Electrochemical model parameter identification method and device based on prediction model
By adopting an iterative update method based on prediction models, the problems of long time consumption and wasted computing power in electrochemical model parameter identification are solved, and fast and accurate electrochemical model parameter identification is achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SHANGHAI MAKESENS ENERGY STORAGE TECH CO LTD
- Filing Date
- 2023-04-14
- Publication Date
- 2026-05-15
AI Technical Summary
Existing methods for identifying parameters of electrochemical models are time-consuming, require a large amount of data, and involve repeated calls to the electrochemical model, resulting in a waste of computing power.
A prediction model-based approach is adopted. By acquiring the training condition dataset, the prediction model generates a prediction model parameter set, and iteratively updates the electrochemical model parameter set until the verification is successful, thus generating the target electrochemical model parameter set.
It reduces the number of times the electrochemical model is called during parameter identification, reduces time consumption, improves the generalization ability of the prediction model parameter set, and can accurately identify electrochemical model parameters under conditions of limited data.
Smart Images
Figure CN116611311B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of electrochemical modeling, and more particularly to a method and apparatus for identifying electrochemical model parameters based on a prediction model. Background Technology
[0002] By establishing an electrochemical model of lithium batteries, we can obtain simulated values of the physicochemical state quantities in the space and time inside the battery. This allows us to more clearly understand and monitor the real-time operating status of the battery, thereby better ensuring the economy, reliability, and safety of lithium-ion batteries.
[0003] Electrochemical models involve numerous coupled partial differential equations, particularly dozens of physical parameters, which limits their practical application. However, with advancements in hardware capabilities, model parameters can be obtained in a data-driven manner using methods such as heuristic algorithms, neural networks, and Kalman filtering.
[0004] Existing data-driven parameter identification methods mainly suffer from the problem of long parameter identification time, often taking several hours. They also require preset battery parameter ranges, as well as charging and discharging data of the battery under different operating conditions (starting from the same state). Furthermore, each identification requires repeated calls to the electrochemical model, resulting in a significant waste of computing power. Summary of the Invention
[0005] To address the aforementioned technical problems, this invention provides a method and apparatus for identifying electrochemical model parameters based on a prediction model.
[0006] Specifically, the technical solution of the present invention is as follows:
[0007] On the one hand, the present invention provides a method for identifying electrochemical model parameters based on a prediction model, comprising:
[0008] Obtain a training operating condition dataset, which includes several sets of operating condition data for the same type of battery cell;
[0009] The training condition dataset is input into the prediction model of the preset electrochemical model parameters to generate the prediction model parameter set of the prediction model. The prediction model is used to generate the corresponding electrochemical model parameter set according to each group of training condition data.
[0010] The electrochemical model verification dataset is input into the preset electrochemical model, and the verification voltage corresponding to each set of the electrochemical model verification dataset is calculated. The electrochemical model verification dataset includes each set of electrochemical model parameters generated according to the prediction model parameter set and its corresponding training condition data.
[0011] Based on the verification voltage of each group and the operating voltage in the corresponding training operating condition data, the electrochemical model parameter set corresponding to the verification voltage of each group is verified.
[0012] If any of the electrochemical model parameter sets fails the verification, the prediction model parameter set is iteratively updated until all the electrochemical model parameter sets generated based on the current prediction model parameter set pass the verification, and then the iteratively updated prediction model parameter set is output.
[0013] Based on the training dataset, the prediction model, and the iteratively updated prediction model parameter set, the target electrochemical model parameter set is generated for the electrochemical model.
[0014] In some implementations, obtaining the training condition dataset includes:
[0015] Obtain the actual working condition dataset and normalize each actual working condition data in the actual working condition dataset to obtain the training working condition dataset, specifically including:
[0016] The actual operating condition data is normalized according to the following formula:
[0017]
[0018] Where x represents the actual working condition data, μ is the data mean, and σ is the data standard deviation. * The training data is the normalized version.
