Battery simulation system and method, and parameter adjustment method and device of simulation model

Abstract

The invention relates to a battery simulation system and method and a parameter adjusting method and device of a simulation model. The battery simulation system comprises the simulation model and a parameter adjusting agent configured in the simulation model. The parameter adjusting agent is used for determining the model parameter adjusting quantity of the simulation model based on at least the working condition parameters of the battery; the simulation model is used for performing model parameter adjustment based on the model parameter adjustment quantity, and the simulation model after model parameter adjustment is used for battery simulation. According to the method, the parameters of the simulation model can be dynamically adjusted according to the actual working condition of the battery based on the parameter adjusting agent, dynamic adjustment and simulation of the model parameters for different working conditions are achieved, and therefore the simulation accuracy and robustness are improved.

Description

Technical Field

[0001] The present disclosure relates to the field of vehicle technology, and in particular to a battery simulation system, method, and simulation model parameter adjustment method and device. Background Art

[0002] A battery simulation model, also referred to as a battery model, is mainly used to simulate the working process of a battery. It can output performance indicators such as battery voltage, power, state of charge (SOC), and state of health (SOH). It can be used to estimate battery conditions such as battery aging. This is of great significance for improving the battery's working state (such as maximum charge and discharge power), extending battery life, and optimizing battery design.

[0003] Currently, there are four main types of battery simulation models used to estimate battery state: pseudo-electrochemical models, equivalent circuit models, algebraic equations, and data-driven intelligent models, which can include decision trees and deep neural networks. These battery simulation models have all been very successful in certain scenarios. However, their accuracy depends on the accuracy of the relevant model parameters.

[0004] Batteries are typically highly nonlinear systems, exhibiting nonlinear characteristics under diverse operating conditions, including high and low currents, charging and discharging, long and short operating times, and high and low temperatures. Current online parameter estimation algorithms are primarily based on the least squares method, which adjusts parameters based on the error between the cell voltage test values ​​and the model-calculated values ​​over a period of time to minimize the model-based prediction error. This parameter estimation method is suitable for situations where operating conditions remain unchanged. However, using this method to determine model parameters under varying operating conditions results in poor accuracy and robustness of the simulation model. Summary of the Invention

[0005] In order to solve the above technical problems, the present disclosure provides a battery simulation system, method, simulation model parameter adjustment method and device.

[0006] In a first aspect, the present disclosure provides a battery simulation system, comprising a simulation model and a parameter adjustment agent configured in the simulation model;

[0007] The parameter adjustment agent is used to determine the model parameter adjustment amount of the simulation model based on at least the operating parameters of the battery;

[0008] The simulation model is used to adjust model parameters based on the model parameter adjustment amount, and the simulation model after adjusting the model parameters is used for battery simulation.

[0009] Optionally, the simulation model is constructed based on an actual battery pack, the battery pack includes a plurality of battery cells, and the simulation model correspondingly includes a plurality of battery cell models, each of which has adjustable model parameters;

[0010] The parameter adjustment agent is constructed and trained based on a reinforcement learning algorithm, wherein the input of the parameter adjustment agent includes model parameters of the simulation model, input and output of the simulation model, and evaluation results provided by an evaluation model, wherein the input of the evaluation model includes simulation values, measurement values, and model parameter adjustment amounts, and the output of the parameter adjustment agent includes the model parameter adjustment amounts of the simulation model;

[0011] Among them, based on the operation of the battery pack and the synchronous status update of the simulation model, the simulation values, measurement values ​​and model parameter adjustment amounts before and after the update are obtained and used to determine the evaluation results; based on multiple groups of simulation values, measurement values, model parameter adjustment amounts and evaluation results before and after the update, the parameter adjustment agent is trained to determine that the parameter adjustment agent corresponding to the output model parameter adjustment amount when the target evaluation result is obtained is the parameter adjustment agent.

[0012] Optionally, the training of the parameter adjustment agent further comprises: for each type of collection cell, training the parameter adjustment agent based on the collection cell respectively;

[0013] The collected cells include cells with the maximum cell voltage, the minimum cell voltage and the average cell voltage among all cell voltages in the battery pack.

[0014] Optionally, the battery pack includes a plurality of different battery packs configured on different carriers, and the carrier includes at least a vehicle;

[0015] The parameter adjustment agent is trained based on the acquisition cell, including: training the parameter adjustment agent based on multiple different acquisition cells of the same type in multiple different battery packs to obtain multiple parameter adjustment amounts; superimposing the multiple parameter adjustment amounts, and adjusting the parameters of the parameter adjustment agent itself based on the superimposed parameter adjustment amounts.

[0016] Optionally, the evaluation result is calculated based on the cell voltage simulation value, cell voltage test value, open circuit voltage simulation value, open circuit voltage test value and upper and lower limits of the cell voltage before and after three adjacent updates.

[0017] Optionally, the adjustment of the model parameters of the simulation model includes: adding the model parameter adjustment amount to the current model parameters of the simulation model to obtain the adjusted model parameters.

[0018] In a second aspect, the present disclosure provides a parameter adjustment method for a battery simulation model, comprising:

[0019] Obtaining the parameter adjustment agent, model parameters of the simulation model, and operating parameters of the battery;

[0020] Determining, based on the parameter adjustment agent, a model parameter adjustment amount corresponding to the model parameter and the operating condition parameter;

[0021] adjusting the model parameters of the simulation model based on the model parameter adjustment amount;

[0022] The simulation model after adjusting the model parameters is used for battery simulation.

[0023] Optionally, the acquiring parameter adjustment agent includes:

[0024] Constructing a simulation model based on an actual battery pack; the battery pack includes a plurality of battery cells, and the simulation model correspondingly includes a plurality of battery cell models, each of the battery cell models having adjustable model parameters;

[0025] Constructing a parameter adjustment agent to be trained based on a reinforcement learning algorithm; the input of the parameter adjustment agent includes model parameters of the simulation model, the input and output of the simulation model, and an evaluation result provided by an evaluation model, the input of the evaluation model includes simulation values, measurement values, and model parameter adjustment amounts, and the output of the parameter adjustment agent includes the model parameter adjustment amounts of the simulation model;

[0026] Running the battery pack and performing synchronous state updates on the simulation model, and determining an evaluation result based on simulation values, measurement values, and model parameter adjustments before and after the update;

[0027] The parameter adjustment agent is trained based on multiple sets of simulation values, measurement values, model parameter adjustment amounts and evaluation results before and after the update, and the parameter adjustment agent corresponding to the output model parameter adjustment amount that obtains the target evaluation result is determined to be the parameter adjustment agent.

