Method, device and energy storage system for predicting the lifetime of a secondary battery

Abstract

The application provides a secondary battery life prediction method, device and energy storage system, relates to the technical field of secondary battery management, and comprises the following steps: obtaining a training data set; performing parameter identification and battery health state calibration every N cycles according to the training data set, obtaining a plurality of parameter identification results and a plurality of battery health state test values, each parameter identification result corresponding to each battery health state test value; constructing a plurality of simulation models based on the plurality of parameter identification results, each simulation model corresponding to each parameter identification result; obtaining actual voltage data of a target secondary battery and a plurality of groups of simulation voltage data; and determining a battery health state estimation value of the target secondary battery according to the plurality of battery health state test values, the actual voltage data of the target secondary battery and the plurality of groups of simulation voltage data. The method provided by the application is helpful to improve the accuracy and reliability of secondary battery life prediction.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of secondary battery management, and in particular to a secondary battery life prediction method and device and an energy storage system. BACKGROUND

[0002] With the increasingly wide application of secondary batteries, how to accurately and effectively predict the life of the secondary batteries has become a key problem. The predicted life of the secondary batteries is not only an important factor for ensuring the safety of the secondary batteries, but also an important basis for formulating the maintenance strategy of the secondary batteries.

[0003] The state of health (SOH) of a battery represents the current internal health state of the battery, and the accuracy of SOH estimation can affect the reliability of battery life prediction. Therefore, accurately estimating the SOH is a top priority for a battery management system. However, the SOH estimation method in the related art is not accurate enough, which affects the accuracy and reliability of the life prediction of the secondary batteries. SUMMARY

[0004] The present application provides a secondary battery life prediction method, device and energy storage system, which helps to improve the accuracy and reliability of the secondary battery life prediction.

[0005] In a first aspect, the present application provides a secondary battery life prediction method, comprising:

[0006] obtaining a training data set, the training data set comprising characteristic data of a secondary battery obtained through aging experiments of a test secondary battery under different working conditions;

[0007] performing parameter identification and battery state of health calibration every N cycles based on the training data set to obtain a plurality of parameter identification results and a plurality of battery state of health test values, the plurality of parameter identification results being used to represent the identification results of the to-be-identified parameters of the simulation model under different cycle numbers of different working conditions, and each parameter identification result corresponding to each battery state of health test value;

[0008] constructing a plurality of simulation models based on the plurality of parameter identification results, each simulation model corresponding to each parameter identification result;

[0009] obtaining actual voltage data of a target secondary battery and a plurality of sets of simulation voltage data, the actual voltage data of the target secondary battery being voltage data actually obtained by the target secondary battery under a target working condition, and each set of simulation voltage data being voltage data obtained by each simulation model under the target working condition;

[0010] The battery health state estimation value of the target secondary battery is determined according to the plurality of battery health state test values, the actual voltage data of the target secondary battery and the plurality of sets of simulation voltage data.

[0011] In one possible implementation, the battery health state estimation value of the target secondary battery is determined according to the plurality of battery health state test values, the actual voltage data of the target secondary battery and the plurality of sets of simulation voltage data, including:

[0012] The actual voltage data of the target secondary battery is compared with the plurality of sets of simulation voltage data respectively to determine a target set of simulation voltage data, the target set of simulation voltage data being a set of simulation voltage data in the plurality of sets of simulation voltage data that has the minimum deviation from the actual voltage data of the target secondary battery.

[0013] The battery health state estimation value of the target secondary battery is determined according to the correspondence between the target set of simulation voltage data and the battery health state test values, the battery health state estimation value of the target secondary battery being one of the plurality of battery health state test values.

[0014] In one possible implementation, the battery health state estimation value of the target secondary battery is determined according to the correspondence between the target set of simulation voltage data and the battery health state test values, including:

[0015] The target simulation model is determined according to the correspondence between the target set of simulation voltage data and the simulation model.

[0016] The target parameter identification result is determined according to the correspondence between the target simulation model and the parameter identification result.

[0017] The battery health state estimation value of the target secondary battery is determined according to the correspondence between the target parameter identification result and the battery health state test values.

[0018] In one possible implementation, the actual voltage data of the target secondary battery is compared with the plurality of sets of simulation voltage data respectively to determine the target set of simulation voltage data, including:

[0019] The target set of simulation voltage data is determined according to the minimum root mean square error of the actual voltage data of the target secondary battery and the simulation voltage data.

[0020] In one possible implementation, the characteristic data of the secondary battery includes current data, voltage data and temperature data at different cycle numbers under different working conditions, and the parameter identification is performed every N cycles according to the training data set, including:

[0021] The first current data, the first voltage data and the first temperature data at the corresponding cycle number are selected from the characteristic data of the secondary battery according to the preset cycle number interval N.

[0022] inputting the first current data, the first voltage data and the first temperature data into a pre-constructed parameter identification model to perform parameter identification.

[0023] In one possible implementation, the pre-constructed parameter identification model is a first P2D model, and a construction process of the first P2D model includes:

[0024] constructing an initial P2D model based on PyBaMM and known battery parameters of the target secondary battery;

[0025] setting to-be-identified parameters and value ranges of the to-be-identified parameters in the initial P2D model based on PyBOP, and setting a cost function and an optimization algorithm to obtain the first P2D model.

[0026] In one possible implementation, the simulation model is a second P2D model, and a plurality of simulation models are constructed based on a plurality of parameter identification results, including:

[0027] constructing a plurality of second P2D models based on PyBaMM, the plurality of parameter identification results and the known battery parameters of the target secondary battery.

