as the first training data.
  • the vehicle characteristics can be selected to be the same as the local characteristics.
  • the training data of the power battery of the second historical vehicle with the same characteristics is used as the second training data.
  • the difference from the above embodiment is that the vehicle characteristics include vehicle type characteristics and geographical characteristics.
  • the training data of the power battery of the first historical vehicle that is the same as the vehicle type characteristics and the same geographical characteristics can be selected as the third
  • the training data of the power battery of the second historical vehicle that is the same as the vehicle model characteristics and the same as the location characteristics can be selected as the second training data.
  • both the first historical vehicle and the second historical vehicle are vehicles in which specified battery faults have occurred, and the specified battery faults are overvoltage faults, undervoltage faults, charging overcurrent faults, discharge overcurrent faults, and high temperature faults. Either a low temperature fault or a specified critical fault.
  • This embodiment can generate a normal attenuation model based on the first training data of the first historical vehicle, so as to use the normal attenuation model to output the normal health state attenuation value, and can generate a fault based on the second training data of the second historical vehicle and the normal attenuation model training. Attenuation model to use the fault attenuation model to output the fault health state attenuation value. Compared with the existing technology, this embodiment can also quickly evaluate the health state of the power battery.
  • the difference from the above embodiments is that the first historical vehicle is a vehicle that has experienced a specified battery failure, and the second historical vehicle is a vehicle that has not experienced a specified battery failure.
  • this embodiment is based on this The first training data and the second training data at that time are used to obtain the normal attenuation model and the fault attenuation model respectively, which can quickly assess the health status of the power battery.
  • the difference from the above embodiment is that the first historical vehicle is a vehicle that has not experienced a specified battery failure, and the second historical vehicle is a vehicle that has experienced a specified battery failure. Since the first historical vehicle is a vehicle without a specified battery failure, the training data of a vehicle without a specified battery failure is selected as the first training data. The impact of the training data of a vehicle with a specified battery failure on the normal attenuation model has been eliminated, and there is It is beneficial to generate a more accurate and reliable normal attenuation model.
  • the second historical vehicle is a vehicle that has experienced a specified battery failure
  • the impact of the training data on the vehicle that has not experienced the specified battery failure on the fault attenuation model has been eliminated, and the training data of the vehicle that has experienced the specified battery failure is selected as The second training data is conducive to generating a more accurate and reliable fault attenuation model.
  • the normal decay model is: Among them, ⁇ t is the normal health state attenuation value, is the normal decay rate, and P t is the real-time battery-related parameter. Among them, the normal decay rate evaluates the decay rate of the power battery from the normal dimensions of the power battery.
  • the server When the server obtains the real-time battery-related parameter P t of the electric vehicle, it can substitute the real-time battery-related parameter P t into the normal attenuation model, and the normal attenuation model can output the normal health state attenuation value ⁇ t .
  • the first training data includes first historical battery-related parameters and first battery health status values of a plurality of first historical vehicles that are the same as the vehicle characteristics.
  • the first historical battery-related parameter may be one of the driving mileage, the cumulative charging capacity or the cumulative discharge capacity of the power battery, and the first battery health status value is to obtain the battery health status value of the first historical vehicle in advance.
  • the first battery health status value can be calculated by any suitable algorithm for evaluating the power battery SOH value.
  • the server directly calls the result provided by the algorithm for evaluating the power battery SOH value.
  • the first battery health status value can also be calculated by this embodiment. The method provided is obtained by evaluating the SOH value of the power battery, and then the existing first battery health status value is subsequently used for iterative updates.
  • the normal decay rate is: is the normal decay rate, ⁇ i is the undetermined decay rate of the i-th first historical vehicle, and n is the total number of first historical vehicles participating in training the normal decay model.
  • the pending decay rate of the i-th first historical vehicle is calculated based on the first historical battery-related parameters and the first battery health state value of the i-th first historical vehicle.
  • the total number n of first historical vehicles participating in training the normal decay model is 100
  • the undetermined decay rate of the first first historical vehicle is eta 1
  • the undetermined decay rate of the second first historical vehicle is eta 2
  • the normal attenuation rate is obtained by averaging the undetermined attenuation rates of multiple first historical vehicles. It is beneficial to obtain a more accurate and reliable normal attenuation rate.
  • the undetermined decay rate of the i-th first historical vehicle is: The unit of eta i is %/10000km, M wi is the mileage of the i-th first historical vehicle, and SOH wi is the first battery health status value of the i-th first historical vehicle.
  • M wi is 100000km
  • SOH wi is 95%
  • eta i is 0.5
  • its unit is %/10000km, that is, %/10,000 kilometers.
  • the normal attenuation rate is obtained is 0.5.
  • the difference from the above embodiment is that when the first historical battery-related parameter can also be the cumulative charging capacity or the cumulative discharge capacity, the undetermined decay rate of the i-th first historical vehicle is: The unit of eta i is %/Ah, and C wi is the cumulative charging capacity or cumulative discharge capacity of the i-th first historical vehicle.
  • the fault decay model is: Among them, ⁇ t is the fault health state attenuation value, x j is the jth real-time fault type parameter, and ⁇ j is the fault attenuation rate of the jth real-time fault type parameter. Among them, the fault attenuation rate ⁇ j of the jth real-time fault type parameter evaluates the attenuation rate of the power battery under the action of the jth real-time fault type from the fault dimension of the power battery.
  • the fault attenuation model has a total of 7 real-time fault type parameters, and the fault attenuation rates of the 7 real-time fault type parameters are 0.244, 0.112, 1.112, 0.774, 0.365, 0.585, and 0.119 respectively.
  • the real-time fault type parameter is one of the number of overvoltage faults, the number of undervoltage faults, the number of charging overcurrent faults, the number of discharge overcurrent faults, the number of high temperature faults, the number of low temperature faults, or the number of specified severe faults.
  • the health status decay value is:
  • the seven real-time fault type parameters of car D3 are input into the fault attenuation model, and the fault health state attenuation value is:
  • the second training data includes second battery health status values, second historical battery-related parameters, and historical fault type parameters of a plurality of second historical vehicles that are the same as the vehicle characteristics.
  • the second historical battery-related parameter may be one of the driving mileage, the cumulative charging capacity or the cumulative discharge capacity of the power battery, and the second battery health status value is to obtain the battery health status value of the second historical vehicle in advance.
  • the second battery health status value can be calculated by any suitable algorithm for evaluating the SOH value of the power battery. The server directly calls the result provided by the algorithm for evaluating the SOH value of the power battery. Alternatively, the second battery health status value can also be calculated by this embodiment.
  • the method provided is obtained by evaluating the SOH value of the power battery, and then the existing second battery health status value is subsequently used for iterative updates.
  • the historical fault type parameter can be one of the number of overvoltage faults, the number of undervoltage faults, the number of charging overcurrent faults, the number of discharge overcurrent faults, the number of high temperature faults, the number of low temperature faults or the number of specified serious faults.
  • the fault attenuation rate of the j-th second historical vehicle is obtained by calculating the determinant of the health differences and historical fault type parameters of multiple second historical vehicles based on the linear regression algorithm. Among them, the health of the j-th second historical vehicle The difference is the difference between the expected health value of the j-th second historical vehicle and the j-th second battery health value. The expected health value of the j-th second historical vehicle is based on the j-th second historical battery associated parameter and Calculated using the normal attenuation model.
  • linear regression algorithm here can choose any suitable type of linear regression algorithm, for example, the linear regression algorithm uses the least squares method.
  • the expected health value of the j-th second historical vehicle is: is the expected health value of the j-th second historical vehicle, Std is the standard battery health value, is the normal attenuation rate, and P yj is the fault history battery-related parameter of the j-th second historical vehicle.
  • the j-th second battery health value can be obtained by the server calling the existing battery health value of the j-th second historical vehicle.
  • the health difference value of the j-th second historical vehicle is: ⁇ yj is the health difference value of the j-th second historical vehicle, and SOH yj is the second battery health value of the j-th second historical vehicle.
  • the fault attenuation rate of the j-th second historical vehicle is:
  • x sj is the j-th historical fault type parameter of the s-th second historical vehicle
  • ⁇ j is the fault attenuation rate of the j-th historical fault type parameter
  • the unit of each fault attenuation rate is %.
  • the seven historical fault type parameters of electric vehicle L1 are ⁇ 4, 6, 8, 3, 1, 5, 4 ⁇ , respectively. is 91.6%, and SOH y1 is 75%.
  • the seven historical fault type parameters of electric vehicle L2 are ⁇ 2, 4, 5, 7, 2, 5, 3 ⁇ , is 95.9% and SOH y2 is 80%.
  • the seven historical fault type parameters of electric vehicle L3 are ⁇ 3,5,6,7,1,6,4 ⁇ respectively. is 89.7% and SOH y3 is 72%.
  • the seven historical fault type parameters of electric vehicle L4 are ⁇ 3, 6, 5, 5, 3, 4, 3 ⁇ , respectively. is 96.6% and SOH y4 is 82%.
  • the seven historical fault type parameters of electric vehicle L5 are ⁇ 2, 6, 8, 4, 1, 4, 4 ⁇ , respectively.
  • ⁇ 1 , ⁇ 2 , ⁇ 3 , ⁇ 4 , ⁇ 5 , ⁇ 6 , and ⁇ 7 respectively: 0.244, 0.112, 1.112, 0.774, 0.365, 0.585, 0.119,
  • the fault attenuation rates of 7 kinds of battery faults can be obtained. It can be understood that this embodiment can also add 8 or 9 kinds of more battery faults, and match a fault attenuation rate for each battery fault, through the least squares method.
  • the failure decay rate for each battery failure is estimated.
  • this embodiment can also add a smaller number of battery faults, and match a fault attenuation rate for each battery fault. For example, only add two types of battery faults, and estimate the probability of these two battery faults through the least squares method. Failure decay rate.
  • this embodiment correlates multiple battery faults, and uses a linear regression algorithm to comprehensively calculate the fault attenuation rate of each battery fault, and then uses the fault attenuation rates of multiple battery faults to substitute Compared with using a single battery fault to calculate the fault health state attenuation value ⁇ t , this embodiment can synthesize and correlate various battery faults with the fault health state attenuation value ⁇ t , thereby reducing the error of a single battery fault on the fault health state attenuation value. The influence of ⁇ t can train a more reliable and accurate fault attenuation model.
  • Figure 1 has shown the practical application process of quickly evaluating the comprehensive health evaluation value of the power battery.
  • the model training process will be detailed below in conjunction with Figure 3. The details are as follows:
  • each first historical vehicle 31 and each second historical vehicle 32 are the same.
  • each first historical vehicle 31 and each second historical vehicle 32 have the same model characteristics and the same geographical characteristics. .
  • the first vehicle communication device 33 is plugged into the OBD interface of the first historical vehicle 31.
  • the first vehicle communication device 33 communicates with the first historical vehicle 31 based on the OBD interface to obtain the first training data of the first historical vehicle 31.
  • a training data includes first battery data and first vehicle characteristics, wherein the battery data includes a first driving mileage, a first battery health state value and a first fault type parameter.
  • the first vehicle communication device 33 sends the first training data to the server 34, and the server 34 trains and generates a normal attenuation model 35 based on the first driving mileage and the first battery health state value.
  • the second vehicle communication device 36 is plugged into the OBD interface of the second historical vehicle 32.
  • the second vehicle communication device 36 communicates with the second historical vehicle 32 based on the OBD interface to obtain the second training data of the second historical vehicle 32.
  • the second training data includes second battery data and second vehicle characteristics, where the battery data includes a second driving mileage, a second battery health state value, and a second fault type parameter.
  • the second vehicle communication device 36 sends the second training data to the server 34 .
  • the server 34 inputs the second driving mileage into the normal decay model 35 to obtain the expected health value. Next, the server 34 subtracts the expected health value and the second battery health state value to obtain the health difference value. Then, the server calculates the fault attenuation rate of each battery fault based on the health difference values of the plurality of second historical vehicles 32 and the second fault type parameters. Finally, the server generates a fault decay model37 based on the fault decay rate of each battery fault.
  • the embodiments of the present application provide a device for determining the health state of a power battery of an electric vehicle.
  • the device for determining the health status of the power battery of the electric vehicle can be a software module.
  • the software module includes a number of instructions, which are stored in a memory.
  • the processor can access the memory and call the instructions for execution to complete the above-mentioned embodiments. elaboration of electricity Method for determining the health status of the power battery of an electric vehicle.
  • the device for determining the health status of the power battery of an electric vehicle can also be built by hardware devices.
  • the device for determining the health status of the power battery of an electric vehicle can be built by one or more than two chips. Each chip can work in coordination with each other to complete the health status determination method of the power battery of the electric vehicle described in each of the above embodiments.
  • the device for determining the health status of the power battery of an electric vehicle can also be built from various types of logic devices, such as general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASICs), and field-programmable gate arrays. (FPGA), microcontroller, ARM (Acorn RISC Machine) or other programmable logic devices, discrete gate or transistor logic, discrete hardware components, or any combination of these components.
  • the health status determination device 400 of the power battery of an electric vehicle includes: a data acquisition module 41 , a normal attenuation calculation module 42 , a fault attenuation calculation module 43 and a health status determination module 44 .
  • the data acquisition module 41 is used to acquire battery data of the power battery and vehicle characteristics of the electric vehicle.
  • the battery data includes real-time battery-related parameters and real-time fault type parameters.
  • the normal decay calculation module 42 is used to calculate the normal health state decay value of the power battery based on the real-time battery-related parameters and the normal decay model corresponding to the vehicle characteristics.
  • the fault attenuation calculation module 43 is used to calculate the fault health state attenuation value of the power battery according to the real-time fault type parameters and the fault attenuation model corresponding to the vehicle characteristics.
  • the health state determination module 44 is used to determine the comprehensive health evaluation value of the power battery based on the normal health state attenuation value and the fault health state attenuation value.
  • this embodiment When evaluating the health status of the power battery, this embodiment does not need to control the charging and discharging of the power battery, nor does it require the first health status value of the power battery.
  • the health status of the power battery can be quickly assessed, thereby improving the existing technology for evaluating the power battery. There are technical inefficiencies in the health state.
  • the real-time battery-related parameter is one of driving mileage, cumulative charging capacity, or cumulative discharge capacity of the power battery.
  • the normal decay model is trained based on first training data of power batteries of a plurality of first historical vehicles that have the same characteristics as the vehicle.
  • the fault attenuation model is trained based on the second training data of the power batteries of multiple second historical vehicles with the same characteristics as the vehicle and the normal attenuation model.
  • the first historical vehicle is a vehicle that did not experience the specified battery failure.
  • the second historical vehicle is a vehicle in which a specified battery failure has occurred.
  • the specified battery fault is any one of overvoltage fault, undervoltage fault, charging overcurrent fault, discharge overcurrent fault, high temperature fault, low temperature fault or specified serious fault.
  • the normal decay model is: Among them, ⁇ t is the normal health state attenuation value, is the normal decay rate, and P t is the real-time battery-related parameter.
  • the first training data includes first historical battery-related parameters and first battery health status values of a plurality of first historical vehicles that are the same as the vehicle characteristics.
  • the normal decay rate is: is the normal decay rate, ⁇ i is the undetermined decay rate of the i-th first historical vehicle, and n is the total number of first historical vehicles participating in training the normal decay model.
  • the pending decay rate of the i-th first historical vehicle is calculated based on the first historical battery-related parameters and the first battery health state value of the i-th first historical vehicle.
  • the undetermined decay rate of the i-th first historical vehicle is: The unit of eta i is %/10000km, M wi is the mileage of the i-th first historical vehicle, and SOH wi is the first battery health status value of the i-th first historical vehicle.
  • the fault decay model is: Among them, ⁇ t is the fault health state attenuation value, x j is the jth real-time fault type parameter, and ⁇ j is the fault attenuation rate of the jth real-time fault type parameter.
  • the real-time fault type parameter is one of the number of overvoltage faults, the number of undervoltage faults, the number of charging overcurrent faults, the number of discharge overcurrent faults, the number of high temperature faults, the number of low temperature faults, or the number of specified severe faults.
  • the second training data includes second battery health status values, second historical battery-related parameters, and historical fault type parameters of a plurality of second historical vehicles that are the same as the vehicle characteristics.
  • the fault attenuation rate of the j-th second historical vehicle is calculated based on the linear regression algorithm by calculating the determinant of the health differences and historical fault type parameters of multiple second historical vehicles.
  • the health difference value of the j-th second historical vehicle is the difference between the expected health value of the j-th second historical vehicle and the j-th second battery health value;
  • the expected health value of the j-th second historical vehicle is calculated based on the j-th second historical battery-related parameters and the normal decay model.
  • the fault decay rate of the jth real-time fault type parameter is:
  • ⁇ yj is the health difference value of the j-th second historical vehicle, is the expected health value of the j-th second historical vehicle, SOH yj is the second battery health value of the j-th second historical vehicle, Std is the standard battery health state value, is the normal decay rate, P yj is the fault history battery-related parameter of the j-th second historical vehicle, x sj is the j-th historical fault type parameter of the s-th second historical vehicle, ⁇ j is the j-th historical fault type The parameter's fault decay rate.
  • the above-mentioned device for determining the health status of the power battery of an electric vehicle can execute the method for determining the health status of the power battery of an electric vehicle provided by the embodiment of the present application, and has functional modules and beneficial effects corresponding to the execution method.
  • the method for determining the health status of the power battery of an electric vehicle provided in the embodiments of this application.
  • server 500 includes one or more processors 51 and memory 52 .
  • processors 51 is taken as an example in FIG. 5 .
  • the processor 51 and the memory 52 may be connected through a bus or other means.
  • the connection through a bus is taken as an example.
  • the memory 52 can be used to store non-volatile software programs, non-volatile computer executable programs and modules, such as program instructions corresponding to the method for determining the health status of the power battery of an electric vehicle in the embodiment of the present application. /module.
  • the processor 51 executes various functional applications and data processing of the health status determination device of the power battery of the electric vehicle by running non-volatile software programs, instructions and modules stored in the memory 52, that is, implementing the method provided by the above embodiments.
  • Memory 52 may include high-speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other non-volatile solid-state storage device.
  • the memory 52 optionally includes memory located remotely relative to the processor 51 , and these remote memories may be connected to the processor 51 through a network. Examples of the above-mentioned networks include but are not limited to the Internet, intranets, local area networks, mobile communication networks and combinations thereof.
  • the program instructions/modules are stored in the memory 52, and when executed by the one or more processors 51, perform the health status determination method of the power battery of the electric vehicle in any of the above method embodiments.
  • Embodiments of the present application also provide a computer-readable storage medium that stores computer-executable instructions, and the computer-executable instructions are executed by one or more processors, such as a process in Figure 5
  • the processor 51 can enable the above one or more processors to execute the method for determining the health status of the power battery of the electric vehicle in any of the above method embodiments.
  • Embodiments of the present application also provide a computer program product.
  • the computer program product includes a computer program stored on a non-volatile computer-readable storage medium.
  • the computer program includes program instructions. When the program instructions are processed by a server When executed, the server is caused to execute any one of the methods for determining the health status of the power battery of the electric vehicle.
  • the device or equipment embodiments described above are only illustrative, in which the unit modules described as separate components may or may not be physically separated, and the components shown as modular units may or may not be physical units. , that is, it can be located in one place, or it can be distributed to multiple network module units. You can select some or all of the modules according to actual needs. achieve the purpose of this embodiment.
  • each embodiment can be implemented by means of software plus a general hardware platform, and of course, it can also be implemented by hardware.
  • the computer software products can be stored in computer-readable storage media, such as ROM/RAM, disks. , optical disk, etc., including a number of instructions to cause a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in various embodiments or certain parts of the embodiments.

