Lithium battery remaining capacity model training method and system applying physical knowledge
By establishing an equivalent circuit model and processing power parameters, and using neural network training, the problem of insufficient accuracy in the residual power prediction of traditional lithium batteries is solved, and a more accurate prediction of residual power of lithium batteries is achieved.
Patent Information
- Application Number
- CN202411135332.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-19
- Publication Date
- 2024-11-19
AI Technical Summary
The traditional lithium battery residual power estimation method is ineffective due to factors such as aging state, and the prediction accuracy is insufficient, making it difficult to accurately estimate the residual power of lithium batteries.
Physical knowledge is used to establish an equivalent circuit model, obtain power parameters by applying pulse current, process and form a data set, and use neural network training to obtain the residual power prediction model of lithium battery. The specific steps include obtaining the terminal voltage and discharge current, constructing a time function under the zero input response and zero state response, calculating the electrochemical parameters, forming a data set, and finally inputting the neural network for training.
By combining physical knowledge and electrochemical characteristics, more accurate power parameters are provided, the generalization ability of neural networks is enhanced, and the accuracy of lithium battery residual power prediction is improved.
Smart Images

Figure CN118980937A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of lithium battery remaining power prediction, and in particular to a lithium battery remaining power model training method and system using physical knowledge. Background Art
[0002] With the widespread application of mobile devices, electric vehicles and renewable energy storage systems, lithium batteries, as a power source with high energy density and long life cycle, have received more and more attention. Its application in various devices makes it crucial to accurately estimate the remaining power (State of Charge, SoC) of lithium batteries. The remaining power of the battery directly affects the use time, performance and safety of the device. Therefore, it is of great practical significance to develop an efficient and accurate lithium battery remaining power prediction model.
[0003] The traditional method of estimating the remaining battery capacity is mainly to estimate the remaining capacity by integrating the discharge current of the lithium battery. However, due to the influence of various factors such as the aging state of the lithium battery, its prediction accuracy is insufficient. Summary of the invention
[0004] To solve the above problems, an embodiment of the present invention aims to provide a method and system for training a lithium battery remaining capacity model using physical knowledge.
[0005] A method for training a lithium battery remaining capacity model using physical knowledge, comprising:
[0006] Step 1: Apply a pulse current to the lithium battery, and then establish an equivalent circuit model for the discharge condition of the lithium battery;
[0007] Step 2: Obtain the power parameters of the lithium battery during the discharge process;
[0008] Step 3: Processing the power parameters based on physical knowledge to form a data set;
[0009] Step 4: Input the data set into a neural network for training to obtain a lithium battery remaining power prediction model;
[0010] Step 5: Use the lithium battery remaining power prediction model to complete the prediction of the remaining power of the target lithium battery.
[0011] Preferably, the step 3: processing the power parameters based on physical knowledge to form a data set includes:
[0012] Step 3.1: Obtain the terminal voltage and discharge current of the lithium battery during the discharge period;
[0013] Step 3.2: Use Kirchhoff's law to solve the node voltage of the RC loop in the equivalent circuit model using the terminal voltage and discharge current;
[0014] Step 3.3: Construct the time function of the terminal voltage under zero input response based on the node voltage of the RC loop;
[0015] Step 3.4: Obtaining a time constant based on the time function of the terminal voltage under the zero input response;
[0016] Step 3.5: Use Kirchhoff's law to construct the time function of the terminal voltage under the zero-state response in the equivalent circuit model using the terminal voltage and discharge current;
[0017] Step 3.6: Obtain electrochemical parameters of the lithium battery according to the time function of the lower terminal voltage of the zero-state response, and construct a data set using the electrochemical parameters of the lithium battery.
[0018] Preferably, in step 3.2, the node voltage of the RC loop is:
[0019]
[0020] Among them, U 01 (t) represents the node voltage of the first RC loop at time t, U 02 (t) represents the node voltage of the second RC loop at time t, U1(t1) represents the terminal voltage of the lithium battery at time t1, U2(t2) represents the terminal voltage of the lithium battery at time t2, I represents the discharge current, R1 represents the electrochemical polarization resistance, R2 represents the electrochemical concentration polarization resistance, C1 represents the electrochemical polarization capacitance, and C2 represents the electrochemical concentration polarization capacitance.
[0021] Preferably, the step 3.3: constructing a time function of the terminal voltage under zero input response according to the node voltage of the RC loop includes:
[0022] When the lithium battery is discharged, the discharge current is 0. At this time, the node voltage of the RC loop of the equivalent circuit model under zero input response is:
[0023]
[0024] The time function of the terminal voltage under zero input response is:
[0025]
[0026] U′(t) represents the time function of the voltage at the lower end of the zero input response. oc represents the open circuit voltage, τ1 represents the first time constant τ1=R1·C1, and τ2 represents the second time constant τ2=R2·C2.
