Method and device for predicting electric quantity of lithium battery and electronic equipment

By combining a second-order capacitor-resistance equivalent circuit model with a neural network model, the inaccuracy problem caused by a single factor in lithium battery power prediction is solved, achieving more accurate and adaptive power prediction.

CN119024176APending Publication Date: 2024-11-26HEFEI GUOXUAN HIGH TECH POWER ENERGY
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411119436.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-13
Publication Date
2024-11-26

AI Technical Summary

Technical Problem

Existing technologies for predicting lithium battery capacity consider only a single factor, leading to inaccurate predictions.

Method used

A lithium battery simulation model is established using a second-order capacitor-resistance equivalent circuit model. Combining the state equation and observation equation, a neural network model is used to predict the battery capacity, taking into account both physical characteristics and data-driven methods.

Benefits of technology

It improves the accuracy and robustness of lithium battery power prediction, and can more comprehensively reflect changes in lithium battery power, adapting to various operating conditions and noise interference.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119024176A_ABST
    Figure CN119024176A_ABST
Patent Text Reader

Abstract

The invention discloses an electric quantity prediction method and device of a lithium battery and electronic equipment. The method comprises the following steps: acquiring a lithium battery simulation model; determining a state equation of the lithium battery simulation model and an observation equation of the lithium battery simulation model; based on the actual input quantity, the state equation and the observation equation, determining first electric quantity estimated values corresponding to the lithium battery simulation model at a plurality of moments; inputting the actual input quantity and the lithium battery simulation model into an electric quantity prediction model to obtain second electric quantity estimated values corresponding to the lithium battery simulation model at multiple moments; and according to the first electric quantity estimation values and the second electric quantity estimation values corresponding to the multiple moments, target electric quantity estimation values corresponding to the lithium ion battery at the multiple moments are determined. According to the invention, the technical problem of inaccurate prediction of the electric quantity of the battery caused by relatively single consideration factor when the electric quantity of the battery is predicted in the prior art is solved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of power prediction, and in particular to a method, device and electronic equipment for predicting power of a lithium battery. Background Art

[0002] Lithium-ion batteries have become a highly regarded and favored energy medium in the fields of electric vehicles and energy storage power stations due to their advantages in energy density, power density, cycle life and memory effect. In order to ensure the safety of the battery, prevent safety accidents such as fire and explosion, and maximize the utilization of the battery and extend its service life, a battery management system (BMS) is usually required. BMS monitors various status indicators inside the battery, such as state of charge (SOC), state of health (SOH), state of power (SOP), etc. Among them, the state of charge describes the remaining power of the battery and is one of the most basic and critical states of the battery.

[0003] The SOC estimation methods include the ampere-hour integration method, the open circuit voltage method, the neural network method and the Kalman filter algorithm. The ampere-hour integration method and the open circuit voltage method have the advantages of small calculation amount and easy implementation in SOC estimation, but are affected by the current accumulation error and the initial value error, and their SOC estimation accuracy is low. For the nonlinear system of the lithium battery model, the extended Kalman filter (EKF) algorithm is used to linearize it, and the SOC estimation of the linear system is performed in combination with the KF algorithm, thereby improving the SOC estimation accuracy. However, in addition to the large amount of calculation, the extended Kalman filter is also affected by linear errors. As a type of neural network, the BP neural network has strong learning ability and complex mapping ability, so it can be trained offline and used to compensate the EKF algorithm online, and this method has a short calculation time. However, when predicting the battery power through the above technology, the factors considered are relatively single, and there are technical problems that lead to inaccurate prediction of battery power.

[0004] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0005] The embodiments of the present invention provide a lithium battery power prediction method, device and electronic device to at least solve the technical problem that when predicting the battery power in the related art, a relatively single factor is considered, resulting in inaccurate prediction of the battery power.

[0006] According to one aspect of an embodiment of the present invention, a method for predicting the power of a lithium battery is provided, comprising: obtaining a lithium battery simulation model, wherein the lithium battery simulation model is obtained by modeling a lithium-ion battery using a second-order capacitor-resistor equivalent circuit model; determining a state equation of the lithium battery simulation model and an observation equation of the lithium battery simulation model, wherein the state equation is an equation that reflects the change of the power of the lithium battery and the polarization voltage of the lithium battery over time, and the observation equation is an equation that reflects the relationship between the terminal voltage of the lithium battery and a predetermined parameter, and the predetermined parameter includes the power of the lithium battery and the polarization voltage of the lithium battery; based on an actual input amount, the state equation and the observation equation, determining first power estimation values ​​corresponding to the lithium battery simulation model at multiple moments; inputting the actual input amount and the lithium battery simulation model into a power prediction model to obtain second power estimation values ​​corresponding to the lithium battery simulation model at the multiple moments, wherein the power prediction model is obtained by training based on sample data, and the power prediction model is a neural network model; determining target power estimation values ​​corresponding to the multiple moments corresponding to the lithium-ion battery according to the first power estimation values ​​and the second power estimation values ​​corresponding to the multiple moments.

[0007] Optionally, obtaining the lithium battery simulation model includes: obtaining the lithium-ion battery; performing a mixed pulse power characteristic experiment on the lithium-ion battery to determine the relationship between the open circuit voltage, power and time of the lithium-ion battery; determining equivalent circuit model parameters based on the relationship, wherein the equivalent circuit model parameters include at least: battery open circuit voltage, equivalent ohmic internal resistance, equivalent resistance generated by electrochemical polarization, equivalent capacitance generated by electrochemical polarization, equivalent resistance generated by concentration polarization, equivalent capacitance generated by concentration polarization; determining the lithium battery simulation model based on the equivalent circuit model parameters.

[0008] Optionally, determining the target power estimation values ​​corresponding to the multiple moments corresponding to the lithium-ion battery based on the first power estimation values ​​and the second power estimation values ​​corresponding to the multiple moments includes: constructing a decision matrix based on the first power estimation values ​​and the second power estimation values ​​corresponding to the multiple moments; determining a first weight corresponding to the first method and a second weight corresponding to the second method, wherein the first method is a method of determining the power estimation value based on the state equation and the observation equation, and the second method is a method of determining the power estimation value based on the lithium battery simulation model; based on the decision matrix, the first weight and the second weight are used to determine the target power estimation values ​​corresponding to the multiple moments corresponding to the lithium-ion battery.

[0009] Optionally, determining the target power estimation values ​​corresponding to the multiple moments corresponding to the lithium-ion battery based on the decision matrix, the first weight and the second weight includes: determining an ideal decision matrix based on the decision matrix, the first weight and the second weight; determining Euclidean distance and grey correlation coefficient based on the ideal decision matrix; determining the target power estimation values ​​corresponding to the multiple moments corresponding to the lithium-ion battery based on the Euclidean distance and the grey correlation coefficient.

[0010] Optionally, determining the state equation of the lithium battery simulation model and the observation equation of the lithium battery simulation model includes: determining a predetermined equivalent circuit model expression equation based on the lithium battery simulation model, and determining a definition equation for the lithium battery power; discretizing the second-order capacitor-resistor equivalent circuit model expression equation and the definition equation for the lithium battery power using a first discrete parameter to determine the state equation, and discretizing the second-order capacitor-resistor equivalent circuit model expression equation and the definition equation for the lithium battery power using a second discrete parameter to determine the observation equation.