[0019] In some embodiments, the step of verifying the electrochemical model parameter set corresponding to each set of verification voltages based on the operating condition voltages in the corresponding training operating condition data includes:
[0020] Calculate the voltage loss function based on the verification voltage and the corresponding operating voltage of each group;
[0021] The electrochemical model parameter set is validated based on the voltage loss function.
[0022] In some implementations, calculating the voltage loss function based on each group of verification voltages and their corresponding operating voltages includes:
[0023] The voltage loss function MSELoss is calculated according to the following formula:
[0024]
[0025] Among them, v cell,i The operating voltage is v. sim,iThe verification voltage is N, where N is the number of voltages and i is the sequence number.
[0026] Wherein, is the operating voltage, is the verification voltage, N is the number of voltages, and i is the sequence number.
[0027] In some implementations, the iterative update of the prediction model parameter set includes updating the prediction model parameter set using a heuristic algorithm.
[0028] On the other hand, the present invention also provides an electrochemical model parameter identification device based on a prediction model, comprising:
[0029] The acquisition module is used to acquire the training operating condition dataset, which includes several sets of operating condition data for the same type of battery cell.
[0030] The prediction module is used to input the training condition dataset into the prediction model of the preset electrochemical model parameters to generate the prediction model parameter set of the prediction model. The prediction model is used to generate the corresponding electrochemical model parameter set according to each group of training condition data.
[0031] The calculation module is used to input the electrochemical model verification dataset into the preset electrochemical model and calculate the verification voltage corresponding to each set of the electrochemical model verification dataset. The electrochemical model verification dataset includes each set of electrochemical model parameters generated according to the prediction model parameter set and its corresponding training condition data.
[0032] The verification module is used to verify the electrochemical model parameter set corresponding to each set of verification voltages based on the verification voltages of each set and the operating voltages in the corresponding training operating condition data.
[0033] The update module is used to iteratively update the prediction model parameter set when any of the electrochemical model parameter sets fails the verification, until all the electrochemical model parameter sets generated based on the current prediction model parameter set pass the verification, and then output the iteratively updated prediction model parameter set.
[0034] The identification module is used to generate the target electrochemical model parameter set of the electrochemical model based on the training condition dataset, the prediction model, and the iteratively updated prediction model parameter set.
[0035] In some embodiments, the acquisition module includes:
[0036] The acquisition unit is used to acquire actual working condition datasets.
[0037] The normalization unit is used to normalize each actual working condition data in the actual working condition dataset to obtain the training working condition dataset.
[0038] The normalization unit normalizes the actual operating condition data according to the following formula:
[0039]
[0040] Where x represents the actual working condition data, μ is the data mean, and σ is the data standard deviation. * The training condition data is the normalized version;
[0041] In some implementations, the verification module includes:
[0042] The calculation unit is used to calculate the voltage loss function based on the verification voltage of each group and its corresponding operating condition voltage;
[0043] A verification unit is used to verify the electrochemical model parameter set based on the voltage loss function.
[0044] In some implementations, the calculation unit calculates the voltage drop function MSELoss according to the following formula:
[0045]
[0046] Among them, v cell,i The operating voltage is v. sim,i The verification voltage is N, where N is the number of voltages and i is the sequence number.
[0047] The present invention also discloses a readable storage medium having a computer program stored thereon, characterized in that, when the computer program is executed by a processor, it implements the steps of the electrochemical model parameter identification method based on the prediction model as described in any of the preceding claims.
[0048] Compared with the prior art, the present invention has at least one of the following beneficial effects:
[0049] (1) This invention is based on a prediction model. It predicts electrochemical model parameters using the acquired training condition dataset, inputs the predicted values into a preset electrochemical model to obtain a verification voltage, and verifies the accuracy of the prediction model parameter set based on this verification voltage and the operating condition voltages in the training condition dataset. If the verification fails, the prediction model parameter set is iteratively updated and used in subsequent electrochemical model parameter identification. This invention helps reduce the number of times the electrochemical model is called during parameter identification, avoiding wasted computing power and significantly reducing the time spent on parameter identification. More importantly, by optimizing the prediction model parameter set, this invention improves the generalization ability of the prediction model parameter set, enabling accurate identification of electrochemical model parameter sets even with limited data.
