Lithium ion battery eigenmap construction method and system based on PINN
Through the PINN-based intrinsic map construction method of lithium-ion batteries, the parameters of electrolyte, positive and negative electrodes are calculated, neural network models are built, battery performance is predicted and two-dimensional maps are drawn, which solves the difficulties in performance and cost evaluation in the development of new lithium-ion batteries, and achieves rapid and economical material evaluation.
Patent Information
- Application Number
- CN202510092209.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-01-21
- Publication Date
- 2025-05-16
AI Technical Summary
When developing new lithium-ion batteries, the existing technology requires deductions, sample testing and pilot projects through a large amount of manpower and material costs, making it difficult to quickly judge the performance and economic benefits of new materials.
Using the PINN-based intrinsic map construction method, a neural network model is built for training by calculating the intrinsic parameters and physical parameters of the electrolyte, the positive electrode and the negative electrode, a neural network model is built to predict the battery performance parameters, and a two-dimensional intrinsic map is drawn through dimensionality reduction processing.
This method can quickly judge the performance improvement of new materials from the material and process level, and evaluate the economic benefits of development from the cost level, reducing manpower and material costs.
Smart Images
Figure CN120015139A_ABST
Abstract
Description
Technical Field
[0001] The present invention specifically relates to a method and system for constructing an intrinsic spectrum of a lithium-ion battery based on PINN. Background Art
[0002] For lithium-ion batteries, the development of new positive and negative electrode materials has always been a top priority. When developing a new lithium-ion battery, if you want to quickly know whether the current development is valuable and whether it can supplement and improve existing products in terms of performance and cost, you need to do battery charging, make battery samples, and conduct battery pilot tests, and finally judge based on the performance of the samples, which results in a lot of manpower and material costs. Summary of the invention
[0003] The present invention provides a method and system for constructing an intrinsic spectrum of a lithium-ion battery based on PINN to solve the above-mentioned technical problems, and specifically adopts the following technical solutions:
[0004] A method for constructing an intrinsic spectrum of a lithium-ion battery based on PINN, comprising:
[0005] Calculate the intrinsic parameters of the electrolyte, cathode, and anode;
[0006] Determine the physical parameters of the electrolyte, cathode, and anode;
[0007] Obtaining battery performance parameters based on the intrinsic parameters and the physical parameters;
[0008] Building a first neural network model, taking the intrinsic parameters and the physical parameters as inputs of the first neural network model, taking the battery performance parameters as outputs of the first neural network model, training the first neural network model, and finally constructing a PINN framework model;
[0009] Inputting the electrolyte, positive electrode and negative electrode combination to be predicted into the PINN framework model, and outputting the predicted performance parameters;
[0010] Calculate new intrinsic parameters for the electrolyte, cathode, and anode combinations to be predicted;
[0011] Identify new physical parameters of the electrolyte, cathode, and anode to be predicted;
[0012] Performing dimensionality reduction processing on the predicted performance parameters, new intrinsic parameters and new physical parameters;
[0013] The two-dimensional intrinsic map of lithium-ion batteries is drawn using the reduced dimension parameters.
[0014] Furthermore, the specific method of performing dimensionality reduction processing on the predicted performance parameters, new intrinsic parameters and new physical parameters is:
[0015] The predicted performance parameters, new intrinsic parameters and new physical parameters are input into a dimensionality reduction network composed of a first neural network model and an autoencoder for dimensionality reduction processing.
[0016] Furthermore, the PINN-based lithium-ion battery intrinsic spectrum construction method further comprises:
[0017] The dimension reduction network composed of the second neural network model and the autoencoder is trained using the intrinsic parameters, the physical parameters and the battery performance parameters.
[0018] Furthermore, the second neural network model is a CNN network.
[0019] Furthermore, the second neural network model is a fully connected neural network.
