Power battery charge state determination method, device and electronic equipment

Abstract

The present invention provides a method, device and electronic device for determining the state of charge of a power battery, which relates to the technical field of data processing, including: obtaining the state data of the battery within a target time period, processing the state data using a local battery model, and obtaining the preset state of charge data of the battery. The model parameters of the local battery model are obtained by directly applying the model parameters of the corresponding battery model to the local battery model after training a variety of cloud battery models, so that the computing power requirements of the local BMS system can be effectively reduced; the state data of the battery characterizes the working condition of the battery, and the local battery model can determine a solution algorithm that matches the working condition of the current battery according to the preset correspondence between the working condition of the battery and the solution algorithm of the battery parameter identification equation, thereby ensuring the identification accuracy of the battery parameters, improving the estimation accuracy of the preset state of charge data, and alleviating the technical problem of limited application scenarios.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a method, device and electronic equipment for determining the state of charge of a power battery. Background Art

[0002] The power battery state of charge (SOX) is an important basis for the battery management system (BMS) to monitor extended-range electric vehicles. The power battery state of charge (SOX) includes: battery remaining charge estimation (SOC), battery health state estimation (SOH), battery power state estimation (SOP), and battery energy state estimation (SOE). The accurate estimation of SOX has important theoretical significance and practical value for the safety and rational use of batteries and the energy optimization matching of the whole vehicle. The estimation of SOX requires the acquisition of battery parameters, and online identification of battery parameters is an important means to improve the accuracy of SOX estimation.

[0003] However, the online identification of battery parameters in the prior art has high requirements for chip computing power, and ordinary BMS systems cannot meet the computing power requirements. In addition, battery parameter identification can only be carried out under some steady-state conditions, and it is difficult to carry out in all complex vehicle conditions. Therefore, the prior art method for determining the state of charge of power batteries has technical problems such as high requirements for chip computing power and limited application scenarios. Summary of the invention

[0004] The purpose of the present invention is to provide a method, device and electronic device for determining the state of charge of a power battery, so as to alleviate the technical problems existing in the prior art of the method for determining the state of charge of a power battery, namely, high requirements on chip computing power and limited application scenarios.

[0005] In a first aspect, the present invention provides a method for determining the state of charge of a power battery, comprising: obtaining state data of a battery within a target time period; wherein the state data comprises: voltage, current and temperature; the state data is used to characterize the operating condition of the battery; processing the state data using a local battery model to obtain preset state of charge data of the battery; wherein the local battery model is a battery parameter determined based on a battery parameter identification equation, and the preset state of charge data is determined based on the battery parameters; the battery parameter identification equation corresponds one-to-one to the battery parameters; the model parameters of the local battery model are determined after a designated cloud platform trains multiple cloud battery models, and the cloud battery model includes a battery model identical to the local battery model; there is a preset correspondence between the solution algorithm of the battery parameter identification equation and the operating condition of the battery.

[0006] In an optional embodiment, it also includes: obtaining an offline training data set; wherein the offline training data set includes: multiple groups of offline training data; each group of offline training data includes: battery voltage within a first preset time period, battery load current within the first preset time period, battery temperature within the first preset time period, battery cell performance parameters within the first preset time period, and preset state of charge data of the battery within the first preset time period; using a Monte Carlo random algorithm and the offline training data set to perform reinforcement learning training on multiple cloud battery models to obtain multiple cloud battery models to be optimized; obtaining an online training data set; wherein the online training data set includes: multiple groups of online training data; each group of online training data includes: battery voltage within a second preset time period, battery load current within the second preset time period, battery temperature within the second preset time period, battery cell performance parameters within the second preset time period, and preset state of charge data of the battery within the second preset time period; using a Monte Carlo random algorithm and the online training data set to perform reinforcement learning training on multiple cloud battery models to be optimized to obtain multiple trained cloud battery models.

[0007] In an optional embodiment, a Monte Carlo random algorithm and the offline training data set are used to perform reinforcement learning training on multiple cloud battery models, including: for target offline training data, determining a target cloud battery model based on the target offline training data; wherein the target offline training data represents any offline training data in the offline training data set; the target cloud battery model represents a cloud battery model that matches the battery cell performance parameters in the target offline training data; determining the solution algorithm for each battery parameter identification equation and the training battery state of charge data output by the target cloud battery model when the target cloud battery model identifies the battery parameters corresponding to the target offline training data; processing the target offline training data using a Monte Carlo random algorithm to obtain target random parameters; wherein the target random parameters are used to characterize the battery operating conditions corresponding to the target offline training data; reverse checking the battery parameter identification equation based on the training battery state of charge data; and performing reinforcement training on the target cloud battery model based on the target random parameters, the training battery state of charge data, and the preset state of charge data in the target offline training data to obtain the target cloud battery model to be optimized.

[0008] In an optional embodiment, when determining that the target cloud-based battery model identifies the battery parameters corresponding to the target offline training data, the solution algorithm for each battery parameter identification equation includes: determining a voltage change curve fitted by a target solution algorithm group based on the target offline training data and the target cloud-based battery model; wherein the target solution algorithm group represents any set of preset solution algorithms for several battery parameter identification equations in the target cloud-based battery model; using the solution algorithm group corresponding to the target voltage change curve as a method for the target cloud-based battery model to identify the battery parameters corresponding to the target offline training data; wherein the target voltage change curve represents a curve with the highest similarity to the voltage change curve reflected by the target offline training data; determining a solution algorithm for each battery parameter identification equation based on the solution algorithm group corresponding to the target voltage change curve.

