Solid-state battery peak power estimation method and device based on PMC-GPR method

Abstract

The invention provides a solid-state battery peak power estimation method and device based on a PMC-GPR method, which are applied to a solid-state battery and comprise the following steps: constructing an electromechanical coupling model comprising an equivalent stress model and an equivalent circuit model; inputting the acquired battery operation data into an electromechanical coupling model to obtain a current state including SOC, branch voltage and expansion stress; under the first time scale, on the basis of the current state, estimating the battery capacity on line by adopting an unscented Kalman filtering algorithm; under the second time scale, based on the battery capacity and the current state, an unscented Kalman filtering algorithm is adopted for online estimation, and an estimated SOC is obtained; wherein the first time scale is larger than the second time scale; and constructing feature data based on the estimated SOC, and inputting the feature data into the trained Gaussian process regression model to obtain predicted peak power. According to the scheme, the mechanical effect of the solid-state battery is fully considered, and the precision, robustness and reliability of peak power estimation of the solid-state battery are effectively improved.

Description

A method and apparatus for estimating peak power of solid-state batteries based on the PMC-GPR method. Technical Field

[0001] This invention relates to the field of battery technology, particularly to the field of solid-state battery technology, and especially to a method and apparatus for estimating the peak power of solid-state batteries based on the PMC-GPR method. Background Technology

[0002] Solid-state batteries, as the most promising next-generation lithium battery technology, possess significant advantages in energy density and safety, and also exhibit higher mechanical strength and stability. During the operation of solid-state batteries, the migration of lithium ions causes changes in electrode volume, which manifests as a mechanical effect on a macroscopic scale. In other words, the operation of solid-state batteries is always accompanied by mechanical effects. The core of the Battery Management System (BMS) lies in estimating the battery's state of charge, primarily including State of Charge (SOC), State of Health (SOH), and State of Power (SOP). SOC characterizes the battery's charge level, while SOH reflects changes in battery capacity. High-accuracy estimation of SOC and SOH helps extend battery life and provides support for thermal management. Furthermore, high-accuracy SOC and capacity estimation also guide peak power estimation. SOP is defined as the ratio of peak power to rated power. Peak power characterizes the battery's extreme charge and discharge capabilities and relates to the energy supply under various steady-state and dynamic operating conditions. Accurate power information helps achieve dynamic matching of the battery at different stages, thereby extending the vehicle's driving range.

[0003] For SOC estimation, current mainstream methods include the ampere-hour integration method, the open-circuit voltage method, model-based combined methods, and data-driven methods. The ampere-hour integration method suffers from initial error accumulation; the open-circuit voltage method requires a relatively long period of battery rest and is not suitable for online estimation; model-based and data-driven methods are highly dependent on model accuracy and the quality of training data. For battery usable capacity estimation, existing methods face challenges related to accelerated battery aging, capacity-parameter sensitivity, and computational cost, making it difficult to accurately estimate usable capacity in real time. Most existing battery power estimation methods are based on limitations of battery terminal voltage and SOC, and for solid-state battery systems, their dynamic mechanical characteristics are often ignored. Furthermore, the scales of capacity and battery state parameters change are inconsistent.

[0004] Therefore, there is an urgent need for a method and device for estimating the peak power of solid-state batteries based on the PMC-GPR method. Summary of the Invention

[0005] To address the aforementioned issues, this invention provides a method and apparatus for estimating the peak power of solid-state batteries based on the PMC-GPR method. This method fully considers the mechanical effects of solid-state batteries and effectively improves the accuracy, robustness, and reliability of peak power estimation for solid-state batteries.

[0006] In a first aspect, embodiments of the present invention provide a peak power estimation method for solid-state batteries based on the PMC-GPR method, applied to solid-state batteries, comprising: constructing an electromechanical coupling model including an equivalent stress model and an equivalent circuit model; inputting acquired battery operating data into the electromechanical coupling model to obtain the current state including SOC, branch voltage, and expansion stress; estimating the battery capacity online using an unscented Kalman filter algorithm based on the current state at a first time scale; estimating the estimated SOC online using an unscented Kalman filter algorithm based on the battery capacity and the current state at a second time scale; wherein the first time scale is larger than the second time scale; constructing feature data based on the estimated SOC, and inputting the feature data into a trained Gaussian process regression model to obtain the predicted peak power.