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Secondary Cells (AREA)
  • Electric Propulsion And Braking For Vehicles (AREA)
  • Charge And Discharge Circuits For Batteries Or The Like (AREA)

Abstract

一种电动车辆的动力电池的健康状态确定方法及服务器。方法包括：获取动力电池的电池数据以及电动车辆的车辆特征（S21），根据实时电池关联参数及与车辆特征对应的正常衰减模型，计算动力电池的正常健康状态衰减值（S22），根据实时故障类型参数及与车辆特征对应的故障衰减模型，计算动力电池的故障健康状态衰减值（S23），根据正常健康状态衰减值及故障健康状态衰减值，确定动力电池的综合健康评价值（S24）。评估动力电池的健康状态时，无需控制动力电池进行充放电，也无需动力电池的首次健康状态值，都可快速地评估动力电池的健康状态，从而改善了现有技术评估动力电池的健康状态时存在效率低下的技术问题。

Description

电动车辆的动力电池的健康状态确定方法及服务器
本申请要求于2022年5月16日提交中国专利局、申请号为2022105291046、申请名称为“电动车辆的动力电池的健康状态确定方法及服务器”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及新能源技术领域，具体涉及一种电动车辆的动力电池的健康状态确定方法及服务器。
背景技术
动力电池作为新能源电动汽车最主要的动力源，动力电池的使用寿命是有限的，动力电池中活性物质随着使用过程而自然损耗，因此，动力电池的满电容量会逐渐降低，业内常采用电池的健康状态(State of Health，SOH)评估动力电池的寿命。
现有评估动力电池的SOH值的方法主要有电化学机理方法和大数据人工智能评估方法。电化学机理方法是通过一定条件的充放电过程，对动力电池的电压、电流等数据进行分析而得到动力电池的SOH值。由于动力电池的充放电过程需要一定充放电时间和充放电条件，因此上述作法评估SOH值的效率比较低。
大数据人工智能评估方法是从大量历史数据中，利用长时间尺度数据分析算法进行得到动力电池的SOH值。由于动力电池的SOH衰减是一个缓慢且不可逆的过程，因此，对于使用一段时间的电动汽车的动力电池进行首次SOH评估时，此方法由于缺少初始值而使得后续利用长时间尺度数据分析算法进行评估时，算法的收敛比较缓慢，无法快速可靠地输出SOH值。
发明内容
本申请实施例的一个目的旨在提供一种电动车辆的动力电池的健康状态确定方法及服务器，用于改善现有技术评估动力电池的健康状态时，存在效率低下的技术问题。
在第一方面，本申请实施例提供一种电动车辆的动力电池的健康状态确定方法，包括：
获取所述动力电池的电池数据以及所述电动车辆的车辆特征，所述电池数据包括实时电池关联参数与实时故障类型参数；
根据所述实时电池关联参数及与所述车辆特征对应的正常衰减模型，计算所述动力电池的正常健康状态衰减值；
根据所述实时故障类型参数及与所述车辆特征对应的故障衰减模型，计算所述动力电池的故障健康状态衰减值；
根据所述正常健康状态衰减值及所述故障健康状态衰减值，确定所述动力电池的综合健康评价值。
可选地，所述实时电池关联参数为行驶里程、动力电池的累计充电容量或累计放电容量中的一个。
可选地，所述根据所述正常衰减值及所述故障衰减值，确定所述动力电池的综合健康评价值包括：
根据以下公式：SOHt＝Std-Δαt-Δεt，确定所述动力电池的综合健康评价值，Std为标准电池健康状态值，SOHt为综合健康评价值，Δαt为正常健康状态衰减值，Δεt为故障健康状态衰减值。
可选地，所述正常衰减模型是根据与所述车辆特征相同的多个第一历史车辆的动力电池的第一训练数据训练得到的；
所述故障衰减模型是根据与所述车辆特征相同的多个第二历史车辆的动力电池的第二训练数据与所述正常衰减模型训练得到的。
可选地，所述第一历史车辆为未发生指定电池故障的车辆；
所述第二历史车辆为已发生指定电池故障的车辆；
所述指定电池故障为过压故障、欠压故障、充电过流故障、放电过流故障、高温故障、低温故障或指定严重故障中的任意一种。
可选地，所述正常衰减模型为：其中，Δαt为正常健康状态衰减值，为正常衰减率，Pt为实时电池关联参数。
可选地，所述第一训练数据包括与所述车辆特征相同的多个第一历史车辆的第一历史电池关联参数与第一电池健康状态值；
所述正常衰减率为：为正常衰减率，ηi为第i个第一历史车辆的待定衰减率，n为参与训练所述正常衰减模型的第一历史车辆的总数；
所述第i个第一历史车辆的待定衰减率是根据第i个第一历史车辆的第一历史电池关联参数和第一电池健康状态值进行计算得到。
可选地，当所述第一历史电池关联参数为行驶里程时，所述第i个第一历史车辆的待定衰减率为：ηi的单位为％/10000km，Mwi为第i个第一历史车辆的行驶里程，SOHwi为第i个第一历史车辆的第一电池健康状态值。
可选地，所述故障衰减模型为：其中，Δεt为故障健康状态衰减值，xj为第j个实时故障类型参数，εj为第j个实时故障类型参数的故障衰减率。
可选地，所述实时故障类型参数为过压故障次数、欠压故障次数、充电过流故障次数、放电过流故障次数、高温故障次数、低温故障次数或指定严重故障次数中的一个。
可选地，所述第二训练数据包括与所述车辆特征相同的多个第二历史车辆的第二电池健康状态值、第二历史电池关联参数及历史故障类型参数；
第j个第二历史车辆的故障衰减率是根据线性回归算法，对多个所述第二历史车辆的健康差值及历史故障类型参数进行行列式计算得到；
第j个第二历史车辆的健康差值为第j个第二历史车辆的期望健康值与第j个第二电池健康值的差值；
第j个第二历史车辆的期望健康值是根据第j个第二历史电池关联参数及所述正常衰减模型进行计算得到的。
可选地，第j个实时故障类型参数的故障衰减率为：