[0028] In one possible implementation, the cost function includes a mean square error function, and the optimization algorithm includes an exponential natural evolution strategy algorithm.

[0029] In one possible implementation, the known battery parameters of the target secondary battery at least include electrode thickness and separator thickness, and the to-be-identified parameters at least include positive electrode diffusion rate, negative electrode diffusion rate, electrolyte diffusion rate, positive electrode conductivity, negative electrode conductivity and electrolyte conductivity.

[0030] In one possible implementation, the aging experiment of the test secondary battery under different working conditions includes:

[0031] performing charge-discharge cycles of the test secondary battery under working conditions of different temperatures and powers until the battery health state of the test secondary battery decays to a preset threshold.

[0032] In a second aspect, the present application provides a life prediction device of a secondary battery, including:

[0033] a first obtaining module, configured to obtain a training data set, the training data set including characteristic data of a secondary battery obtained through an aging experiment of a test secondary battery under different working conditions;

[0034] The parameter identification and battery health state calibration module is configured to perform parameter identification and battery health state calibration every N cycles according to the training data set, to obtain a plurality of parameter identification results and a plurality of battery health state test values, wherein the plurality of parameter identification results are used to represent the identification results of the to-be-identified parameters of the simulation model at different cycle numbers under different working conditions, and each parameter identification result corresponds to each battery health state test value;

[0035] The model construction module is configured to construct a plurality of simulation models based on the plurality of parameter identification results, and each simulation model corresponds to each parameter identification result.

[0036] The second acquisition module is configured to acquire actual voltage data of the target secondary battery and a plurality of sets of simulation voltage data, wherein the actual voltage data of the target secondary battery is voltage data actually obtained by running the target secondary battery under a target working condition, and each set of simulation voltage data is voltage data obtained by simulating and running each simulation model under the target working condition.

[0037] The determination module is configured to determine a battery health state estimation value of the target secondary battery according to the plurality of battery health state test values, the actual voltage data of the target secondary battery, and the plurality of sets of simulation voltage data.

[0038] In a third aspect, the present application provides a storage system, comprising: a BMS, the BMS comprising a processor and a memory, the memory being configured to store a computer program, and the processor being configured to run the computer program to implement the secondary battery life prediction method of the first aspect.

[0039] The present application has the following beneficial effects:

[0040] The application provides a secondary battery life prediction method, device and energy storage system. The application obtains characteristic data of a secondary battery through aging experiments of a test secondary battery under different working conditions to form a training data set. The application performs parameter identification and battery health state calibration every N cycles based on the training data set to obtain multiple parameter identification results and multiple battery health state test values. The application constructs multiple simulation models based on the multiple parameter identification results. The application obtains actual voltage data of a target secondary battery and multiple sets of simulation voltage data. The actual voltage data of the target secondary battery is voltage data actually obtained by the target secondary battery under a target working condition. Each set of simulation voltage data is voltage data obtained by each simulation model under the target working condition. The application determines a battery health state estimation value of the target secondary battery based on the multiple battery health state test values, the actual voltage data of the target secondary battery and the multiple sets of simulation voltage data. The application sets aging experiments under different working conditions to obtain comprehensive characteristic data of the secondary battery, fully considers the complex characteristic changes of the secondary battery under different working conditions, performs parameter identification based on the characteristic data of the secondary battery, realizes dynamic updating of to-be-identified parameters of the secondary battery, and then builds a simulation model with strong applicability, enhances the dynamic tracking capability of the simulation model on the aging state of the secondary battery, and improves the accuracy and reliability of the secondary battery life prediction. BRIEF DESCRIPTION OF DRAWINGS

[0041] Figure 1 A flowchart of a secondary battery life prediction method provided by an embodiment of the application is shown in the figure.

[0042] Figure 2 A flowchart of an SOH estimation method provided by an embodiment of the application is shown in the figure.

[0043] Figure 3 A structure diagram of a secondary battery life prediction device provided by an embodiment of the application is shown in the figure. DETAILED DESCRIPTION

[0044] In the embodiments of the application, unless otherwise specified, the character " / " represents a relationship of one or the other between the associated objects before and after. For example, A / B can represent A or B. The "and / or" describes the relationship between the associated objects, which means that there can be three relationships. For example, A and / or B can represent three cases of A alone, A and B together, and B alone.

[0045] It should be noted that the "first", "second" and the like in the embodiments of the application are only used for distinguishing purposes of description, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features, nor can it be understood as indicating or implying an order.

[0046] In the embodiments of the present application, "at least one" means one or more, and "multiple" means two or more. In addition, "at least one of the following" or the like means any combination of the items, which can include any combination of single item or multiple items. For example, at least one of A, B or C can mean A, B, C, A and B, A and C, B and C, or A, B and C. Each of A, B and C can be an element or a set containing one or more elements.

[0047] In the embodiments of the present application, "example", "in some embodiments", "in another embodiment" and the like are used to represent an example, illustration or description. Any embodiment or design scheme described as "example" in the present application should not be interpreted as more preferred or more advantageous than other embodiments or design schemes. Rather, the word "example" is intended to present the concept in a specific manner.

[0048] In the embodiments of the present application, "of", "corresponding" and "corresponding" can be used interchangeably at times. It should be pointed out that the meanings to be expressed are consistent when the distinction is not emphasized. In the embodiments of the present application, communication and transmission can be used interchangeably at times. It should be pointed out that the meanings to be expressed are consistent when the distinction is not emphasized. For example, transmission can include sending and / or receiving, and can be a noun or a verb.