[0011] Optionally, the step of inputting the actual input quantity and the lithium battery simulation model into a power prediction model to obtain second power estimation values ​​corresponding to the lithium battery simulation model at the multiple moments includes: inputting the actual input quantity and the lithium battery simulation model into a hidden layer of the power prediction model to obtain a hidden layer output; inputting the hidden layer output into an output layer of the power prediction model to obtain an output layer output; inputting the output layer output into a neural network output error layer of the power prediction model to obtain an output error; and adjusting the output layer output according to the output error to obtain second power estimation values ​​corresponding to the lithium battery simulation model at the multiple moments.

[0012] Optionally, the method of determining the first power estimation values ​​corresponding to the lithium battery simulation model at multiple moments based on the actual input quantity, the state equation and the observation equation includes: determining a system state equation and an error covariance matrix equation; determining a predicted value of the state variable at the second moment based on the system state equation and the estimated value of the state variable at the first moment, and determining a predicted value of the error covariance matrix at the second moment based on the error covariance matrix equation and the estimated value of the error covariance matrix at the first moment, wherein the first moment is the moment before the second moment; determining a Kalman filter gain matrix based on the predicted value of the state variable at the second moment and the predicted value of the error covariance matrix at the second moment; updating the estimated value of the state variable at the first moment based on the Kalman filter gain matrix to obtain an updated state variable estimated value, and updating the estimated value of the error covariance matrix at the first moment based on the Kalman filter gain matrix to obtain an updated error covariance matrix estimated value; determining the first power estimation values ​​corresponding to the lithium battery simulation model at multiple moments based on the updated state variable estimated value and the updated error covariance matrix estimated value.

[0013] According to one aspect of an embodiment of the present invention, a lithium battery power prediction device is provided, comprising: an acquisition module, used to acquire a lithium battery simulation model, wherein the lithium battery simulation model is obtained by modeling a lithium-ion battery using a second-order capacitor-resistor equivalent circuit model; a first determination module, used to determine a state equation of the lithium battery simulation model, and an observation equation of the lithium battery simulation model, wherein the state equation is an equation that reflects the change of the lithium battery power and the lithium battery polarization voltage over time, and the observation equation is an equation that reflects the relationship between the lithium battery terminal voltage and a predetermined parameter, wherein the predetermined parameter includes the lithium battery power and the lithium battery polarization voltage; a second determination module, used to determine a state equation of the lithium battery simulation model, and an observation equation of the lithium battery simulation model, wherein the state equation is an equation that reflects the change of the lithium battery power and the lithium battery polarization voltage over time, and the observation equation is an equation that reflects the relationship between the lithium battery terminal voltage and a predetermined parameter, and the predetermined parameter includes the lithium battery power and the lithium battery polarization voltage; A module is used to determine the first power estimation values ​​corresponding to the lithium battery simulation model at multiple moments based on the actual input, the state equation and the observation equation; a third determination module is used to input the actual input and the lithium battery simulation model into a power prediction model to obtain the second power estimation values ​​corresponding to the lithium battery simulation model at the multiple moments, wherein the power prediction model is trained based on sample data; a fourth determination module is used to determine the target power estimation values ​​corresponding to the multiple moments corresponding to the lithium-ion battery based on the first power estimation values ​​and the second power estimation values ​​corresponding to the multiple moments.

[0014] According to one aspect of an embodiment of the present invention, an electronic device is provided, comprising: a processor; and a memory for storing instructions executable by the processor; wherein the processor is configured to execute the instructions to implement any of the above-mentioned lithium battery power prediction methods.

[0015] According to one aspect of an embodiment of the present invention, a computer-readable storage medium is provided. When instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device can execute any of the above-mentioned lithium battery power prediction methods.

[0016] In an embodiment of the present invention, a lithium battery simulation model is obtained, wherein the lithium battery simulation model is obtained by modeling a lithium ion battery using a second-order capacitor-resistor equivalent circuit model. The state equation of the lithium battery simulation model and the observation equation of the lithium battery simulation model are determined, wherein the state equation is an equation that reflects the change of the lithium battery charge and the lithium battery polarization voltage over time, and the observation equation is an equation that reflects the relationship between the lithium battery terminal voltage and the predetermined parameters, and the predetermined parameters include the lithium battery charge and the lithium battery polarization voltage. Based on the actual input, the state equation and the observation equation, the first charge estimation value corresponding to the lithium battery simulation model at multiple moments is determined. The actual input and the lithium battery simulation model are input into the charge prediction model to obtain the second charge estimation value corresponding to the lithium battery simulation model at multiple moments, wherein the charge prediction model is obtained by training based on sample data, and the charge prediction model is a neural network model. According to the first charge estimation value and the second charge estimation value corresponding to multiple moments, the target charge estimation value corresponding to the lithium ion battery at multiple moments is determined. It can be seen that the embodiment of the present invention obtains corresponding estimation values ​​through state estimation operation and neural network model operation, respectively, thereby taking into account the characteristics of the two methods, and comprehensively determining the target power estimation values ​​corresponding to multiple moments corresponding to the lithium-ion battery, thereby solving the technical problem of inaccurate battery power prediction caused by relatively single factors when predicting battery power in related technologies. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0018] Figure 1 is a flow chart of a method for predicting power of a lithium battery according to an embodiment of the present invention;

[0019] Figure 2 It is a flow chart of a method for estimating SOC of a lithium-ion battery in an energy storage system based on multi-attribute fusion provided in an optional embodiment of the present invention;

[0020] Figure 3 It is a second-order RC equivalent circuit model of a lithium battery provided by an optional embodiment of the present invention;

[0021] Figure 44 is a structural block diagram of a lithium battery power prediction device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0022] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme 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 only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0023] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0024] First, some nouns or terms that appear in the description of the embodiments of the present application are subject to the following explanations:

[0025] KF algorithm: KF algorithm is an efficient recursive filter (autoregressive filter) proposed by Hungarian-American mathematician Rudolf Emil Kalman in 1960. It can estimate the state of a dynamic system from a series of measurements containing statistical noise.

[0026] EKF algorithm: It is an extension of KF algorithm and is used to process nonlinear systems. It introduces extension terms in the linearization process of KF algorithm to effectively estimate the state of nonlinear systems.

[0027] BP neural network, full name Backpropagation Neural Network, is a common and widely used artificial neural network model. It is trained based on the error back propagation algorithm to solve various problems such as classification, regression, pattern recognition, data mining, etc.