[0028] Optionally, the parameter adjustment method of the battery simulation model further includes:

[0029] Obtaining the voltages of all cells in the battery pack, and determining the cells with the maximum cell voltage, the minimum cell voltage, and the average cell voltage as the collected cells;

[0030] The training of the parameter adjustment agent includes: for each type of collection cell, respectively training the parameter adjustment agent based on the collection cell.

[0031] Optionally, the battery pack includes a plurality of different battery packs configured on different carriers, and the carrier includes at least a vehicle;

[0032] Training a parameter adjustment agent based on the acquisition cell includes:

[0033] The parameter adjustment agent is trained based on multiple different acquisition cells of the same type in multiple different battery packs to obtain multiple parameter adjustment values;

[0034] The multiple parameter adjustment amounts are superimposed, and the parameters of the parameter adjustment agent itself are adjusted based on the superimposed parameter adjustment amounts.

[0035] Optionally, determining the evaluation result based on the simulation values, measurement values, and model parameter adjustment amounts before and after the update includes:

[0036] The evaluation result is calculated based on the cell voltage simulation values, cell voltage test values, open circuit voltage simulation values, open circuit voltage test values, and upper and lower limits of the cell voltage before and after three adjacent updates.

[0037] Optionally, adjusting the model parameters of the simulation model based on the model parameter adjustment amount includes:

[0038] For the model parameters of the simulation model, the model parameter adjustment amount is added to the current model parameters to obtain the adjusted model parameters.

[0039] In a third aspect, the present disclosure further provides a battery simulation method, comprising the steps of any one of the battery simulation model parameter adjustment methods provided in the second aspect;

[0040] The battery simulation method further includes:

[0041] The battery is simulated based on the simulation model after parameter adjustment to obtain simulation performance parameters.

[0042] Optionally, the battery simulation method further includes:

[0043] The aging state of the battery is determined based on the model parameter adjustment amount and / or the adjusted model parameter.

[0044] In a fourth aspect, the present disclosure further provides a parameter adjustment device for a battery simulation model, comprising:

[0045] An acquisition module is used to obtain the model parameters of the parameter adjustment agent, the simulation model, and the operating parameters of the battery;

[0046] a determination module, configured to determine, based on the parameter adjustment agent, a model parameter adjustment amount corresponding to the model parameter and the operating condition parameter;

[0047] An adjustment module is used to adjust the model parameters of the simulation model based on the model parameter adjustment amount; wherein the simulation model after adjusting the model parameters is used for battery simulation.

[0048] The technical solution provided by the present disclosure has the following advantages compared with the existing technology:

[0049] In the battery simulation system, method, simulation model parameter adjustment method and device provided by the present disclosure, the battery simulation system is set to include a simulation model and a parameter adjustment agent configured in the simulation model; the parameter adjustment agent is used to determine the model parameter adjustment amount of the simulation model based on at least the operating parameters of the battery; the simulation model is used to adjust the model parameters based on the model parameter adjustment amount, and the simulation model after adjusting the model parameters is used for battery simulation. Based on the parameter adjustment agent, the parameters of the simulation model can be dynamically adjusted according to the actual operating conditions of the battery, thereby realizing dynamic adjustment and simulation of the model parameters for different operating conditions, thereby improving the simulation accuracy and robustness. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the present disclosure.

[0051] In order to more clearly illustrate the embodiments of the present disclosure or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0052] Figure 1 A flow chart of a parameter adjustment method for a battery simulation model provided in an embodiment of the present disclosure;

[0053] Figure 2 A schematic diagram of a system architecture for implementing a parameter adjustment method for a battery simulation model provided by an embodiment of the present disclosure;

[0054] Figure 3 A schematic diagram of a training process for a battery simulation model provided in an embodiment of the present disclosure;

[0055] Figure 4 A schematic structural diagram of a battery cell simulation model provided in an embodiment of the present disclosure;

[0056] Figure 5 A flowchart of a battery simulation method provided by an embodiment of the present disclosure;

[0057] Figure 6 A schematic structural diagram of a battery simulation model parameter adjustment device provided in an embodiment of the present disclosure;

[0058] Figure 7 A schematic diagram of the structure of a battery simulation system provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0059] In order to more clearly understand the above-mentioned objectives, features and advantages of the present disclosure, the scheme of the present disclosure will be further described below. It should be noted that the embodiments of the present disclosure and the features therein can be combined with each other in the absence of conflict.

[0060] In the following description, many specific details are set forth to facilitate a full understanding of the present disclosure, but the present disclosure may also be implemented in other ways different from those described herein; it is obvious that the embodiments in the specification are only part of the embodiments of the present disclosure, rather than all of the embodiments.

[0061] In related technologies, since the model parameters of the battery simulation model are usually applicable to situations where the operating conditions remain unchanged, that is, the model parameters do not change with changes in the operating conditions, the fixedness and static nature of the model parameters limit the accuracy and robustness of the simulation model under complex operating conditions.

[0062] Battery simulation models can be divided into battery pack models and cell models based on the simulation objectives. The battery pack model studies the overall battery pack's output voltage, SOC, SOH, and other state indicators; the cell model studies the cell's output voltage, SOC, SOH, and other state indicators. Because the cell model has fewer interfering factors, the simulation results are relatively accurate. Separating the battery pack from the cell model is primarily intended to reduce the amount of computational data and improve the model's real-time performance. However, this separation further compromises the simulation model's compatibility with multiple operating scenarios and its ability to discern differences between cells. Furthermore, a battery pack typically consists of approximately 100 cells. Simulating each cell with a separate cell model would significantly increase the computing power and memory requirements of the battery management system (BMS), significantly increasing costs. Furthermore, the fixed and static nature of the model parameters often makes it difficult to monitor and adjust the varying states and characteristics of multiple cells within the battery pack in real time.

[0063] In response to at least one of the above technical problems, the embodiments of the present disclosure provide a battery simulation system, method, simulation model parameter adjustment method and device, wherein, by setting a parameter adjustment intelligent agent, the model parameter adjustment amount corresponding to the model parameters and the operating condition parameters can be determined, and based on the model parameter adjustment amount, the model parameters of the simulation model are adjusted so that the adjusted model parameters can be applicable to the operating condition parameters, that is, the model parameters change with the change of the operating conditions / working scenarios, thereby meeting the simulation requirements under different working conditions, realizing the dynamic adjustment of the model parameters of the simulation model in multiple scenarios and long life cycles, and improving the accuracy and robustness of the simulation under complex working conditions.