[0049] In the embodiments of the present application, equal to can be used with greater than, which is applicable to the technical solutions adopted when greater than is used. Equal to can also be used with less than, which is applicable to the technical solutions adopted when less than is used. It should be noted that when equal to is used with greater than, it cannot be used with less than; when equal to is used with less than, it cannot be used with greater than.

[0050] Secondary batteries are prone to aging due to the influence of factors such as current, temperature, internal resistance, etc. During the aging process, the chemical characteristic parameters inside the battery are not easy to measure, thereby increasing the difficulty of battery life prediction.

[0051] In the related art, the life prediction method of the secondary battery usually adopts a relatively simplified empirical model or an electrochemical model, but these models have many deficiencies in accuracy and applicability. On the one hand, the battery characteristic data on which the battery life prediction method in the related art is based is not comprehensive enough, and the complex characteristic changes of the battery under different working conditions are not fully considered, which leads to limited estimation accuracy of the SOH in actual application, and it is difficult to meet the demand of high-precision battery management. On the other hand, the related art has defects in parameter identification, and cannot update the key parameters in the battery aging process in time and accurately, so that the model has poor dynamic tracking ability for the battery state, and is difficult to adapt to the performance evolution of the battery in the long-term cycle process. In addition, some methods in the related art have strong dependence on experimental conditions, and the parameter identification is performed offline under experimental conditions, so that the model lacks universality and flexibility, and has poor adaptability under different battery types or working conditions, which limits its wide application.

[0052] Based on the above problems, the embodiment of the present application proposes a life prediction method of a secondary battery, which helps to improve the accuracy and reliability of the life prediction of the secondary battery.

[0053] Now combined Figure 1 and Figure 2 The life prediction method of the secondary battery provided by the embodiment of the present application is described.

[0054] Figure 1 The flowchart of the life prediction method of the secondary battery provided by the embodiment of the present application is shown, which specifically includes the following steps:

[0055] Step S11, obtaining a training data set.

[0056] The training data set includes characteristic data of the secondary battery obtained by aging experiment of the test secondary battery under different working conditions. In this embodiment, the common different operating conditions of the target secondary battery (i.e. the secondary battery for which the life prediction is performed) are first determined, and then the aging experiment is designed for the test secondary battery according to different operating conditions. Different working conditions include different temperatures and powers, for example, working condition 1: 25℃, 0.5P, working condition 2: 25℃, 0.3P, working condition 3: 30℃, 0.5P. It can be understood that the operating conditions of the secondary battery can also include current, humidity, pressure and other factors, and different working conditions refer to at least one of temperature, power, current and other factors being different.

[0057] Optionally, the test secondary battery is subjected to an aging experiment under different working conditions, including: the test secondary battery is subjected to a charge-discharge cycle under different temperature and power working conditions until the state of health of the test secondary battery decays to a preset threshold. In this embodiment, a plurality of test secondary batteries can be selected, and a long cycle aging experiment is designed based on different working conditions. The different test secondary batteries are subjected to a charge-discharge cycle under different working conditions until the SOH of the test secondary battery reaches a retirement standard (i.e., a preset threshold). The retirement standard represents an SOH threshold at which the battery cannot continue to be used due to performance decay or safety risk. The threshold can be set according to the application field, battery type, and battery management strategy, etc. For example, when the SOH of a certain test secondary battery is ≤70%, the aging experiment is exited. The secondary battery characteristic data of the test secondary battery in the aging experiment under each working condition and each cycle is recorded to form a training data set. The characteristic data of the secondary battery includes current data, voltage data, temperature data, charge-discharge capacity, etc. under different cycles under different working conditions.

[0058] In the present application, comprehensive characteristic data of the secondary battery is obtained by setting aging experiments under different working conditions, and the complex characteristic changes of the secondary battery under different working conditions are fully considered, which helps to improve the accuracy and reliability of SOH estimation.

[0059] Optionally, in order to ensure the accuracy of the life prediction result of the battery, a battery of the same model as the target secondary battery can be selected as the test secondary battery.

[0060] Step S12: According to the training data set, parameter identification is performed every N cycles and the state of health of the battery is calibrated every N cycles to obtain a plurality of parameter identification results and a plurality of state of health test values.

[0061] The plurality of parameter identification results are used to represent the identification results of the to-be-identified parameters of the simulation model under different cycles under different working conditions. Each parameter identification result corresponds to each state of health test value. The identification result refers to the numerical value of the to-be-identified parameter.

[0062] In some optional embodiments, according to the training data set, parameter identification is performed every N cycles, including: selecting first current data, first voltage data, and first temperature data under a corresponding cycle from the characteristic data of the secondary battery according to a preset cycle interval N; and inputting the first current data, the first voltage data, and the first temperature data into a pre-constructed parameter identification model for parameter identification.

[0063] For example, the No. 1 test secondary battery is subjected to an aging experiment under the working condition 1, and reaches a preset threshold after 50 cycles of cyclic charging and discharging, and the aging experiment is stopped. The cycle interval N is set to 10, and parameter identification is performed once every 10 cycles. In the training data set, the battery characteristic data of the 10th cycle, the 20th cycle, the 30th cycle, the 40th cycle and the 50th cycle under the working condition 1 are obtained, including current data, voltage data and temperature data, and the battery characteristic data of the five cycles are input into the pre-constructed parameter identification model for parameter identification, respectively, to obtain the parameter identification results of the 10th cycle, the 20th cycle, the 30th cycle, the 40th cycle and the 50th cycle under the working condition 1, that is, five parameter identification results, each of which corresponds to a different cycle number (10, 20, 30, 40, 50) of the working condition 1.