[0009] In an optional embodiment, obtaining an online training data set includes: repeatedly performing the following steps until a specified amount of online training data is obtained: utilizing the specified cloud platform to simulate a battery usage environment; determining the battery status data within a second preset time period under the battery usage environment, the battery cell performance parameters, and the preset charge status data of the battery within the second preset time period, and using them as a set of online training data.

[0010] In an optional embodiment, the cloud-based battery model includes: a multi-order Thevenin circuit model, a fractional-order electrochemical model, an electrolyte single particle model, and a reduced-order electrochemical model.

[0011] In a second aspect, the present invention provides a device for determining the state of charge of a power battery, comprising: a first acquisition module, used to obtain state data of a battery within a target time period; wherein the state data comprises: voltage, current and temperature; the state data is used to characterize the operating condition of the battery; a processing module, used to process the state data using a local battery model to obtain preset state of charge data of the battery; wherein the local battery model is a battery parameter determined based on a battery parameter identification equation, and the battery parameter identification equation has a one-to-one correspondence with the battery parameters; the model parameters of the local battery model are determined after a designated cloud platform trains multiple cloud battery models, and the cloud battery model includes a battery model identical to the local battery model; there is a preset correspondence between the solution algorithm of the battery parameter identification equation and the operating condition of the battery.

[0012] In an optional embodiment, it also includes: a second acquisition module, used to acquire an offline training data set; wherein the offline training data set includes: multiple groups of offline training data; each group of offline training data includes: a battery voltage within a first preset time period, a battery load current within the first preset time period, a battery temperature within the first preset time period, a battery cell performance parameter within the first preset time period, and a preset state of charge data of the battery within the first preset time period; a first training module, used to perform reinforcement learning training on multiple cloud battery models using a Monte Carlo random algorithm and the offline training data set to obtain multiple cloud battery models to be optimized; a third acquisition module, used to acquire an online training data set; wherein the online training data set includes: multiple groups of online training data; each group of online training data includes: a battery voltage within a second preset time period, a battery load current within the second preset time period, a battery temperature within the second preset time period, a battery cell performance parameter within the second preset time period, and a preset state of charge data of the battery within the second preset time period; a second training module, used to perform reinforcement learning training on multiple cloud battery models to be optimized using a Monte Carlo random algorithm and the online training data set to obtain multiple trained cloud battery models.

[0013] In a third aspect, the present invention provides an electronic device, comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and when the processor executes the computer program, the steps of the method for determining the state of charge of a power battery described in any one of the aforementioned embodiments are implemented.

[0014] In a fourth aspect, the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, and when the computer instructions are executed by a processor, the method for determining the state of charge of a power battery described in any one of the aforementioned embodiments is implemented.

[0015] The method for determining the state of charge of a power battery provided by the present invention comprises: obtaining state data of a battery within a target time period; wherein the state data comprises: voltage, current and temperature; the state data is used to characterize the working condition of the battery; processing the state data using a local battery model to obtain preset state of charge data of the battery; wherein the local battery model is a battery parameter determined based on a battery parameter identification equation, and preset state of charge data determined based on the battery parameters; the battery parameter identification equation corresponds one-to-one to the battery parameters; the model parameters of the local battery model are determined after a designated cloud platform trains a plurality of cloud battery models, and the cloud battery model includes a battery model identical to the local battery model; there is a preset corresponding relationship between the solution algorithm of the battery parameter identification equation and the working condition of the battery.

[0016] The local battery model used in the method of the present invention has model parameters that are obtained by directly applying the model parameters of the corresponding battery models to the local battery model after training a variety of cloud battery models, so that the computing power requirements of the local BMS system can be effectively reduced; and the battery status data is used to characterize the operating condition of the battery. When the local battery model determines the battery parameters based on the battery parameter identification equation, it can determine a solution algorithm that matches the current battery operating condition based on the preset correspondence between the battery operating condition and the solution algorithm of the battery parameter identification equation, thereby ensuring the identification accuracy of the battery parameters, which not only improves the estimation accuracy of the preset charge state data, but also alleviates the technical problem of limited application scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 A flow chart of a method for determining the state of charge of a power battery provided by an embodiment of the present invention;

[0019] Figure 2 A flow chart of a method for performing reinforcement learning training on multiple cloud battery models using a Monte Carlo random algorithm and an offline training data set provided in an embodiment of the present invention;

[0020] Figure 3 A functional module diagram of a device for determining the state of charge of a power battery provided by an embodiment of the present invention;

[0021] Figure 4 A schematic diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0023] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0024] Some embodiments of the present invention are described in detail below in conjunction with the accompanying drawings. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0025] The power battery state of charge (SOX) is an important basis for the battery management system (BMS) to monitor extended-range electric vehicles. The accurate estimation of SOX has important theoretical significance and practical value for the safe and rational use of the battery and the energy optimization matching of the whole vehicle. The estimation of SOX requires the acquisition of battery parameters, and online identification of battery parameters is an important means to improve the accuracy of SOX estimation.