[0028] Example 1

[0029] According to an embodiment of the present invention, an embodiment of a lithium battery power prediction method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0030] Figure 1 is a flow chart of a method for predicting the power of a lithium battery according to an embodiment of the present invention. Figure 1 As shown, the method comprises the following steps:

[0031] Step S102, obtaining a lithium battery simulation model, wherein the lithium battery simulation model is obtained by modeling a lithium-ion battery using a second-order capacitor-resistance equivalent circuit model;

[0032] In step S102 provided in the present application, a lithium battery simulation model is obtained. The simulation model involved in the embodiment of the present invention is created by modeling the lithium-ion battery using a second-order capacitor-resistor (RC) equivalent circuit model. This modeling method is achieved by simplifying the complex electrochemical reaction and charge transfer process inside the lithium battery into a circuit network composed of resistors and capacitors. The second-order RC model consists of two resistors (R1, R2) and two capacitors (C1, C2). In this model, the resistor represents various resistances inside the battery, such as the ion transfer resistance in the electrolyte, the contact resistance between the battery plates, etc.; while the capacitor represents the charge storage capacity inside the battery, as well as the charge accumulation and release in the electrochemical process. R1 and C1 are connected in series to simulate the main electrochemical process and charge storage capacity of the battery. R1 represents the main resistance inside the battery, and C1 represents the electrochemical capacitance of the battery, reflecting the voltage change of the battery during the charge and discharge process. The second RC link (R2 and C2) is used to more accurately simulate the charge transfer and dissipation process inside the battery, such as electrolyte dissipation, capacitance effect between the plates, etc. When modeling, you can first determine the basic parameters of the lithium battery, such as open circuit voltage, internal resistance, capacity, etc., and then set the resistance and capacitance values ​​in the circuit according to these parameters. Next, by adjusting the circuit parameters and simulation conditions, you can simulate the charging and discharging process of the battery under different working conditions and observe the dynamic changes of its parameters such as voltage, current and power.

[0033] Step S104, determining a state equation of the lithium battery simulation model and an observation equation of the lithium battery simulation model, wherein the state equation is an equation that reflects the change of the lithium battery power and the lithium battery polarization voltage over time, and the observation equation is an equation that reflects the relationship between the lithium battery terminal voltage and predetermined parameters, and the predetermined parameters include the lithium battery power and the lithium battery polarization voltage;

[0034] In step S104 provided in the present application, the state equation of the lithium battery simulation model and the observation equation of the lithium battery simulation model are determined.

[0035] Among them, the state equation describes the law of change of the internal state of the lithium battery (such as charge and polarization voltage) over time. The charge usually refers to the remaining capacity of the battery (SOC, State of Charge), while the polarization voltage reflects the additional voltage drop caused by electrochemical polarization, concentration polarization and other phenomena during the charge and discharge process of the battery. The change in charge (SOC) is usually related to the actual input of the battery, follows a certain integral relationship, and takes into account the efficiency loss and capacity decay of the battery. The change in polarization voltage may be determined by multiple factors such as the battery's charge and discharge rate, temperature, and battery aging, and can usually be described by empirical formulas or more complex electrochemical models.

[0036] Among them, the observation equation establishes the relationship between the external observables of the lithium battery (such as terminal voltage) and the internal state (such as charge and polarization voltage). In the lithium battery model, the terminal voltage is an important indicator of battery performance, which directly affects the battery's performance and safety.

[0037] In order to more accurately describe the dynamic behavior of lithium batteries, the state equation and observation equation may need to adopt nonlinear forms to take into account the nonlinear characteristics of the battery at different SOC, temperature and charge and discharge rates.

[0038] It should be noted that in order to more comprehensively describe the behavior of lithium batteries, a multi-time scale model can also be introduced to consider the battery's fast dynamic response (such as rapid changes in polarization voltage) and slow dynamic response (such as long-term capacity decay) respectively.

[0039] Through this setting, the performance of lithium batteries under different operating conditions, such as remaining capacity, charge and discharge time, voltage fluctuation, etc., can be more accurately predicted through precise state equations and observation equations. Based on the prediction results of the model, battery management strategies such as charge and discharge control, balance control, thermal management, etc. can be optimized to improve the overall performance and safety of the battery system. In addition, by comparing the actual observation data with the model prediction results, abnormal battery behavior can be discovered in time, such as too fast capacity decay, increased internal resistance, severe polarization, etc., so as to perform fault diagnosis and early warning.

[0040] Step S106, determining first power estimation values ​​corresponding to the lithium battery simulation model at multiple times based on the actual input quantity, the state equation and the observation equation;

[0041] In step S106 provided in the present application, based on the actual input quantity, the state equation and the observation equation, the first power estimation value corresponding to each of the lithium battery simulation model at multiple moments is determined. Wherein, the actual input quantity refers to the actual data generated during the use of the lithium battery in a real environment, such as charging and discharging current, voltage, temperature, etc. These data reflect the actual working state of the lithium battery. Optionally, the state equation can be used to calculate the battery state estimation value at the current moment according to the battery state at the previous moment and the actual input quantity at the current moment. Substitute the actual input quantity observed at the current moment, such as the battery parameter (such as voltage), into the observation equation to obtain the observed value of the battery state. Compare the observed value with the state estimation value, and correct the state estimation value according to the difference between the two. This correction process can be implemented by algorithms such as Kalman filtering and particle filtering, which can comprehensively consider the measurement noise and system noise to improve the accuracy of state estimation. The corrected state estimation value is used as the initial value of the next moment, and the above process is repeated until the simulation ends. At each simulation moment, the first power estimation value of the corresponding battery state (such as SOC) can be obtained. By combining the state equation and the observation equation and collecting battery parameters in real time for correction, the accuracy of battery state estimation can be significantly improved.

[0042] Step S108, inputting the actual input amount and the lithium battery simulation model into the power prediction model to obtain the second power estimation values ​​corresponding to the lithium battery simulation model at multiple moments, wherein the power prediction model is obtained by training based on the sample data, and the power prediction model is a neural network model;

[0043] In step S108 provided in the present application, the actual input amount and the lithium battery simulation model are input into the power prediction model, and the second power estimation values ​​corresponding to the lithium battery simulation model at multiple moments are obtained. The above content describes a process of using the actual input amount and the output of the lithium battery simulation model, combined with a neural network model (i.e., power prediction model) that has been trained with sample data, to predict the second power estimation values ​​corresponding to the lithium battery at multiple moments. This process combines physical simulation with data-driven methods, which can improve the accuracy and robustness of power prediction. As the lithium battery is used for an increasing period of time and the operating conditions change, its performance will also change. The power prediction model can also be updated and adapted online to adapt to these changes, thereby making better predictions. The relationship between the characteristics of the lithium battery's use time, operating conditions, etc. and the results can be taken into account, and a second power estimation value taking this aspect into account can be obtained.

[0044] It should be noted that due to the large computing power of the neural network model, in addition to basic data such as current, voltage and temperature, data from other sources (such as battery aging data, ambient temperature change data, etc.) can also be considered to be included in the power prediction model to improve the comprehensiveness and accuracy of the prediction.

[0045] Step S110 , determining target power estimation values ​​corresponding to the lithium-ion battery at the multiple moments according to the first power estimation values ​​and the second power estimation values ​​corresponding to the multiple moments.

[0046] In step S110 provided in the present application, the state equation and the observation equation are based on the physical characteristics and electrochemical principles of the battery, and can more accurately describe the changes in the internal state of the battery. The neural network model has strong generalization ability and adaptability, and can handle various complex working conditions and noise interference. When the battery is in extreme conditions or affected by external factors, the neural network model can provide a more stable power estimate. The state equation and observation equation focus on the physical characteristics and dynamic change process of the battery, and can provide a more accurate prediction of the power change trend. The neural network model focuses on the nonlinear processing and pattern recognition capabilities of data, and can capture subtle features and abnormal phenomena in the changes in battery power. Combining the two can achieve complementary advantages and improve the comprehensiveness and accuracy of power estimation.