[0064] Based on the battery characteristic data of different cycles under different working conditions in the training data set, parameter identification can be performed to obtain multiple parameter identification results. For example, five working conditions are designed in the present application, and each working condition is subjected to 50 cycles, and parameter identification is performed once every 10 cycles, that is, 25 parameter identification results can be obtained.

[0065] The present application also calibrates SOH every N cycles according to the training data set, which is used to detect the aging degree of the test secondary battery and also provides a reference basis for subsequent SOH estimation of the target secondary battery. Alternatively, the capacity method can be used to calibrate SOH. For example, the No. 1 test secondary battery is subjected to an aging experiment under the working condition 1, and after the 10th cycle, the ratio of the current maximum available capacity of the No. 1 test secondary battery to the initial rated capacity is calculated to obtain the SOH test value of the 10th cycle under the working condition 1. Similarly, the SOH test values of the 20th cycle, the 30th cycle, the 40th cycle and the 50th cycle under the working condition 1 can be calculated.

[0066] Taking the above example as an example, five working conditions are designed in the present application, and each working condition is subjected to 50 cycles, and SOH calibration is performed once every 10 cycles, that is, 25 SOH calibration values can be obtained, and each parameter identification result corresponds to each SOH test value.

[0067] Alternatively, the value range of the cycle interval N can be 10≤N≤20, and N is a positive integer. If N is too small, it will cause high time cost, high resource consumption and low efficiency. If N is too large, it will affect the accuracy of parameter identification and SOH calibration. Setting N to an integer between 10 and 20 can balance the accuracy and efficiency.

[0068] In the present application, the characteristic data of the secondary battery with different aging degrees under different working conditions is obtained, and the characteristic data of the secondary battery is used for parameter identification, so that the dynamic update of the to-be-identified parameters of the secondary battery is realized.

[0069] In some optional embodiments, the pre-constructed parameter identification model is a first P2D model, and a construction process of the first P2D model includes: constructing an initial P2D model based on PyBaMM and known battery parameters of the target secondary battery; setting to-be-identified parameters and value ranges of the to-be-identified parameters in the initial P2D model based on PyBOP, and setting a cost function and an optimization algorithm to obtain the first P2D model. Wherein, the input of the first P2D model is current data, voltage data and temperature data in the training data set, and the output is a numerical value of the to-be-identified parameter.

[0070] It should be noted that the P2D model (Pseudo-Two-Dimensional Model) is an electrochemical model of a battery, which is used to describe the electrochemical reaction, ion transport and polarization behavior inside the battery. The parameters of the P2D model can represent the characteristics of the electrode or electrolyte material, and can be used to study the battery health state characteristics. By identifying the model parameters using battery characteristic data (current, voltage, temperature, etc.), the parameters affecting the aging degree of the battery can be obtained. PyBaMM (Python Battery Mathematical Modelling) is a battery simulation tool used for fast simulation and optimization of electrochemical models, which has multiple battery models (such as P2D, SPM single particle model) built-in. PyBOP (Python Battery Optimization Package) is an extension toolkit based on PyBaMM, which is used for battery parameter identification and optimization.

[0071] In the present application, the first P2D model is established by using PyBaMM and PyBOP tools and known battery parameters of the target secondary battery, and the parameter identification is performed according to the battery characteristic data of the training data set by using the first P2D model, so as to obtain the test battery parameter identification result under different working conditions and different cycle numbers, to obtain the parameter change of the battery under different aging degrees, to realize the dynamic update of the to-be-identified parameters of the test secondary battery, to improve the dynamic tracking ability of the simulation model constructed by using the to-be-identified parameters on the battery state in the subsequent, and to improve the accuracy and reliability of SOH estimation.

[0072] In some optional embodiments, the cost function of the first P2D model includes a mean squared error function (MSE, Mean Squared Error), and the optimization algorithm of the first P2D model includes an exponential natural evolution strategy algorithm (XNES, Exponential Natural Evolution Strategy).

[0073] The cost function is also called the objective function or loss function, which optimizes the parameters of the model by calculating the deviation between the predicted value and the true value. The optimization goal is to minimize the cost function, that is, to minimize the deviation. Optionally, the voltage data in the training data set is used as the cost function of the first P2D model.

[0074] The calculation formula of MSE is as follows:

[0075]

[0076] Wherein, represents the i-th moment, represents the predicted value, represents the true value, and n represents the total number of data points.

[0077] The parameter update rule of XNES algorithm is as follows:

[0078] (1) Mean update:

[0079] (2) Linear transformation matrix update:

[0080] (3) Global step update:

[0081] Wherein, represents the mean, represents the learning rate of the mean, represents the global step, represents the linear transformation matrix, represents the component of the natural gradient in the mean direction, represents the learning rate of the covariance, represents the component of the natural gradient in the covariance direction, represents the learning rate of the global step, represents the matrix exponential, represents the trace of the matrix.

[0082] Optionally, the known battery parameters of the target secondary battery include electrode thickness, separator thickness, solid / liquid volume fraction, etc. The known parameters of the target secondary battery refer to the set values or nominal values, or parameters that can be obtained by direct measurement; the to-be-identified parameters include positive electrode diffusion rate, negative electrode diffusion rate, electrolyte diffusion rate, positive electrode conductivity, negative electrode conductivity, electrolyte conductivity, positive electrode reaction activation energy, negative electrode reaction activation energy, positive electrode reaction rate constant and negative electrode reaction rate constant, etc. The to-be-identified parameters refer to parameters that cannot be directly measured and need to be identified by experimental data.

[0083] Step S13, constructing a plurality of simulation models based on a plurality of parameter identification results.