[0050] (2) This invention does not require that several sets of operating condition data of the same type of battery cell in the training operating condition dataset be charging and discharging data under the same starting state and different operating conditions. Therefore, this invention does not need to obtain a large dataset in advance, further reducing the time for parameter identification.
[0051] (3) The present invention uses a heuristic algorithm to optimize and update the identified electrochemical model parameters, which can solve the problem that the electrochemical model is difficult to backpropagate. Attached Figure Description
[0052] The preferred embodiments will now be described in a clear and easy-to-understand manner, in conjunction with the accompanying drawings, to further explain the above-mentioned characteristics, technical features, advantages, and implementation methods of the present invention.
[0053] Figure 1 This is a flowchart of an embodiment of the parameter identification method for the lithium battery electrochemical model of the present invention;
[0054] Figure 2 This is a schematic diagram of the parameter identification process of another embodiment of the parameter identification method for the lithium battery electrochemical model of the present invention;
[0055] Figure 3 This is a schematic diagram of an embodiment of the parameter identification device for the lithium battery electrochemical model of the present invention.
[0056] Explanation of icon numbers:
[0057] The module consists of: acquisition module 10, prediction module 20, calculation module 30, verification module 40, update module 50, and identification module 60. Detailed Implementation
[0058] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application can also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.
[0059] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the specific implementation methods of the present invention will be described below with reference to the accompanying drawings. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings and other implementation methods can be obtained based on these drawings without any creative effort.
[0060] To keep the drawings concise, each figure only schematically shows the parts relevant to the invention, and these do not represent the actual structure of the product. Furthermore, to facilitate understanding, in some figures, only one of components with the same structure or function is schematically depicted, or only one is labeled. In this document, "one" not only means "only one," but can also mean "more than one."
[0061] It should be understood that, when used in this specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or sets.
[0062] Furthermore, in the description of this application, the terms "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.
[0063] One embodiment of the present invention, such as Figure 1 As shown, a method for identifying electrochemical model parameters based on a prediction model includes the following steps:
[0064] S100 acquires the training operating condition dataset, which includes several sets of operating condition data for the same type of battery cell.
[0065] Specifically, training data sets are obtained from actual operating conditions of energy storage power stations or batteries; "same model" refers to the same product model, and each set of data includes operating conditions data of the same cell under different charge and discharge modes. It is important to note that these operating conditions data do not need to be obtained under the same initial conditions, thus avoiding the problem of excessively long acquisition times for training datasets.
[0066] S200 inputs the training condition dataset into the prediction model of the preset electrochemical model parameters to generate the prediction model parameter set of the prediction model. The prediction model is used to generate the corresponding electrochemical model parameter set based on each set of training condition data.
[0067] Specifically, training condition data for each set of battery cells of the same model are input into a pre-set prediction model to obtain an electrochemical model parameter set corresponding to each set of training condition data. This electrochemical model parameter set represents predicted values, which typically differ from actual values. The prediction model parameter set includes some variables within the prediction model, which are used for prediction.
[0068] The S300 inputs the electrochemical model verification dataset into the preset electrochemical model and calculates the verification voltage corresponding to each set of electrochemical model verification datasets. The electrochemical model verification dataset includes each electrochemical model parameter set generated based on the prediction model parameter set and its corresponding training condition data.
[0069] S400 verifies the electrochemical model parameter set corresponding to each set of verification voltages based on the verification voltages of each set and the corresponding operating voltages in the training operating condition data.