[0047] Through the above steps S102-S110, a lithium battery simulation model is obtained, wherein the lithium battery simulation model is obtained by modeling the lithium ion battery using a second-order capacitor-resistor equivalent circuit model. Determine the state equation of the lithium battery simulation model and the observation equation of the lithium battery simulation model, wherein the state equation is an equation that reflects the change of the lithium battery power and the lithium battery polarization voltage over time, and the observation equation is an equation between the lithium battery terminal voltage and the predetermined parameters, and the predetermined parameters include the lithium battery power and the lithium battery polarization voltage. Based on the actual input, the state equation and the observation equation, determine the first power estimation value corresponding to the lithium battery simulation model at multiple moments. Input the actual input and the lithium battery simulation model into the power prediction model to obtain the second power estimation value corresponding to the lithium battery simulation model at multiple moments, wherein the power prediction model is obtained by training based on sample data, and the power prediction model is a neural network model. According to the first power estimation value and the second power estimation value corresponding to multiple moments, determine the target power estimation value corresponding to the lithium ion battery at multiple moments. It can be seen that the embodiment of the present invention obtains corresponding estimation values ​​through state estimation operation and neural network model operation, respectively, thereby taking into account the characteristics of the two methods, and comprehensively determining the target power estimation values ​​corresponding to multiple moments corresponding to the lithium-ion battery, thereby solving the technical problem of inaccurate battery power prediction caused by relatively single factors when predicting battery power in related technologies.

[0048] As an optional embodiment, obtaining a lithium battery simulation model includes: obtaining a lithium-ion battery; performing a mixed pulse power characteristic experiment on the lithium-ion battery to determine the relationship between the open circuit voltage, power and time of the lithium-ion battery; determining equivalent circuit model parameters based on the relationship, wherein the equivalent circuit model parameters include at least: battery open circuit voltage, equivalent ohmic internal resistance, equivalent resistance generated by electrochemical polarization, equivalent capacitance generated by electrochemical polarization, equivalent resistance generated by concentration polarization, equivalent capacitance generated by concentration polarization; determining the lithium battery simulation model based on the equivalent circuit model parameters.

[0049] In this embodiment, a lithium-ion battery is obtained, and a representative lithium-ion battery sample is selected or prepared to ensure the universality and accuracy of the experimental results. A hybrid pulse power characteristic experiment HPPC is performed on the lithium-ion battery. This experiment applies a series of mixed pulse currents (including charging and discharging pulses) to the lithium-ion battery to observe and record the voltage response of the battery under different current excitations. These pulse currents simulate the dynamic working conditions of the battery in actual use, which helps to reveal the electrochemical process inside the battery. By analyzing the experimental data, the law of the change of the open circuit voltage (OCV) of the battery over time (or power) can be extracted, and the experimental curves of the HPPC experiment and the single pulse are obtained. The parameters of the second-order RC equivalent circuit model of the lithium battery are identified offline by fitting using the MATLAB toolbox. That is, the equivalent parameters are determined, including the open circuit voltage (OCV) U of the lithium battery. OC ; Lithium battery equivalent ohmic internal resistance R0; equivalent resistance, capacitance R1, C1 generated by lithium battery electrochemical polarization; equivalent resistance, capacitance R2, C2 generated by concentration polarization. Then a lithium battery simulation model is constructed. When conducting mixed pulse power characteristic experiments, the experimental design can be further optimized, such as adjusting the amplitude, duration, and interval of the pulse current, so as to more comprehensively cover the working range of the battery and improve the reliability of the data.

[0050] It should be noted that in order to simulate the performance of the battery more accurately, it is possible to consider coupling multiple physical fields such as electrochemistry, thermal and mechanical fields for modeling. This can not only more comprehensively reflect the complex processes inside the battery, but also predict the battery's performance parameters such as temperature distribution and stress changes under different conditions.

[0051] In this way, by establishing an accurate lithium battery simulation model, the performance of the battery under different operating conditions can be predicted more accurately.

[0052] As an optional embodiment, based on the first power estimation value and the second power estimation value corresponding to the multiple moments respectively corresponding to the lithium-ion battery, the target power estimation value corresponding to the multiple moments respectively corresponding to the lithium-ion battery is determined, including: constructing a decision matrix based on the first power estimation value and the second power estimation value corresponding to the multiple moments respectively; determining a first weight corresponding to the first method and a second weight corresponding to the second method, wherein the first method is a method of determining the power estimation value based on the state equation and the observation equation, and the second method is a method of determining the power estimation value based on the lithium battery simulation model; determining the target power estimation value corresponding to the multiple moments respectively corresponding to the lithium-ion battery based on the decision matrix, the first weight and the second weight.

[0053] In this embodiment, the process of determining the target power estimation values ​​corresponding to multiple moments corresponding to the lithium-ion battery is described. First, according to the first power estimation values ​​and the second power estimation values ​​corresponding to the multiple moments, a decision matrix is ​​constructed. In this scenario, each column of the decision matrix represents a moment, and each row corresponds to the first power estimation value based on the physical model and the second power estimation value based on the neural network model at the moment. Among them, the first weight and the second weight are also determined. The first weight is assigned to the power estimation method based on the state equation and the observation equation, reflecting the accuracy and reliability of the method in this scenario. The second weight is assigned to the power estimation method based on the prediction of the neural network model, reflecting the adaptability and learning ability of the method in this application. The objective weights of the two methods can be calculated by the entropy weight method. The entropy weight method is an objective weighting method based on the information amount of the data itself. It determines the weight by calculating the entropy value and information utility value of each indicator, avoiding the subjective interference of human factors. In power estimation, this means that whether it is a power estimation value based on a physical model or a neural network model, its weight is determined according to the variability of the data itself, thereby ensuring the objectivity and fairness of the weight. Thereby, a more accurate target power estimation value corresponding to multiple moments corresponding to the lithium-ion battery is obtained.

[0054] As an optional embodiment, according to the decision matrix, the first weight and the second weight, the target power estimation values ​​corresponding to the multiple moments corresponding to the lithium-ion battery are determined, including: according to the decision matrix, the first weight and the second weight, an ideal decision matrix is ​​determined; according to the ideal decision matrix, the Euclidean distance and the grey correlation coefficient are determined; according to the Euclidean distance and the grey correlation coefficient, the target power estimation values ​​corresponding to the multiple moments corresponding to the lithium-ion battery are determined.

[0055] In this embodiment, a specific method for determining the target power estimation values ​​corresponding to multiple moments corresponding to the lithium-ion battery is described. After constructing the decision matrix, an ideal decision matrix can be constructed, and the ideal decision matrix is ​​a matrix including positive ideal decision schemes and negative ideal decision schemes. Among them, the positive ideal decision scheme is a solution or scheme in which all evaluation indicators reach the optimal level. The positive ideal solution represents the "perfect" state among all possible schemes and is the goal pursued by the decision maker. The negative ideal decision scheme is a solution or scheme in which all evaluation indicators reach the worst level. In contrast to the positive ideal solution, the negative ideal solution represents the "worst" state among all possible schemes, which is the state that the decision maker wants to avoid. By determining the optimal value and the worst value on each evaluation indicator, the positive ideal solution and the negative ideal solution are constructed. After that, the Euclidean distance and the gray correlation coefficient can be determined based on the ideal decision matrix. That is, an appropriate distance measurement method or similarity measurement method can be used to calculate the distance or similarity between each alternative solution and the positive ideal solution and the negative ideal solution. Then, based on the Euclidean distance and the gray correlation coefficient, the target power estimation values ​​corresponding to the multiple moments corresponding to the lithium-ion battery are determined. The alternative solutions are sorted according to the calculation results of the distance or similarity. Generally, the solution closest to the positive ideal solution and farthest from the negative ideal solution is considered the optimal solution or the preferred solution. This helps determine the target power estimation values ​​corresponding to multiple moments corresponding to the lithium-ion battery.