[0084] In this embodiment, a plurality of simulation models are constructed using the plurality of parameter identification results obtained in step S12, i.e., the values of the plurality of to-be-identified parameters of the simulation model. Each simulation model corresponds to each parameter identification result.

[0085] In this application, the plurality of parameter identification results are used to represent the identification results of the to-be-identified parameters of the simulation model under different cycle numbers of different working conditions, and the plurality of simulation models constructed based on the plurality of parameter identification results represent the simulation model under different cycle numbers of different working conditions. The application identifies the parameters of the simulation model using comprehensive battery data in the training data set, which helps to comprehensively and accurately reflect the parameter changes in the battery aging process, and at the same time enhances the adaptability of the simulation model to different working conditions, thereby realizing high-precision and reliable estimation of the SOH of the secondary battery, and providing strong support for efficient management and safe use of the secondary battery.

[0086] It can be understood that each parameter identification result corresponds to each SOH battery health state test data value, and each simulation model corresponds to each parameter identification result. Therefore, each simulation model corresponds to each SOH battery health state test data value.

[0087] In some optional embodiments, the simulation model is a second P2D model, and the plurality of simulation models are constructed based on the plurality of parameter identification results, including: constructing a plurality of second P2D models based on PyBaMM, the plurality of parameter identification results, and known battery parameters of the target secondary battery. The input of the second P2D model is current data and temperature data, and the output is voltage data.

[0088] Similarly, the cost function and the optimization algorithm of the second P2D model can also be set, for example, the cost function of the second P2D model includes a mean squared error function (MSE, Mean Squared Error), and the optimization algorithm of the second P2D model includes an exponential natural evolution strategy algorithm (XNES, Exponential Natural Evolution Strategy).

[0089] In step S14, actual voltage data of the target secondary battery and a plurality of sets of simulation voltage data are obtained.

[0090] The actual voltage data of the target secondary battery is voltage data (i.e., data of the change of actual voltage with time) actually obtained by the target secondary battery under a target working condition, and each set of simulation voltage data is voltage data (i.e., data of the change of simulation voltage with time) obtained by each simulation model under the target working condition. The target working condition includes temperature and current information.

[0091] In this embodiment, the actual voltage data of the target secondary battery is obtained under the actual working condition (i.e., the target working condition) of the target secondary battery, and is recorded as the actual voltage data of the target secondary battery. The plurality of simulation models are simulated under the actual working condition to obtain the simulation voltage data of each model, thereby forming a simulation voltage data set, and the simulation voltage data set includes a plurality of simulation voltage data.

[0092] In step S15, the SOH estimation value of the target secondary battery is determined according to the plurality of battery health state test values, the actual voltage data of the target secondary battery, and the plurality of simulation voltage data.

[0093] Figure 2 The flowchart of the SOH estimation method provided in the embodiments of the present application is shown in FIG. 1. As shown in FIG. 1, the actual voltage data of the target secondary battery is obtained by actually running the target secondary battery under the target working condition, the simulation voltage data set is obtained by simulating the plurality of simulation models under the target working condition, the actual voltage data of the target secondary battery and the simulation voltage data set are input into the SOH estimation model, and the SOH estimation value of the target secondary battery is output. Figure 2

[0094] In some optional embodiments, the actual voltage data of the target secondary battery and the simulation voltage data set are input into the SOH estimation model, and the SOH estimation value of the target secondary battery is output, including: determining the SOH estimation value of the target secondary battery according to the plurality of battery health state test values, the actual voltage data of the target secondary battery, and the plurality of simulation voltage data.

[0095] Optionally, the SOH estimation value of the target secondary battery is determined according to the plurality of battery health state test values, the actual voltage data of the target secondary battery, and the plurality of simulation voltage data, including:

[0096] The actual voltage data of the target secondary battery is compared with the plurality of simulation voltage data respectively to determine a target simulation voltage data group, the target simulation voltage data group being the simulation voltage data group with the smallest deviation from the actual voltage data of the target secondary battery in the plurality of simulation voltage data; the SOH estimation value of the target secondary battery is determined according to the corresponding relationship between the target simulation voltage data group and the battery health state test value, and the SOH estimation value of the target secondary battery is one of the plurality of battery health state test values.

[0097] ​In this embodiment, by comparing the actual voltage data of the target secondary battery with a plurality of sets of simulation voltage data respectively, the set of simulation voltage data with the smallest deviation from the actual voltage data of the target secondary battery is determined, i.e., the target set of simulation voltage data is determined, and then the SOH estimation value of the target secondary battery is determined according to the corresponding relationship between the target set of simulation voltage data and the SOH test value. The above steps show that the trend of the voltage of the target set of simulation voltage data changing with time is most close to the trend of the voltage of the target secondary battery changing with time, and the SOH test value corresponding to the target set of simulation voltage data can be taken as the SOH estimation value of the target secondary battery.

[0098] Optionally, determining the battery state of health estimation value of the target secondary battery according to the corresponding relationship between the target set of simulation voltage data and the battery state of health test value comprises:

[0099] determining a target simulation model according to the corresponding relationship between the target set of simulation voltage data, the simulation voltage data and the simulation model;

[0100] determining a target parameter identification result according to the corresponding relationship between the target simulation model, the simulation model and the parameter identification result;

[0101] determining the battery state of health estimation value of the target secondary battery according to the corresponding relationship between the target parameter identification result, the parameter identification result and the battery state of health test value.

[0102] In this application, each set of simulation voltage data corresponds to a simulation model, each simulation model corresponds to a parameter identification result, and each parameter identification result corresponds to an SOH test value. Therefore, the SOH test value corresponding to the target set of simulation voltage data can be obtained by using the corresponding relationship between the parameters, and the SOH estimation value of the target secondary battery can be determined.