[0064] In some embodiments, the battery simulation model may be a battery pack model, which includes multiple cell models, and simulates the cells with maximum cell voltage, minimum cell voltage and average cell voltage in the battery pack without simulating all the cells in the battery pack. It can accurately simulate the cells in the battery pack while reducing the amount of data processing for the battery pack simulation, reducing the computing power and memory requirements for the battery management system, and reducing costs; at the same time, the model parameters are dynamically adjustable for different working conditions, so that real-time monitoring and adjustment of different states and characteristics of multiple cells inside the battery pack are highly feasible.

[0065] The parameter adjustment method and device of the battery simulation model and the battery simulation method and system provided by the embodiments of the present disclosure are exemplarily described below with reference to the accompanying drawings.

[0066] For example, Figure 1 A flow chart of a parameter adjustment method for a battery simulation model provided by an embodiment of the present disclosure. Figure 1 , the parameter adjustment method of the battery simulation model may include the following steps.

[0067] S110 , obtaining the parameter adjustment agent, model parameters of the simulation model, and operating parameters of the battery.

[0068] Among them, the parameter adjustment intelligent agent is an intelligent agent used to adjust the model parameters of the simulation model, which can also be called a battery model parameter adjuster, or a model parameter adjuster. It can output the model parameter adjustment amount according to at least the operating condition parameters, thereby realizing the parameter adjustment of the simulation model and then realizing dynamic parameter simulation.

[0069] For example, the parameter adjustment agent itself may include a neural network model with adjustable parameters. This parameter adjustment agent can be implemented based on the principles of reinforcement learning and possess self-learning capabilities. As the amount of data and exposure to scenarios increases, the accuracy of model parameter adjustment will gradually improve. The acquisition process of this parameter adjustment agent is described in detail below.

[0070] Among them, the simulation model is a battery simulation model, which can also be called a battery model, which can simulate and obtain battery-related states, and the battery-related states can be characterized based on the performance indicators of the battery. In the embodiment of the present disclosure, the battery simulation model can be a battery pack model, which includes cell models with connection relationships; wherein the connection relationship between the cell models is determined based on the actual connection relationship between different cells in the actual battery pack. The adjustment of model parameters can be understood as adjusting the parameters of different cell models in the battery pack model, such as adjusting the cell model parameters of the maximum cell voltage, the minimum cell voltage and the average cell voltage, so as to achieve parameter adjustment of the battery pack model as a whole, and then realize the simulation of the cell based on the variable parameter model, obtain the cell state, and then obtain the battery pack state based on the connection relationship between the cells.

[0071] Exemplarily, the simulation model may be an equivalent circuit model (ECM), an electrochemical reaction model (Electrochemical Models), an empirical model, or other simulation models capable of simulating the state of a battery cell, which is not limited here.

[0072] Among them, the simulation model has adjustable model parameters to adapt to different working conditions. For example, it is assumed that the battery cell model is y=a×x+b; wherein x is the input variable, y is the output variable, and the equation coefficients a and b are adjustable parameters. The parameter adjustment method provided in the embodiment of the present disclosure includes how to dynamically adjust the parameters a and b based on the input variable x and output variable y obtained by measurement. In other embodiments, when the battery cell model is other models, the model parameters can be understood as coefficients other than the input variables and output variables, which are not described here.

[0073] The battery operating parameters are used to characterize the battery operating conditions or battery operating scenarios. For example, the battery operating parameters may include current, temperature, or other operating parameters that affect the model parameters, which are not described in detail here.

[0074] In this step, the parameter adjustment intelligent agent can be obtained through preliminary training and applied to the parameter adjustment method of the battery simulation model; the model parameters of the simulation model can be the model parameters obtained after the last parameter adjustment, and the operating parameters of the battery can be the parameters corresponding to the current operating conditions of the battery, which can be obtained based on monitoring / updating the operating conditions, or based on other methods known to technical personnel in this field, and are not limited here.

[0075] S120 . Determine the model parameter adjustment amount corresponding to the model parameter and the operating condition parameter based on the parameter adjustment agent.

[0076] Among them, the model parameter adjustment amount is the amount of model parameter adjustment required to compensate for the difference between the battery pack state obtained by the simulation model and the actual state of the battery pack based on the current model parameters and operating parameters, so that the simulation model can more accurately simulate the battery state under the current operating conditions.

[0077] The parameter adjustment agent can output the corresponding model parameter adjustment value based on the input operating condition parameters and model parameters. Specifically, the simulation model can adjust the model parameters based on the model parameter adjustment value, simulate the battery cell state based on the adjusted model parameters, and output performance indicators to represent the battery cell state. (See below for details.)

[0078] S130 , adjusting the model parameters of the simulation model based on the model parameter adjustment amount; wherein the simulation model after the model parameters are adjusted is used for battery simulation.

[0079] Among them, the model parameter adjustment amount is used to adjust the model parameters of the simulation model. The simulation model after adjustment has model parameters suitable for the current working conditions and can accurately simulate the battery cell state and battery pack state under the current working conditions.

[0080] In the parameter adjustment method of the battery simulation model provided in the embodiment of the present disclosure, by setting a parameter adjustment intelligent agent, the model parameter adjustment amount corresponding to the model parameter and the operating condition parameter can be determined, and based on the model parameter adjustment amount, the model parameters of the simulation model are adjusted so that the adjusted model parameters can be applicable to the operating condition parameters, that is, the model parameters change with the change of the operating condition / working scenario, thereby meeting the simulation requirements under different operating conditions, realizing the dynamic adjustment of the model parameters of the simulation model in multiple scenarios and long life cycles, and improving the accuracy and robustness of the simulation under complex working conditions.

[0081] In some embodiments, Figure 1 Based on the above, obtaining the parameter adjustment agent in S110 can be understood as a process of training and updating the initialized agent to obtain a parameter adjustment agent that meets the requirements. For example, obtaining the parameter adjustment agent may include:

[0082] Step 1: Build a simulation model based on the actual battery pack; the battery pack includes multiple battery cells, and the simulation model correspondingly includes multiple battery cell models, and the battery cell models have adjustable model parameters.

[0083] Among them, by adjusting the model parameters of the battery cell model, the overall parameter adjustment of the battery pack model can be achieved, thereby realizing variable parameter model simulation.

[0084] It should be noted that the cell models in the same battery pack may use the same cell model or different cell models, which is not limited here.