[0056] As an optional embodiment, determining the state equation of a lithium battery simulation model and the observation equation of the lithium battery simulation model includes: determining a predetermined equivalent circuit model expression equation based on the lithium battery simulation model, and determining a definition equation for the lithium battery power; using a first discrete parameter to discretize the second-order capacitor-resistor equivalent circuit model expression equation and the definition equation for the lithium battery power to determine the state equation, and using a second discrete parameter to discretize the second-order capacitor-resistor equivalent circuit model expression equation and the definition equation for the lithium battery power to determine the observation equation.

[0057] In this embodiment, the process of determining the observation equation is described. First, a predetermined equivalent circuit model expression equation can be determined based on the lithium battery simulation model. The equivalent circuit model simulates the behavior of the battery by simplifying the internal characteristics of the battery into a combination of a series of circuit elements (such as resistors, capacitors, etc.). The definition equation of the lithium battery power is also determined. The state equation is determined by discretizing the second-order capacitor-resistor equivalent circuit model expression equation and the definition equation of the lithium battery power using the first discrete parameter, and the second-order capacitor-resistor equivalent circuit model expression equation and the definition equation of the lithium battery power are discretized using the second discrete parameter to determine the observation equation. Among them, the first and second discretization parameters are parameters that convert the state equation of the continuous time system into a discrete time form. The discretized state equation describes the change of the battery state (such as capacitor voltage) at discrete time points. In the discretized second-order capacitor-resistor equivalent circuit model, the observation equation will involve the measured value of the battery voltage and the discrete value of the circuit state variable (such as capacitor voltage) related thereto. Discretization processing helps to reduce the impact of noise and interference in the digital system on the model accuracy, thereby improving the stability and reliability of the system. In addition, the discretized model provides a basis for implementing complex control algorithms and state estimation algorithms in digital systems, such as Kalman filtering, particle filtering, etc. Compared with the model in the continuous time domain, the discretized model is computationally simpler and more efficient.

[0058] As an optional embodiment, the actual input quantity and the lithium battery simulation model are input into the power prediction model to obtain the second power estimation values ​​corresponding to the lithium battery simulation model at multiple moments, including: inputting the actual input quantity and the lithium battery simulation model into the hidden layer of the power prediction model to obtain the hidden layer output; inputting the hidden layer output into the output layer of the power prediction model to obtain the output layer output; inputting the output layer output into the neural network output error layer of the power prediction model to obtain the output error; adjusting the output layer output according to the output error to obtain the second power estimation values ​​corresponding to the lithium battery simulation model at multiple moments.

[0059] In this embodiment, the process of determining the second power estimate is described. In this process, the actual input quantity (such as the current, voltage, temperature, etc. of the lithium battery) and the lithium battery simulation model are used as inputs and sent to the hidden layer of the power prediction model. The hidden layer is a key part of the neural network. It is responsible for learning the complex relationship between input and output and extracting features through a series of nonlinear transformations. The output of the hidden layer is then sent to the output layer, which generates the final prediction result based on these features, that is, the power estimate of the lithium battery at a specific moment. This estimate represents the model's judgment on the current power state of the lithium battery. In order to evaluate the accuracy of the prediction, the output of the output layer (i.e., the power estimate) will be sent to the neural network output error layer. In this layer, the power estimate is compared with a reliable reference value to calculate the output error. The output error is an important indicator for measuring the prediction performance of the model, which reflects the difference between the model prediction value and the actual value. Based on this output error, the neural network will adjust its internal parameters (such as weights and biases) to optimize the prediction performance of the model. This process is usually implemented through the back-propagation algorithm, which uses the output error to update the weights and biases of the neural network so that the model's predicted values ​​are closer to the actual values.

[0060] As an optional embodiment, based on actual input quantity, state equation and observation equation, the first power estimation value corresponding to the lithium battery simulation model at multiple moments is determined, including: determining the system state equation and the error covariance matrix equation; determining the predicted value of the state variable at the second moment according to the system state equation and the estimated value of the state variable at the first moment, and determining the predicted value of the error covariance matrix at the second moment according to the error covariance matrix equation and the estimated value of the error covariance matrix at the first moment, wherein the first moment is the moment before the second moment; determining the Kalman filter gain matrix according to the predicted value of the state variable at the second moment and the predicted value of the error covariance matrix at the second moment; updating the estimated value of the state variable at the first moment according to the Kalman filter gain matrix to obtain the updated state variable estimated value, and updating the estimated value of the error covariance matrix at the first moment according to the Kalman filter gain matrix to obtain the updated error covariance matrix estimated value; determining the first power estimation value corresponding to the lithium battery simulation model at multiple moments according to the updated state variable estimated value and the updated error covariance matrix estimated value.

[0061] In this embodiment, the process of determining the first power estimate is described. The application process of Kalman filtering in the power prediction of the lithium battery simulation model is described in detail. Kalman filtering is an efficient recursive filter (autoregressive filter) that can estimate the state of a dynamic system from a series of measurements containing statistical noise. Among them, the system state equation is involved in this step, which describes how the system state changes over time, including the dynamic characteristics of the system itself and the influence of external inputs. It also involves the error covariance matrix equation, which describes how the uncertainty of the state estimate propagates over time. Therefore, in this step, the value of the state variable at the second moment (current moment) is predicted according to the system state equation and the estimated value of the state variable at the first moment (previous moment). At the same time, the value of the error covariance matrix at the second moment is predicted according to the error covariance matrix equation and the estimated value of the error covariance matrix at the first moment. Using the measured value and the predicted state variable value at the second moment, the estimated value of the state variable is updated through the Kalman gain matrix to obtain a more accurate state estimate. At the same time, the estimated value of the error covariance matrix is ​​also updated to reflect the uncertainty of the updated state estimate. Based on the updated estimated values ​​of the state variables, the corresponding estimated values ​​of the lithium battery at multiple times can be determined. The Kalman filter optimizes the state estimation by fusing the system model and the measurement data, thereby improving the accuracy of the prediction.

[0062] Based on the above embodiments and optional embodiments, an optional implementation is provided, which is described in detail below.