其中，Δεyj为第j个第二历史车辆的健康差值，为第j个第二历史车辆的期望健康值，SOHyj为第j个第二历史车辆的第二电池健康值，Std为标准电池健康状态值，为正常衰减率，Pyj为第j个第二历史车辆的故障历史电池关联参数，xsj为第s个第二历史车辆的第j个历史故障类型参数，εj为第j个历史故障类型参数的故障衰减率。
在第二方面，本申请实施例提供一种服务器，包括：
至少一个处理器；以及，
与所述至少一个处理器通信连接的存储器；其中，
所述存储器存储有可被所述至少一个处理器执行的指令，所述指令被所述至少一个处理器执行，以使所述至少一个处理器能够执行上述的电动车辆的动力电池的健康状态确定方法。
在第三方面，本申请实施例提供一种计算机可读存储介质，所述计算机可读存储介质存储有计算机可执行指令，所述计算机可执行指令用于使计算机执行上述的电动车辆的动力电池的健康状态确定方 法。
在第四方面，本申请实施例提供一种计算机程序产品，所述计算机程序产品包括存储在非易失性计算机可读存储介质上的计算机程序，所述计算机程序包括程序指令，当所述程序指令被电子设备执行时，使电子设备执行上述的电动车辆的动力电池的健康状态确定方法。
本申请与现有技术相比至少具有以下有益效果：在本申请实施例提供的动力电池的健康状态确定方法中，获取动力电池的电池数据以及电动车辆的车辆特征，电池数据包括实时电池关联参数与实时故障类型参数，根据实时电池关联参数及与车辆特征对应的正常衰减模型，计算动力电池的正常健康状态衰减值，根据实时故障类型参数及与车辆特征对应的故障衰减模型，计算动力电池的故障健康状态衰减值，根据正常健康状态衰减值及故障健康状态衰减值，确定动力电池的综合健康评价值，因此，本实施例评估动力电池的健康状态时，无需控制动力电池进行充放电，也无需动力电池的首次健康状态值，都可快速地评估动力电池的健康状态，从而改善了现有技术评估动力电池的健康状态时存在效率低下的技术问题。
附图说明
一个或多个实施例通过与之对应的附图中的图片进行示例性说明，这些示例性说明并不构成对实施例的限定，附图中具有相同参考数字标号的元件表示为类似的元件，除非有特别申明，附图中的图不构成比例限制。
图1为本申请实施例提供的一种动力电池的健康状态确定系统的结构示意图；
图2为本申请实施例提供的一种电动车辆的动力电池的健康状态确定方法的流程示意图；
图3为本申请实施例提供的训练正常衰减模型和故障衰减模型的场景示意图；
图4为本申请实施例提供的一种电动车辆的动力电池的健康状态确定装置的结构示意图；
图5为本申请实施例提供的一种服务器的电路结构示意图。
具体实施方式
为了使本申请的目的、技术方案及优点更加清楚明白，以下结合附图及实施例，对本申请进行进一步详细说明。应当理解，此处所描述的具体实施例仅用以解释本申请，并不用于限定本申请。基于本申请中的实施例，本领域普通技术人员在没有作出创造性劳动前提下所获得的所有其他实施例，都属于本申请保护的范围。
需要说明的是，如果不冲突，本申请实施例中的各个特征可以相互结合，均在本申请的保护范围之内。另外，虽然在装置示意图中进行了功能模块划分，在流程图中示出了逻辑顺序，但是在某些情况下，可以以不同于装置中的模块划分，或流程图中的顺序执行所示出或描述的步骤。再者，本申请所采用的“第一”、“第二”、“第三”等字样并不对数据和执行次序进行限定，仅是对功能和作用基本相同的相同项或相似项进行区分。
本申请实施例提供一种动力电池的健康状态确定系统，请参阅图1，健康状态确定系统100包括车辆通信设备11(Vehicle Communication Interface,VCI)与服务器12，服务器12与车辆通信设备11通信连接，其中，通信连接包括有线通信连接或无线通信连接，有线通信连接包括利用金属导线、光纤等有形媒质传送信息的各类通信连接。无线通信连接包括5G通讯、4G通讯、3G通讯、2G通讯、CDMA、蓝牙、无线宽带、超宽带通信、近场通信、CDMA2000、GSM、ISM、RFID、UMTS/3GPPw/HSDPA、WiMAX、Wi-Fi或ZigBee等。
车辆通信设备11用于插接在电动车辆13的OBD接口(On Board Diagnostics,OBD)，车辆通信设备11基于OBD接口与电动车辆13进行通信，以获取电动车辆13的车辆数据，车辆数据包括故障码、电池数据或车辆特征。故障码用于表示电动车辆的故障类型。电池数据为与电动车辆的动力电池关联的数据。车辆特征为用于表示电动车辆的车型特征和/或属地特征。
如前所述，车辆通信设备11获得电动车辆13的电池数据和车辆特征，于是将电池数据和车辆特征打包发送给服务器12，服务器12根据车辆特征，选择相应的正常衰减模型14与故障衰减模型15，并将电池数据分别输入正常衰减模型14与故障衰减模型15，并根据正常衰减模型14的输出结果与故障衰减模型15的输出结果，确定动力电池的综合健康评价值。
可以理解的是，此处服务器可以是一个物理服务器或者多个物理服务器虚拟而成的一个逻辑服务器。服务器也可以是多个可互联通信的服务器组成的服务器群，且各个功能模块可分别分布在服务器群中的各个服务器上。
作为本申请实施例另一方面，本申请实施例提供一种电动车辆的动力电池的健康状态确定方法，本实施例提供的方法应用在汽车后市场诊断、电池维修、公共充电、保险定损、残值评估等多个应用场景。请参阅图2，电动车辆的动力电池的健康状态确定方法包括：
S21.获取动力电池的电池数据以及电动车辆的车辆特征，电池数据包括实时电池关联参数与实时故障类型参数。
本步骤中，车辆通信设备与电动车辆进行通信，电动车辆将动力电池的电池数据和车辆特征发送至车辆通信设备，车辆通信设备将电池数据和车辆特征进行打包并发送至服务器。
如前所述，电池数据包括实时电池关联参数，实时电池关联参数为用于表示动力电池的健康状态的参数，实时故障类型参数为用于表示动力电池发生相应故障类型的次数。
在一些实施例中，实时电池关联参数为行驶里程、动力电池的累计充电容量或累计放电容量中的一个。
行驶里程为电动车辆驾驶行走的公里数，行驶里程的单位可以为每万公里(10000公里)。行驶里程与动力电池的健康状态存在关联性，行驶里程越大，动力电池的衰减程度较强，行驶里程越小，动力电池的衰减程度较弱，因此，行驶里程可反映出动力电池的健康状态。
累计充电容量为动力电池在多次充电中的充电容量的总和，累计放电容量为动力电池在多次放电中的放电容量的总和，比如动力电池第一次进行充电时，充电容量为C1。动力电池第二次进行充电时，充电容量为C2，累计充电容量＝C1+C2。动力电池第三次进行充电时，充电容量为C3，累计充电容量＝C1+C2+C3。累计充电容量与动力电池的健康状态存在关联性，累计充电容量越大，动力电池的衰减程度较强，累计充电容量越小，动力电池的衰减程度较弱，因此，累计充电容量可反映出动力电池的健康状态。
如前所述，电池数据包括实时故障类型参数，在一些实施例中，实时故障类型参数为过压故障次数、欠压故障次数、充电过流故障次数、放电过流故障次数、高温故障次数、低温故障次数或其他严重故障次数中的一个。
过压故障次数为过压故障对应的总次数，其中，过压故障e1包括每个单体电池过压故障或系统总压过压故障等，比如，在时间t1，动力电池发生过压故障，过压故障次数＝1。在时间t2，动力电池发生过压故障，过压故障次数＝2。在时间t3，动力电池发生过压故障，过压故障次数＝3。
欠压故障次数为欠压故障对应的总次数，其中，欠压故障e2包括每个单体电池欠压故障、单体电池内短路故障或系统总压欠压故障等。
充电过流故障次数为充电过流故障类型对应的总次数，其中，充电过流故障e3包括快慢充充电过流故障或充电电流异常故障等。
放电过流故障次数为放电过流故障对应的总次数，其中，放电过流故障e4包括放电过流故障、放电电流异常故障或短路故障等。
高温故障次数为高温故障对应的总次数，其中，高温故障e5包括电池充电高温故障、电池放电高温故障、环境高温故障、热管理高温故障或控制器高温故障等。
低温故障次数为低温故障对应的总次数，其中，低温故障e6包括电池充电低温故障、电池放电低温故障、环境低温故障或热管理低温故障等。
指定严重故障次数为指定严重故障对应的总次数，其中，指定严重故障e7为除去上述6中故障类型之外的故障。
可以理解的是，在电动汽车的电池管理系统中，与动力电池相关的电池故障有成百上千项，本实施例根据动力电池的特性和电化学机理综合分析，将影响动力电池的健康状态的电池故障划分为上述7类。但是，可以理解的是，本领域技术人员还可根据其它分类目的，对与动力电池相关的电池故障进行分类，上文提供的7类电池故障并不对本申请的保护范围造成任何不当的限定。
在一些实施例中，车辆特征包括电动车辆的车型特征，车型特征用于电动车辆的车型，其中，车型特征可采用MMYB信息进行表示，MMYB信息为新能源车辆的生产商(Make)、车型(Model)、年款(Year)、电池版本(Battery)等信息的统称。
在一些实施例中，车辆特征包括电动车辆的属地特征，属地特征用于表示驾驶电动车辆的地域，其中，属地特征可采用地理位置信息进行表示，地理位置信息可由电动车辆的定位系统进行定位生成，比如地理位置信息为广东深圳南山区。
S22.根据实时电池关联参数及与车辆特征对应的正常衰减模型，计算动力电池的正常健康状态衰减值。
本步骤中，正常衰减模型为用于计算正常健康状态衰减值的模型，正常健康状态衰减值为从动力 电池的正常维度上，评价动力电池的衰减值。本实施例将实时电池关联参数输入正常衰减模型后，正常衰减模型可输出动力电池的正常健康状态衰减值。
可以理解的是，电动车辆的车辆特征不同，电动车辆之间的电池健康衰减曲线的相似度比较小，或者，电动车辆的车辆特征相同，电动车辆之间的电池健康衰减曲线的相似度比较大，因此，本实施例可根据车辆特征，构建正常衰减模型，不同车辆特征对应不同正常衰减模型，有利于提高计算正常健康状态衰减值的准确性。
在一些实施例中，本实施例获取车辆特征，根据车辆特征与正常特征标签，在正常模型库中搜索与车辆特征匹配的正常衰减模型，其中，正常模型库包括多个正常特征标签及与正常特征标签对应的正常衰减模型。
S23.根据实时故障类型参数及与车辆特征对应的故障衰减模型，计算动力电池的故障健康状态衰减值。
本步骤中，故障衰减模型为用于计算故障健康状态衰减值的模型，故障健康状态衰减值为从动力电池的故障维度上，评价动力电池的衰减值。本实施例将实时故障类型参数输入故障衰减模型后，故障衰减模型可输出动力电池的故障健康状态衰减值。
如前所述，电动车辆的车辆特征不同，电动车辆之间的电池健衰减曲线的相似度比较小，因此，本实施例可根据车辆特征，构建故障衰减模型，不同车辆特征对应不同故障衰减模型，有利于提高计算故障健康状态衰减值的准确性。
在一些实施例中，本实施例获取车辆特征，根据车辆特征与故障特征标签，在故障模型库中搜索与车辆特征匹配的故障衰减模型，其中，故障模型库包括多个故障特征标签及与故障特征标签对应的故障衰减模型。