[0020] Furthermore, the PINN-based lithium-ion battery intrinsic spectrum construction method further comprises:
[0021] Obtaining cost parameters of the electrolyte, cathode and anode combination to be predicted;
[0022] A three-dimensional lithium-ion battery intrinsic spectrum is drawn based on the two-dimensional lithium-ion battery intrinsic spectrum and the cost parameter.
[0023] Furthermore, the specific method for calculating the intrinsic parameters of the electrolyte, the positive electrode and the negative electrode is:
[0024] The intrinsic parameters of the electrolyte, cathode, and anode were calculated based on density functional theory and molecular dynamics.
[0025] Furthermore, the specific method for obtaining the battery performance parameters based on the intrinsic parameters and the physical parameters is:
[0026] The intrinsic parameters and the physical parameters are input into a P2D model to obtain the battery performance parameters.
[0027] A PINN-based lithium-ion battery intrinsic spectrum construction system, comprising:
[0028] A calculation unit for calculating the intrinsic parameters of the electrolyte, the positive electrode and the negative electrode;
[0029] An acquisition unit, used to acquire physical parameters of the electrolyte, the positive electrode and the negative electrode;
[0030] A processing unit, configured to obtain a battery performance parameter based on the intrinsic parameter and the physical parameter;
[0031] A PINN framework model, using the intrinsic parameters and the physical parameters as inputs of a first neural network model, using the battery performance parameters as outputs of the first neural network model, training the first neural network model, and finally constructing the PINN framework model, inputting the electrolyte, positive electrode and negative electrode combination to be predicted into the PINN framework model, and outputting predicted performance parameters;
[0032] The calculation unit is also used to calculate new intrinsic parameters of the electrolyte, positive electrode and negative electrode combination to be predicted, and the acquisition unit is also used to acquire new physical parameters of the electrolyte, positive electrode and negative electrode to be predicted;
[0033] A dimensionality reduction processing unit, used for performing dimensionality reduction processing on the predicted performance parameter, the new intrinsic parameter and the new physical parameter;
[0034] A drawing unit is used to draw a two-dimensional lithium-ion battery intrinsic spectrum through the parameters after dimensionality reduction.
[0035] The benefit of the present invention lies in that the provided method and system for constructing an intrinsic spectrum of a lithium-ion battery based on PINN can determine the improvement of existing products by the development of new materials from the material level and the process level.
[0036] The benefit of the present invention is that the provided method and system for constructing the intrinsic spectrum of a lithium-ion battery based on PINN can also determine whether the current development is economically beneficial from a cost perspective. BRIEF DESCRIPTION OF THE DRAWINGS
[0037] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.
[0038] Figure 1 It is a flow chart of a method for constructing an intrinsic spectrum of a lithium-ion battery based on PINN of the present invention;
[0039] Figure 2 It is another schematic diagram of a process of constructing a lithium-ion battery intrinsic spectrum based on PINN of the present invention;
[0040] Figure 3 It is a schematic diagram of the process of dimensionality reduction through a neural network and an autoencoder of the present invention. DETAILED DESCRIPTION
[0041] The embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.
[0042] like Figure 1-2 As shown, the present application discloses a method for constructing an intrinsic map of a lithium-ion battery based on PINN (Physics Informed Neural Networks), comprising: S1: calculating the intrinsic parameters of the electrolyte, the positive electrode and the negative electrode. S2: determining the physical parameters of the electrolyte, the positive electrode and the negative electrode. S3: obtaining the battery performance parameters based on the intrinsic parameters and the physical parameters. S4: building a first neural network model, taking the intrinsic parameters and the physical parameters as the input of the first neural network model, taking the battery performance parameters as the output of the first neural network model, training the first neural network model, and finally constructing a PINN framework model. S5: inputting the electrolyte, positive electrode and negative electrode combination to be predicted into the PINN framework model, and outputting the predicted performance parameters. S6: calculating the new intrinsic parameters of the electrolyte, positive electrode and negative electrode combination to be predicted. S7: determining the new physical parameters of the electrolyte, positive electrode and negative electrode to be predicted. S8: performing dimensionality reduction processing on the predicted performance parameters, the new intrinsic parameters and the new physical parameters. S9: drawing a two-dimensional lithium-ion battery intrinsic map through the parameters after dimensionality reduction. Through the PINN-based lithium-ion battery intrinsic spectrum construction method of the present application, it is possible to determine from the material level and process level whether the current development can supplement and improve the existing products. The above steps are described in detail below.