[0085] Step 2: Construct a parameter adjustment agent to be trained based on the reinforcement learning algorithm; the input of the parameter adjustment agent includes the model parameters of the simulation model, the input and output of the simulation model, and the evaluation results provided by the evaluation model. The input of the evaluation model includes simulation values, measurement values, and model parameter adjustment amounts. The output of the parameter adjustment agent includes the model parameter adjustment amounts of the simulation model.

[0086] The parameter adjustment agent is implemented based on the principles of reinforcement learning and has self-learning capabilities. As the amount of data and the scenarios it encounters increase, the accuracy of parameter adjustment will gradually improve. For example, the reinforcement learning algorithm can adopt a deep reinforcement learning algorithm, such as the Deep Deterministic Policy Gradient (DDPG) algorithm.

[0087] Specifically, the parameter adjustment agent is implemented based on reinforcement learning, and can dynamically adjust the model parameters according to the current state of the battery cell and the corresponding simulation model, such as current, temperature and other operating parameters. It has the following advantages: 1) The parameter adjustment agent not only learns model parameters online, but can also be applied in simulation to dynamically adjust model parameters. It can be used as part of the simulation system to achieve variable parameter simulation. 2) The parameter adjustment agent has a memory function. With the accumulation of training data and the increase of usage time, the accuracy of the simulation model will continue to improve. 3) Learning is a kind of ability. When the parameter adjustment agent encounters new scene parameters, it can still have good adaptability, which can make the simulation model have better simulation accuracy.

[0088] For example, Figure 2 A schematic diagram of a system architecture for implementing a parameter adjustment method for a battery simulation model according to an embodiment of the present disclosure. Figure 2 The system architecture includes a cell model 001, a parameter adjustment agent 002 (referred to as agent 002), an evaluation model 003, a battery pack 004, a load unit 005 and a charger unit 006.

[0089] Among them, the battery pack 004, the load unit 005 and the charger unit 006 belong to the relevant physical hardware structure of the new energy power system.

[0090] For example, the battery pack 004 may be a high voltage battery pack, and the battery pack 004 may also be configured with a battery management system ( Figure 2 (not shown); the battery pack is used to store electrical energy, and the battery management system is used to collect parameters of the battery pack and battery cells, such as battery cell voltage, battery pack current, battery temperature and other parameters; the battery management system is also used for the execution of software such as diagnosis, communication, and logical calculation, which are not limited here.

[0091] For example, the load unit 005 may be a high-voltage power-consuming component, which may include a drive motor, an electric air conditioner, etc. These high-voltage power-consuming components draw power from the high-voltage battery pack, causing the high-voltage battery pack to be in a discharged state and the battery SOC to gradually decrease.

[0092] For example, the charger unit 006 may include an onboard charger, an off-board charging station, and other charging components. These charging components charge the battery pack 004, causing the battery SOC to gradually increase.

[0093] Exemplarily, the cell model 001 may constitute a battery pack model.

[0094] Exemplarily, the agent 002 is a parameter-adjustable agent 002 based on an intelligent algorithm (i.e., an intelligent model or algorithm model). The algorithm model may include a deep neural network model (DNN), a recursive neural network model (RNN), a support vector machine (SVM), a genetic algorithm (GA), or other algorithm models that support adjustable parameters, without limitation herein.

[0095] The input of the parameter adjustment agent 002 is the model parameters of the cell model 001, the input and output of the cell model 001, and the evaluation results provided by the evaluation module 003. The output of the parameter adjustment agent 002 is the adjustment value step of the parameters of the cell model 001, that is, the model parameter adjustment amount.

[0096] In some embodiments, adjusting the model parameters of the simulation model based on the model parameter adjustment amount may specifically include:

[0097] For the model parameters of the simulation model, the model parameter adjustment amount is added to the current model parameters to obtain the adjusted model parameters.

[0098] For example, taking the resistance R0 of the ECM model as an example, the parameter adjustment agent 002 outputs the model parameter adjustment amount ▽R0, and the parameter used for the next simulation of the battery cell model is R0+▽R0.

[0099] The parameter adjustment agent 002 has the ability to adjust its own parameters so that the obtained evaluation results gradually increase. The algorithm for adjusting its own parameters varies based on the algorithm model adopted by the parameter adjustment agent 002. For example, the algorithm for adjusting its own parameters can be based on a parameter adjustment algorithm such as gradient descent, Monte Carlo gradient strategy, or genetic algorithm, which is not limited here.

[0100] The evaluation model 003 is used to evaluate whether the action performed by the parameter adjustment agent 002 is good. Specifically, if the action performed by the parameter adjustment agent 002 is a good operation, that is, corresponds to a correct adjustment, a high evaluation score is given; if the action performed by the parameter adjustment agent 002 is not a good adjustment, for example, resulting in a large difference between the model simulation results and the actual measured values ​​of the battery cell, a low evaluation score is given.

[0101] For example, Figure 3 This is a schematic diagram of the training process of a battery simulation model provided by an embodiment of the present disclosure, which also shows the parameter update process of the battery model. Figure 3 The input of evaluation model 003 can be multiple groups (State-Action-Updated State + ), the output is the evaluation score Reward. Among them, State refers to the current state, such as the simulated battery cell voltage Equal simulation value, current current i i And the measured voltage u i The measured value is the model parameter adjustment value output by the parameter adjustment agent 002; the updated state is State + This refers to the state of the simulation model at the next time step after the model parameter adjustment is performed using the model parameter adjustment value output by the parameter adjustment agent 002, i.e., the simulated value and measured value of the next state. In some embodiments, the evaluation model may also include custom rules to determine the quality of the adjustment results based on the simulation error, which is not limited here.

[0102] The agent training and model parameter updating methods based on this system architecture are explained in conjunction with subsequent steps three and four.

[0103] Step 3: Run the battery pack and perform a synchronous status update on the simulation model, and determine the evaluation results based on the simulation values, measurement values, and model parameter adjustments before and after the update.

[0104] During operation, the battery pack and its cells are monitored using the battery management system, generating measured values. Simultaneously, the simulation model simulates the battery status based on the adjusted parameters, updating the simulated battery state and generating simulated values. The evaluation model outputs an evaluation score based on the simulated and measured values ​​before and after the update, as well as the model parameter adjustments.

[0105] It should be noted that Figure 3 In the State, the state includes the simulation value and the measured value before the update. +Includes updated simulation values ​​and measurement values, and the action corresponds to the model parameter adjustment amount.