在一些实施例中，车辆特征为车型特征，本实施例分别搜索与车型特征匹配的正常衰减模型与故障衰减模型，亦即，单一的车型特征可分别对应有正常衰减模型与故障衰减模型。
在一些实施例中，车辆特征为属地特征，本实施例分别搜索与属地特征匹配的正常衰减模型与故障衰减模型，亦即，单一的属地特征可分别对应有正常衰减模型与故障衰减模型。
在一些实施例中，车辆特征为车型特征和属地特征，本实施例分别搜索与车型特征匹配且与属地特征匹配的正常衰减模型与故障衰减模型，亦即，正常衰减模型与故障衰减模型都由车型特征与属地特征共同确定。
S24.根据正常健康状态衰减值及故障健康状态衰减值，确定动力电池的综合健康评价值。
本步骤中，综合健康评价值为用于评价动力电池的健康的数值，其中，综合健康评价值包括电池健康状态值，比如，电池健康状态值可采用如前所述的SOH值进行表示，本实施例根据以下公式：SOHt＝Std-Δαt-Δεt，确定动力电池的综合健康评价值，Std为标准电池健康状态值，SOHt为综合健康评价值，Δαt为正常健康状态衰减值，Δεt为故障健康状态衰减值。在一些实施例中，标准电池健康状态值Std可由工程师根据业务需求自定义，比如标准电池健康状态值Std为100％。
在一些实施例中，电动车辆的动力电池的健康状态确定方法还包括：判断综合健康评价值是否小于或等于第一预设阈值，若是，则生成电池警示信息，若否，则根据综合健康评价值，生成电池健康评价信息，其中，电池警示信息用于提示用户更换或维护动力电池，电池警示信息可以为任意形式的警示信息，比如电池警示信息为文字警示信息、语音警示信息或闪光警示信息。电池健康评价信息为评价动力电池的健康状态所处的健康等级信息，比如电池健康评价信息包括“很好”、“较好”或“一般”等。
可以理解的是，第一预设阈值可由工程师根据业务需求自定义，比如第一预设阈值为70％。
在一些实施例中，与上述实施例不同点在于，综合健康评价值包括电池健康衰减值，本实施例可将正常健康状态衰减值与故障健康状态衰减值进行相加，从而得到电池健康衰减值，本实施例根据以下公式：ΔSOHt＝Δαt+Δεt，确定动力电池的电池健康衰减值。
在一些实施例中，电动车辆的动力电池的健康状态确定方法还包括：判断综合健康评价值是否大于或等于第二预设阈值，若是，则生成电池警示信息，若否，则根据综合健康评价值，生成电池健康评价信息。
可以理解的是，第二预设阈值可由工程师根据业务需求自定义，比如第二预设阈值为80％。
如前所述，区别于现有技术，本实施例评估动力电池的健康状态时，无需控制动力电池进行充放电，亦即节省用户需要设置用于控制动力电池进行充放电的装备，在电动车辆处于停止工作状态下都可快速评估动力电池的健康状态，从而改善了现有技术评估动力电池的健康状态时存在效率低下的技术问题，也扩大本方法的应用范围。
如前所述，区别于现有技术，本实施例评估动力电池的健康状态时，无需动力电池的首次健康状态值，亦即，对无历史数据的电动车辆可一次性检测即可输出该电动车辆的综合健康评价值，从而改善了现有技术评估动力电池的健康状态时存在效率低下的技术问题。另外，后续再采用电化学机理方法或大数据人工智能评估方法，进行跟踪评估电动车辆的综合健康评价值而需要初始SOH值时，采用本实施例能够提供出初始SOH值，由此也可看出本实施例提供的方法的应用范围更加灵活和广泛。
另外，采用电化学机理方法或大数据人工智能评估方法往往缺少初始值，导致算法收敛比较缓慢，如前所述，本实施例能够提供出初始SOH值，后续再采用电化学机理方法或大数据人工智能评估方法时，有利于提高上述两种算法的收敛速度和精度。
在一些实施例中，正常衰减模型是根据与车辆特征相同的多个第一历史车辆的动力电池的第一训练数据训练得到的。故障衰减模型是根据与车辆特征相同的多个第二历史车辆的动力电池的第二训练数据与正常衰减模型训练得到的。
第一训练数据为用于训练生成正常衰减模型的数据。第二训练数据为用于配合正常衰减模型以训练生成故障衰减模型的数据。第一历史车辆为：训练生成正常衰减模型之前的电动车辆。第二历史车辆为：训练生成故障衰减模型之前的电动车辆。举例而言，在时间点t11训练生成正常衰减模型，本实施例选择多个第一电动车辆的训练数据作为第一训练数据，此处第一电动车辆为第一历史车辆。在时间点t22训练生成故障衰减模型，本实施例选择多个第二电动车辆的训练数据作为第二训练数据，此处第二电动车辆为第二历史车辆。
本实施例先训练得到正常衰减模型，后续再在正常衰减模型的基础上，结合第二训练数据进行训练得到故障衰减模型，因此，本实施例能够利用正常衰减模型演化生成故障衰减模型，相对于抛开正常衰减模型而另行构建故障衰减模型的作法，本实施例提供的作法能够提高正常衰减模型与故障衰减模型之间的融合度，有利于生成更为可靠准确的故障衰减模型。
在一些实施例中，车辆特征为车型特征，本实施例可选择与车型特征相同的第一历史车辆的动力电池的训练数据作为第一训练数据，可选择与车型特征相同的第二历史车辆的动力电池的训练数据作为第二训练数据。
在一些实施例中，与上述实施例不同点在于，车辆特征为属地特征，本实施例可选择与属地特征相同的第一历史车辆的动力电池的训练数据作为第一训练数据，可选择与属地特征相同的第二历史车辆的动力电池的训练数据作为第二训练数据。
在一些实施例中，与上述实施例不同点在于，车辆特征包括车型特征和属地特征，本实施例可选择与车型特征相同且与属地特征相同的第一历史车辆的动力电池的训练数据作为第一训练数据，可选择与车型特征相同且与属地特征相同的第二历史车辆的动力电池的训练数据作为第二训练数据。
在一些实施例中，第一历史车辆和第二历史车辆都为已发生指定电池故障的车辆，指定电池故障为过压故障、欠压故障、充电过流故障、放电过流故障、高温故障、低温故障或指定严重故障中的任意一种。
本实施例可根据第一历史车辆的第一训练数据训练生成正常衰减模型，以利用正常衰减模型输出正常健康状态衰减值，可根据第二历史车辆的第二训练数据和正常衰减模型训练生成故障衰减模型，以利用故障衰减模型输出故障健康状态衰减值，相对于现有技术，本实施例也可快速评估动力电池的健康状态。
在一些实施例中，与上述实施例不同点在于，第一历史车辆为已发生指定电池故障的车辆，第二历史车辆为未发生指定电池故障的车辆，相对现有技术，本实施例根据此时的第一训练数据与第二训练数据分别得到正常衰减模型和故障衰减模型，可快速评估动力电池的健康状态。
在一些实施例中，与上述实施例不同点在于，第一历史车辆为未发生指定电池故障的车辆，第二历史车辆为已发生指定电池故障的车辆。由于第一历史车辆是未发生指定电池故障的车辆，选择未发生指定电池故障的车辆的训练数据作为第一训练数据，已剔除发生指定电池故障的车辆的训练数据对正常衰减模型的影响，有利于生成更为精确可靠的正常衰减模型。
同理可得，由于第二历史车辆是已发生指定电池故障的车辆，已剔除未发生指定电池故障的车辆的训练数据对故障衰减模型的影响，选择已发生指定电池故障的车辆的训练数据作为第二训练数据，有利于生成更为精确可靠的故障衰减模型。
在一些实施例中，正常衰减模型为：其中，Δαt为正常健康状态衰减值，为正常衰减率，Pt为实时电池关联参数。其中，正常衰减率为从动力电池的正常维度上，评价动力电池的衰减速率。
当服务器获得电动车辆的实时电池关联参数Pt，便可将实时电池关联参数Pt代入正常衰减模型，正常衰减模型可输出正常健康状态衰减值Δαt。
在一些实施例中，第一训练数据包括与车辆特征相同的多个第一历史车辆的第一历史电池关联参数与第一电池健康状态值。如前所述，第一历史电池关联参数可为行驶里程、动力电池的累计充电容量或累计放电容量中的一个，第一电池健康状态值为提前获取第一历史车辆的电池健康状态值。可以理解的是，第一电池健康状态值可以由任意合适评估动力电池SOH值算法计算得到，服务器直接调用评估动力电池SOH值算法提供的结果，或者，第一电池健康状态值也可以由本实施例提供的方法评估动力电池SOH值而得到，后续再利用已有的第一电池健康状态值进行迭代更新。
正常衰减率为：为正常衰减率，ηi为第i个第一历史车辆的待定衰减率，n为参与训练正常衰减模型的第一历史车辆的总数。第i个第一历史车辆的待定衰减率是根据第i个第一历史车辆的第一历史电池关联参数和第一电池健康状态值进行计算得到。
举例而言，参与训练正常衰减模型的第一历史车辆的总数n为100，第1个第一历史车辆的待定衰减率为η1，第2个第一历史车辆的待定衰减率为η2，以此类推，最后正常衰减率本实施例通过求取多个第一历史车辆的待定衰减率的平均值，以得到正常衰减率有利于得到更为准确可靠的正常衰减率。
在一些实施例中，当第一历史电池关联参数为行驶里程时，第i个第一历史车辆的待定衰减率为：ηi的单位为％/10000km，Mwi为第i个第一历史车辆的行驶里程，SOHwi为第i个第一历史车辆的第一电池健康状态值。举例而言，Mwi为100000km，SOHwi为95％，则ηi为0.5,其单位为％/10000km，亦即％/1万公里。
再举例而言，假设经过求取多个待定衰减率的平均值后，得到的正常衰减率为0.5。假设实时电池关联参数Pt为60000km＝60000/10000＝6万公里，将其代入正常衰减模型则有
在一些实施例中，与上述实施例不同点在于，第一历史电池关联参数还可为累计充电容量或累计放电容量时，第i个第一历史车辆的待定衰减率为：ηi的单位为％/Ah，Cwi为第i个第一历史车辆的累计充电容量或累计放电容量。
在一些实施例中，故障衰减模型为：其中，Δεt为故障健康状态衰减值，xj为第j个实时故障类型参数，εj为第j个实时故障类型参数的故障衰减率。其中，第j个实时故障类型参数的故障衰减率εj为从动力电池的故障维度上，评价动力电池在第j个实时故障类型的故障作用下的衰减速率。举例而言，故障衰减模型一共设有7种实时故障类型参数，7个实时故障类型参数的故障衰减率依序分别为0.244、0.112、1.112、0.774、0.365、0.585、0.119。
在一些实施例中，实时故障类型参数为过压故障次数、欠压故障次数、充电过流故障次数、放电过流故障次数、高温故障次数、低温故障次数或指定严重故障次数中的一个。假设电动汽车D1的7个实时故障类型参数依序分别为{f1,f2,f3,f4,f5,f6,f7}＝{4,6,8,3,1,5,4}，将电动汽车D1的7个实时故障类型参数输入故障衰减模型，则故障健康状态衰减值为：
Δεt＝0.244*4+0.112*6+1.112*8+0.774*3+0.365*1+0.585*5+0.119*4＝16.632。
再假设电动汽车D2的7个实时故障类型参数依序分别为 {f1,f2,f3,f4,f5,f6,f7}＝{3,6,5,5,3,4,3}，将电动汽车D2的7个实时故障类型参数输入故障衰减模型，则故障健康状态衰减值为：
Δεt＝0.244*3+0.112*6+1.112*5+0.774*5+0.365*3+0.585*4+0.119*3＝14.626.632。