[0007] Secondly, embodiments of the present invention also provide a solid-state battery peak power estimation device based on the PMC-GPR method, comprising: a construction module for constructing an electromechanical coupling model including an equivalent stress model and an equivalent circuit model; a coupling calculation module for inputting acquired battery operating data into the electromechanical coupling model to obtain a current state including SOC, branch voltage, and expansion stress; an estimation module for estimating the battery capacity online using an unscented Kalman filter algorithm based on the current state at a first time scale; and for estimating the estimated SOC online using an unscented Kalman filter algorithm based on the battery capacity and the current state at a second time scale; wherein the first time scale is greater than the second time scale; and a prediction module for constructing feature data based on the estimated SOC and inputting the feature data into a trained Gaussian process regression model to obtain a predicted peak power.

[0008] Thirdly, embodiments of the present invention also provide a computing device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, it implements the solid-state battery peak power estimation method based on the PMC-GPR method described above.

[0009] Fourthly, embodiments of the present invention also provide a computer-readable storage medium having a computer program stored thereon, which, when executed in a computer, causes the computer to execute the solid-state battery peak power estimation method based on the PMC-GPR method described above.

[0010] Fifthly, embodiments of the present invention also provide a computer program product, including computer instructions, which, when executed by a processor, implement the steps of the method described in any of the first aspects of this specification.

[0011] This invention provides a peak power estimation method for solid-state batteries based on the PMC-GPR method. Applied to solid-state batteries, this method constructs an electromechanical coupling model including an equivalent circuit model and an equivalent stress model. This model accurately describes the stress-electrochemical coupling effect unique to solid-state batteries, namely the inverse effect of mechanical stress changes caused by volume changes during charging and discharging on circuit model parameters (such as internal resistance), thus achieving accurate estimation of SOC and expansion stress. Then, a dual unscented particle filter algorithm is used to achieve high-precision joint estimation of SOC and battery capacity at both macroscopic and microscopic time scales, obtaining the estimated SOC. Finally, based on the particle filter results corresponding to the estimated SOC, the Particle Monte Carlo-Gaussian Process Regression (PMC-GPR) method is used to achieve probabilistic prediction and uncertainty quantification of peak power. Thus, this invention effectively improves the accuracy, robustness, and reliability of solid-state battery state estimation. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 is a flowchart of a solid-state battery peak power estimation method based on the PMC-GPR method according to an embodiment of the present invention; Figure 2 is a second-order fractional equivalent circuit according to an embodiment of the present invention; Figure 3 is a hardware architecture diagram of a computing device according to an embodiment of the present invention; Figure 4 is a structural diagram of a solid-state battery peak power estimation device based on the PMC-GPR method according to an embodiment of the present invention. Detailed Implementation

[0014] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0015] Solid-state batteries are an important direction for the next generation of power batteries. During operation, due to the insertion and extraction of lithium ions, the volume of the electrode will change significantly. The volume change rate of silicon-based negative electrode can reach more than 300%. This is manifested as a change in force on a macroscopic scale. In other words, solid-state batteries are always accompanied by significant mechanical effects. The state of power estimation of solid-state batteries needs to focus on their electromechanical coupling characteristics.

[0016] The following is the concept of the present invention. As shown in Figure 1, an embodiment of the present invention provides a peak power estimation method for solid-state batteries based on the PMC-GPR method. The method includes: Step 100, constructing an electromechanical coupling model including an equivalent stress model and an equivalent circuit model; Step 102, inputting the acquired battery operating data into the electromechanical coupling model to obtain the current state including SOC, branch voltage, and expansion stress; Step 104, estimating the battery capacity online using an unscented Kalman filter algorithm based on the current state at a first time scale; Step 106, estimating the estimated SOC online using an unscented Kalman filter algorithm based on the battery capacity and the current state at a second time scale; wherein the first time scale is larger than the second time scale; Step 108, constructing feature data based on the estimated SOC and inputting the feature data into a trained Gaussian process regression model to obtain the predicted peak power.