[0026] In the prior art, online identification of battery parameters has the following problems: 1. The chip computing power requirement is high, and ordinary BMS systems cannot meet the computing power requirement; 2. Battery parameter identification can only open some steady-state conditions, and it is difficult to carry out in the vehicle's full complex working conditions; 3. Parameter identification itself is limited by the identification formula, and the identification accuracy is subject to certain restrictions; 4. If you want to combine cloud platform big data for online identification, the real-time nature of the identification results cannot meet actual needs. In view of this, an embodiment of the present invention provides a method for determining the state of charge of a power battery to reduce the computing power requirements of the local BMS system, ensure the accuracy of battery parameter identification, improve the estimation accuracy of preset state of charge data, and alleviate the technical problem of limited application scenarios.

[0027] Embodiment 1

[0028] Figure 1 A flow chart of a method for determining the state of charge of a power battery provided by an embodiment of the present invention, such as Figure 1 As shown, the method specifically comprises the following steps:

[0029] Step S102, obtaining battery status data within a target time period.

[0030] Among them, the status data includes: voltage, current and temperature; the status data is used to characterize the working condition of the battery.

[0031] Step S104: Process the state data using the local battery model to obtain preset state of charge data of the battery.

[0032] The local battery model is a battery parameter determined based on a battery parameter identification equation and a preset state of charge data determined based on the battery parameter; the battery parameter identification equation corresponds one-to-one to the battery parameter.

[0033] The model parameters of the local battery model are determined after a designated cloud platform trains multiple cloud battery models. The cloud battery model includes a battery model that is identical to the local battery model. There is a preset correspondence between the solution algorithm of the battery parameter identification equation and the operating condition of the battery.

[0034] Specifically, it is known that the state of charge SOX of a power battery includes parameters such as SOC, SOH, SOP and SOE. The calculation of the state of charge SOX of a power battery requires the use of battery parameters, which include but are not limited to internal resistance, polarization capacitance, impedance and OCV (Open Circuit Voltage) parameters, etc. The prior art has clearly defined the calculation method for each state of charge data. If the user determines the specific type of state of charge data to be identified (that is, preset state of charge data), it is only necessary to configure the battery parameter identification equation corresponding to the battery parameters required to calculate the preset state of charge data in the battery model during model training.

[0035] In the embodiment of the present invention, to determine the preset state of charge data, the required source data is the state data of the battery to be tested within the target time period, wherein the state data includes: voltage, current and temperature. The embodiment of the present invention does not specifically limit the duration of the target time period. According to domain knowledge, the longer the time represented by the target time period, the larger the amount of battery state data obtained, and the more accurate the final output of the model will be, but the corresponding data processing time will also be longer.

[0036] Further, the battery status data can characterize the battery's operating conditions, that is, the battery voltage, load current and temperature within the target time period can reflect the battery's operating conditions, for example, whether the battery's operating conditions are stable / unstable, and whether the operating conditions are complex or not. The battery parameter identification method in the prior art is only applicable to the identification of battery parameters under stable conditions, and cannot accurately identify complex conditions. In view of this, in an embodiment of the present invention, the relationship between the battery's operating conditions and the solution algorithm of the battery parameter identification equation is studied, wherein the solution algorithm of the battery parameter identification equation includes but is not limited to: FFRLS, Kalman, Polynomial, Exponential, Fourier, Gaussian. During model training, each solution algorithm can obtain a battery parameter value, but by reversely bringing the solved battery parameter value in, the similarity between the battery performance inferred by different solution algorithms and the actual battery performance can be compared, and then the optimal solution algorithm under the battery condition can be determined. By learning the optimal solution algorithm of the battery parameter identification equation under different battery conditions, the battery model can be applied to the identification of battery parameters under complex conditions.

[0037] In an embodiment of the present invention, if the calculation of the preset state of charge data requires the use of 5 battery parameters, then the local battery model should include 5 battery parameter identification equations, each battery parameter identification equation is used to identify a battery parameter, that is, the battery parameter identification equation corresponds one-to-one to the battery parameter.

[0038] In order to reduce the computing power requirements for the local BMS system, the embodiment of the present invention uses a designated cloud platform to train multiple cloud battery models in advance. Each cloud battery model corresponds to the recognition of preset charge data of a battery model / type. When performing local applications, it is only necessary to apply the model parameters of the cloud battery model corresponding to the model / type of the local battery after training to the local battery model. The direct application of the model parameters after cloud training avoids the local training process, thereby effectively reducing the computing power requirements for the local BMS system.

[0039] The local battery model used in the method of the present invention has model parameters that are obtained by directly applying the model parameters of the corresponding battery models to the local battery model after training a variety of cloud battery models, so that the computing power requirements of the local BMS system can be effectively reduced; and the battery status data is used to characterize the operating condition of the battery. When the local battery model determines the battery parameters based on the battery parameter identification equation, it can determine a solution algorithm that matches the current battery operating condition based on the preset correspondence between the battery operating condition and the solution algorithm of the battery parameter identification equation, thereby ensuring the identification accuracy of the battery parameters, which not only improves the estimation accuracy of the preset charge state data, but also alleviates the technical problem of limited application scenarios.

[0040] The above describes the method for determining the state of charge of the power battery, and it is understood that the training of the cloud battery model is a prerequisite for determining the state of charge of the power battery. Therefore, the following will introduce in detail how to train the cloud battery model.