[0103] In some optional embodiments, comparing the actual voltage data of the target secondary battery with a plurality of sets of simulation voltage data respectively to determine the target set of simulation voltage data comprises:

[0104] determining the target set of simulation voltage data according to the minimum root mean square error of the actual voltage data of the target secondary battery and the simulation voltage data.

[0105] Specifically, the root mean square error of the target voltage data and a plurality of sets of simulation voltage data is calculated to determine the target set of simulation voltage data, and the root mean square error of the target set of simulation voltage data and the target voltage data is the smallest.

[0106] The calculation formula of the root mean square is:

[0107]

[0108] wherein, represents the i th moment, represents the predicted value, represents the true value, and n represents the total number of data points.

[0109] In this embodiment, by calculating the root mean square error of the target voltage data and a plurality of sets of simulation voltage data, the set of simulation voltage data with the smallest root mean square error is determined, the trend of the voltage of the set of simulation voltage data changing with time is most close to the trend of the voltage of the target secondary battery changing with time, and the SOH test value corresponding to the set of simulation voltage data can be taken as the SOH estimation value of the target secondary battery.

[0110] In other embodiments, the mean absolute error (MAE), the maximum absolute error (Max Error), the coefficient of determination (R²), etc. can also be used to determine the set of simulation voltage data with the smallest deviation from the actual voltage data of the target secondary battery.

[0111] The method for predicting the life of a secondary battery provided by the present application comprises the following steps: obtaining characteristic data of a test secondary battery through aging experiments under different working conditions to form a training data set; performing parameter identification and battery health state calibration every N cycles based on the training data set to obtain a plurality of parameter identification results and a plurality of battery health state test values; constructing a plurality of simulation models based on the plurality of parameter identification results; obtaining actual voltage data of a target secondary battery and a plurality of sets of simulation voltage data, wherein the actual voltage data of the target secondary battery is voltage data actually obtained by the target secondary battery under a target working condition, and each set of simulation voltage data is voltage data obtained by each simulation model under the target working condition; and determining a battery health state estimation value of the target secondary battery based on the plurality of battery health state test values, the actual voltage data of the target secondary battery, and the plurality of sets of simulation voltage data. By setting aging experiments under different working conditions to obtain comprehensive characteristic data of the secondary battery, the complex characteristic changes of the secondary battery under different working conditions are fully considered, and parameter identification is performed using the characteristic data of the secondary battery, so that the dynamic update of the to-be-identified parameters of the secondary battery is realized, and then a simulation model with strong applicability is built, the dynamic tracking capability of the simulation model for the aging state of the secondary battery is enhanced, and the accuracy and reliability of the life prediction of the secondary battery are improved.

[0112] In some optional embodiments, the method provided by the present application further comprises: formulating a maintenance strategy for the target secondary battery according to the SOH estimation value of the target secondary battery.

[0113] Optionally, when the estimation value of SOH > 90%, the SOH is detected once every 3 months, and the number of charge-discharge cycles is recorded; or

[0114] When 80% < the estimation value of SOH ≤ 90%, the SOH is detected once a month, and the charge-discharge strategy is optimized;

[0115] When the estimated value of SOH is 70% < SOH ≤ 80%, the SOH is detected once every 2 weeks, and the equalization charging is implemented.

[0116] When the estimated value of SOH is ≤ 70%, the on-site inspection is stopped, the safety diagnosis (such as internal resistance, lithium extraction detection) is performed, and the battery is prepared to be replaced.

[0117] The application formulates the maintenance strategy of the battery by using the SOH estimated value, can early warning, guarantees the safety of the battery and the system, reduces the maintenance time and cost.

[0118] Based on the same idea, the application embodiment also provides a life prediction device of a secondary battery, such as Figure 3 A structure schematic diagram of a life prediction device of a secondary battery provided by the application embodiment, the life prediction device of the secondary battery 30 mainly includes:

[0119] The first acquisition module 31 is used to acquire a training data set, and the training data set includes characteristic data of the secondary battery obtained by performing an aging experiment on a test secondary battery under different working conditions;

[0120] The parameter identification and battery health state calibration module 32 is used to perform parameter identification and battery health state calibration every N cycles according to the training data set, to obtain a plurality of parameter identification results and a plurality of battery health state test values, the plurality of parameter identification results are used to represent the identification results of the to-be-identified parameters of the simulation model under different cycle numbers of different working conditions, and each parameter identification result corresponds to each battery health state test value;

[0121] The model construction module 33 is used to construct a plurality of simulation models based on the plurality of parameter identification results, and each simulation model corresponds to each parameter identification result;

[0122] The second acquisition module 34 is used to acquire actual voltage data of a target secondary battery and a plurality of groups of simulation voltage data, the actual voltage data of the target secondary battery is voltage data actually obtained by the target secondary battery under a target working condition, and each group of simulation voltage data is voltage data obtained by simulating and running each simulation model under the target working condition;

[0123] The determination module 35 is used to determine a battery health state estimated value of the target secondary battery according to the plurality of battery health state test values, the actual voltage data of the target secondary battery and the plurality of groups of simulation voltage data.

[0124] In one possible implementation manner, the determination module 35 can also be used to:

[0125] The actual voltage data of the target secondary battery is compared with each of the plurality of sets of simulation voltage data respectively to determine a target set of simulation voltage data, the target set of simulation voltage data being the set of simulation voltage data that has the smallest deviation from the actual voltage data of the target secondary battery among the plurality of sets of simulation voltage data;

[0126] The battery state of health estimation value of the target secondary battery is determined according to the correspondence between the target set of simulation voltage data and the battery state of health test values, the battery state of health estimation value of the target secondary battery being one of the plurality of battery state of health test values.