[0063] An optional implementation of the present invention provides a battery power prediction method, which can also be called a SOC estimation method for lithium-ion batteries in energy storage systems based on multi-attribute fusion. Figure 2 : is a flow chart of a method for estimating SOC of a lithium-ion battery in an energy storage system based on multi-attribute fusion according to an optional embodiment of the present invention, such as Figure 2 As shown, two of the attributes are extended Kalman filtering and BP neural network, and the estimation method includes: using a second-order RC equivalent circuit model to model the lithium battery, and performing parameter identification on the lithium battery model; constructing the state equation and observation equation of the battery model, and using the EKF algorithm to determine the SOC value of the battery model; training the BP neural network model for SOC estimation, and inputting current and voltage data to obtain the predicted value of the model; after obtaining the predicted values ​​of the two methods, using the entropy weight method to calculate the weights of the two methods, and then using the TOPSIS method and grey correlation analysis to calculate the relative closeness to obtain the final SOC estimation value. The optional implementation method of the present invention combines the advantages of the extended Kalman filter and the BP neural network methods to ensure the SOC estimation accuracy of the lithium-ion battery of the energy storage system. It is introduced below:

[0064] A method for estimating SOC of a lithium-ion battery of an energy storage system based on multi-attribute fusion comprises the following steps:

[0065] S1: Use the second-order RC equivalent circuit model to model the lithium-ion battery and identify the parameters of the model;

[0066] Specifically, in step S1, a second-order RC equivalent circuit model is selected as the lithium battery model. The second-order RC equivalent circuit model of the lithium battery includes the open circuit voltage (OCV) U OC ; Lithium battery equivalent ohmic internal resistance R0; equivalent resistance and capacitance R1, C1 generated by lithium battery electrochemical polarization; equivalent resistance and capacitance R2, C2 generated by concentration polarization;

[0067] Figure 3 is a second-order RC equivalent circuit model of a lithium battery provided by an optional embodiment of the present invention. The second-order RC equivalent circuit model of a lithium battery is as follows: Figure 3 As shown. OC is the open circuit voltage (OCV) of the lithium battery; R0 is the equivalent ohmic internal resistance of the lithium battery; R1 and C1 are the equivalent resistance and capacitance respectively generated by the electrochemical polarization of the lithium battery; R2 and C2 are the equivalent resistance and capacitance respectively generated by the concentration polarization of the lithium battery; I is the charge and discharge current (negative when charging and positive when discharging); U1 and U2 are the polarization voltages of the lithium battery; UL is the terminal voltage of the lithium battery.

[0068] After the model is built, the parameter identification results of the model are determined. The purpose of the experiment is to determine the U OC The relationship between the HPPC experiment and the SOC is obtained, and the experimental curves of the single pulse are fitted using the MATLAB toolbox to identify the parameters of the second-order RC equivalent circuit model of the lithium battery offline.

[0069] S2: constructing the state equation and observation equation of the battery model, and using the improved EKF algorithm to determine the SOC value of the model;

[0070] Optionally, in step S2, the state equation and observation equation of the battery model are constructed, and the specific steps of using the EKF algorithm to determine the SOC value of the model are as follows: discretize the definition equation of the lithium battery SOC and the electrical characteristic expression equation of the second-order RC equivalent circuit model to obtain the state equation and observation equation in the EKF algorithm, and then use the EKF algorithm to calculate the model SOC.

[0071] (1) Initial values ​​of x(0), P(0), Q, and R are assigned to the system, where Q and R are obtained based on engineering experience;

[0072] (2) Status prediction;

[0073] (3) Error covariance matrix prediction;

[0074] (4) Kalman filter gain matrix;

[0075] (5) Update of state estimate;

[0076] (6) Error covariance matrix update.

[0077] Finally, repeat steps 2 and 6 to recursively calculate the estimated SOC value of the lithium battery.

[0078] Specifically, according to the definition equation of the lithium battery SOC in step 1, the definition equation is SOC=QC / Q1×100%.

[0079] Where: QC represents the remaining power of the lithium battery (Ah); Q1 represents the rated capacity of the lithium battery (Ah).

[0080] And, the electrical characteristic expression equation of the second-order RC equivalent circuit model is discretized.

[0081] Among them, the electrical characteristic expression equation of the second-order RC equivalent circuit model is:

[0082]

[0083] The battery's open circuit voltage Uocv, ohmic internal resistance R0, polarized internal resistances R1, R2, and polarized capacitors C1, C2. The polarized internal resistance R1 and polarized capacitor C1 are connected in parallel, and their terminal voltage is U1; the polarized internal resistance R2 and polarized capacitor C2 are connected in parallel, and their terminal voltage is U2. The ohmic internal resistance R0, the equivalent circuit with the terminal voltage U1, and the equivalent circuit with the terminal voltage U1 are connected in series to the positive electrode of the battery's open circuit voltage Uocv. The terminal voltage between the positive output terminal of the series RC equivalent circuit and the negative output terminal of the battery's open circuit voltage Uocv is U, and the current passing through the circuit is I;

[0084] After discretization, the state equation and observation equation in the EKF algorithm are obtained as follows:

[0085]

[0086] U L (k) = U ocv [SOC(k)]-U1(k)-U2(k)-R0*I(k)+v k

[0087] In the formula, τ1=R1C1 is the time constant, and τ2=R2C2 is also the time constant, which reflects how fast the lithium battery reaches a steady state after charging and discharging. W1(k), W2(k), and W3(k) represent the random disturbance or noise of the system at different times.

[0088] The state equation and observation equation of the lithium battery nonlinear system can be described as:

[0089]

[0090] In the formula, x(k) represents the state vector at time k; x(k+1) represents the state vector at time k+1, y(k) represents the observation at time k; f(x(k),u(k)) is the state function; h(x(k),u(k)) is the observation function; u(k) is the input vector at time k; w(k) is the process noise at time k, and its variance is Q; v(k) is the observation noise at time k, and its variance is R.

[0091] Where x(k) = [SOC(k), U1(k), U2(k)] T ; y(k)=U L (k); u(k) = I(k);

[0092] Based on the above state equation and observation equation, the specific steps of EKF algorithm to calculate SOC are as follows:

[0093] (1) The system’s x(0), P(0), Q, and R are assigned initial values, where Q and R are obtained based on engineering experience.

[0094] (2) Status prediction:

[0095]

[0096] is the predicted value of the state variable at time k+1; is the estimated value of the state variable at time k; A(k) and B(k) are the coefficient matrices of the system state equation at time k.

[0097] in,

[0098]

[0099] (3) Error covariance matrix prediction:

[0100]

[0101] is the predicted value of the error covariance matrix at time k+1; is the estimated value of the error covariance matrix at time k;

[0102] (4) Kalman filter gain matrix:

[0103]

[0104] K(k+1) is the Kalman gain matrix at time k+1; C(k) is the coefficient matrix of the system observation equation at time k.

[0105] in,

[0106]

[0107] (5) Update of state estimate:

[0108]

[0109] Where: is the estimated value of the state variable at time k+1; y(k) is the battery U at time k L ; D(k) is the coefficient matrix of the system observation equation at time k.

[0110] (6) Error covariance matrix update:

[0111]

[0112] In the formula, is the positive estimate of the error covariance at time k+1.

[0113] Finally, repeat steps 2 and 6 to recursively calculate the estimated SOC value of the lithium battery.

[0114] S3: training the SOC estimation model of the improved BP neural network model, and inputting current and voltage data to obtain the SOC value of the model;

[0115] Optionally, in step S3, the specific steps of performing SOC estimation on the BP neural network model are as follows:

[0116] (1) Neural network initialization. When constructing a neural network model, the number of neurons in the input layer, hidden layer, and output layer of the neural network is determined based on the output variables of the system. Then, the connection weights between the input layer, hidden layer, and output layer are initialized using a random initialization method, and the activation function between each neuron and the network learning rate is determined.