[0017] In this embodiment of the invention, an electromechanical coupling model comprising an equivalent circuit model and an equivalent stress model is constructed in the solid-state battery. This model can accurately describe the stress-electrochemical coupling effect unique to solid-state batteries, namely, the inverse effect of mechanical stress changes caused by volume changes during charging and discharging on circuit model parameters (such as internal resistance), thereby achieving accurate estimation of SOC and expansion stress. Then, a dual unscented particle filter algorithm is used to achieve high-precision joint estimation of SOC and battery capacity at both macroscopic and microscopic time scales, obtaining the estimated SOC. Finally, based on the particle filter results corresponding to the estimated SOC, the particle Monte Carlo-Gaussian process regression (PMC-GPR) method is used to achieve probabilistic prediction and uncertainty quantification of peak power. Thus, this invention effectively improves the accuracy, robustness, and reliability of solid-state battery state estimation.

[0018] The following describes how each step shown in Figure 1 is performed.

[0019] In step 100, the equivalent stress model is constructed as follows: a swelling and shrinkage strain model is constructed based on the swelling and shrinkage strain, SOC, and battery temperature of the solid-state battery; the swelling and shrinkage strain model is determined by the following formula: in, The expansion and contraction strain; The SOC deformation coefficient; Temperature coefficient; For the current SOC; This is the initial SOC; Current battery temperature; The initial battery temperature is given; an expansion stress model is constructed based on the expansion stress and contraction strain of the solid-state battery; the expansion stress model is determined by the following formula: in, The expansion stress; This refers to the battery stress coefficient. This refers to the battery expansion and contraction stress coefficient. The rate of change of the expansion stress over time; The rate of change of the expansion and contraction strain over time is given; the equivalent stress model is determined based on the expansion and contraction strain model, the expansion stress model, and the ampere-hour integration method.

[0020] In step 100, the equivalent stress model is determined by the following formula: Where k is the k-th sample, The sampling interval; , These represent the expansion stress of the solid-state battery at the (k+1)th and kth sampling times, respectively. The first coefficient; The second coefficient; This represents the current at the k-th sampling point; This refers to the battery stress coefficient. This is the SOC coefficient; Let be the battery capacity at the k-th sampling.

[0021] It should be noted that the current at the kth sampling point is the input current of the equivalent circuit model; , SOC deformation coefficient, The coefficients of battery expansion and contraction stress can all be obtained through experimental calibration.

[0022] In this invention, the expansion and contraction strain model is used to characterize the relationship between expansion and contraction strain and SOC (State of Charge) and battery temperature; the expansion stress model is used to characterize the relationship between expansion stress and expansion and contraction strain. Specifically, for the mechanical effects of solid-state batteries, this invention proposes a modeling method for an equivalent stress model of solid-state batteries based on the rigid clamp assumption and the expansion and contraction properties of the battery: the mechanical effects of the solid-state battery working process are equivalent to a preload force F. ini and expansion force F expThe sum of these factors is used to model the macroscopic expansion effect caused by the electrode volume change due to lithium-ion migration, where the expansion force is the product of the expansion stress and the area. Since solid-state batteries exhibit insignificant temperature fluctuations during operation, this invention assumes that the temperature change of solid-state batteries is slow, and thus the expansion / contraction strain rate is expressed by the following formula: Based on the ampere-hour integration method: at this time, Q is the input current in the equivalent circuit model; n Given the battery capacity, the expansion and contraction strain rate can be simplified to: Therefore, the expansion stress model can be simplified to: Where SOC coefficient The simplified expansion stress model is solved using zero-order preservation, assuming a sampling interval of... The discretized equivalent stress model is obtained: .

[0023] In step 102, in the electromechanical coupling model: the battery operating data and expansion stress of the previous moment are substituted into the equivalent circuit model to obtain the SOC and terminal voltage of the previous moment; the SOC of the previous moment is substituted into the equivalent stress model to obtain the current expansion stress, and the current expansion stress is fed back into the equivalent circuit model to obtain the current SOC.