[0127] In one possible implementation, the battery state of health estimation value of the target secondary battery is determined according to the correspondence between the target set of simulation voltage data and the battery state of health test values, including:

[0128] The target simulation model is determined according to the correspondence between the target set of simulation voltage data, the simulation voltage data, and the simulation model.

[0129] The target parameter identification result is determined according to the correspondence between the target simulation model, the simulation model, and the parameter identification result.

[0130] The battery state of health estimation value of the target secondary battery is determined according to the correspondence between the target parameter identification result, the parameter identification result, and the battery state of health test values.

[0131] In one possible implementation, the actual voltage data of the target secondary battery is compared with each of the plurality of sets of simulation voltage data respectively to determine the target set of simulation voltage data, including:

[0132] The target set of simulation voltage data is determined according to the smallest root mean square error of the actual voltage data of the target secondary battery and the simulation voltage data.

[0133] In one possible implementation, the characteristic data of the secondary battery includes current data, voltage data, and temperature data at different cycle numbers under different working conditions, and the parameter identification and battery state of health calibration module 32 can be further configured to:

[0134] The first current data, the first voltage data, and the first temperature data at the corresponding cycle number are selected from the characteristic data of the secondary battery according to a preset cycle number interval N.

[0135] The first current data, the first voltage data, and the first temperature data are input into the pre-constructed parameter identification model for parameter identification.

[0136] In one possible implementation, the pre-constructed parameter identification model is a first P2D model, and the construction process of the first P2D model includes:

[0137] constructing an initial P2D model based on PyBaMM and known battery parameters of the target secondary battery;

[0138] setting the to-be-identified parameters and the value ranges of the to-be-identified parameters in the initial P2D model based on PyBOP, setting a cost function and an optimization algorithm, and obtaining a first P2D model.

[0139] In one possible implementation, the simulation model is a second P2D model, and the plurality of simulation models are constructed based on the plurality of parameter identification results, including:

[0140] constructing a plurality of second P2D models based on PyBaMM, the plurality of parameter identification results, and the known battery parameters of the target secondary battery.

[0141] In one possible implementation, the cost function includes a mean square error function, and the optimization algorithm includes an exponential natural evolution strategy algorithm.

[0142] In one possible implementation, the known battery parameters of the target secondary battery at least include electrode thickness and separator thickness, and the to-be-identified parameters at least include positive electrode diffusion rate, negative electrode diffusion rate, electrolyte diffusion rate, positive electrode conductivity, negative electrode conductivity, and electrolyte conductivity.

[0143] In one possible implementation, the aging experiment of the test secondary battery under different working conditions includes:

[0144] The test secondary battery is subjected to charge-discharge cycles under working conditions of different temperatures and powers until the battery health state of the test secondary battery decays to a preset threshold.

[0145] Figure 3 The secondary battery life prediction device 30 provided in the embodiments can be used to execute the technical solutions of the method embodiments, and the implementation principles and technical effects can be further referred to the related descriptions in the method embodiments.

[0146] It should be understood that the above Figure 3The division of each module of the secondary battery life prediction device 30 shown is only a logical functional division, and in actual implementation, all or part of the modules can be integrated into one physical entity or physically separated. Moreover, the modules can all be implemented in the form of software invoked by a processing element; all be implemented in the form of hardware; or part of the modules be implemented in the form of software invoked by a processing element and part of the modules be implemented in the form of hardware. For example, the first acquisition module can be a separately established processing element or integrated in a certain chip of an electronic device. The implementation of other modules is similar. Moreover, all or part of the modules can be integrated together or independently implemented. In the implementation process, each step of the above method or each of the above modules can be completed by an integrated logic circuit of hardware in a processor element or an instruction in the form of software.

[0147] For example, the above modules can be one or more integrated circuits configured to implement the above method, such as one or more application specific integrated circuits (ASICs), or one or more digital signal processors (DSPs), or one or more field programmable gate arrays (FPGAs), etc. For another example, the modules can be integrated together to implement in the form of a system on a chip (SOC).

[0148] In the above embodiments, the processor can include a CPU, a DSP, a microcontroller, or a digital signal processor, and can further include a GPU, an embedded neural network processing unit (NPU), and an image signal processor (ISP). The processor can further include necessary hardware accelerators or logic processing hardware circuits, such as an ASIC, or one or more integrated circuits for controlling the execution of programs of the technical solutions of the present application. In addition, the processor can have the function of operating one or more software programs, and the software programs can be stored in a storage medium.

[0149] The embodiments of the present application also provide a storage system, including a BMS (Battery Management System), the BMS including a processor and a memory, the memory being configured to store a computer program, and the processor being configured to execute the computer program to implement the secondary battery life prediction method provided by the embodiments of the present application.

[0150] The above description is only specific embodiments of the present application, and any skilled in the art can easily think of changes or replacements within the technical scope disclosed by the present application, which should be covered by the protection scope of the present application. The protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A life prediction method of a secondary battery, characterized by, The method comprises: obtaining a training data set comprising characteristic data of secondary batteries obtained through aging experiments of the secondary batteries under different working conditions; performing parameter identification and battery health state calibration every N cycles according to the training data set to obtain a plurality of parameter identification results and a plurality of battery health state test values, wherein the plurality of parameter identification results are used to represent the identification results of the to-be-identified parameters of the simulation model under different cycle numbers of different working conditions, and each parameter identification result corresponds to each battery health state test value; constructing a plurality of simulation models based on the plurality of parameter identification results, wherein each simulation model corresponds to each parameter identification result; obtaining actual voltage data of a target secondary battery and a plurality of sets of simulation voltage data, wherein the actual voltage data of the target secondary battery is voltage data actually obtained through actual operation of the target secondary battery under a target working condition, and each set of simulation voltage data is voltage data obtained through simulation operation of each simulation model under the target working condition; determining a battery health state estimation value of the target secondary battery according to the plurality of battery health state test values, the actual voltage data of the target secondary battery, and the plurality of sets of simulation voltage data.