[0106] Step 4: Train the parameter adjustment agent based on multiple sets of simulation values, measurement values, model parameter adjustment amounts, and evaluation results before and after the update, and determine that the parameter adjustment agent corresponding to the output model parameter adjustment amount when the target evaluation result is obtained is the parameter adjustment agent.

[0107] Among them, the target evaluation result is the evaluation result that meets the training requirements. When the evaluation result output by the evaluation model meets the target evaluation result, the parameter adjustment agent is a trained parameter adjustment agent.

[0108] It should be noted that Figure 2 The figure only shows that the battery cell model includes two RC circuits connected in series. In other embodiments, the battery cell model may also adopt other simulation models, which are not limited here.

[0109] The following combination Figure 3 The specific process steps of agent training are illustrated.

[0110] Figure 3 In the equation, State represents the current state, State+ represents the state of the next time step, the superscript + represents the content of the next time step, Action represents the action of the parameter adjustment agent, which can also be called the model parameter adjustment amount Action, and Reward represents the evaluation parameter of the evaluation model output.

[0111] refer to Figure 3 , given the current state State at time t, the parameter adjustment agent outputs the model parameter adjustment amount Action.

[0112] The current state can include the following:

[0113] 1) Current I: I i-2 , I i-1 , I i Where, the subscript i represents the time step, I i-2 Indicates the current value of the previous 2 steps, I i-1 Indicates the current value of the previous step, I i Indicates the current value of the current step.

[0114] 2) Cell voltage measurement value: u i-2 ,u i-1 ,u i Where, the subscript i represents the time step, u i-2 Indicates the cell voltage measurement value of the previous 2 steps, u i-1 Indicates the cell voltage measurement value of the previous step, u iIndicates the cell voltage measurement value of the current step.

[0115] 3) Cell voltage simulation value: Where, the subscript i represents the time step, Indicates the simulated value of the cell voltage in the previous 2 steps. Indicates the simulated value of the cell voltage in the previous step. Indicates the cell voltage simulation value of the current step.

[0116] 4) Battery cell temperature measurement value: T i .

[0117] 5) Parameter values ​​of simulation model: MOCV i , R 0i , R 1i , C 1i (Take a set of RC equivalent circuit models as an example), M represents the simulation model, MOCV represents the OCV used in the simulation model, R 0i , R 1i , C 1i Represents the resistance and capacitance values ​​in the simulation model.

[0118] 6) Actual SOC of the battery cell: SOC i ,

[0119] 7) Actual open circuit voltage (OCV) of the cell: OCV i =f(SOC i ), where f() represents OCV as a function of SOC, which can be obtained based on cell test measurements and is not limited here.

[0120] Among them, the model parameter adjustment amount Action may include: ▽MOCVi, ▽R0i, ▽R1i, ▽C1i (taking a group of RC equivalent circuit models as an example); among them, ▽ represents the change value relative to the original model parameter based on the current state State, that is, the adjustment amount.

[0121] Wait for a time step. During the waiting period, the battery simulation model updates its state. The battery physical model runs for the same amount of time and then collects the updated state. + .

[0122] The specific method of updating the simulation model state is to use an enhanced Thevenin equivalent circuit model (such as Figure 4 As shown in the figure, the following is an example:

[0123] According to the following formula for calculating the cell voltage, after inputting the current I, the cell voltage can be calculated:

[0124]

[0125] In the formula Represents the voltage across resistor R1. Indicates the simulated value of the voltage across the battery cell, corresponding to the measurable voltage u across the battery cell.

[0126] After discretization, we get:

[0127]

[0128] Where Ts is the discretized sampling time.

[0129] The physical state of the battery cell is updated, and the updated actual state parameters of the battery cell are obtained, that is, the updated measurement values ​​are obtained. Specifically, the actual battery cell sampling can be performed according to the sampling time interval Ts calculated by the simulation model, so that the simulation model and the actual battery cell can run synchronously.

[0130] According to the current state State at time t, the model parameter adjustment Action and the state State of the next time step after the update + , the evaluation model outputs the evaluation score Reward.

[0131] Among them, the evaluation score Reward is used to represent the state State of the simulation system in the next time step under the action of the model parameter adjustment amount Action when the current state State + validity, and other demands on model or system performance.

[0132] In some embodiments, determining an evaluation result based on the simulation values, measurement values, and model parameter adjustment amounts before and after the update may specifically include:

[0133] The evaluation result is calculated based on the cell voltage simulation values, cell voltage test values, open circuit voltage simulation values, open circuit voltage test values, and upper and lower limits of the cell voltage before and after three adjacent updates.

[0134] For example, the evaluation score Reward can be calculated by the following formula:

[0135]

[0136] In the formula, γ1, γ2 and γ3 represent the influence weights of the corresponding deviation items on the evaluation score Reward, u lim Indicates the upper and lower limits of the cell voltage, used to ensure the simulated cell voltage Within a certain voltage range.

[0137] In the embodiment of the present disclosure, multiple groups (Statei , Action i , State i + , Reward i ), training and updating the parameters of the parameter adjustment agent so that in the current state State, the model parameter adjustment amount Action output by the parameter adjustment agent can correspond to the maximum evaluation score Reward.

[0138] Exemplarily, the parameter adjustment agent itself is also a model containing adjustable parameters θ, such as a deep learning model, such as a DNN or a long short-term memory (LSTM) recursive neural network model, etc., which is not limited here.

[0139] For example, the method for updating the adjustable parameter θ may adopt a policy gradient method, which may be implemented according to the following formula: θ←θ+η▽Rθ.

[0140] Where η represents the ratio of each parameter update θ, and ▽Rθ represents the effect of the change in the adjustable parameter θ on the reward change, which can be calculated as follows:

[0141]

[0142] Where p θ Indicates that in State i When an action is generated i The probability can be obtained by running the following data set multiple times.

[0143] (State1, Action1, State1 + , Reward1)

[0144] (State2, Action2, State2 + , Reward2)

[0145] (State3, Action3, State3 + , Reward3)

[0146] (State4, Action4, State4 + , Reward4)

[0147] (……,……,……,……)

[0148] (State i , Action i , State i+ , Reward i )

[0149] Alternatively, ▽Rθ can also be calculated directly according to the gradient descent formula:

[0150]

[0151] In other implementations, ▽Rθ may also be obtained in other ways, which are not limited here.

[0152] implement Figure 3 The cyclic process shown can obtain the parameter adjustment agent corresponding to the model parameter adjustment amount that satisfies the output and obtains the target evaluation result, and can be applied to the battery simulation method.