[0043] For step S1: calculate the intrinsic parameters of the electrolyte, the positive electrode and the negative electrode.
[0044] In the embodiment of the present application, the specific method for calculating the intrinsic parameters of the electrolyte, the positive electrode and the negative electrode is:
[0045] The intrinsic parameters of the electrolyte, positive electrode and negative electrode, such as voltage and power, are calculated based on density functional theory (DFT) and molecular dynamics (MD).
[0046] For step S2: determine the physical parameters of the electrolyte, the positive electrode and the negative electrode.
[0047] In the embodiments of the present application, the physical parameters of the electrolyte, the positive electrode and the negative electrode include but are not limited to the physical parameters of the material such as particle size, electrolyte concentration, thickness, etc.
[0048] For step S3: obtain battery performance parameters based on intrinsic parameters and physical parameters.
[0049] In the implementation manner of the present application, the specific method for obtaining the battery performance parameters based on the intrinsic parameters and the physical parameters is:
[0050] The intrinsic parameters and physical parameters are input into the P2D (Pseudo Two Dimensional) model to obtain the battery performance parameters.
[0051] For step S4: build a first neural network model, use intrinsic parameters and physical parameters as inputs of the first neural network model, use battery performance parameters as outputs of the first neural network model, train the first neural network model, and finally construct a PINN framework model.
[0052] For step S5: input the electrolyte, positive electrode and negative electrode combination to be predicted into the PINN framework model, and output the predicted performance parameters.
[0053] The trained PINN framework model can automatically output predicted performance parameters, such as the battery's charge and discharge performance, current-voltage curve, power performance, etc., when the material combination of the electrolyte, positive electrode, and negative electrode to be predicted is input.
[0054] For step S6: calculate new intrinsic parameters of the electrolyte, positive electrode and negative electrode combination to be predicted.
[0055] The calculation method refers to step S1.
[0056] For step S7: determine new physical parameters of the electrolyte, the positive electrode and the negative electrode to be predicted.
[0057] Refer to step S2.
[0058] For step S8: perform dimensionality reduction processing on the predicted performance parameters, new intrinsic parameters and new physical parameters.
[0059] like Figure 3 As shown, in the implementation manner of the present application, the specific method of performing dimensionality reduction processing on the predicted performance parameters, the new intrinsic parameters and the new physical parameters is:
[0060] The predicted performance parameters, new intrinsic parameters and new physical parameters are input into a dimensionality reduction network composed of a second neural network model and an autoencoder for dimensionality reduction processing.
[0061] In an embodiment of the present application, the second neural network model is a CNN network (convolutional neural network) or a fully connected neural network.
[0062] Specifically, taking the use of a fully connected neural network and an autoencoder as an example, suppose that the intrinsic parameters calculated by DFT, the obtained physical parameters, and the output results of the P2D model have a total of 120 parameters and data. Build a fully connected neural network with 120 neurons in the output layer, and then through the encoding operation, the number of neurons in the hidden layer is 2, and then through the decoding operation, the number of neurons in the output layer is also 120. At this time, it is necessary to ensure that the input layer and the output layer are completely consistent. That is, what parameters are input, what parameters are output. This ensures that this set of encoding and decoding operations is lossless to the original parameters. That is, the operation from 120 dimensions to 2 dimensions is reversible. Then, select the left half of the neural network, that is, the part including the input layer and the hidden layer. This part can be regarded as a dimensionality reduction operation using an autoencoder to reduce the 120 dimensions to 2 dimensions. The above is to use fully connected neural network and autoencoder for dimensionality reduction, replace the fully connected neural network with convolutional neural network, and replace the fully connected layer with convolutional layer and pooling layer, so as to select features more accurately, retain the information extracted by the previous convolutional layers and pooling layers, and speed up the training efficiency.