[0024] In this invention, the electrical effects of solid-state batteries are captured using an equivalent circuit model. Furthermore, for the dispersion effect caused by relaxation in solid-state batteries, this invention employs a second-order fractional-order equivalent circuit for high-precision modeling, as shown in Figure 2. The output voltage of the equivalent circuit model is as follows: in, Terminal voltage, For input current, Open circuit voltage, For ohm resistance, For the voltage of the first RCPE branch, Voltage of the second RCPE branch; expansion stress By influencing the model parameters, the terminal voltage output is affected. This couples the equivalent stress model and the equivalent circuit model, resulting in an electromechanical coupling model. In this model, at the k-th sampling time, the battery operating data from the previous moment (including input current, open-circuit voltage, etc.) and the expansion stress are substituted into the equivalent circuit model to obtain the SOC and terminal voltage at the k-th sampling time. Then, the SOC and input current at the k-th sampling time are passed to the equivalent stress model, which, based on the SOC, input current, and battery capacity, calculates the expansion stress at the (k+1)-th sampling time. Repeating this cycle allows for the accurate determination of the terminal voltage and SOC at any given sampling time.

[0025] It should be noted that offline parameter identification is performed on the electromechanical coupling model to obtain the model parameters.

[0026] For step 104, the online estimation of battery capacity and model parameters using the unscented Kalman filter algorithm includes: obtaining the current time in the current state, determining the current first time scale to which the current time belongs, and using the sum of the capacitance and Gaussian noise under the previous first time scale as the online estimated battery capacity at the current time.

[0027] To achieve accurate estimation of the peak power of a solid-state battery, it is necessary to first estimate the SOC and battery capacity with high precision. In this invention, a filtering method is used. However, due to the multiplicative coupling between SOC and battery capacity, this relationship leads to a non-Gaussian multi-peaked posterior, causing the Gaussian filtering method to fail. Therefore, a particle filtering algorithm is required. Furthermore, to avoid the accuracy degradation caused by particle resampling, unscented particle filtering is chosen to update the macroscopic battery capacity and microscopic SOC.

[0028] In this invention, battery capacity change is a macroscopic, slow, time-varying process. Step 104 is initiated when the sampling count k reaches the set update step, i.e., when entering the next first time scale. Its purpose is to obtain the latest battery capacity, and the model equation is: here, This refers to the battery capacity at the current stage or the (l+1)th first time scale. This refers to the battery capacity of the previous stage or the lth first time scale; It is a Gaussian noise with a mean of 0, representing random fluctuations in capacity; the final estimated capacity As the model parameters for the current stage of microscopic SOC estimation, since the first time scale is larger than the second time scale, information transfer from the macroscopic scale to the microscopic scale is completed. After the next charge-discharge cycle ends, the process re-enters this process. Each first time scale includes several sampling moments, and the sampling period between adjacent sampling moments is the second time scale. In this way, the macroscopic scale is represented by the first time scale and the microscopic scale is represented by the second time scale.

[0029] For step 106, the online estimation using the unscented Kalman filter algorithm to obtain the estimated SOC includes: for any sampling moment within the first time scale, performing the following: substituting the battery capacity at that first time scale into the electromechanical coupling model to obtain a first electromechanical coupling model; generating several particles and initializing the initial state, number of particles, and weights of each particle; wherein, different particles correspond to different states; updating the state prediction value of each particle through unscented transformation based on the first electromechanical coupling model and the current state; and updating the weight of each particle based on the state prediction value; normalizing the weight of each particle and resampling to obtain a set of particles with equal weights; performing a weighted summation on the set of particles with equal weights to obtain the estimated SOC; wherein, when the sampling moment is the last sampling moment of the first time scale, the estimated SOC of that sampling moment is passed to the next first time scale.

[0030] In this invention, the estimation of SOC at the microscale is performed at each sampling time. The purpose is to utilize the battery capacity at the current stage provided at the macroscale to estimate the SOC at the current moment in real time. The state-space equation and observation equation are obtained from the first electromechanical coupling model. The state-space equation consists of the state-space equation of a second-order fractional-order equivalent circuit model with added process noise terms and an equivalent stress model with added process noise terms. Specifically, the state prediction value of each particle is updated through an unscented transformation; and the observation prediction value of the corresponding particle is generated based on the state prediction value and the observation equation. The observation residual is calculated based on the observation prediction value and the actual observation value, and the weight of each particle is calculated based on the observation residual. Thus, this invention achieves joint online estimation of state and parameters through an unscented Kalman filter algorithm, and corrects the inherent degradation of the solid-state battery through dynamic learning, thereby maintaining high accuracy and high robustness in long-term operation.