[0070] Specifically, a set of training condition data corresponds to an electrochemical model parameter set. The electrochemical model parameter sets obtained in step S200 and their corresponding training condition data are used as verification datasets to verify the predicted model parameter sets. Since the electrochemical model parameter sets generated in step S200 are predicted values, not actual values (i.e., the values to be identified), it is difficult to determine the error between the predicted and actual values. Therefore, this scheme uses the output voltage of the electrochemical model as the verification voltage, while the operating voltage in the training condition data is obtained from the actual operating condition data. This allows the error between the predicted and actual electrochemical model parameter values to be transformed into the error between the verification voltage and the operating voltage. The predicted electrochemical model parameter sets are then verified based on the verification voltage and its corresponding operating voltage. In this scheme, the output voltage value of the electrochemical model can be used for verification; however, current or temperature data can also be used.
[0071] When any electrochemical model parameter set fails verification, S500 iteratively updates the prediction model parameter set until all electrochemical model parameter sets generated based on the current prediction model parameter set pass verification, at which point it outputs the iteratively updated prediction model parameter set.
[0072] Specifically, during validation, if any electrochemical model parameter set fails validation, it indicates that the currently used prediction model parameter set is inaccurate and requires further iterative updates. After each update, the updated prediction model parameter set is used to re-input the training data from each set into the prediction model, resulting in several new electrochemical model parameter sets. These new sets are then validated. This process is repeated until all electrochemical model parameter sets pass validation, at which point the final prediction model parameter set is determined. When subsequent electrochemical model parameter identification is required, the iteratively updated prediction model parameters are directly used for identification.
[0073] S600 generates the target electrochemical model parameter set for the electrochemical model based on the training condition dataset, the prediction model, and the iteratively updated prediction model parameter set.
[0074] Specifically, when cleaning operating condition data from energy storage power stations or power batteries, multiple data points cleaned over a period of time may have similar values, resulting in a limited amount of data to be identified. To obtain a large amount of data, more time is needed to acquire the training operating condition dataset. In this solution, the iteratively updated prediction model parameter set has better generalization ability, improving the generalization ability of the prediction model parameter set. Even when the amount of data to be identified is small, it can still identify an accurate electrochemical model parameter set.
[0075] In this embodiment, based on a prediction model, electrochemical model parameters are predicted using the acquired training condition dataset. The predicted values are input into a preset electrochemical model to obtain a verification voltage. Based on this verification voltage and the operating voltages in the training condition dataset, the parameter set of the prediction model is verified. If verification fails, the prediction model parameter set is iteratively updated and used in subsequent electrochemical model parameter identification. Compared with conventional parameter identification methods, this approach reduces the number of times the electrochemical model is called, avoids wasting computational resources, and significantly reduces the time spent on parameter identification.
[0076] In one embodiment, step S100 includes:
[0077] Obtain the actual working condition dataset and normalize each actual working condition data in the dataset to obtain the training working condition dataset, which specifically includes:
[0078] The actual operating data is normalized using the following formula:
[0079]
[0080] Where x represents the actual operating data, μ is the data mean, and σ is the data standard deviation. * This is the normalized training data.
[0081] Specifically, data under battery operating conditions (i.e., current I is not zero) is extracted from the data of energy storage power stations or power batteries (including data on the relaxation process, i.e., data for a period of time after the current changes from non-zero to zero, specifically a few minutes to tens of minutes), including voltage (V), current (I), temperature (T), etc. Since different data have different units, and to accelerate the gradient descent solution during model training, this embodiment normalizes the above-mentioned actual operating condition data.
[0082] An embodiment of the present invention provides a method for identifying electrochemical model parameters based on a prediction model, comprising the following steps:
[0083] S100 acquires the training operating condition dataset, which includes several sets of operating condition data for the same type of battery cell.
[0084] S200 inputs the training condition dataset into the prediction model of the preset electrochemical model parameters to generate the prediction model parameter set of the prediction model. The prediction model is used to generate the corresponding electrochemical model parameter set based on each set of training condition data.
[0085] The S300 inputs the electrochemical model verification dataset into the preset electrochemical model and calculates the verification voltage corresponding to each set of electrochemical model verification datasets. The electrochemical model verification dataset includes each electrochemical model parameter set generated based on the prediction model parameter set and its corresponding training condition data.