[0063] It can be understood that the dimensionality reduction network composed of the second neural network model and the autoencoder can be trained by the intrinsic parameters, physical parameters and battery performance parameters obtained in the previous steps S1-S3.
[0064] For step S9: a two-dimensional lithium-ion battery intrinsic spectrum is drawn using the parameters after dimensionality reduction.
[0065] In the present application, the predicted performance parameters, new intrinsic parameters and new physical parameters are reduced to two dimensions. Specifically, a two-dimensional lithium-ion battery intrinsic map is drawn using the reduced two-dimensional parameters.
[0066] In an embodiment of the present application, the method for constructing a lithium-ion battery intrinsic spectrum based on PINN further comprises:
[0067] The cost parameters of the electrolyte, positive electrode and negative electrode combination to be predicted are obtained, and then a three-dimensional lithium-ion battery intrinsic spectrum is drawn based on the two-dimensional lithium-ion battery intrinsic spectrum and the cost parameters.
[0068] It is understandable that the bill of materials (BOM) of the battery is collected, the total cost of the battery is calculated as S, and the battery cost is divided into three aspects, namely, positive electrode cost, negative electrode cost and other costs. These three costs are converted into a three-dimensional array [A, B, C] through Soft-max and Auto-Encoder, and [A, B, C] is matched with the three primary colors [R, G, B], so as to represent which item of the battery cost accounts for the most by color.
[0069] The present application also discloses a lithium-ion battery intrinsic map construction system based on PINN, which is used to implement the aforementioned lithium-ion battery intrinsic map construction method based on PINN. The lithium-ion battery intrinsic map construction system based on PINN includes: a calculation unit, an acquisition unit, a processing unit, a PINN framework model, a dimensionality reduction processing unit and a drawing unit.
[0070] The calculation unit is used to calculate the intrinsic parameters of the electrolyte, the positive electrode and the negative electrode. The acquisition unit is used to obtain the physical parameters of the electrolyte, the positive electrode and the negative electrode. The processing unit is used to obtain the battery performance parameters based on the intrinsic parameters and the physical parameters. The PINN framework model is constructed by the following process. The intrinsic parameters and the physical parameters are used as the input of the first neural network model, and the battery performance parameters are used as the output of the first neural network model. The first neural network model is trained and finally constructed into a PINN framework model. The electrolyte, positive electrode and negative electrode combination to be predicted are input into the PINN framework model, and the predicted performance parameters are output. The calculation unit is also used to calculate the new intrinsic parameters of the electrolyte, positive electrode and negative electrode combination to be predicted, and the acquisition unit is also used to obtain the new physical parameters of the electrolyte, positive electrode and negative electrode to be predicted. The dimensionality reduction processing unit is used to perform dimensionality reduction processing on the predicted performance parameters, the new intrinsic parameters and the new physical parameters. The drawing unit is used to draw a two-dimensional lithium-ion battery intrinsic map through the parameters after dimensionality reduction. The technical details of the PINN-based lithium-ion battery intrinsic spectrum construction system refer to the corresponding parts of the aforementioned PINN-based lithium-ion battery intrinsic spectrum construction method, which will not be repeated here.
[0071] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the above embodiments do not limit the present invention in any form, and any technical solution obtained by equivalent replacement or equivalent transformation falls within the protection scope of the present invention.