[0041] In an optional embodiment, the method of the present invention further comprises the steps of:

[0042] Step S201, obtaining an offline training data set.

[0043] Among them, the offline training data set includes: multiple groups of offline training data; each group of offline training data includes: battery voltage within a first preset time period, battery load current within a first preset time period, battery temperature within a first preset time period, battery cell performance parameters within the first preset time period and preset charge state data of the battery within the first preset time period.

[0044] When training multiple cloud battery models, the embodiment of the present invention adopts a combination of offline and online training methods. Specifically, the multiple cloud battery models are first preliminarily trained using an offline training data set, and then the preliminarily trained cloud battery models are further trained using an online training data set, ultimately obtaining multiple trained cloud battery models.

[0045] The above introduces the identification of preset charge data corresponding to a battery model / type for each cloud-based battery model. Therefore, in addition to the battery voltage within the first preset time period, the battery load current within the first preset time period, the battery temperature within the first preset time period, and the preset state of charge data within the first preset time period (that is, the actual state of charge data of the training sample data) that affect the battery parameter identification, each group of offline training data in the offline training data set must also include the battery cell performance parameters, wherein the battery cell performance parameters are used to indicate the battery model / type, and then when a group of offline training data is input, a cloud-based battery model that matches the battery model / type can be selected according to its battery cell performance parameters.

[0046] Step S202, using a Monte Carlo random algorithm and an offline training data set to perform reinforcement learning training on multiple cloud battery models to obtain multiple cloud battery models to be optimized.

[0047] In order to improve the identification accuracy of the identification equation in the cloud battery model, the embodiment of the present invention is based on deep reinforcement learning and Monte Carlo random algorithm, and trains multiple cloud battery models through offline training data sets to obtain multiple cloud battery models to be optimized. It can be understood that the data in the offline training data set is obtained after sampling multiple actual battery data, and the vehicle operating conditions that can be covered are limited. In order to enable the local battery model to be applied to the determination of the battery state of charge under various operating conditions, the embodiment of the present invention also needs to use online training data to train multiple cloud battery models to be optimized.

[0048] Step S203, obtaining an online training data set.

[0049] Among them, the online training data set includes: multiple groups of online training data; each group of online training data includes: battery voltage within a second preset time period, battery load current within a second preset time period, battery temperature within a second preset time period, battery cell performance parameters within the second preset time period, and preset charge state data of the battery within the second preset time period.

[0050] Step S204, using a Monte Carlo random algorithm and an online training data set to perform reinforcement learning training on multiple cloud battery models to be optimized, to obtain multiple trained cloud battery models.

[0051] Specifically, the embodiment of the present invention does not specifically limit the duration of the first preset time period and the second preset time period. The user can set them according to actual needs. The duration of the first preset time period can be equal to the duration of the second preset time period, or the two can be unequal.

[0052] The data types included in the online training data set and the offline training data set are the same, but the data sources are different. The offline training data is obtained after sampling the historical data, and the online training data is obtained by online simulation. Therefore, the data simulation environment can be configured to simulate a variety of complex working conditions (battery usage environment), and then data collection can be performed. By using the online training data set covering a variety of complex working conditions, multiple cloud battery models to be optimized are trained based on deep reinforcement learning and Monte Carlo random algorithms to obtain multiple trained cloud battery models, that is, multiple cloud battery models with improved battery parameter identification capabilities, and multiple cloud battery models that can improve the accuracy of power battery state of charge estimation.

[0053] In an optional embodiment, if Figure 2 As shown, the above step S202 uses the Monte Carlo random algorithm and the offline training data set to perform reinforcement learning training on multiple cloud battery models, which specifically includes the following steps:

[0054] Step S2021, for the target offline training data, determine the target cloud battery model based on the target offline training data.

[0055] Among them, the target offline training data represents any offline training data in the offline training data set; the target cloud battery model represents a cloud battery model that matches the battery cell performance parameters in the target offline training data.

[0056] Specifically, according to the description above, there are multiple cloud battery models in the cloud platform, and each cloud battery model corresponds to the recognition of preset charge data of a battery model / type. Therefore, for the target offline training data, it is first necessary to determine the target cloud battery model that matches it based on the battery cell performance parameters contained therein. In other words, the cloud battery model X can only recognize the preset charge data of the set battery model / type, and the battery data of this model / type must also be used during model training.

[0057] Step S2022, determining the solution algorithm of each battery parameter identification equation and the training battery state of charge data output by the target cloud battery model when the target cloud battery model identifies the battery parameters corresponding to the target offline training data.

[0058] After selecting the corresponding target cloud battery model for the target offline training data, it is known that the cloud battery model is based on the battery parameter identification equation to determine the battery parameters, and each battery parameter identification equation has multiple solution algorithms, and one solution algorithm corresponds to one identification result. Therefore, the next step is to determine the solution algorithm for each battery parameter identification equation when the target cloud battery model identifies the battery parameters, and then determine the training battery state of charge data output by the target cloud battery model, that is, the predicted output of the target cloud battery model.

[0059] Step S2023, using the Monte Carlo random algorithm to process the target offline training data to obtain target random parameters.