[0027] Preferably, the step 3.5: using Kirchhoff's law to construct a time function of the terminal voltage under the zero-state response in the equivalent circuit model using the terminal voltage and the discharge current, comprises:
[0028] Using the formula:
[0029]
[0030] Construct a time function of the zero-state response lower terminal voltage; wherein U″(t) represents the time function of the zero-state response lower terminal voltage, and R0 represents the internal resistance of the battery.
[0031] Preferably, the step 3.6: obtaining the electrochemical parameters of the lithium battery according to the time function of the lower terminal voltage of the zero-state response, and constructing a data set using the electrochemical parameters of the lithium battery, comprises:
[0032] Step 3.6.1: Perform parameter identification on the time function of the lower terminal voltage of the zero-state response to obtain the electrochemical polarization resistance and the electrochemical concentration polarization resistance;
[0033] Step 3.6.2: Calculate the electrochemical polarization capacitance and the electrochemical concentration polarization capacitance using the electrochemical polarization resistance and the electrochemical concentration polarization resistance; wherein the calculation formula of the electrochemical polarization capacitance is:
[0034]
[0035] The calculation formula of electrochemical concentration polarization capacitance is:
[0036]
[0037] Step 3.6.3: Construct a data set using the electrochemical polarization resistance, electrochemical concentration polarization resistance, electrochemical polarization capacitance, electrochemical concentration polarization capacitance, battery internal resistance, and remaining capacity of the lithium battery.
[0038] Preferably, the step 4: inputting the data set into a neural network for training to obtain a lithium battery remaining capacity prediction model comprises:
[0039] Multiple loss functions are used to measure the difference between the predicted value and the true value of the lithium battery remaining power prediction model. When the difference value is within the preset range, the training is completed; the loss function is:
[0040]
[0041] Among them, MAE represents the difference value output by the first loss function, MRE represents the difference value output by the second loss function, RMSE represents the difference value output by the third loss function, and n represents the number of sample points. Represents the predicted value of the lithium battery remaining power prediction model, yi Represents the true value.
[0042] The present invention also provides a lithium battery remaining capacity model training system using physical knowledge, including:
[0043] An equivalent circuit model building module is used to apply a pulse current to a lithium battery and then establish an equivalent circuit model for the discharge condition of the lithium battery;
[0044] A power parameter acquisition module is used to obtain the power parameters of the lithium battery during the discharge process;
[0045] A parameter processing module, used for processing the power parameters based on physical knowledge to form a data set;
[0046] A model training module, used to input the data set into a neural network for training to obtain a lithium battery remaining power prediction model;
[0047] The remaining power prediction module is used to use the lithium battery remaining power prediction model to complete the prediction of the remaining power of the target lithium battery.
[0048] The present invention also provides an electronic device, comprising a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the transceiver, the memory, and the processor are connected via the bus, and wherein the computer program, when executed by the processor, implements the steps in the above-mentioned method for training a lithium battery remaining power model using physical knowledge.
[0049] The present invention also provides a computer-readable storage medium having a computer program stored thereon, characterized in that when the computer program is executed by a processor, the steps in the above-mentioned method for training a lithium battery remaining power model using physical knowledge are implemented.
[0050] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:
[0051] The present invention relates to a method and device for training a lithium battery remaining capacity model using physical knowledge. Compared with the prior art, the present invention combines physical knowledge with the electrochemical characteristics of the battery to more accurately simulate the discharge process of the lithium battery and provide more realistic power parameters. These power parameters based on the physical model can enhance the generalization ability of the neural network for the remaining capacity of the lithium battery, thereby improving the accuracy of the prediction model.
[0052] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS
[0053] In order to more clearly illustrate the embodiments of the present invention 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 invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0054] Figure 1 A flow chart of a lithium battery remaining capacity model training method using physical knowledge provided by the present invention;
[0055] Figure 2 A schematic diagram of an equivalent circuit model provided by the present invention;
[0056] Figure 3 A schematic diagram of the changes in voltage and current parameters during the discharge process of the lithium battery provided by the present invention;
[0057] Figure 4 Schematic diagram of the power parameter data set constructed for the present invention. DETAILED DESCRIPTION
[0058] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.
[0059] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.
[0060] In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be an indirect connection through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.