Claims
1. A method for constructing an intrinsic spectrum of a lithium-ion battery based on PINN, characterized in that: Include: Calculate the intrinsic parameters of the electrolyte, cathode, and anode; Determine the physical parameters of the electrolyte, cathode, and anode; Obtaining battery performance parameters based on the intrinsic parameters and the physical parameters; Building a first neural network model, taking the intrinsic parameters and the physical parameters as inputs of the first neural network model, taking the battery performance parameters as outputs of the first neural network model, training the first neural network model, and finally constructing a PINN framework model; Inputting the electrolyte, positive electrode and negative electrode combination to be predicted into the PINN framework model, and outputting the predicted performance parameters; Calculate new intrinsic parameters for the electrolyte, cathode, and anode combinations to be predicted; Identify new physical parameters of the electrolyte, cathode, and anode to be predicted; Performing dimensionality reduction processing on the predicted performance parameter, the new intrinsic parameter and the new physical parameter; The two-dimensional intrinsic map of lithium-ion batteries is drawn using the reduced dimension parameters.
2. The method for constructing a lithium-ion battery intrinsic spectrum based on PINN according to claim 1, characterized in that: The specific method of performing dimensionality reduction processing on the predicted performance parameters, new intrinsic parameters and new physical parameters is: The predicted performance parameters, new intrinsic parameters and new physical parameters are input into a dimensionality reduction network composed of a second neural network model and an autoencoder for dimensionality reduction processing.
3. The method for constructing a lithium-ion battery intrinsic spectrum based on PINN according to claim 2, characterized in that: The PINN-based lithium-ion battery intrinsic spectrum construction method further comprises: The dimension reduction network composed of the second neural network model and the autoencoder is trained using the intrinsic parameters, the physical parameters and the battery performance parameters.
4. The method for constructing a lithium-ion battery intrinsic spectrum based on PINN according to claim 3, characterized in that: The second neural network model is a CNN network.
5. The method for constructing a lithium-ion battery intrinsic spectrum based on PINN according to claim 3, characterized in that: The second neural network model is a fully connected neural network.
6. The method for constructing a lithium-ion battery intrinsic spectrum based on PINN according to claim 1, characterized in that: The PINN-based lithium-ion battery intrinsic spectrum construction method further comprises: Obtaining cost parameters of the electrolyte, cathode and anode combination to be predicted; A three-dimensional lithium-ion battery intrinsic spectrum is drawn based on the two-dimensional lithium-ion battery intrinsic spectrum and the cost parameter.
7. The method for constructing a lithium-ion battery intrinsic spectrum based on PINN according to claim 1, characterized in that: The specific method for calculating the intrinsic parameters of the electrolyte, the positive electrode and the negative electrode is: The intrinsic parameters of the electrolyte, cathode, and anode were calculated based on density functional theory and molecular dynamics.
8. The method for constructing a lithium-ion battery intrinsic spectrum based on PINN according to claim 1, characterized in that: The specific method for obtaining the battery performance parameters based on the intrinsic parameters and the physical parameters is: The intrinsic parameters and the physical parameters are input into a P2D model to obtain the battery performance parameters.
9. A PINN-based lithium-ion battery intrinsic spectrum construction system, characterized in that: Include: A calculation unit for calculating the intrinsic parameters of the electrolyte, the positive electrode and the negative electrode; An acquisition unit, used to acquire physical parameters of the electrolyte, the positive electrode and the negative electrode; A processing unit, configured to obtain a battery performance parameter based on the intrinsic parameter and the physical parameter; A PINN framework model, using the intrinsic parameters and the physical parameters as inputs of a first neural network model, using the battery performance parameters as outputs of the first neural network model, training the first neural network model, and finally constructing the PINN framework model, inputting the electrolyte, positive electrode and negative electrode combination to be predicted into the PINN framework model, and outputting predicted performance parameters; The calculation unit is also used to calculate new intrinsic parameters of the electrolyte, positive electrode and negative electrode combination to be predicted, and the acquisition unit is also used to acquire new physical parameters of the electrolyte, positive electrode and negative electrode to be predicted; A dimensionality reduction processing unit, used for performing dimensionality reduction processing on the predicted performance parameter, the new intrinsic parameter and the new physical parameter; A drawing unit is used to draw a two-dimensional lithium-ion battery intrinsic spectrum through the parameters after dimensionality reduction.