[0060] In an embodiment of the present invention, not only a deep reinforcement learning training method is applied, but also a Monte Carlo random algorithm is combined. The target offline training data is processed using the Monte Carlo random algorithm to obtain target random parameters corresponding to the target offline training data, wherein the target random parameters are used to characterize the battery operating conditions corresponding to the target offline training data; that is, during the training process, the target offline training data is processed to determine the target random parameters that can be used to characterize the battery operating conditions, so that the correspondence between the battery operating conditions and the solution algorithm of the battery parameter identification equation can be established based on the target random parameters.

[0061] Step S2024, reverse-checking the battery parameter identification equation based on the training battery state of charge data.

[0062] In order to improve the identification accuracy of the battery parameter identification equation for the battery parameters, after the target cloud battery model outputs the training battery state of charge data, the training battery state of charge data is further used to reversely check the battery parameter identification equation to adjust the coefficients of the battery parameter identification equation, thereby achieving the effect of improving its identification accuracy.

[0063] Step S2025, based on the target random parameters, the training battery state of charge data and the preset state of charge data in the target offline training data, the target cloud battery model is subjected to intensive training to obtain the target cloud battery model to be optimized.

[0064] In an optional embodiment, the above step S2022, determining the solution algorithm of each battery parameter identification equation when the target cloud battery model identifies the battery parameters corresponding to the target offline training data, specifically includes the following steps:

[0065] Step S20221, based on the target offline training data and the target cloud battery model, determine the voltage change curve fitted by the target solution algorithm group.

[0066] The target solution algorithm group represents any set of preset solution algorithms for several battery parameter identification equations in the target cloud battery model.

[0067] Specifically, each battery parameter identification equation is used to identify a battery parameter, but each solution algorithm of the battery parameter identification equation can solve a battery parameter value. For example, the identification equation of internal resistance can obtain 6 optional resistance values ​​of internal resistance through FFRLS, Kalman, Polynomial, Exponential, Fourier, and Gaussian algorithms. By analogy, assuming that there are 5 types of battery parameters to be identified, and each identification equation has 6 solution algorithms, then there are 6 optional resistance values ​​of 5 types of battery parameters. 5 Group solution method (solution algorithm), that is, 6 5 Each group of solutions can fit the voltage change curve of the corresponding battery voltage.

[0068] Step S20222, using the solution algorithm group corresponding to the target voltage change curve as a target cloud battery model to identify the battery parameters corresponding to the target offline training data.

[0069] The target voltage variation curve represents a curve having the highest similarity to the voltage variation curve reflected by the target offline training data.

[0070] The target offline training data includes the battery voltage within the first preset time period. Therefore, the corresponding voltage change curve can be fitted according to the battery voltage within the first preset time period, that is, the voltage change curve reflected by the target offline training data. After the execution of step S20221 is completed, the battery parameter group corresponding to the combination of solution algorithms of different battery parameter identification equations and the voltage change curves fitted by each battery parameter group can be obtained. The embodiment of the present invention uses the curve with the highest similarity to the voltage change curve reflected by the target offline training data among multiple voltage change curves as the target voltage change curve, and uses the solution algorithm group of the battery parameter identification equation corresponding to the target voltage change curve as the target cloud battery model to identify the battery parameters of the target offline training data.

[0071] Step S20223, determining a solution algorithm for each battery parameter identification equation based on the solution algorithm group corresponding to the target voltage variation curve.

[0072] For ease of understanding, the following example is given. If the battery parameters to be identified are: internal resistance, polarization capacitance and OCV, then a total of three battery parameter identification equations are set in the cloud battery model. Assume that the solution algorithm group corresponding to the target voltage change curve is {Polynomial, Kalman, Gaussian}, that is, when the identification equation of the internal resistance is solved using the Polynomial algorithm, the identification equation of the polarization capacitance is solved using the Kalman algorithm, and the identification equation of the OCV is solved using the Gaussian algorithm, the voltage change curves fitted by the internal resistance values, polarization capacitance values ​​and OCV values ​​obtained by the above three solutions have the highest similarity to the voltage change curve reflected by the target offline training data. Therefore, after determining the target voltage change curve, the solution algorithm for each battery parameter identification equation can be determined.

[0073] In an optional implementation, the above step S203, obtaining an online training data set, specifically includes the following contents:

[0074] Repeat the following steps until you have obtained the specified amount of online training data:

[0075] Use the designated cloud platform to simulate the battery usage environment;

[0076] The state data of the battery within a second preset time period under the battery use environment, the battery cell performance parameters, and the preset state of charge data of the battery within the second preset time period are determined and used as a set of online training data.

[0077] In order to obtain an online training data set, an embodiment of the present invention specifies a cloud platform big data to simulate a random online environment (i.e., a battery usage environment), and then determines the battery status data within a second preset time period, the battery cell performance parameters, and the preset charge state data of the battery within the second preset time period from the simulated environment. Relying on the simulation capability of the cloud platform, a variety of complex working conditions can be randomly simulated, thereby improving the diversity of training data.

[0078] The above describes in detail how to use the Monte Carlo random algorithm and offline training data sets to perform reinforcement learning training on multiple cloud battery models. For the online training data set, the method of using the Monte Carlo random algorithm and the online training data set to perform reinforcement learning training on multiple cloud battery models to be optimized is the same as the training method using the offline training data set, which will not be repeated here. For details, please refer to the relevant introduction to step S202 above.