[0153] In some embodiments, the parameter adjustment method of the battery simulation model further includes:

[0154] Obtain the voltages of all cells in the battery pack and determine the cells with the maximum cell voltage, minimum cell voltage, and average cell voltage as the collected cells.

[0155] Among them, based on the cell voltage in the battery pack, the collection cell for collecting training data is determined. Specifically, by measuring the voltage of all cells in the battery pack, the cells with the maximum cell voltage, the minimum cell voltage and the average cell voltage are selected as feature cells and used as the development object for online data training. It can be understood that the cell voltage will change with the operation of the battery pack. In each time step, the feature cells are determined based on the measured voltage and learned. The feature cells in different time steps can be the same cell or different cells, which is not limited here.

[0156] Among them, the cell voltage can represent the actual state of the cell, and the data of abnormal cells with maximum and minimum cell voltages can be modeled, which increases the coverage of training data and improves the accuracy of parameter adjustment intelligent agents and simulation models.

[0157] Based on this, the parameter adjustment agent is trained, which may specifically include: for each type of collected battery cell, the parameter adjustment agent is trained based on the collected battery cell respectively.

[0158] The same type of collected cells refers to a type of cells with the maximum cell voltage, or a type of cells with the minimum cell voltage, or a type of cells with the average cell voltage.

[0159] In the disclosed embodiment, based on all the measured battery cell voltages, the above-mentioned parameter adjustment intelligent agent training is performed on the battery cell with the largest measured battery cell voltage, the battery cell with the smallest measured battery cell voltage, and the battery cell with the average battery cell voltage. For the same type of battery cells, one parameter adjustment intelligent agent is used for training respectively, so that the trained parameter adjustment intelligent agent can be applicable to the simulation of different types of battery cells, thereby improving the simulation accuracy.

[0160] In some embodiments, the battery pack includes a plurality of different battery packs disposed on different carriers, and the carrier includes at least a vehicle.

[0161] The battery pack can be used in the carrier as a power component, providing power for the carrier's operation. For example, the battery pack can serve as the vehicle's power source, providing power for driving or other actions, thereby meeting the vehicle's operational needs. In other embodiments, the carrier may also include other structural components, which are not limited here.

[0162] Based on this, the parameter adjustment agent is trained based on the collected cells, which may include:

[0163] The parameter adjustment agent is trained based on multiple different acquisition cells of the same type in multiple different battery packs to obtain multiple parameter adjustment values; the parameter adjustment values ​​in this embodiment are used to adjust and update the parameters of the parameter adjustment agent itself;

[0164] Multiple parameter adjustment amounts are superimposed, and the parameters of the parameter adjustment agent itself are adjusted based on the superimposed parameter adjustment amounts.

[0165] In this disclosed embodiment, for the same type of battery cell, all vehicles use a single parameter adjustment agent to adjust model parameters. The learning results of all parameter adjustment agents are summed together, thereby increasing the number of battery cells by increasing the amount of training data, shortening the training cycle and improving training efficiency.

[0166] The specific training process can be implemented in the cloud or on the vehicle side, which is not limited here.

[0167] In the parameter adjustment method of the battery simulation model provided in the embodiment of the present disclosure, a parameter adjustment intelligent agent is constructed based on reinforcement learning to automatically adjust the parameters of the battery cell model. The parameter adjustment intelligent agent can output the numerical value that needs to be adjusted for the simulation model parameters based on different working conditions, that is, output the model parameter adjustment amount, so that the battery cell model is more closely matched with the working conditions after being adjusted according to the model parameter adjustment amount, so that the output of the battery cell model can accurately characterize the working state of the battery cell, and then the simulation model can output the performance parameters of the battery that matches the working conditions.

[0168] Based on the above embodiments, an embodiment of the present disclosure further provides a battery simulation method, which may include the steps of any parameter adjustment method of the battery simulation model provided in the above embodiments, and has corresponding beneficial effects, which will not be repeated here.

[0169] For example, Figure 5 A flow chart of a battery simulation method provided by an embodiment of the present disclosure. Figure 5 , the battery simulation method may include the following steps:

[0170] S410: Execute a parameter adjustment step of a battery simulation model.

[0171] The model parameter adjustment amount for the simulation model is obtained.

[0172] S420 , simulating the battery based on the simulation model after parameter adjustment to obtain simulation performance parameters.

[0173] Specifically, the model parameters of the simulation model are adjusted and updated using the model parameter adjustment amount to obtain a simulation model that matches the operating conditions; the battery is simulated using the simulation model that matches the operating conditions after the model parameter adjustment and update to obtain simulation performance parameters suitable for the current operating conditions, thereby improving the simulation accuracy.

[0174] In some embodiments, the battery simulation method may further include:

[0175] The aging state of the battery is determined based on the model parameter adjustment amount and / or the adjusted model parameter.

[0176] Among them, the model parameters can characterize the status of different components in the model, and the overall aging status of the battery can be known based on the status of different components in the model.

[0177] For example, the cell status, parameter adjustment agent, and evaluation model can be set up in the battery management system and run continuously as the battery operates. As the battery ages, the model parameter adjustment output by the parameter adjustment agent will also change, thereby determining the actual aging state of the battery.

[0178] In the embodiments of the present disclosure, the aging state of the battery can be determined based on the model parameter adjustment amount; or, the aging state of the battery can be determined based on the model parameters after adjustment; or, the aging state of the battery can be determined based on the model parameter adjustment amount and the model parameters after adjustment, thereby achieving simulation verification of the battery aging state.

[0179] On the basis of the above embodiments, the embodiment of the present disclosure further provides a parameter adjustment device for a battery simulation model, which can execute the steps of any parameter adjustment method provided in the above embodiments and has corresponding beneficial effects. For details, please refer to the above and will not be repeated here.

[0180] For example, Figure 6 This is a schematic diagram of the structure of a battery simulation model parameter adjustment device provided by an embodiment of the present disclosure. Figure 6 The parameter adjustment device 60 may include: an acquisition module 610, used to obtain the parameter adjustment agent, model parameters of the simulation model and the operating parameters of the battery; a determination module 620, used to determine the model parameter adjustment amounts corresponding to the model parameters and the operating parameters based on the parameter adjustment agent; an adjustment module 630, used to adjust the model parameters of the simulation model based on the model parameter adjustment amounts; wherein, the simulation model after adjusting the model parameters is used for battery simulation.