[0061] See also Figure 1 , a lithium battery remaining capacity model training method using physical knowledge, including:
[0062] Step 1: Apply pulse current to the lithium battery, and then establish an equivalent circuit model for the discharge condition of the lithium battery, such as Figure 2 As shown;
[0063] Step 2: Obtain the power parameters of the lithium battery during the discharge process;
[0064] Step 3: Processing the power parameters based on physical knowledge to form a data set;
[0065] Furthermore, the step 3 comprises:
[0066] Step 3.1: Obtain the terminal voltage and discharge current of the lithium battery during the discharge period, such as Figure 3 As shown, the terminal voltage of lithium batteries will change during the discharge process;
[0067] Step 3.2: Use Kirchhoff's law to solve the node voltage of the RC loop in the equivalent circuit model using the terminal voltage and discharge current; where the node voltage of the RC loop is:
[0068]
[0069] Among them, U 01 (t) represents the node voltage of the first RC loop at time t, U 02 (t) represents the node voltage of the second RC loop at time t, U1(t1) represents the terminal voltage of the lithium battery at time t1, U2(t2) represents the terminal voltage of the lithium battery at time t2, I represents the discharge current, R1 represents the electrochemical polarization resistance, R2 represents the electrochemical concentration polarization resistance, C1 represents the electrochemical polarization capacitance, and C2 represents the electrochemical concentration polarization capacitance.
[0070] Step 3.3: Construct the time function of the terminal voltage under zero input response based on the node voltage of the RC loop;
[0071] Among them, step 3.3 includes:
[0072] like Figure 3 As shown in the figure, when the lithium battery is discharged, the discharge current is 0. At this time, the node voltage of the RC loop of the equivalent circuit model under zero input response is:
[0073]
[0074] The time function of the terminal voltage under zero input response is:
[0075]
[0076] U′(t) represents the time function of the voltage at the lower end of the zero input response. oc represents the open circuit voltage, τ1 represents the first time constant τ1=R1·C1, and τ2 represents the second time constant τ2=R2·C2.
[0077] In the embodiment of the present invention, the least square method may be used to fit the time function of the terminal voltage under zero input response to obtain the values of the first time constant and the second time constant.
[0078] Step 3.4: Obtaining a time constant based on the time function of the terminal voltage under the zero input response;
[0079] Step 3.5: Use Kirchhoff's law to construct the time function of the terminal voltage under the zero-state response in the equivalent circuit model using the terminal voltage and discharge current;
[0080] In the U1-U2 time period, the voltage can be regarded as a zero-state response. Based on Kirchhoff's law, the terminal voltage and discharge current are used to construct the time function of the terminal voltage under the zero-state response in the equivalent circuit model:
[0081]
[0082] Wherein, U″(t) represents the time function of the lower terminal voltage of the zero-state response, and R0 represents the internal resistance of the battery.
[0083] Step 3.6: Obtaining electrochemical parameters of the lithium battery according to the time function of the lower terminal voltage of the zero-state response, and constructing a data set using the electrochemical parameters of the lithium battery;
[0084] Wherein, the step 3.6 comprises:
[0085] Step 3.6.1: Perform parameter identification on the time function of the lower terminal voltage of the zero-state response to obtain the electrochemical polarization resistance and the electrochemical concentration polarization resistance;
[0086] In the embodiment of the present invention, the least square method can be used to fit the time function of the terminal voltage under zero input response to complete the calculation of the electrochemical polarization resistance and the electrochemical concentration polarization resistance.
[0087] Step 3.6.2: Calculate the electrochemical polarization capacitance and the electrochemical concentration polarization capacitance using the electrochemical polarization resistance and the electrochemical concentration polarization resistance; wherein the calculation formula of the electrochemical polarization capacitance is:
[0088]
[0089] The calculation formula of electrochemical concentration polarization capacitance is:
[0090]
[0091] Step 3.6.3: Construct a data set using the electrochemical polarization resistance, electrochemical concentration polarization resistance, electrochemical polarization capacitance, electrochemical concentration polarization capacitance, battery internal resistance, and remaining capacity of the lithium battery, such as Figure 4 shown.
[0092] Step 4: Input the data set into a neural network for training to obtain a lithium battery remaining power prediction model;
[0093] In step 4, multiple loss functions are used to measure the difference between the predicted value and the true value of the lithium battery remaining power prediction model. When the difference value is within a preset range, the training is completed; wherein the loss function is:
[0094]
[0095] Among them, MAE represents the difference value output by the first loss function, MRE represents the difference value output by the second loss function, RMSE represents the difference value output by the third loss function, and n represents the number of sample points. Represents the predicted value of the lithium battery remaining power prediction model, y i Represents the true value.