[0079] Optionally, the cloud battery model includes: a multi-order Thevenin circuit model, a fractional-order electrochemical model, an electrolyte single particle model, and a reduced-order electrochemical model. In an embodiment of the present invention, each cloud battery model includes the above four models, that is, the local / cloud battery model is essentially a hybrid model based on the fusion of the physical characteristics of the battery circuit and the electrochemical mechanism inside the battery cell.

[0080] The method for determining the state of charge of a power battery provided in an embodiment of the present invention integrates and applies a variety of battery models; a variety of battery parameter identification equations are trained through a neural network, which effectively improves the ability to identify battery parameters; a designated cloud platform with high computing power is used to simulate online training data, and a full-process training of the cloud-based battery model is performed, which can effectively improve the ability to identify battery parameters and the estimation accuracy of preset state of charge data. The trained battery model can be deployed locally with low computing power, which can reduce the computing power requirements for the local BMS system, thereby reducing the application cost of the method.

[0081] Embodiment 2

[0082] An embodiment of the present invention further provides a device for determining the state of charge of a power battery. The device for determining the state of charge of a power battery is mainly used to execute the method for determining the state of charge of a power battery provided in the first embodiment above. The device for determining the state of charge of a power battery provided in the embodiment of the present invention is specifically introduced below.

[0083] Figure 3 is a functional module diagram of a power battery charge state determination device provided by an embodiment of the present invention, such as Figure 3 As shown, the device mainly includes: a first acquisition module 10, a processing module 20, wherein:

[0084] The first acquisition module 10 is used to acquire the status data of the battery within a target time period; wherein the status data includes: voltage, current and temperature; and the status data is used to characterize the working condition of the battery.

[0085] The processing module 20 is used to process the state data using the local battery model to obtain preset state of charge data of the battery.

[0086] Among them, the local battery model is a battery parameter determined based on a battery parameter identification equation, and the battery parameter identification equation corresponds one-to-one to the battery parameters.

[0087] The model parameters of the local battery model are determined after a designated cloud platform trains multiple cloud battery models. The cloud battery model includes a battery model that is identical to the local battery model. There is a preset correspondence between the solution algorithm of the battery parameter identification equation and the operating condition of the battery.

[0088] The local battery model used in the power battery state of charge determination device provided in the embodiment of the present invention has model parameters that are obtained by directly applying the model parameters of the corresponding battery models to the local battery model after training multiple cloud battery models, so that the computing power requirements of the local BMS system can be effectively reduced; and the battery status data is used to characterize the operating condition of the battery. When the local battery model determines the battery parameters based on the battery parameter identification equation, it can determine a solution algorithm that matches the current battery operating condition based on the preset correspondence between the battery operating condition and the solution algorithm of the battery parameter identification equation, thereby ensuring the identification accuracy of the battery parameters, which not only improves the estimation accuracy of the preset state of charge data, but also alleviates the technical problem of limited application scenarios.

[0089] Optionally, the device further comprises:

[0090] The second acquisition module is used to acquire an offline training data set; wherein the offline training data set includes: multiple groups of offline training data; each group of offline training data includes: battery voltage within a first preset time period, battery load current within a first preset time period, battery temperature within a first preset time period, battery cell performance parameters within the first preset time period, and preset charge state data of the battery within the first preset time period.

[0091] The first training module is used to perform reinforcement learning training on multiple cloud battery models using a Monte Carlo random algorithm and an offline training data set to obtain multiple cloud battery models to be optimized.

[0092] The third acquisition module is used to acquire an online training data set; wherein the online training data set includes: multiple groups of online training data; each group of online training data includes: the battery voltage within a second preset time period, the battery load current within the second preset time period, the battery temperature within the second preset time period, the battery cell performance parameters within the second preset time period, and the preset charge state data of the battery within the second preset time period.

[0093] The second training module is used to perform reinforcement learning training on multiple cloud battery models to be optimized using a Monte Carlo random algorithm and an online training data set to obtain multiple trained cloud battery models.

[0094] Optionally, the first training module includes:

[0095] The first determination unit is used to determine a target cloud-based battery model based on the target offline training data; wherein the target offline training data represents any offline training data in the offline training data set; and the target cloud-based battery model represents a cloud-based battery model that matches the battery cell performance parameters in the target offline training data.

[0096] The second determination unit is used to determine the solution algorithm of each battery parameter identification equation and the training battery state of charge data output by the target cloud battery model when the target cloud battery model identifies the battery parameters corresponding to the target offline training data.

[0097] The processing unit is used to process the target offline training data using a Monte Carlo random algorithm to obtain target random parameters; wherein the target random parameters are used to characterize the battery operating conditions corresponding to the target offline training data.

[0098] The calibration unit is used to reversely calibrate the battery parameter identification equation based on the training battery state of charge data.

[0099] The training unit is used to perform intensive training on the target cloud battery model based on the target random parameters, the training battery state of charge data and the preset state of charge data in the target offline training data to obtain the target cloud battery model to be optimized.

[0100] Optionally, the second determining unit is specifically configured to:

[0101] Based on the target offline training data and the target cloud battery model, determine the voltage variation curve fitted by the target solution algorithm group; wherein the target solution algorithm group represents any set of preset solution algorithms for several battery parameter identification equations in the target cloud battery model.

[0102] A method for identifying battery parameters corresponding to target offline training data using a solution algorithm group corresponding to a target voltage change curve as a target cloud-based battery model; wherein the target voltage change curve represents a curve with the highest similarity to a voltage change curve reflected by the target offline training data.