[0181] In some embodiments, the acquisition module 610 is used to acquire a parameter adjustment agent, which may specifically include the acquisition module 610 being used to: construct a simulation model based on an actual battery pack; the battery pack includes multiple battery cells, and the simulation model correspondingly includes multiple battery cell models, and the battery cell model has adjustable model parameters; construct a parameter adjustment agent to be trained based on a reinforcement learning algorithm; the input of the parameter adjustment agent includes the model parameters of the simulation model, the input and output of the simulation model, and the evaluation results provided by the evaluation model, the input of the evaluation model includes simulation values, measurement values, and model parameter adjustment amounts, and the output of the parameter adjustment agent includes the model parameter adjustment amounts of the simulation model; run the battery pack, and the simulation model performs a synchronous status update, and determines the evaluation result based on the simulation values, measurement values, and model parameter adjustment amounts before and after the update; train the parameter adjustment agent based on multiple groups of simulation values, measurement values, model parameter adjustment amounts, and evaluation results before and after the update, and determine that the parameter adjustment agent corresponding to the output model parameter adjustment amount when the target evaluation result is obtained is the parameter adjustment agent.

[0182] In some embodiments, the parameter adjustment device of the battery simulation model may also include: a cell acquisition module, used to obtain the voltages of all cells in the battery pack, and determine the cells with the maximum cell voltage, the minimum cell voltage and the average cell voltage as the collection cells; wherein, the acquisition module 610 is used to train the parameter adjustment intelligent agent, which may specifically include: for each type of collection cell, training the parameter adjustment intelligent agent based on the collection cell.

[0183] In some embodiments, the battery pack includes multiple different battery packs configured on different carriers, and the carriers include at least vehicles; the acquisition module 610 is used to train the parameter adjustment agent based on the acquisition cells, which may specifically include: training the parameter adjustment agent based on multiple different acquisition cells of the same type in multiple different battery packs to obtain multiple parameter adjustment amounts; superimposing the multiple parameter adjustment amounts, and adjusting the parameter adjustment agent's own parameters based on the superimposed parameter adjustment amounts.

[0184] In some embodiments, the acquisition module 610 is used to determine the evaluation results based on the simulation values, measurement values ​​and model parameter adjustment values ​​before and after the update, which may specifically include: calculating the evaluation results based on the cell voltage simulation values, cell voltage test values, open circuit voltage simulation values, open circuit voltage test values ​​and upper and lower limits of the cell voltage before and after three adjacent updates.

[0185] In some embodiments, the adjustment module 630 is used to adjust the model parameters of the simulation model based on the model parameter adjustment amount, which may specifically include: for the model parameters of the simulation model, adding the model parameter adjustment amount to the current model parameters to obtain the adjusted model parameters.

[0186] Based on the above-mentioned implementation manner, the embodiment of the present disclosure also provides a battery simulation system, including a simulation model and a parameter adjustment intelligent agent configured in the simulation model; the parameter adjustment intelligent agent is used to determine the model parameter adjustment amount of the simulation model based on at least the operating parameters of the battery; the simulation model is used to adjust the model parameters based on the model parameter adjustment amount, and the simulation model after adjusting the model parameters is used for battery simulation.

[0187] For example, Figure 7 This is a schematic diagram of the structure of a battery simulation system provided by an embodiment of the present disclosure. Figure 7 The battery simulation system 70 includes a simulation model 001 and a parameter adjustment agent 002 configured in the simulation model 002 .

[0188] For example, simulation model 001 may be a battery cell model. When simulating the battery cell model, the parameter adjustment agent and the battery cell model together constitute a simulation system. Parameter adjustment agent 002 can output model parameter adjustment values ​​based on operating condition parameters / scenario parameters to adjust the parameters of simulation model 001. Simulation model 001 performs numerical calculation simulation based on the system after the model parameters have been changed, thereby obtaining performance indicators such as the simulated voltage, internal resistance, SOH, and SOC of the battery cell. In some embodiments, if the battery cell is damaged, the performance indicators may also use the aforementioned performance indicators, but the numerical values ​​of the performance indicators will be different.

[0189] For example, when applying the simulation system for battery simulation, the set current I, temperature and other data representing the working conditions / scenario can be input, and the parameter adjustment intelligent agent 002 outputs the model parameter adjustment amount. The battery cell model can then be simulated according to the adjusted parameters and output performance indicators such as battery cell voltage, battery cell internal resistance, and battery charge SOC.

[0190] In some embodiments, the simulation model is constructed based on an actual battery pack, which includes multiple battery cells. The simulation model correspondingly includes multiple battery cell models, and the battery cell models have adjustable model parameters.

[0191] The parameter adjustment agent is constructed and trained based on a reinforcement learning algorithm. The input of the parameter adjustment agent includes the model parameters of the simulation model, the input and output of the simulation model, and the evaluation results provided by the evaluation model. The input of the evaluation model includes simulation values, measurement values, and model parameter adjustment amounts. The output of the parameter adjustment agent includes the model parameter adjustment amounts of the simulation model.

[0192] Among them, based on the battery pack operation, the simulation model is updated synchronously to obtain the simulation values, measurement values ​​and model parameter adjustment amounts before and after the update, and are used to determine the evaluation results; the parameter adjustment agent is trained based on multiple groups of simulation values, measurement values, model parameter adjustment amounts and evaluation results before and after the update, and the parameter adjustment agent corresponding to the output model parameter adjustment amount when the target evaluation result is obtained is determined to be the parameter adjustment agent.

[0193] In some embodiments, the training of the parameter adjustment agent further includes: for each type of collected battery cell, training the parameter adjustment agent based on the collected battery cell respectively;

[0194] The collected cells include the cells with the maximum cell voltage, the minimum cell voltage and the average cell voltage among all the cell voltages in the battery pack.

[0195] In some embodiments, the battery pack includes a plurality of different battery packs configured on different carriers, the carrier including at least a vehicle;

[0196] The parameter adjustment agent is trained based on the collected battery cells, including: training the parameter adjustment agent based on multiple different collected battery cells of the same type in multiple different battery packs to obtain multiple parameter adjustment amounts; superimposing the multiple parameter adjustment amounts, and adjusting the parameters of the parameter adjustment agent itself based on the superimposed parameter adjustment amounts.

[0197] In some embodiments, the evaluation result is calculated based on the cell voltage simulation value, cell voltage test value, open circuit voltage simulation value, open circuit voltage test value, and upper and lower limits of the cell voltage before and after three adjacent updates.

[0198] In some embodiments, adjusting the model parameters of the simulation model includes: adding a model parameter adjustment amount to the model parameters of the simulation model on the basis of the current model parameters to obtain the adjusted model parameters.