[0086] Specifically, for example, the prediction model uses an LSTM (Long Short-Term Memory Recurrent Neural Network) model, such as... Figure 2 As shown, voltage, current, and temperature data under battery operating conditions are extracted from data from energy storage power stations or power batteries. The extracted battery data is preprocessed using an encoder component to obtain a training condition dataset. Then, each training condition dataset is converted into an encoded state with a fixed shape (i.e., the corresponding electrochemical model parameter set) through the hidden layers and fully connected layers of an LSTM. The electrochemical model parameter set is input into the electrochemical model layer in the decoder component to calculate the output voltage (i.e., the verification voltage) of the electrochemical model.
[0087] S410 calculates the voltage loss function based on each group of verification voltages and their corresponding operating voltages.
[0088] Specifically, the voltage drop function is calculated to determine the error between the verification voltage and the operating voltage. Preferably, the voltage drop function MSELoss is calculated using the following formula:
[0089]
[0090] Among them, v cell,i The operating voltage is v. sim,i For voltage verification, N is the number of voltages and i is the sequence number.
[0091] S420 verifies the parameter set of the electrochemical model based on the voltage loss function.
[0092] Specifically, it is determined whether the voltage loss function is within the preset range. If it is, the verification is considered to have passed; otherwise, the verification is considered to have failed.
[0093] When any electrochemical model parameter set fails verification, S500 iteratively updates the prediction model parameter set until all electrochemical model parameter sets generated based on the current prediction model parameter set pass verification. Then, it outputs the iteratively updated prediction model parameter set so that it can be directly called when identifying electrochemical model parameters through the prediction model next time.
[0094] Specifically, it is preferable to use heuristic algorithms (such as genetic algorithms, particle swarm optimization, etc.) to update the prediction model parameter set, which helps to solve the problem of backpropagation difficulties in electrochemical models. When all electrochemical model parameter sets have passed validation, predicting subsequent electrochemical model parameters based on the current prediction model parameter set can yield an accurate electrochemical model parameter set. Figure 2 As shown, the verification is performed based on the verification voltage and the corresponding operating condition voltage in the training operating condition dataset. When all electrochemical model parameter sets pass the verification, the training operating condition dataset is input into the prediction model based on the current prediction model parameter set, and the electrochemical model parameter set is directly output through the encoder component.
[0095] Compared to conventional data-driven approaches, this solution improves the generalization ability of the prediction model parameter set. For the same type of battery cell, after determining the final prediction model parameter set, the training operating condition dataset is input into the prediction model, directly outputting an accurate electrochemical model parameter set without repeatedly calling the electrochemical model. This accelerates the electrochemical model parameter identification process and improves parameter identification efficiency. Moreover, the electrochemical model parameter identification method used in this solution can identify the required electrochemical model parameters without disassembling the battery cell.
[0096] One embodiment of the present invention, such as Figure 3 As shown, an electrochemical model parameter identification device based on a prediction model is characterized by comprising an acquisition module 10, a prediction module 20, a calculation module 30, a verification module 40, and an update module 50, wherein:
[0097] The acquisition module 10 is used to acquire the training operating condition dataset, which includes several sets of operating condition data for the same type of battery cell.
[0098] The prediction module 20 is used to input the training condition dataset into the prediction model of the preset electrochemical model parameters to generate the prediction model parameter set of the prediction model. The prediction model is used to generate the corresponding electrochemical model parameter set based on each group of training condition data.
[0099] The calculation module 30 is used to input the electrochemical model verification dataset into the preset electrochemical model and calculate the verification voltage corresponding to each set of electrochemical model verification datasets. The electrochemical model verification dataset includes each electrochemical model parameter set generated based on the prediction model parameter set and its corresponding training condition data.
[0100] The verification module 40 is used to verify the electrochemical model parameter set corresponding to each set of verification voltages based on the operating voltages in each set of verification voltages and their corresponding training operating condition data.
[0101] The update module 50 is used to iteratively update the prediction model parameter set when any electrochemical model parameter set fails the verification, until all electrochemical model parameter sets generated based on the current prediction model parameter set pass the verification, and then output the iteratively updated prediction model parameter set.