2. The life prediction method of the secondary battery according to claim 1, characterized by, The determination of the battery health state estimation value of the target secondary battery according to the plurality of battery health state test values, the actual voltage data of the target secondary battery, and the plurality of sets of simulation voltage data comprises: comparing the actual voltage data of the target secondary battery with the plurality of sets of simulation voltage data respectively to determine a target set of simulation voltage data, wherein the target set of simulation voltage data is the set of simulation voltage data with the smallest deviation from the actual voltage data of the target secondary battery in the plurality of sets of simulation voltage data; determining the battery health state estimation value of the target secondary battery according to the correspondence between the target set of simulation voltage data and the battery health state test value, wherein the battery health state estimation value of the target secondary battery is one of the plurality of battery health state test values.

3. The life prediction method of the secondary battery according to claim 2, characterized by, The determination of the battery health state estimation value of the target secondary battery according to the correspondence between the target set of simulation voltage data and the battery health state test value comprises: determining a target simulation model according to the correspondence between the target set of simulation voltage data, simulation voltage data, and the simulation model; determining a target parameter identification result according to the correspondence between the target simulation model, the simulation model, and the parameter identification result; determining the battery health state estimation value of the target secondary battery according to the correspondence between the target parameter identification result, the parameter identification result, and the battery health state test value.

4. The life prediction method of the secondary battery according to claim 2 or 3, characterized by, The comparison of the actual voltage data of the target secondary battery with the plurality of sets of simulation voltage data respectively to determine the target set of simulation voltage data comprises: determining the target set of simulation voltage data according to the smallest root mean square error of the actual voltage data of the target secondary battery and the simulation voltage data.

5. The life prediction method of the secondary battery according to claim 1, characterized by, The characteristic data of the secondary batteries comprises current data, voltage data, and temperature data under different cycle numbers of different working conditions, and the parameter identification every N cycles according to the training data set comprises: selecting first current data, first voltage data and first temperature data at corresponding cycle numbers from the characteristic data of the secondary battery according to a preset cycle number interval N; inputting the first current data, the first voltage data and the first temperature data into a pre-constructed parameter identification model for parameter identification.

6. The life prediction method of the secondary battery according to claim 5, characterized by, The pre-constructed parameter identification model is a first P2D model, and the construction process of the first P2D model comprises: constructing an initial P2D model based on PyBaMM and known battery parameters of the target secondary battery; setting to-be-identified parameters and value ranges of the to-be-identified parameters in the initial P2D model based on PyBOP, and setting a cost function and an optimization algorithm to obtain the first P2D model.

7. The life prediction method of the secondary battery according to claim 1, characterized by, The simulation model is a second P2D model, and the constructing a plurality of simulation models based on the plurality of parameter identification results comprises: constructing a plurality of second P2D models based on PyBaMM, the plurality of parameter identification results and the known battery parameters of the target secondary battery respectively.

8. The life prediction method of the secondary battery according to claim 6, characterized by, The cost function comprises a mean square error function, and the optimization algorithm comprises an exponential natural evolution strategy algorithm.

9. The life prediction method of the secondary battery according to claim 6, characterized by, The known battery parameters of the target secondary battery at least include electrode thickness and separator thickness; and the to-be-identified parameters at least include positive electrode diffusion rate, negative electrode diffusion rate, electrolyte diffusion rate, positive electrode conductivity, negative electrode conductivity and electrolyte conductivity.

10. The life prediction method of the secondary battery according to claim 1, characterized by, The test secondary battery is subjected to an aging experiment under different working conditions, comprising: The test secondary battery is subjected to charge-discharge cycles under working conditions of different temperatures and powers until the battery health state of the test secondary battery decays to a preset threshold.

11. A life prediction device of a secondary battery, characterized by comprising: comprising: a first acquisition module configured to acquire a training data set, the training data set comprising characteristic data of a test secondary battery obtained by an aging experiment of the test secondary battery under different working conditions; a parameter identification and battery health state calibration module configured to perform parameter identification and battery health state calibration every N cycles according to the training data set to obtain a plurality of parameter identification results and a plurality of battery health state test values, the plurality of parameter identification results being used to represent identification results of to-be-identified parameters of a simulation model at different cycle numbers under different working conditions, and each parameter identification result corresponding to each battery health state test value; a model construction module configured to construct a plurality of simulation models based on the plurality of parameter identification results, each simulation model corresponding to each parameter identification result; a second acquisition module configured to acquire actual voltage data of a target secondary battery and a plurality of sets of simulation voltage data, the actual voltage data of the target secondary battery being voltage data actually obtained by actual operation of the target secondary battery under a target working condition, and each set of simulation voltage data being voltage data obtained by simulation operation of each simulation model under the target working condition; a determination module configured to determine a battery health state estimation value of the target secondary battery according to the plurality of battery health state test values, the actual voltage data of the target secondary battery and the plurality of sets of simulation voltage data.

12. An energy storage system characterized by, comprising: A battery management system (BMS) including a processor and a memory for storing a computer program; the processor is configured to run the computer program to implement the method for predicting the life of a secondary battery according to any one of claims 1-10.

Images