再假设电动汽车D3的7个实时故障类型参数依序分别为{f1,f2,f3,f4,f5,f6,f7}＝{1,2,1,0,0,1,3}，将电动汽车D3的7个实时故障类型参数输入故障衰减模型，则故障健康状态衰减值为：
Δεt＝0.244*1+0.112*2+1.112*4+0.774*0+0.365*0+0.585*1+0.119*3＝5.858。
在一些实施例中，第二训练数据包括与车辆特征相同的多个第二历史车辆的第二电池健康状态值、第二历史电池关联参数及历史故障类型参数。如前所述，第二历史电池关联参数可为行驶里程、动力电池的累计充电容量或累计放电容量中的一个，第二电池健康状态值为提前获取第二历史车辆的电池健康状态值。可以理解的是，第二电池健康状态值可以由任意合适评估动力电池SOH值算法计算得到，服务器直接调用评估动力电池SOH值算法提供的结果，或者，第二电池健康状态值也可以由本实施例提供的方法评估动力电池SOH值而得到，后续再利用已有的第二电池健康状态值进行迭代更新。历史故障类型参数可为过压故障次数、欠压故障次数、充电过流故障次数、放电过流故障次数、高温故障次数、低温故障次数或指定严重故障次数中的一个。
第j个第二历史车辆的故障衰减率是根据线性回归算法，对多个第二历史车辆的健康差值及历史故障类型参数进行行列式计算得到，其中，第j个第二历史车辆的健康差值为第j个第二历史车辆的期望健康值与第j个第二电池健康值的差值，第j个第二历史车辆的期望健康值是根据第j个第二历史电池关联参数及正常衰减模型进行计算得到的。
可以理解的是，此处线性回归算法可选择任意合适类型的线性回归算法，比如线性回归算法采用最小二乘法。
第j个第二历史车辆的期望健康值为：为第j个第二历史车辆的期望健康值，Std为标准电池健康状态值，为正常衰减率，Pyj为第j个第二历史车辆的故障历史电池关联参数。
如前所述，第j个第二电池健康值可由服务器调用第j个第二历史车辆的已有电池健康值而得到。
第j个第二历史车辆的健康差值为：Δεyj为第j个第二历史车辆的健康差值，SOHyj为第j个第二历史车辆的第二电池健康值。
第j个第二历史车辆的故障衰减率为：
其中，xsj为第s个第二历史车辆的第j个历史故障类型参数，εj为第j个历史故障类型参数的故障衰减率，每个故障衰减率的单位为％。
举例而言，假设故障衰减模型一共设有7种实时故障类型参数，亦即r＝7。再假设取5个第二历史车辆的7个历史故障类型参数。
电动车辆L1的7个历史故障类型参数分别为{4,6,8,3,1,5,4}，为91.6％，SOHy1为75％。电动车辆L2的7个历史故障类型参数分别为{2,4,5,7,2,5,3}，为95.9％，SOHy2为80％。电动车辆L3的7个历史故障类型参数分别为{3,5,6,7,1,6,4}，为89.7％，SOHy3为72％。电动车辆L4的7个历史故障类型参数分别为{3,6,5,5,3,4,3}，为96.6％，SOHy4为82％。电动车辆L5的7个历史故障类型参数分别为{2,6,8,4,1,4,4}，为94.3％，SOHy5为78％。电动车辆 L6的7个历史故障类型参数分别为{0,0,1,0,1,0,2}，为94.7％，SOHy5为93％。电动车辆L7的7个历史故障类型参数分别为{1,2,1,0,0,1,3}，为92.5％，SOHy5为90％。则有：
Δεy1＝16.6％，Δεy2＝15.9％，Δεy3＝17.7％，Δεy4＝14.6％，Δεy5＝16.3％，Δεy6＝1.7％，Δεy7＝2.5％
根据最小二乘法对上式子进行辨识,可得到ε1、ε2、ε3、ε4、ε5、ε6、ε7分别为：0.244、0.112、1.112、0.774、0.365、0.585、0.119，可得到7种电池故障的故障衰减率，可以理解的是，本实施例还可通过增设8种或9种等更多电池故障，并为每个电池故障匹配一个故障衰减率，通过最小二乘法估计出每个电池故障的故障衰减率。
还可以理解的是，本实施例也可增设更少数目的电池故障，并为每个电池故障匹配一个故障衰减率，比如只增设2种电池故障，通过最小二乘法估计出此两个电池故障的故障衰减率。
如前所述，本实施例是将多种电池故障进行关联，并通过线性回归算法，综合计算出每个电池故障的故障衰减率，后续再利用多个电池故障的故障衰减率代入相对于使用单一电池故障进行计算故障健康状态衰减值Δεt，本实施例能够将各种电池故障对故障健康状态衰减值Δεt进行综合和关联，降低单一电池故障的误差对故障健康状态衰减值Δεt的影响，从而能够训练得到更加可靠准确的故障衰减模型。
如前所述，本实施例分为模型训练过程和实际应用过程，其中，图1已展示了快速评估动力电池的综合健康评价值的实际应用过程，下文结合图3对模型训练过程再作出详细阐述，具体如下：
如图3所示，各个第一历史车辆31和各个第二历史车辆32的车辆特征都是相同的，比如各个第一历史车辆31和各个第二历史车辆32都是相同车型特征和相同属地特征。
第一车辆通信设备33插接在第一历史车辆31的OBD接口，第一车辆通信设备33基于OBD接口与第一历史车辆31进行通信，以获取第一历史车辆31的第一训练数据，第一训练数据包括第一电池数据和第一车辆特征，其中，电池数据包括第一行驶里程、第一电池健康状态值及第一故障类型参数。第一车辆通信设备33将第一训练数据发送给服务器34，服务器34根据第一行驶里程与第一电池健康状态值，训练生成正常衰减模型35。
第二车辆通信设备36插接在第二历史车辆32的OBD接口，第二车辆通信设备36基于OBD接口与第二历史车辆32进行通信，以获取第二历史车辆32的第二训练数据，第二训练数据包括第二电池数据和第二车辆特征，其中，电池数据包括第二行驶里程、第二电池健康状态值及第二故障类型参数。第二车辆通信设备36将第二训练数据发送给服务器34。
服务器34将第二行驶里程输入正常衰减模型35，得到期望健康值。接着，服务器34将期望健康值与第二电池健康状态值进行相减，得到健康差值。再接着，服务器根据多个第二历史车辆32的健康差值及第二故障类型参数，计算得到每种电池故障的故障衰减率。最后，服务器根据每种电池故障的故障衰减率，生成故障衰减模型37。
需要说明的是，在上述各个实施方式中，上述各步骤之间并不必然存在一定的先后顺序，本领域普通技术人员，根据本申请实施方式的描述可以理解，不同实施方式中，上述各步骤可以有不同的执行顺序，亦即，可以并行执行，亦可以交换执行等等。
作为本申请实施方式的另一方面，本申请实施方式提供一种电动车辆的动力电池的健康状态确定装置。其中，电动车辆的动力电池的健康状态确定装置可以为软件模块，所述软件模块包括若干指令，其存储在存储器内，处理器可以访问该存储器，调用指令进行执行，以完成上述各个实施方式所阐述的电 动车辆的动力电池的健康状态确定方法。
在一些实施方式中，电动车辆的动力电池的健康状态确定装置亦可以由硬件器件搭建成的，例如，电动车辆的动力电池的健康状态确定装置可以由一个或两个以上的芯片搭建而成，各个芯片可以互相协调工作，以完成上述各个实施方式所阐述的电动车辆的动力电池的健康状态确定方法。再例如，电动车辆的动力电池的健康状态确定装置还可以由各类逻辑器件搭建而成，诸如由通用处理器、数字信号处理器(DSP)、专用集成电路(ASIC)、现场可编程门阵列(FPGA)、单片机、ARM(Acorn RISC Machine)或其它可编程逻辑器件、分立门或晶体管逻辑、分立的硬件组件或者这些部件的任何组合而搭建成。
请参阅图4，电动车辆的动力电池的健康状态确定装置400包括：数据获取模块41、正常衰减计算模块42、故障衰减计算模块43及健康状态确定模块44。数据获取模块41用于获取动力电池的电池数据以及电动车辆的车辆特征，电池数据包括实时电池关联参数与实时故障类型参数。正常衰减计算模块42用于根据实时电池关联参数及与车辆特征对应的正常衰减模型，计算动力电池的正常健康状态衰减值。故障衰减计算模块43用于根据实时故障类型参数及与车辆特征对应的故障衰减模型，计算动力电池的故障健康状态衰减值。健康状态确定模块44用于根据正常健康状态衰减值及故障健康状态衰减值，确定动力电池的综合健康评价值。
本实施例评估动力电池的健康状态时，无需控制动力电池进行充放电，也无需动力电池的首次健康状态值，都可快速地评估动力电池的健康状态，从而改善了现有技术评估动力电池的健康状态时存在效率低下的技术问题。
在一些实施例中，实时电池关联参数为行驶里程、动力电池的累计充电容量或累计放电容量中的一个。
在一些实施例中，健康状态确定模块44具体用于：根据以下公式：SOHt＝Std-Δαt-Δεt，确定动力电池的综合健康评价值，Std为标准电池健康状态值，SOHt为综合健康评价值，Δαt为正常健康状态衰减值，Δεt为故障健康状态衰减值。
在一些实施例中，正常衰减模型是根据与车辆特征相同的多个第一历史车辆的动力电池的第一训练数据训练得到的。故障衰减模型是根据与车辆特征相同的多个第二历史车辆的动力电池的第二训练数据与正常衰减模型训练得到的。
在一些实施例中，第一历史车辆为未发生指定电池故障的车辆。第二历史车辆为已发生指定电池故障的车辆。指定电池故障为过压故障、欠压故障、充电过流故障、放电过流故障、高温故障、低温故障或指定严重故障中的任意一种。
在一些实施例中，正常衰减模型为：其中，Δαt为正常健康状态衰减值，为正常衰减率，Pt为实时电池关联参数。
在一些实施例中，第一训练数据包括与车辆特征相同的多个第一历史车辆的第一历史电池关联参数与第一电池健康状态值。正常衰减率为：为正常衰减率，ηi为第i个第一历史车辆的待定衰减率，n为参与训练所正常衰减模型的第一历史车辆的总数。第i个第一历史车辆的待定衰减率是根据第i个第一历史车辆的第一历史电池关联参数和第一电池健康状态值进行计算得到。
在一些实施例中，当第一历史电池关联参数为行驶里程时，第i个第一历史车辆的待定衰减率为：ηi的单位为％/10000km，Mwi为第i个第一历史车辆的行驶里程，SOHwi为第i个第一历史车辆的第一电池健康状态值。
在一些实施例中，故障衰减模型为：其中，Δεt为故障健康状态衰减值，xj为第j个实时故障类型参数，εj为第j个实时故障类型参数的故障衰减率。
在一些实施例中，实时故障类型参数为过压故障次数、欠压故障次数、充电过流故障次数、放电过流故障次数、高温故障次数、低温故障次数或指定严重故障次数中的一个。
在一些实施例中，第二训练数据包括与所述车辆特征相同的多个第二历史车辆的第二电池健康状态值、第二历史电池关联参数及历史故障类型参数。
第j个第二历史车辆的故障衰减率是根据线性回归算法，对多个第二历史车辆的健康差值及历史故障类型参数进行行列式计算得到。
第j个第二历史车辆的健康差值为第j个第二历史车辆的期望健康值与第j个第二电池健康值的差值；
第j个第二历史车辆的期望健康值是根据第j个第二历史电池关联参数及正常衰减模型进行计算得到的。
在一些实施例中，第j个实时故障类型参数的故障衰减率为：