[0117] (2) Calculate the hidden layer output (H j ).

[0118] (3) Calculate the output layer output (O k ).

[0119] (4) Calculate the output layer output (e k ).

[0120] (6) Iterative training. Determine whether the BP neural network training is completed. If not, return to (2) to continue iterative training until the training setting conditions are met and the neural network training is terminated.

[0121] (7) After the BP neural network SOC estimation model is trained, the input current is used to obtain the model's predicted value.

[0122] In detail, in step S3, the specific steps of performing SOC estimation on the BP neural network model are as follows:

[0123] (1) Neural network initialization. When constructing a neural network model, the number of neurons in the input layer (n), hidden layer (l), and output layer (m) of the neural network is determined based on the output variables of the system. Then, the connection weights between the input layer, hidden layer, and output layer are initialized using a random initialization method, and the activation function between each neuron and the network learning rate is determined.

[0124] (2) Calculate the hidden layer output (H j ). The hidden layer output calculation formula is as follows:

[0125]

[0126] Among them, w ij is the connection weight between the input layer and the hidden layer, f is the activation function between the neuron and the network learning rate, x i is the input vector, x i is the threshold of the hidden layer neurons, a i is the threshold of the hidden layer neurons. Common activation functions include Purelin function, Tanh function, Sigmoid function, etc.

[0127] (3) Calculate the output layer output (O k ). The output calculation formula of the output layer is as follows:

[0128]

[0129] Among them, b k is the threshold value, w jk is the connection weight between the hidden layer and the output layer, that is, the weight by which each unit in the previous network layer is connected to each unit in the next network layer, and the weight is generally initialized to a random decimal in [-0.1, 0.1].

[0130] (4) Calculate the output layer output (e k ). The output calculation formula of the output layer is as follows:

[0131] e k =Y k -O k

[0132] Among them, O k is the predicted output of the neural network, Y k is the expected output of the neural network.

[0133] (5) Error back propagation to correct the threshold. According to the output error, the threshold is corrected using the gradient descent method.

[0134]

[0135] w jk =w jk +ηH j e k (20)

[0136]

[0137] b k =b k +e k (27)

[0138] Among them, η is the learning rate of the neural network, and x(i) is the input sample.

[0139] (6) Iterative training. Determine whether the BP neural network training is completed. If not, return to (2) to continue iterative training until the training setting conditions are met and the neural network training is terminated.

[0140] (7) After the BP neural network SOC estimation model is trained, the current, voltage and temperature data are input to obtain the model's predicted value.

[0141] S4: After obtaining the predicted values ​​of the two methods, the weights of the two methods are calculated according to the entropy weight method, and the relative closeness is calculated using the TOPSIS method and grey correlation analysis to obtain the final SOC estimation value.

[0142] Optionally, in step S4, the specific steps of obtaining the final SOC estimation value by using the multi-attribute fusion method are as follows:

[0143] (1) Construct a decision matrix:

[0144] The two methods are regarded as different attributes, and the obtained SOC value is used as a single indicator to form a decision matrix:

[0145]

[0146] Where, SOC1(k1) is the value obtained by the EKF algorithm at time k1; SOC2(k2) is the value obtained by the BP neural network algorithm at time k1;

[0147] (2) Entropy weight method calculates the objective weights of the two methods:

[0148] First, the entropy of the jth method is calculated as follows:

[0149]

[0150]

[0151] where e j is the entropy of the jth method, p ij is the weight of the jth method at the i-th moment in this method, and finally the weights are obtained as

[0152]

[0153] Among them, 1-e j is the information entropy redundancy, and the weight is determined in this way.

[0154] (3) Relative closeness calculation:

[0155] According to the matrix (y ij )=(w j r ij ) Determine the positive ideal decision plan and the negative ideal decision plan, and then calculate the Euclidean distance of each plan to the positive ideal plan and the negative ideal plan according to the following formula:

[0156]

[0157] Based on the weighted standardization matrix, the grey correlation coefficient between the i-th sample and the positive ideal sample with respect to the j-th method is calculated:

[0158] Calculate the grey correlation coefficient between the i-th sample and the negative ideal sample with respect to the j-th method:

[0159]

[0160] Where: Δy ij is the minimum difference between the two levels; is the maximum difference between the two levels; is the resolution coefficient, which is 0.5. Then the grey correlation coefficient matrix of each sample and the positive ideal sample is

[0161]

[0162] The grey correlation degrees of the i-th sample with the positive ideal sample and the negative ideal sample are:

[0163]

[0164] Considering the Euclidean distance and grey relational degree comprehensively, we get:

[0165]

[0166]

[0167] Among them, e1+e2=1 can be set according to your own preferences, and finally the relative closeness is calculated, and the final power evaluation result can be obtained using the following formula.

[0168]

[0169] It should be noted that, for the above-mentioned method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should know that the present invention is not limited by the described action sequence, because according to the present invention, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present invention.

[0170] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, a magnetic disk, or an optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods of various embodiments of the present invention.

[0171] Example 2

[0172] According to an embodiment of the present invention, a device for implementing the above-mentioned lithium battery power prediction method is also provided. Figure 4 is a structural block diagram of a lithium battery power prediction device according to an embodiment of the present invention. Figure 4 As shown, the device includes: an acquisition module 402, a first determination module 404, a second determination module 406, a third determination module 408 and a fourth determination module 410. The device is described in detail below.

[0173] An acquisition module 402 is used to acquire a lithium battery simulation model, wherein the lithium battery simulation model is obtained by modeling a lithium-ion battery using a second-order capacitor-resistance equivalent circuit model;

[0174] A first determination module 404 is connected to the acquisition module 402 and is used to determine a state equation of a lithium battery simulation model and an observation equation of the lithium battery simulation model, wherein the state equation is an equation that reflects the change of the lithium battery power and the lithium battery polarization voltage over time, and the observation equation is an equation that reflects the relationship between the lithium battery terminal voltage and predetermined parameters, and the predetermined parameters include the lithium battery power and the lithium battery polarization voltage;

[0175] A second determination module 406, connected to the first determination module 404, is used to determine first power estimation values ​​corresponding to the lithium battery simulation model at multiple time points based on the actual input quantity, the state equation and the observation equation;

[0176] A third determination module 408, connected to the second determination module 406, is used to input the actual input amount and the lithium battery simulation model into the power prediction model to obtain the second power estimation values ​​corresponding to the lithium battery simulation model at multiple moments, wherein the power prediction model is obtained by training based on the sample data;

[0177] The fourth determination module 410 is connected to the third determination module 408 and is used to determine the target power estimation values ​​corresponding to the lithium-ion battery at multiple moments according to the first power estimation values ​​and the second power estimation values ​​corresponding to the multiple moments.

[0178] It should be noted here that the above-mentioned acquisition module 402, first determination module 404, second determination module 406, third determination module 408 and fourth determination module 410 correspond to steps S102 to S110 in the implementation of the lithium battery power prediction method, and the examples and application scenarios implemented by the multiple modules and the corresponding steps are the same, but are not limited to the contents disclosed in the above-mentioned embodiment 1.