[0103] The solution algorithm for each battery parameter identification equation is determined based on the solution algorithm group corresponding to the target voltage change curve.

[0104] Optionally, the third acquisition module is specifically used for:

[0105] Repeat the following steps until you have obtained the specified amount of online training data:

[0106] Use the specified cloud platform to simulate the battery usage environment.

[0107] The state data of the battery within a second preset time period under the battery use environment, the battery cell performance parameters, and the preset state of charge data of the battery within the second preset time period are determined and used as a set of online training data.

[0108] Optionally, the cloud battery model includes: a multi-order Thevenin circuit model, a fractional-order electrochemical model, an electrolyte single particle model, and a reduced-order electrochemical model.

[0109] Embodiment 3

[0110] See also Figure 4 An embodiment of the present invention provides an electronic device, which includes: a processor 60, a memory 61, a bus 62 and a communication interface 63, wherein the processor 60, the communication interface 63 and the memory 61 are connected via the bus 62; the processor 60 is used to execute an executable module stored in the memory 61, such as a computer program.

[0111] The memory 61 may include a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk memory. The communication connection between the system network element and at least one other network element is realized through at least one communication interface 63 (which may be wired or wireless), and the Internet, wide area network, local area network, metropolitan area network, etc. may be used.

[0112] The bus 62 may be an ISA bus, a PCI bus, or an EISA bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 4 Only one bidirectional arrow is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0113] Among them, the memory 61 is used to store programs, and the processor 60 executes the program after receiving the execution instruction. The method executed by the device defined by the process disclosed in any embodiment of the above-mentioned embodiment of the present invention can be applied to the processor 60 or implemented by the processor 60.

[0114] The processor 60 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit or software instructions in the processor 60. The above processor 60 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present invention can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor. The steps of the method disclosed in the embodiments of the present invention can be directly embodied as a hardware decoding processor to be executed, or the hardware and software modules in the decoding processor can be executed. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory 61, and the processor 60 reads the information in the memory 61 and completes the steps of the above method in combination with its hardware.

[0115] A computer program product of a power battery state of charge determination method, device, and electronic device provided in an embodiment of the present invention includes a computer-readable storage medium storing a non-volatile program code executable by a processor. The instructions included in the program code can be used to execute the method described in the previous method embodiment. The specific implementation can be found in the method embodiment, which will not be repeated here.

[0116] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0117] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium that is executable by a processor. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0118] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0119] In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inside", "outside", etc. indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, or the positions or positional relationships in which the product of the invention is usually placed when in use. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific position, be constructed and operated in a specific position, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", "third", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.

[0120] In addition, the terms "horizontal", "vertical", "overhanging" and the like do not mean that the components are required to be absolutely horizontal or overhanging, but can be slightly tilted. For example, "horizontal" only means that its direction is more horizontal than "vertical", and does not mean that the structure must be completely horizontal, but can be slightly tilted.

[0121] In the description of the present invention, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms "set", "install", "connect", and "connect" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0122] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for determining the state of charge of a power battery, characterized in that: include: Acquire status data of the battery within a target time period; wherein the status data includes: voltage, current and temperature; the status data is used to characterize the working condition of the battery; Processing the state data using a local battery model to obtain preset state of charge data of the battery; Wherein, the local battery model is the battery parameters determined based on the battery parameter identification equation, and the preset state of charge data is determined based on the battery parameters; the battery parameter identification equation corresponds one-to-one to the battery parameters; The model parameters of the local battery model are determined after a designated cloud platform trains a plurality of cloud battery models, wherein the cloud battery model includes a battery model identical to the local battery model; the solution algorithm of the battery parameter identification equation has a preset corresponding relationship with the operating condition of the battery; Also includes: Acquire an offline training data set; wherein the offline training data set includes: multiple groups of offline training data; each group of offline training data includes: battery voltage within a first preset time period, battery load current within the first preset time period, battery temperature within the first preset time period, battery cell performance parameters within the first preset time period, and preset state of charge data of the battery within the first preset time period; Perform reinforcement learning training on multiple cloud battery models using a Monte Carlo random algorithm and the offline training data set to obtain multiple cloud battery models to be optimized; Acquire an online training data set; wherein the online training data set includes: multiple groups of online training data; each group of online training data includes: battery voltage within a second preset time period, battery load current within the second preset time period, battery temperature within the second preset time period, battery cell performance parameters within the second preset time period, and preset state of charge data of the battery within the second preset time period; Performing reinforcement learning training on the plurality of cloud battery models to be optimized using a Monte Carlo random algorithm and the online training data set to obtain a plurality of trained cloud battery models; Among them, the Monte Carlo random algorithm and the offline training data set are used to perform reinforcement learning training on multiple cloud battery models, including: For target offline training data, determine a target cloud-based battery model based on the target offline training data; wherein the target offline training data represents any offline training data in the offline training data set; and the target cloud-based battery model represents a cloud-based battery model that matches the battery cell performance parameters in the target offline training data; Determine the solution algorithm of each battery parameter identification equation and the training battery state of charge data output by the target cloud battery model when the target cloud battery model identifies the battery parameters corresponding to the target offline training data; The target offline training data is processed using a Monte Carlo random algorithm to obtain a target random parameter; wherein the target random parameter is used to characterize a battery operating condition corresponding to the target offline training data; Reverse-checking the battery parameter identification equation based on the training battery state of charge data; Based on the target random parameter, the training battery state of charge data and the preset state of charge data in the target offline training data, the target cloud battery model is subjected to intensive training to obtain a target cloud battery model to be optimized; Wherein, when determining that the target cloud battery model identifies the battery parameters corresponding to the target offline training data, the solution algorithm of each battery parameter identification equation includes: Based on the target offline training data and the target cloud battery model, determine the voltage change curve fitted by the target solution algorithm group; wherein the target solution algorithm group represents any set of preset solution algorithms for several battery parameter identification equations in the target cloud battery model; A method for identifying battery parameters corresponding to the target offline training data by using a solution algorithm group corresponding to a target voltage change curve as the target cloud-based battery model; wherein the target voltage change curve represents a curve having the highest similarity to a voltage change curve reflected by the target offline training data; A solution algorithm for each battery parameter identification equation is determined based on the solution algorithm group corresponding to the target voltage variation curve.