[0031] Specifically, for online estimation at the second time scale, the process includes: (1) initialization, first setting the initial SOC and covariance, and generating a set of particles related to the system state. ;in, Initialize the weights for particle i. The state corresponding to particle i includes four dimensions. Then, the UKF parameters are initialized. For each particle, an importance proposal distribution is generated using an unscented Kalman filter. By updating the particles over time and through measurement, an optimal proposal distribution considering the current observation data is generated for each particle, allowing the particles to more efficiently approximate the posterior probability distribution of the true state. The particle weights are then adjusted based on the importance of the optimal proposal distribution. After normalizing the particle weights, the particles are resampled based on these weights to avoid particle degeneration, ultimately resulting in a new set of equally weighted particles, i.e., an equally weighted particle set. This equally weighted particle set is then used for SOC estimation at the next sampling time. At this point, the corresponding equation for estimating SOC is: in, Let SOC be the estimated SOC at sampling time k+1; N is the total number of particles in the equally weighted particle set. Let i be the state corresponding to particle i in a set of particles with equal weights. Let SOC be the state value of particle i in the equally weighted particle set at sampling time k+1. In other words, the estimated SOC is obtained by weighted summation of the equally weighted particle sets. It should be noted that if the sampling count reaches the predetermined update step number, the voltage data at that sampling time is passed to the macroscopic-scale battery capacity estimation process (i.e., step 104) to update the battery capacity. Simultaneously, the microscopic-scale process enters a waiting or reset phase, awaiting the start of the next cycle (at which point the capacity has been updated to the new battery capacity).

[0032] For step 108, the estimated SOC is obtained by weighted summation of equal-weighted particle sets, which include particles corresponding to different states; the construction of feature data based on the estimated SOC includes: using the equal-weighted particle sets corresponding to the estimated SOC as feature data; wherein, the feature data is a matrix of dimension N×M, where N is the number of particles in the equal-weighted particle sets, and M is the number of dimensions in the current state.

[0033] More preferably, following the previous example, the number of dimensions in the current state is 4, and the current state includes SOC, first branch voltage, second branch voltage, and expansion stress.

[0034] In this invention, due to the multiplicative coupling between battery SOC and capacity, the system exhibits a non-Gaussian multi-peak posterior. If the model parameters and current and voltage data at the same moment are directly used as feature data for training the Gaussian process trajectory model, the Gaussian Process Regression (GPR) method will fail (because the GPR method requires a Gaussian single-peak prior). Therefore, to solve this problem, this invention combines the particle sampling characteristics of the aforementioned dual unscented particle filtering method and uses the particle Monte Carlo method to construct feature data to capture the posterior multi-peak structure, ensuring that the feature data meets the prior requirements of the GPR method.

[0035] In a preferred embodiment, the Gaussian process regression model is trained by the following method: acquiring physical experimental data of solid-state batteries; generating several sets of sample sets based on the physical experimental data, each set of sample sets including feature data as input and the true peak power corresponding to the feature data as output; and training the Gaussian process regression model using the several sets of sample sets to obtain a trained Gaussian process regression model.

[0036] It should be noted that the physical experimental data were obtained using current sensors, voltage sensors, temperature sensors, and stress sensors. The current sensors used high-precision shunts and dedicated amplifiers to measure the charging and discharging current. The voltage sensors used high-precision differential ADCs to directly measure the battery terminal voltage, which was used as an observation for updating the filtering algorithm. Temperature sensors were placed at multiple points to monitor the battery temperature for temperature compensation. Patch-type strain gauges were used to measure the battery expansion stress, providing direct mechanical parameter verification for the electromechanical coupling model and improving the model accuracy.

[0037] In this embodiment of the invention, physical experimental data containing the true peak power is acquired. Then, the aforementioned unscented Kalman filter algorithm is used to obtain an equally weighted particle set corresponding to the physical experimental data, which includes SOC, branch voltage, and expansion stress. This equally weighted particle set and its corresponding true peak power are then used as the training sample set to train a Gaussian process regression model (GPR model), learning the relationship between peak power and the equally weighted particle set (i.e., feature data). Thus, the trained GPR model can directly output the predicted peak power and the corresponding confidence interval. The confidence interval is used to quantify the uncertainty of the peak power estimation, thereby evaluating the reliability of the results.