[0179] Example 3

[0180] According to another aspect of an embodiment of the present invention, there is also provided an electronic device, comprising: a processor; and a memory for storing instructions executable by the processor, wherein the processor is configured to execute instructions to implement any of the above-mentioned lithium battery power prediction methods.

[0181] Example 4

[0182] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is provided. When instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device can execute any of the above-mentioned lithium battery power prediction methods.

[0183] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0184] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0185] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0186] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0187] 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. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0188] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for 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 method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.

[0189] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for predicting the power of a lithium battery, characterized in that: include: Acquire a lithium battery simulation model, wherein the lithium battery simulation model is obtained by modeling a lithium-ion battery using a second-order capacitance-resistance equivalent circuit model; Determine the state equation of the lithium battery simulation model and the observation equation of the lithium battery simulation model, wherein the state equation is an equation that reflects the change of the lithium battery power and the lithium battery polarization voltage over time, and the observation equation is an equation that reflects the relationship between the lithium battery terminal voltage and the predetermined parameters, and the predetermined parameters include the lithium battery power and the lithium battery polarization voltage; Based on the actual input quantity, the state equation and the observation equation, determining first power estimation values ​​corresponding to the lithium battery simulation model at multiple time points; Input the actual input amount and the lithium battery simulation model into a power prediction model to obtain second power estimation values ​​corresponding to the lithium battery simulation model at the multiple moments, wherein the power prediction model is obtained by training based on sample data, and the power prediction model is a neural network model; According to the first power estimation values ​​and the second power estimation values ​​respectively corresponding to the multiple moments, target power estimation values ​​respectively corresponding to the multiple moments corresponding to the lithium-ion battery are determined.

2. The method according to claim 1, characterized in that The step of obtaining a lithium battery simulation model comprises: Obtaining the lithium-ion battery; Conducting a mixed pulse power characteristic experiment on the lithium-ion battery to determine the relationship between the open circuit voltage, power and time of the lithium-ion battery; Determine the equivalent circuit model parameters according to the relationship, wherein the equivalent circuit model parameters include at least: battery open circuit voltage, equivalent ohmic internal resistance, equivalent resistance generated by electrochemical polarization, equivalent capacitance generated by electrochemical polarization, equivalent resistance generated by concentration polarization, and equivalent capacitance generated by concentration polarization; The lithium battery simulation model is determined according to the equivalent circuit model parameters.

3. The method according to claim 1, characterized in that The determining, based on the first power estimation values ​​and the second power estimation values ​​respectively corresponding to the multiple moments, the target power estimation values ​​respectively corresponding to the multiple moments corresponding to the lithium-ion battery comprises: Constructing a decision matrix according to the first power estimation values ​​and the second power estimation values ​​respectively corresponding to the multiple moments; Determine a first weight corresponding to a first method and a second weight corresponding to a second method, wherein the first method is a method of determining the power estimation value according to the state equation and the observation equation, and the second method is a method of determining the power estimation value according to the lithium battery simulation model; According to the decision matrix, the first weight and the second weight, the target power estimation values ​​corresponding to the multiple moments corresponding to the lithium-ion battery are determined.

4. The method according to claim 3, characterized in that: The determining, based on the decision matrix, the first weight and the second weight, target power estimation values ​​corresponding to the multiple moments corresponding to the lithium-ion battery, respectively, includes: Determine an ideal decision matrix according to the decision matrix, the first weight and the second weight; According to the ideal decision matrix, determining the Euclidean distance and the grey correlation coefficient; According to the Euclidean distance and the grey correlation coefficient, target power estimation values ​​corresponding to the plurality of moments corresponding to the lithium-ion battery are determined.

5. The method according to claim 1, characterized in that The state equation of the lithium battery simulation model and the observation equation of the lithium battery simulation model are determined, including: Determine a predetermined equivalent circuit model expression equation based on the lithium battery simulation model, and determine a definition equation for the lithium battery power; The state equation is determined by discretizing the second-order capacitor-resistance equivalent circuit model expression equation and the definition equation of the lithium battery power using a first discrete parameter, and the observation equation is determined by discretizing the second-order capacitor-resistance equivalent circuit model expression equation and the definition equation of the lithium battery power using a second discrete parameter.

6. The method according to claim 1, characterized in that The step of inputting the actual input amount and the lithium battery simulation model into a power prediction model to obtain second power estimation values ​​corresponding to the lithium battery simulation model at the multiple time points respectively includes: Inputting the actual input quantity and the lithium battery simulation model into the hidden layer of the power prediction model to obtain the hidden layer output; Inputting the hidden layer output to the output layer of the power prediction model to obtain the output layer output; Inputting the output of the output layer into the neural network output error layer of the power prediction model to obtain an output error; The output layer output is adjusted according to the output error to obtain second power estimation values ​​corresponding to the lithium battery simulation model at the multiple moments.

7. The method according to any one of claims 1 to 6, characterized in that: The determining, based on the actual input quantity, the state equation and the observation equation, first power estimation values ​​corresponding to the lithium battery simulation model at multiple moments respectively includes: Determine the system state equation and error covariance matrix equation; Determine the predicted value of the state variable at the second moment according to the system state equation and the estimated value of the state variable at the first moment, and determine the predicted value of the error covariance matrix at the second moment according to the error covariance matrix equation and the estimated value of the error covariance matrix at the first moment, wherein the first moment is the moment before the second moment; Determining a Kalman filter gain matrix according to the predicted value of the state variable at the second moment and the predicted value of the error covariance matrix at the second moment; According to the Kalman filter gain matrix, the estimated value of the state variable at the first moment is updated to obtain the updated state variable estimated value, and according to the Kalman filter gain matrix, the estimated value of the error covariance matrix at the first moment is updated to obtain the updated error covariance matrix estimated value; According to the updated state variable estimate value and the updated error covariance matrix estimate value, first power estimate values ​​corresponding to the lithium battery simulation model at multiple time points are determined.

8. A lithium battery power prediction device, characterized in that: include: An acquisition module is used to acquire a lithium battery simulation model, wherein the lithium battery simulation model is obtained by modeling a lithium-ion battery using a second-order capacitor-resistance equivalent circuit model; A first determination module is used to determine the state equation of the lithium battery simulation model and the observation equation of the lithium battery simulation model, wherein the state equation is an equation that reflects the change of the lithium battery power and the lithium battery polarization voltage over time, and the observation equation is an equation that reflects the relationship between the lithium battery terminal voltage and the predetermined parameters, and the predetermined parameters include the lithium battery power and the lithium battery polarization voltage; A second determination module is used to determine first power estimation values ​​corresponding to the lithium battery simulation model at multiple time points based on the actual input quantity, the state equation and the observation equation; A third determination module is used to input the actual input amount and the lithium battery simulation model into a power prediction model to obtain second power estimation values ​​corresponding to the lithium battery simulation model at the multiple moments, wherein the power prediction model is trained based on sample data; The fourth determination module is used to determine the target power estimation values ​​corresponding to the multiple moments corresponding to the lithium-ion battery according to the first power estimation values ​​and the second power estimation values ​​corresponding to the multiple moments.

9. An electronic device, characterized in that: include: processor; a memory for storing instructions executable by the processor; The processor is configured to execute the instructions to implement the lithium battery power prediction method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: When the instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device is enabled to execute the lithium battery power prediction method as described in any one of claims 1 to 7.