[0102] The identification module 60 is used to generate the target electrochemical model parameter set of the electrochemical model based on the training condition dataset, the prediction model, and the iteratively updated prediction model parameter set.
[0103] In this embodiment, based on the prediction model, the parameters of the electrochemical model are predicted using the acquired training condition dataset. The predicted values are input into the preset electrochemical model to obtain the verification voltage. The parameter set of the prediction model is verified based on this verification voltage and the operating condition voltages in the training condition dataset. If the verification fails, the parameter set of the prediction model is iteratively updated and used in subsequent electrochemical model parameter identification. This approach helps reduce the number of times the electrochemical model is called, avoids wasting computational resources, and significantly reduces the time spent on parameter identification.
[0104] In one embodiment, the acquisition module 10 includes: an acquisition unit for acquiring an actual working condition dataset; and a normalization unit for normalizing each actual working condition data in the actual working condition dataset to obtain a training working condition dataset.
[0105] Specifically, the normalization unit normalizes the actual operating condition data according to the following formula:
[0106]
[0107] Where x represents the actual operating data, μ is the data mean, and σ is the data standard deviation. * This is the normalized training data.
[0108] In this embodiment, normalization is used to unify the units of different data, thereby reducing the complexity of calculations and accelerating the gradient descent solution during model training.
[0109] In one embodiment, the verification module 40 includes: a calculation unit for calculating a voltage drop function based on each set of verification voltages and their corresponding operating voltages; and a verification unit for verifying the electrochemical model parameter set based on the voltage drop function.
[0110] Specifically, the calculation unit calculates the voltage drop function MSELoss according to the following formula:
[0111]
[0112] Among them, v cell,i The operating voltage is v. sim,i For voltage verification, N is the number of voltages and i is the sequence number.
[0113] In this embodiment, the electrochemical model parameter set is verified by calculating the voltage loss function. The error between the electrochemical model parameter set and the actual electrochemical model parameter set is converted into the error between the verification voltage and the operating voltage, so as to verify the accuracy of the electrochemical model parameter set.
[0114] It should be noted that the embodiments of the electrochemical model parameter identification device based on the prediction model provided by this invention and the embodiments of the electrochemical model parameter identification method based on the prediction model provided above are all based on the same inventive concept and can achieve the same technical effect. Therefore, other specific details of the embodiments of the electrochemical model parameter identification device based on the prediction model can be referred to the description of the embodiments of the electrochemical model parameter identification method based on the prediction model.
[0115] Finally, this application also discloses a readable storage medium storing a computer program thereon, wherein the computer program, when executed, implements the steps of the electrochemical model parameter identification method based on a prediction model according to any of the above embodiments of this application. That is, when part or all of the technical solutions of the foregoing embodiments of the present invention that contribute to the prior art are embodied in the form of a computer software product, the aforementioned computer software product is stored in a readable storage medium. The readable storage medium can be any physical device or equipment capable of carrying computer program code, such as a USB flash drive, portable hard disk, magnetic disk, optical disk, computer memory, read-only memory, random access memory, etc.
[0116] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.
[0117] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.
[0118] Furthermore, the functional units in the various embodiments of this application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The integrated unit described above can be implemented in hardware or as a software functional unit.
[0119] It should be noted that the above embodiments can be freely combined as needed. The above are merely preferred embodiments of the present invention. It should be pointed out that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.