[0038] As shown in Figures 3 and 4, this embodiment of the invention provides a solid-state battery peak power estimation device based on the PMC-GPR method. The device embodiment can be implemented through software, hardware, or a combination of both. From a hardware perspective, Figure 3 shows a hardware architecture diagram of the computing device housing the solid-state battery peak power estimation device based on the PMC-GPR method provided in this embodiment. Besides the processor, memory, network interface, and non-volatile memory shown in Figure 3, the computing device in this embodiment may also include other hardware, such as a forwarding chip responsible for processing packets. Taking software implementation as an example, as shown in Figure 4, as a logical device, it is formed by the CPU of the computing device reading the corresponding computer program from the non-volatile memory into memory and running it. This embodiment provides a solid-state battery peak power estimation device based on the PMC-GPR method, comprising: a construction module 400 for constructing an electromechanical coupling model including an equivalent stress model and an equivalent circuit model; a coupling calculation module 402 for inputting acquired battery operating data into the electromechanical coupling model to obtain the current state including SOC, branch voltage, and expansion stress; an estimation module 404 for estimating the battery capacity online using an unscented Kalman filter algorithm based on the current state at a first time scale; and for estimating the estimated SOC online using an unscented Kalman filter algorithm based on the battery capacity and the current state at a second time scale; wherein the first time scale is larger than the second time scale; and a prediction module 406 for constructing feature data based on the estimated SOC and inputting the feature data into a trained Gaussian process regression model to obtain the predicted peak power.

[0039] In some specific implementations, the construction module 400 can be used to perform the above step 100, the coupling calculation module 402 can be used to perform the above step 102, the estimation module 404 can be used to perform the above steps 104 and 106, and the prediction module 406 can be used to perform the above step 108.

[0040] In a preferred embodiment, the above-mentioned devices can also be connected to the vehicle main control (CAN bus), local display, cloud, and alarm module respectively; the vehicle main control is used to send key information such as estimated SOC, battery capacity, and predicted peak power to the vehicle controller; the local display is used to display battery status information for on-site observation; the cloud is used to upload data to the cloud platform for remote monitoring, historical data analysis, and fault diagnosis; the alarm module is used to trigger indicator lights or relays to provide a safety alarm when the estimated SOC result is abnormal (such as SOC jump or predicted peak power being too low).

[0041] Since the contents of the above-described apparatus are based on the same concept as the method embodiments of the present invention, the specific contents can be found in the descriptions in the method embodiments of the present invention, and will not be repeated here.

[0042] Compared with the prior art, the advantages of this invention are: (1) This invention constructs a high-precision electromechanical coupling model of an equivalent circuit model (electrical effect) and an equivalent stress model (mechanical effect). This electromechanical coupling model can accurately describe the stress-electrochemical coupling effect unique to solid-state batteries, that is, the reverse influence of expansion stress changes on circuit model parameters (such as internal resistance), so that the model parameters can be dynamically adjusted according to stress feedback, thereby realizing the synchronous and accurate estimation of terminal voltage, SOC and expansion stress, laying the foundation for accurate peak power prediction.

[0043] (2) To address the problem that the multiplicative coupling between SOC and battery capacity leads to a non-Gaussian, multi-modal distribution of the system's posterior probability, thus rendering traditional Gaussian assumption methods such as Kalman filtering ineffective, this invention employs a dual unscented particle filter algorithm. This algorithm estimates battery capacity (slowly varying parameters) and SOC (rapidly varying state) at both macroscopic and microscopic time scales, respectively. It then uses unscented Kalman filtering to generate an optimal proposal distribution, guiding particles to move towards the high-likelihood region, ultimately obtaining high-precision estimates of battery capacity and SOC. This method effectively handles the nonlinearity and non-Gaussianity of the system, avoids particle degeneration, and ensures the algorithm's robustness and estimation stability when facing complex coupling relationships.