其中，Δεyj为第j个第二历史车辆的健康差值，为第j个第二历史车辆的期望健康值，SOHyj为第j个第二历史车辆的第二电池健康值，Std为标准电池健康状态值，为正常衰减率，Pyj为第j个第二历史车辆的故障历史电池关联参数，xsj为第s个第二历史车辆的第j个历史故障类型参数，εj为第j个历史故障类型参数的故障衰减率。
需要说明的是，上述电动车辆的动力电池的健康状态确定装置可执行本申请实施方式所提供的电动车辆的动力电池的健康状态确定方法，具备执行方法相应的功能模块和有益效果。未在电动车辆的动力电池的健康状态确定装置实施方式中详尽描述的技术细节，可参见本申请实施方式所提供的电动车辆的动力电池的健康状态确定方法。
请参阅图5，图5为本申请实施例提供的一种服务器的电路结构示意图。如图5所示，服务器500包括一个或多个处理器51以及存储器52。其中，图5中以一个处理器51为例。
处理器51和存储器52可以通过总线或者其他方式连接，图5中以通过总线连接为例。
存储器52作为一种存储介质，可用于存储非易失性软件程序、非易失性计算机可执行程序以及模块，如本申请实施例中的电动车辆的动力电池的健康状态确定方法对应的程序指令/模块。处理器51通过运行存储在存储器52中的非易失性软件程序、指令以及模块，从而执行电动车辆的动力电池的健康状态确定装置的各种功能应用以及数据处理，即实现上述方法实施例提供的电动车辆的动力电池的健康状态确定方法以及上述装置实施例的各个模块或单元的功能。
存储器52可以包括高速随机存取存储器，还可以包括非易失性存储器，例如至少一个磁盘存储器件、闪存器件、或其他非易失性固态存储器件。在一些实施例中，存储器52可选包括相对于处理器51远程设置的存储器，这些远程存储器可以通过网络连接至处理器51。上述网络的实例包括但不限于互联网、企业内部网、局域网、移动通信网及其组合。
所述程序指令/模块存储在所述存储器52中，当被所述一个或者多个处理器51执行时，执行上述任意方法实施例中的电动车辆的动力电池的健康状态确定方法。
本申请实施例还提供了一种计算机可读存储介质，所述计算机可读存储介质存储有计算机可执行指令，该计算机可执行指令被一个或多个处理器执行，例如图5中的一个处理器51，可使得上述一个或多个处理器可执行上述任意方法实施例中的电动车辆的动力电池的健康状态确定方法。
本申请实施例还提供了一种计算机程序产品，所述计算机程序产品包括存储在非易失性计算机可读存储介质上的计算机程序，所述计算机程序包括程序指令，当所述程序指令被服务器执行时，使所述服务器执行任一项所述的电动车辆的动力电池的健康状态确定方法。
以上所描述的装置或设备实施例仅仅是示意性的，其中所述作为分离部件说明的单元模块可以是或者也可以不是物理上分开的，作为模块单元显示的部件可以是或者也可以不是物理单元，即可以位于一个地方，或者也可以分布到多个网络模块单元上。可以根据实际的需要选择其中的部分或者全部模块来 实现本实施例方案的目的。
通过以上的实施方式的描述，本领域的技术人员可以清楚地了解到各实施方式可借助软件加通用硬件平台的方式来实现，当然也可以通过硬件。基于这样的理解，上述技术方案本质上或者说对相关技术做出贡献的部分可以以软件产品的形式体现出来，该计算机软件产品可以存储在计算机可读存储介质中，如ROM/RAM、磁碟、光盘等，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行各个实施例或者实施例的某些部分所述的方法。
最后应说明的是：以上实施例仅用以说明本申请的技术方案，而非对其限制；在本申请的思路下，以上实施例或者不同实施例中的技术特征之间也可以进行组合，步骤可以以任意顺序实现，并存在如上所述的本申请的不同方面的许多其它变化，为了简明，它们没有在细节中提供；尽管参照前述实施例对本申请进行了详细的说明，本领域的普通技术人员应当理解：其依然可以对前述各实施例所记载的技术方案进行修改，或者对其中部分技术特征进行等同替换；而这些修改或者替换，并不使相应技术方案的本质脱离本申请各实施例技术方案的范围。
以上所述，仅为本申请的具体实施方式，但本申请的保护范围并不局限于此，任何熟悉本技术领域的技术人员在本申请揭露的技术范围内，可轻易想到变化或替换，都应涵盖在本申请的保护范围之内。因此，本申请的保护范围应以所述权利要求的保护范围为准。