2. The method for determining the state of charge of a power battery according to claim 1, characterized in that: Get online training datasets, including: Repeat the following steps until you have obtained the specified amount of online training data: Using the designated cloud platform to simulate a battery usage environment; The state data of the battery within a second preset time period under the battery use environment, the cell performance parameters of the battery and the preset state of charge data of the battery within the second preset time period are determined and used as a set of online training data.

3. The method for determining the state of charge of a power battery according to claim 1, characterized in that: The cloud battery model includes: a multi-order Thevenin circuit model, a fractional-order electrochemical model, an electrolyte single particle model and a reduced-order electrochemical model.

4. A device for determining the state of charge of a power battery, characterized in that: include: A first acquisition module is used to acquire the status data of the battery within a target time period; wherein the status data includes: voltage, current and temperature; and the status data is used to characterize the working condition of the battery; A processing module, used for processing the state data using a local battery model to obtain preset state of charge data of the battery; Wherein, the local battery model is a battery parameter determined based on a battery parameter identification equation, and the battery parameter identification equation corresponds one-to-one to the battery parameter; The model parameters of the local battery model are determined after a designated cloud platform trains a plurality of cloud battery models, wherein the cloud battery model includes a battery model identical to the local battery model; the solution algorithm of the battery parameter identification equation has a preset corresponding relationship with the operating condition of the battery; Also includes: A second acquisition module is used to acquire an offline training data set; wherein the offline training data set includes: multiple groups of offline training data; each group of offline training data includes: battery voltage within a first preset time period, battery load current within the first preset time period, battery temperature within the first preset time period, battery cell performance parameters within the first preset time period, and preset state of charge data of the battery within the first preset time period; A first training module is used to perform reinforcement learning training on multiple cloud battery models using a Monte Carlo random algorithm and the offline training data set to obtain multiple cloud battery models to be optimized; A third acquisition module is used to acquire an online training data set; wherein the online training data set includes: multiple groups of online training data; each group of online training data includes: the battery voltage within the second preset time period, the battery load current within the second preset time period, the battery temperature within the second preset time period, the battery cell performance parameters within the second preset time period, and the preset state of charge data of the battery within the second preset time period; A second training module is used to perform reinforcement learning training on the multiple cloud battery models to be optimized using a Monte Carlo random algorithm and the online training data set to obtain multiple trained cloud battery models; The first training module includes: A first determination unit is used to determine a target cloud-based battery model based on the target offline training data, wherein the target offline training data represents any offline training data in the offline training data set; and the target cloud-based battery model represents a cloud-based battery model that matches the battery cell performance parameters in the target offline training data; A second determination unit is used to determine, when the target cloud-based battery model identifies the battery parameters corresponding to the target offline training data, a solution algorithm for each battery parameter identification equation and the training battery state of charge data output by the target cloud-based battery model; A processing unit, used to process the target offline training data using a Monte Carlo random algorithm to obtain a target random parameter; wherein the target random parameter is used to characterize a battery operating condition corresponding to the target offline training data; A calibration unit, configured to reversely calibrate the battery parameter identification equation based on the training battery state of charge data; A training unit, configured to perform intensive training on the target cloud battery model based on the target random parameter, the training battery state of charge data, and preset state of charge data in the target offline training data, so as to obtain a target cloud battery model to be optimized; The second determining unit is specifically used for: Based on the target offline training data and the target cloud battery model, determine the voltage change curve fitted by the target solution algorithm group; wherein the target solution algorithm group represents any set of preset solution algorithms for several battery parameter identification equations in the target cloud battery model; A method for identifying battery parameters corresponding to the target offline training data by using a solution algorithm group corresponding to a target voltage change curve as the target cloud-based battery model; wherein the target voltage change curve represents a curve having the highest similarity to a voltage change curve reflected by the target offline training data; A solution algorithm for each battery parameter identification equation is determined based on the solution algorithm group corresponding to the target voltage variation curve.

5. An electronic device comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, wherein: When the processor executes the computer program, the steps of the method for determining the state of charge of a power battery as described in any one of claims 1 to 3 are implemented.

6. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and when the computer instructions are executed by a processor, the method for determining the state of charge of a power battery according to any one of claims 1 to 3 is implemented.

Images