[0044] (3) In this invention, after obtaining high-precision SOC and battery capacity, a PMC-GPR (Particle Monte Carlo-Gaussian Process Regression) method is further proposed for peak power prediction. This method cleverly utilizes the equally weighted particle set generated by resampling using a dual unscented particle filter algorithm, thereby transforming non-Gaussian posterior information into training samples suitable for GPR processing. Thus, the GPR model can not only output point predictions of peak power, but more importantly, it can provide its prediction confidence interval, realizing the quantification of the uncertainty in peak power estimation. This provides a probabilistic basis for battery safety management, greatly improving the reliability and practicality of peak power prediction.

[0045] It is understood that the structures illustrated in the embodiments of the present invention do not constitute a specific limitation on a solid-state battery peak power estimation device based on the PMC-GPR method. In other embodiments of the present invention, a solid-state battery peak power estimation device based on the PMC-GPR method may include more or fewer components than illustrated, or combine some components, or split some components, or have different component arrangements. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.

[0046] The information interaction and execution process between the modules in the above-mentioned device are based on the same concept as the method embodiment of the present invention, and the specific details can be found in the description of the method embodiment of the present invention, and will not be repeated here.

[0047] This invention also provides a computing device, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it implements a solid-state battery peak power estimation method based on the PMC-GPR method according to any embodiment of this invention.

[0048] This invention also provides a computer-readable storage medium storing a computer program. When executed by a processor, the computer program causes the processor to perform a solid-state battery peak power estimation method based on the PMC-GPR method according to any embodiment of this invention.

[0049] Embodiments of this application also provide a computer program product, which includes a computer program. A processor of a computer device reads the computer program from a computer-readable storage medium and executes the computer program, causing the computer device to perform a solid-state battery peak power estimation method based on the PMC-GPR method described in any of the above embodiments.

[0050] Specifically, a system or apparatus equipped with a storage medium may be provided, on which software program code implementing the functions of any of the embodiments described above is stored, and the computer (or CPU or MPU) of the system or apparatus may read and execute the program code stored in the storage medium.

[0051] In this case, the program code read from the storage medium can itself implement the function of any of the above embodiments, and therefore the program code and the storage medium storing the program code constitute part of the present invention.

[0052] Storage media embodiments for providing program code include floppy disks, hard disks, magneto-optical disks, optical disks (such as CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RAM, DVD-RW, DVD+RW), magnetic tapes, non-volatile memory cards, and ROMs. Alternatively, program code can be downloaded from a server computer via a communication network.

[0053] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, which can send, propagate, or transmit programs for use by or in conjunction with an instruction execution system, system, or device.

[0054] The program code contained on a computer-readable medium may be transmitted using any suitable medium, including, but not limited to, wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0055] Computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as "C" or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0056] Furthermore, it should be clear that not only can the program code read by the computer be executed, but also the operating system or other components operating on the computer can be instructed based on the program code to perform some or all of the actual operations, thereby realizing the function of any of the embodiments described above.

[0057] Furthermore, it is understood that the program code read from the storage medium is written to the memory set in the expansion board inserted into the computer or to the memory set in the expansion module connected to the computer. Then, based on the instructions of the program code, the CPU or other components installed on the expansion board or expansion module execute some and all of the actual operations, thereby realizing the function of any of the above embodiments.

[0058] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0059] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media that can store program code, such as ROM, RAM, magnetic disk, or optical disk.

[0060] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for estimating the peak power of a solid-state battery based on the PMC-GPR method, characterized in that, The method, applied to solid-state batteries, includes: constructing an electromechanical coupling model comprising an equivalent stress model and an equivalent circuit model; inputting acquired battery operating data into the electromechanical coupling model to obtain the current state including SOC, branch voltage, and expansion stress; estimating the battery capacity online using an unscented Kalman filter algorithm based on the current state at a first time scale; estimating the estimated SOC online using an unscented Kalman filter algorithm based on the battery capacity and the current state at a second time scale; wherein the first time scale is larger than the second time scale; constructing feature data based on the estimated SOC, and inputting the feature data into a trained Gaussian process regression model to obtain the predicted peak power.

2. The method according to claim 1, characterized in that, The equivalent stress model is constructed as follows: a swelling and shrinkage strain model is constructed based on the swelling and shrinkage strain, SOC, and battery temperature of the solid-state battery; the swelling and shrinkage strain model is determined by the following formula: in, The expansion and contraction strain is the strain described above. The SOC deformation coefficient; Temperature coefficient; For the current SOC; This is the initial SOC; Current battery temperature; The initial battery temperature is given; an expansion stress model is constructed based on the expansion stress and contraction strain of the solid-state battery; the expansion stress model is determined by the following formula: in, The expansion stress; This refers to the battery stress coefficient. The coefficient of stress for battery expansion and contraction is given; the equivalent stress model is determined based on the expansion and contraction strain model, the expansion stress model, and the ampere-hour integration method.