Claims
1. A method for identifying electrochemical model parameters based on a prediction model, characterized in that, include: Obtain a training operating condition dataset, which includes several sets of operating condition data for the same type of battery cell; The training condition dataset is input into the prediction model of the preset electrochemical model parameters to generate the prediction model parameter set of the prediction model. The prediction model is used to generate the corresponding electrochemical model parameter set according to each group of training condition data. The electrochemical model verification dataset is input into the preset electrochemical model, and the verification voltage corresponding to each set of the electrochemical model verification dataset is calculated. The electrochemical model verification dataset includes each set of electrochemical model parameters generated according to the prediction model parameter set and its corresponding training condition data. Based on the verification voltage of each group and the operating voltage in the corresponding training operating condition data, the electrochemical model parameter set corresponding to the verification voltage of each group is verified. If any of the electrochemical model parameter sets fails the verification, the prediction model parameter set is iteratively updated until all the electrochemical model parameter sets generated based on the current prediction model parameter set pass the verification, and then the iteratively updated prediction model parameter set is output. Based on the training condition dataset, the prediction model, and the iteratively updated prediction model parameter set, the target electrochemical model parameter set of the electrochemical model is generated. The step of verifying the electrochemical model parameter set corresponding to each set of verification voltages based on the verification voltages of each set and the corresponding operating condition voltages in the training operating condition data includes: Calculate the voltage loss function based on the verification voltage and the corresponding operating voltage of each group; The electrochemical model parameter set is validated based on the voltage loss function; The calculation of the voltage loss function based on the verification voltages of each group and their corresponding operating voltages includes: The voltage loss function is calculated using the following formula. : ； in, The operating voltage is... The verification voltage is N, where N is the number of voltages and i is the sequence number.
2. The method for identifying electrochemical model parameters based on a prediction model according to claim 1, characterized in that, The acquisition of the training condition dataset includes: Obtain the actual working condition dataset and normalize each actual working condition data in the actual working condition dataset to obtain the training working condition dataset, specifically including: The actual operating condition data is normalized according to the following formula: ； in, The actual working condition data is as described. The mean of the data. The standard deviation of the data. The training data is the normalized version.
3. The method for identifying electrochemical model parameters based on a prediction model according to any one of claims 1-2, characterized in that, The iterative update of the prediction model parameter set includes: The parameter set of the prediction model is updated using a heuristic algorithm.
4. A device for identifying electrochemical model parameters based on a prediction model, characterized in that, include: The acquisition module is used to acquire the training operating condition dataset, which includes several sets of operating condition data for the same type of battery cell. The prediction module is used to input the training condition dataset into the prediction model of the preset electrochemical model parameters to generate the prediction model parameter set of the prediction model. The prediction model is used to generate the corresponding electrochemical model parameter set according to each group of training condition data. The calculation module is used to input the electrochemical model verification dataset into the preset electrochemical model and calculate the verification voltage corresponding to each set of the electrochemical model verification dataset. The electrochemical model verification dataset includes each set of electrochemical model parameters generated according to the prediction model parameter set and its corresponding training condition data. The verification module is used to verify the electrochemical model parameter set corresponding to each set of verification voltages based on the verification voltages of each set and the operating voltages in the corresponding training operating condition data. The update module is used to iteratively update the prediction model parameter set when any of the electrochemical model parameter sets fails the verification, until all the electrochemical model parameter sets generated based on the current prediction model parameter set pass the verification, and then output the iteratively updated prediction model parameter set. The identification module is used to generate the target electrochemical model parameter set of the electrochemical model based on the training condition dataset, the prediction model, and the iteratively updated prediction model parameter set; The verification module includes: The calculation unit is used to calculate the voltage loss function based on the verification voltage of each group and its corresponding operating condition voltage; A verification unit is used to verify the electrochemical model parameter set according to the voltage loss function; The calculation unit calculates the voltage loss function according to the following formula. : ； in, The operating voltage is... The verification voltage is N, where N is the number of voltages and i is the sequence number.
5. The electrochemical model parameter identification device based on a prediction model according to claim 4, characterized in that, The acquisition module includes: The acquisition unit is used to acquire actual working condition datasets. The normalization unit is used to normalize each actual working condition data in the actual working condition dataset to obtain the training working condition dataset. The normalization unit normalizes the actual operating condition data according to the following formula: ； in, The actual working condition data is as described. The mean of the data. The standard deviation of the data. The training data is the normalized version.
6. A readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the method for identifying electrochemical model parameters based on a prediction model as described in any one of claims 1 to 3.