[0199] In the above embodiment, the contents corresponding to the parameter adjustment method of the simulation model and the battery simulation method in the battery simulation system can be understood by referring to the above text and will not be repeated here.

[0200] The simulation system provided by the embodiment of the present disclosure is a variable parameter simulation system, which can realize battery model parameter estimation under multiple working conditions / scenarios and long life cycles, wherein the parameter adjustment intelligent agent 002 can output the numerical size of the model parameters of the simulation model 001 that need to be adjusted according to different working conditions, that is, output the model parameter adjustment amount, so as to optimize the model parameters of the simulation model, so that the output of the simulation model can accurately simulate the working state of the battery under different working conditions.

[0201] The simulation system provided by the embodiments of the present disclosure also has the following advantages:

[0202] 1) Dynamic adjustment capability: Ability to dynamically adjust the parameters of the simulation model according to real-time working conditions, realize variable parameter simulation, and improve simulation prediction accuracy.

[0203] 2) Continuous Improvement: As training data accumulates, the accuracy of the simulation system will continue to improve. Specifically, the parameters of agents for the same type of battery cell can be summed together. This means that the same agent can be continuously trained based on data from different scenarios, different vehicles, and different battery cells to continuously improve accuracy.

[0204] 3) Enhanced system adaptability and robustness: The intelligent agent has memory and reasoning capabilities. Even for scenarios that have not occurred, the intelligent agent will have relatively accurate output results, thereby improving the system's adaptability and robustness for different scenarios.

[0205] 4) Reduced maintenance costs and extended service life: Accurate parameter estimation helps to provide early warning of potential failures and optimize charging strategies, thereby reducing unnecessary maintenance work and extending battery life.

[0206] 5) Assemble a battery pack model based on the cell model: Take the cell model as the smallest simulation unit, and compare it with the cell connection relationship in the actual battery pack to obtain the simulation results of the battery pack model, while improving the simulation accuracy of the battery pack and realizing independent management of the cells.

[0207] 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 entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. The terms "comprise", "include" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises a..." does not exclude the presence of other identical elements in the process, method, article or device that includes the element.

[0208] The foregoing description is intended only to provide specific embodiments of the present disclosure, intended to enable those skilled in the art to understand and implement the present disclosure. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present disclosure. Therefore, the present disclosure is not intended to be limited to the embodiments described herein, but rather to be construed in the broadest manner consistent with the principles and novel features disclosed herein.

Claims

1. A battery simulation system, characterized in that: It includes a simulation model and a parameter adjustment agent configured in the simulation model; The parameter adjustment agent is used to determine the model parameter adjustment amount of the simulation model based on at least the operating parameters of the battery; The simulation model is used to adjust model parameters based on the model parameter adjustment amount, and the simulation model after adjusting the model parameters is used for battery simulation.

2. The battery simulation system according to claim 1, characterized in that: The simulation model is constructed based on an actual battery pack, the battery pack includes a plurality of battery cells, and the simulation model correspondingly includes a plurality of battery cell models, each of which has adjustable model parameters; The parameter adjustment agent is constructed and trained based on a reinforcement learning algorithm, wherein the input of the parameter adjustment agent includes model parameters of the simulation model, input and output of the simulation model, and evaluation results provided by an evaluation model, wherein the input of the evaluation model includes simulation values, measurement values, and model parameter adjustment amounts, and the output of the parameter adjustment agent includes the model parameter adjustment amounts of the simulation model; wherein, based on the operation of the battery pack, the simulation model is synchronously updated to obtain simulation values, measurement values, and model parameter adjustment amounts before and after the update, and to determine the evaluation result; The parameter adjustment agent is trained based on multiple sets of simulation values, measurement values, model parameter adjustment amounts and evaluation results before and after the update, and the parameter adjustment agent corresponding to the output model parameter adjustment amount that obtains the target evaluation result is determined to be the parameter adjustment agent.

3. The battery simulation system according to claim 2, characterized in that: The training of the parameter adjustment agent further includes: for each type of collection cell, training the parameter adjustment agent based on the collection cell respectively; The collected cells include cells with the maximum cell voltage, the minimum cell voltage and the average cell voltage among all cell voltages in the battery pack.

4. The battery simulation system according to claim 3, characterized in that: The battery pack includes a plurality of different battery packs configured on different carriers, and the carrier includes at least a vehicle; The parameter adjustment agent is trained based on the acquisition cell, including: training the parameter adjustment agent based on multiple different acquisition cells of the same type in multiple different battery packs to obtain multiple parameter adjustment amounts; The multiple parameter adjustment amounts are superimposed, and the parameters of the parameter adjustment agent itself are adjusted based on the superimposed parameter adjustment amounts.

5. The battery simulation system according to claim 2, characterized in that: The evaluation result is calculated based on the cell voltage simulation value, cell voltage test value, open circuit voltage simulation value, open circuit voltage test value and upper and lower limits of the cell voltage before and after three adjacent updates.

6. The battery simulation system according to claim 1, characterized in that: The adjustment of the model parameters of the simulation model includes: adding the model parameter adjustment amount to the current model parameters to obtain the adjusted model parameters.

7. A parameter adjustment method for a battery simulation model, characterized in that: include: Obtaining the parameter adjustment agent, model parameters of the simulation model, and operating parameters of the battery; Determining, based on the parameter adjustment agent, model parameter adjustment amounts corresponding to the model parameters and the operating condition parameters; adjusting the model parameters of the simulation model based on the model parameter adjustment amount; The simulation model after adjusting the model parameters is used for battery simulation.

8. A battery simulation method, characterized in that: The method comprises the steps of adjusting the parameters of the battery simulation model according to claim 7; The battery simulation method further includes: The battery is simulated based on the simulation model after parameter adjustment to obtain simulation performance parameters.

9. The battery simulation method according to claim 8, characterized in that: Also includes: The aging state of the battery is determined based on the model parameter adjustment amount and / or the adjusted model parameter.

10. A parameter adjustment device for a battery simulation model, characterized in that: include: An acquisition module is used to obtain the model parameters of the parameter adjustment agent, the simulation model, and the operating parameters of the battery; a determination module, configured to determine, based on the parameter adjustment agent, a model parameter adjustment amount corresponding to the model parameter and the operating condition parameter; An adjustment module is used to adjust the model parameters of the simulation model based on the model parameter adjustment amount; wherein the simulation model after adjusting the model parameters is used for battery simulation.

Images