Claims (14)

  1. 一种电动车辆的动力电池的健康状态确定方法，包括：
    获取所述动力电池的电池数据以及所述电动车辆的车辆特征，所述电池数据包括实时电池关联参数与实时故障类型参数；
    根据所述实时电池关联参数及与所述车辆特征对应的正常衰减模型，计算所述动力电池的正常健康状态衰减值；
    根据所述实时故障类型参数及与所述车辆特征对应的故障衰减模型，计算所述动力电池的故障健康状态衰减值；
    根据所述正常健康状态衰减值及所述故障健康状态衰减值，确定所述动力电池的综合健康评价值。
  2. 根据权利要求1所述的方法，所述实时电池关联参数为行驶里程、动力电池的累计充电容量或累计放电容量中的一个。
  3. 根据权利要求1所述的方法，所述根据所述正常衰减值及所述故障衰减值，确定所述动力电池的综合健康评价值包括：
    根据以下公式：SOHt＝Std-Δαt-Δεt，确定所述动力电池的综合健康评价值，Std为标准电池健康状态值，SOHt为综合健康评价值，Δαt为正常健康状态衰减值，Δεt为故障健康状态衰减值。
  4. 根据权利要求1所述的方法，
    所述正常衰减模型是根据与所述车辆特征相同的多个第一历史车辆的动力电池的第一训练数据训练得到的；
    所述故障衰减模型是根据与所述车辆特征相同的多个第二历史车辆的动力电池的第二训练数据与所述正常衰减模型训练得到的。
  5. 根据权利要求4所述的方法，
    所述第一历史车辆为未发生指定电池故障的车辆；
    所述第二历史车辆为已发生指定电池故障的车辆；
    所述指定电池故障为过压故障、欠压故障、充电过流故障、放电过流故障、高温故障、低温故障或指定严重故障中的任意一种。
  6. 根据权利要求4所述的方法，所述正常衰减模型为：其中，Δαt为正常健康状态衰减值，为正常衰减率，Pt为实时电池关联参数。
  7. 根据权利要求4所述的方法，
    所述第一训练数据包括与所述车辆特征相同的多个第一历史车辆的第一历史电池关联参数与第一电池健康状态值；
    所述正常衰减率为：为正常衰减率，ηi为第i个第一历史车辆的待定衰减率，n为参与训练所述正常衰减模型的第一历史车辆的总数；
    所述第i个第一历史车辆的待定衰减率是根据第i个第一历史车辆的第一历史电池关联参数和第一电池健康状态值进行计算得到。
  8. 根据权利要求7所述的方法，当所述第一历史电池关联参数为行驶里程时，所述第i个第一历史车辆的待定衰减率为：ηi的单位为％/10000km，Mwi为第i个第一历史车辆的行驶里程，SOHwi为第i个第一历史车辆的第一电池健康状态值。
  9. 根据权利要求4所述的方法，所述故障衰减模型为：其中，Δεt为故障健康状态衰减值，xj为第j个实时故障类型参数，εj为第j个实时故障类型参数的故障衰减率。
  10. 根据权利要求9所述的方法，所述实时故障类型参数为过压故障次数、欠压故障次数、充电过流故障次数、放电过流故障次数、高温故障次数、低温故障次数或指定严重故障次数中的一个。
  11. 根据权利要求9所述的方法，
    所述第二训练数据包括与所述车辆特征相同的多个第二历史车辆的第二电池健康状态值、第二历史电池关联参数及历史故障类型参数；
    第j个第二历史车辆的故障衰减率是根据线性回归算法，对多个所述第二历史车辆的健康差值及历史故障类型参数进行行列式计算得到；
    第j个第二历史车辆的健康差值为第j个第二历史车辆的期望健康值与第j个第二电池健康值的差值；
    第j个第二历史车辆的期望健康值是根据第j个第二历史电池关联参数及所述正常衰减模型进行计算得到的。
  12. 根据权利要求11所述的方法，第j个实时故障类型参数的故障衰减率为：