3. The method according to claim 1 or 2, characterized in that, The equivalent stress model is determined by the following formula: Where k is the k-th sample, The sampling interval; 、 These represent the expansion stress of the solid-state battery at the (k+1)th and kth sampling times, respectively. The first coefficient; The second coefficient; This represents the current at the k-th sampling point; This refers to the battery stress coefficient. This is the SOC coefficient; Let be the battery capacity at the k-th sampling point; and / or, inputting the acquired battery operating data into the electromechanical coupling model to obtain the current state including SOC, branch voltage, and expansion stress includes: substituting the battery operating data and expansion stress from the previous moment into the equivalent circuit model to obtain the SOC and terminal voltage from the previous moment; substituting the SOC from the previous moment into the equivalent stress model to obtain the current expansion stress, and feeding the current expansion stress back into the equivalent circuit model to update the current SOC.

4. The method according to claim 1, characterized in that, The online estimation of battery capacity and model parameters using the unscented Kalman filter algorithm includes: obtaining the current moment in the current state, determining the current first time scale to which the current moment belongs, and using the sum of the capacitance and Gaussian noise under the previous first time scale as the online estimated battery capacity at the current moment; and / or, the online estimation using the unscented Kalman filter algorithm to obtain the estimated SOC includes: for any sampling moment within the first time scale, performing the following: substituting the battery capacity under the first time scale into the electromechanical coupling model to obtain the first electromechanical coupling model; generating... Several particles are initialized, including their initial states, number, and weights. Different particles correspond to different states. Based on the first electromechanical coupling model and the current state, the predicted state value of each particle is updated through an unscented transformation. The weight of each particle is updated based on the predicted state value. The weight of each particle is normalized and resampled to obtain a set of particles with equal weights. The set of particles with equal weights is weighted and summed to obtain the estimated SOC. When the sampling time is the last sampling time of the first time scale, the estimated SOC of that sampling time is passed to the next first time scale.

5. The method according to claim 1, characterized in that, The estimated SOC is obtained by weighted summation of equally weighted particle sets, which include particles corresponding to different states; the construction of feature data based on the estimated SOC includes: using the equally weighted particle set corresponding to the estimated SOC as feature data; wherein, the feature data is a matrix of dimension N×M, where N is the number of particles in the equally weighted particle set, and M is the number of dimensions in the current state; preferably, M=4.

6. The method according to any one of claims 1 to 5, characterized in that, The Gaussian process regression model is trained by the following method: acquiring physical experimental data of solid-state batteries; generating several sets of sample data based on the physical experimental data, each set of sample data including feature data as input and the true peak power corresponding to the feature data as output; and training the Gaussian process regression model using the several sets of sample data to obtain the trained Gaussian process regression model.

7. A device for estimating the peak power of a solid-state battery based on the PMC-GPR method, characterized in that, include: The building module is used to construct an electromechanical coupling model that includes an equivalent stress model and an equivalent circuit model; The coupling calculation module is used to input the acquired battery operating data into the electromechanical coupling model to obtain the current state including SOC, branch voltage and expansion stress; The estimation module is used to estimate the battery capacity online based on the current state at a first time scale using an unscented Kalman filter algorithm. And at the second time scale, based on the battery capacity and the current state, an unscented Kalman filter algorithm is used for online estimation to obtain the estimated SOC; wherein, the first time scale is greater than the second time scale; The prediction module is used to construct feature data based on the estimated SOC and input the feature data into the trained Gaussian process regression model to obtain the predicted peak power.

8. A computing device comprising a memory and a processor, wherein the memory stores a computer program, and the processor, when executing the computer program, implements the method as described in any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, which, when executed in a computer, causes the computer to perform the method of any one of claims 1-6.

10. A computer program product, characterized in that, Includes computer instructions that, when executed by a processor, implement the steps of the method as described in any one of claims 1-6.

Images