    其中，Δεyj为第j个第二历史车辆的健康差值，为第j个第二历史车辆的期望健康值，SOHyj为第j个第二历史车辆的第二电池健康值，Std为标准电池健康状态值，为正常衰减率，Pyj为第j个第二历史车辆的故障历史电池关联参数，xsj为第s个第二历史车辆的第j个历史故障类型参数，εj为第j个历史故障类型参数的故障衰减率。
  13. 一种服务器，包括：
    至少一个处理器；以及，
    与所述至少一个处理器通信连接的存储器；其中，
    所述存储器存储有可被所述至少一个处理器执行的指令，所述指令被所述至少一个处理器执行，以使所述至少一个处理器能够执行如权利要求1至12任一项所述的电动车辆的动力电池的健康状态确定方法。
  14. 一种计算机可读存储介质，所述计算机可读存储介质存储有计算机可执行指令，所述计算机可执行指令用于使计算机执行如权利要求1-12任意一项所述的电动车辆的动力电池的健康状态确定方法。
PCT/CN2023/077917 2022-05-16 2023-02-23 电动车辆的动力电池的健康状态确定方法及服务器 WO2023221587A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202210529104.6 2022-05-16
CN202210529104.6A CN115598556A (zh) 2022-05-16 2022-05-16 电动车辆的动力电池的健康状态确定方法及服务器

Publications (1)

Publication Number Publication Date
WO2023221587A1 true WO2023221587A1 (zh) 2023-11-23

Family

ID=84841848

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/077917 WO2023221587A1 (zh) 2022-05-16 2023-02-23 电动车辆的动力电池的健康状态确定方法及服务器

Country Status (2)

Country Link
CN (1) CN115598556A (zh)
WO (1) WO2023221587A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117805663A (zh) * 2024-02-28 2024-04-02 常州拜特测控技术有限公司 基于运行状态的电池测试方法、装置、设备和介质

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115598556A (zh) * 2022-05-16 2023-01-13 深圳市道通合创新能源有限公司（Cn） 电动车辆的动力电池的健康状态确定方法及服务器
CN116552321B (zh) * 2023-07-07 2023-10-13 徐州徐工汽车制造有限公司 用于车辆的动力电池的管理方法、装置和车辆
CN117368742A (zh) * 2023-11-23 2024-01-09 安徽来福电子科技有限公司 新能源汽车的电池状态评估系统

Citations (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
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、装置及系统
CN111968268A (zh) * 2020-06-29 2020-11-20 南斗六星系统集成有限公司 一种新能源车辆健康状况远程评估方法和系统
US10942223B1 (en) * 2019-07-31 2021-03-09 Cox Automotive, Inc. Systems and methods for determining vehicle battery health
CN112505570A (zh) * 2020-12-03 2021-03-16 摩登汽车(盐城)有限公司 电动汽车的电池健康状态的估计方法
US20210132151A1 (en) * 2019-10-07 2021-05-06 Samsung Sdi Co., Ltd. Method and apparatus for estimating state of health of battery
CN112946483A (zh) * 2021-02-05 2021-06-11 重庆长安新能源汽车科技有限公司 电动汽车电池健康的综合评估方法及存储介质
CN113219357A (zh) * 2021-04-28 2021-08-06 东软睿驰汽车技术（沈阳）有限公司 电池包健康状态计算方法、系统及电子设备
CN113306449A (zh) * 2021-06-15 2021-08-27 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系统
CN113806171A (zh) * 2021-09-12 2021-12-17 济南浪潮数据技术有限公司 一种服务器健康评估方法、系统、设备以及介质
CN113933732A (zh) * 2021-10-13 2022-01-14 中国汽车工程研究院股份有限公司 新能源汽车动力电池健康状态分析方法、系统及存储介质
CN114290960A (zh) * 2021-12-30 2022-04-08 中国第一汽车股份有限公司 获取动力电池的电池健康度的方法、装置和车辆
CN115598556A (zh) * 2022-05-16 2023-01-13 深圳市道通合创新能源有限公司（Cn） 电动车辆的动力电池的健康状态确定方法及服务器

Patent Citations (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
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、装置及系统
US10942223B1 (en) * 2019-07-31 2021-03-09 Cox Automotive, Inc. Systems and methods for determining vehicle battery health
US20210132151A1 (en) * 2019-10-07 2021-05-06 Samsung Sdi Co., Ltd. Method and apparatus for estimating state of health of battery
CN111968268A (zh) * 2020-06-29 2020-11-20 南斗六星系统集成有限公司 一种新能源车辆健康状况远程评估方法和系统
CN112505570A (zh) * 2020-12-03 2021-03-16 摩登汽车(盐城)有限公司 电动汽车的电池健康状态的估计方法
CN112946483A (zh) * 2021-02-05 2021-06-11 重庆长安新能源汽车科技有限公司 电动汽车电池健康的综合评估方法及存储介质
CN113219357A (zh) * 2021-04-28 2021-08-06 东软睿驰汽车技术（沈阳）有限公司 电池包健康状态计算方法、系统及电子设备
CN113306449A (zh) * 2021-06-15 2021-08-27 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系统
CN113806171A (zh) * 2021-09-12 2021-12-17 济南浪潮数据技术有限公司 一种服务器健康评估方法、系统、设备以及介质
CN113933732A (zh) * 2021-10-13 2022-01-14 中国汽车工程研究院股份有限公司 新能源汽车动力电池健康状态分析方法、系统及存储介质
CN114290960A (zh) * 2021-12-30 2022-04-08 中国第一汽车股份有限公司 获取动力电池的电池健康度的方法、装置和车辆
CN115598556A (zh) * 2022-05-16 2023-01-13 深圳市道通合创新能源有限公司（Cn） 电动车辆的动力电池的健康状态确定方法及服务器

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
王安晨 (WANG, ANCHEN): "基于多源信息融合的车载电池健康状态评估方法研究 (Non-official translation: Vehicle-Mounted Battery State-of-Health Evaluation Method based on Multi-Source Information Fusion)", 中国博士学位论文全文数据库 (CHINA DOCTORAL DISSERTATIONS FULL-TEXT DATABASE), no. 2, 15 February 2022 (2022-02-15) *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117805663A (zh) * 2024-02-28 2024-04-02 常州拜特测控技术有限公司 基于运行状态的电池测试方法、装置、设备和介质
CN117805663B (zh) * 2024-02-28 2024-05-10 常州拜特测控技术有限公司 基于运行状态的电池测试方法、装置、设备和介质

Also Published As

Publication number Publication date
CN115598556A (zh) 2023-01-13

Similar Documents

Publication Publication Date Title
WO2023221587A1 (zh) 电动车辆的动力电池的健康状态确定方法及服务器
Park et al. LSTM-based battery remaining useful life prediction with multi-channel charging profiles
Wenzl et al. Life prediction of batteries for selecting the technically most suitable and cost effective battery
CN112068004A (zh) 一种电池异常、电池充电剩余时间的确定方法及装置
WO2022213597A1 (zh) 一种新能源车辆三电系统安全特征数据库构建方法
CN114372417A (zh) 基于充电网的电动汽车电池健康状态和剩余寿命评估方法
KR102239040B1 (ko) 운송 수단의 부품에 대한 고장예지 및 건전성 관리 시스템 및 방법
CN111680848A (zh) 基于预测模型融合的电池寿命预测方法及存储介质
CN110412470B (zh) 电动汽车动力电池soc估计方法
US11201486B2 (en) Battery control device, charge and discharge system, parking lot system, secondary battery reuse system, battery control method, and non-transitory storage medium
CN115330275B (zh) 一种退役电池的梯次利用方法及装置
CN116154900B (zh) 一种电池储能电站的主动安全三级防控系统及方法
CN108876052A (zh) 电动汽车充电负荷预测方法、装置和计算机设备
CN107340476A (zh) 电池的电气状态监测系统和电气状态监测方法
CN107634277A (zh) 一种基于无线通讯电芯的汽车云端电池管理系统
CN108009585A (zh) 基于局部信息融合的铅酸电池健康状态预测方法
Zheng et al. A novel method for lithium-ion battery remaining useful life prediction using time window and gradient boosting decision trees
De Sousa et al. Comparison of different approaches for lead acid battery state of health estimation based on artificial neural networks algorithms
CN113791351A (zh) 基于迁移学习和差值概率分布的锂电池寿命预测方法
CN117665620A (zh) 基于新能源汽车数据的电池健康度评估方法
CN112765726A (zh) 一种寿命预测方法及装置
CN116008834A (zh) 移动式储能系统的电池监测运维方法、系统、设备及介质
El Shahat Neural network storage unit parameters modelling
Hemdani et al. Prediction of aging of battery for electric vehicles based on a modified version of neural networks
CN115825790B (zh) 电池绝缘故障的预警方法、装置、系统和计算机设备

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: 23806548

Country of ref document: EP

Kind code of ref document: A1