[0096] Step 5: Use the lithium battery remaining power prediction model to complete the prediction of the remaining power of the target lithium battery.
[0097] The present invention combines physical knowledge with the electrochemical characteristics of batteries to more accurately simulate the discharge process of lithium batteries and provide more realistic power parameters. These power parameters based on physical models can enhance the generalization ability of neural networks for the remaining power of lithium batteries, thereby improving the accuracy of the prediction model.
[0098] The present invention also provides a lithium battery remaining capacity model training system using physical knowledge, including:
[0099] An equivalent circuit model building module is used to apply a pulse current to a lithium battery and then establish an equivalent circuit model for the discharge condition of the lithium battery;
[0100] A power parameter acquisition module is used to obtain the power parameters of the lithium battery during the discharge process;
[0101] A parameter processing module, used for processing the power parameters based on physical knowledge to form a data set;
[0102] A model training module, used to input the data set into a neural network for training to obtain a lithium battery remaining power prediction model;
[0103] The remaining power prediction module is used to use the lithium battery remaining power prediction model to complete the prediction of the remaining power of the target lithium battery.
[0104] Compared with the prior art, the beneficial effects of a lithium battery remaining power model training system using physical knowledge provided by the present invention are the same as the beneficial effects of a lithium battery remaining power model training method using physical knowledge described in the above technical scheme, which will not be repeated here.
[0105] The present invention also provides an electronic device, comprising a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the transceiver, the memory, and the processor are connected via the bus, and wherein the computer program, when executed by the processor, implements the steps in the above-mentioned method for training a lithium battery remaining power model using physical knowledge.
[0106] The present invention also provides a computer-readable storage medium having a computer program stored thereon, characterized in that when the computer program is executed by a processor, the steps in the above-mentioned method for training a lithium battery remaining power model using physical knowledge are implemented.
[0107] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technical solution that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.
Claims
1. A lithium battery remaining capacity model training method using physical knowledge, characterized in that: include: Step 1: Apply a pulse current to the lithium battery, and then establish an equivalent circuit model for the discharge condition of the lithium battery; Step 2: Obtain the power parameters of the lithium battery during the discharge process; Step 3: Processing the power parameters based on physical knowledge to form a data set; Step 4: Input the data set into a neural network for training to obtain a lithium battery remaining power prediction model; Step 5: Use the lithium battery remaining power prediction model to complete the prediction of the remaining power of the target lithium battery.
2. The method for training a lithium battery remaining capacity model using physical knowledge according to claim 1, characterized in that: The step 3: processing the power parameters based on physical knowledge to form a data set includes: Step 3.1: Obtain the terminal voltage and discharge current of the lithium battery during the discharge period; Step 3.2: Use Kirchhoff's law to solve the node voltage of the RC loop in the equivalent circuit model using the terminal voltage and discharge current; Step 3.3: Construct the time function of the terminal voltage under zero input response based on the node voltage of the RC loop; Step 3.4: Obtaining a time constant based on the time function of the terminal voltage under the zero input response; Step 3.5: Use Kirchhoff's law to construct the time function of the terminal voltage under the zero-state response in the equivalent circuit model using the terminal voltage and discharge current; Step 3.6: Obtain electrochemical parameters of the lithium battery according to the time function of the lower terminal voltage of the zero-state response, and construct a data set using the electrochemical parameters of the lithium battery.
3. The method for training a lithium battery residual capacity model using physical knowledge according to claim 2, characterized in that: In step 3.2, the node voltage of the RC loop is: Among them, U 01 (t) represents the node voltage of the first RC loop at time t, U 02 (t) represents the node voltage of the second RC loop at time t, U1(t1) represents the terminal voltage of the lithium battery at time t1, U2(t2) represents the terminal voltage of the lithium battery at time t2, I represents the discharge current, R1 represents the electrochemical polarization resistance, R2 represents the electrochemical concentration polarization resistance, C1 represents the electrochemical polarization capacitance, and C2 represents the electrochemical concentration polarization capacitance.
4. The method for training a lithium battery remaining capacity model using physical knowledge according to claim 3, characterized in that: The step 3.3: constructing a time function of the terminal voltage under zero input response according to the node voltage of the RC loop, comprises: When the lithium battery is discharged, the discharge current is 0. At this time, the node voltage of the RC loop of the equivalent circuit model under zero input response is: The time function of the terminal voltage under zero input response is: U′(t) represents the time function of the voltage at the lower end of the zero input response. oc represents the open circuit voltage, τ1 represents the first time constant τ1=R1·C1, and τ2 represents the second time constant τ2=R2·C2.
5. The method for training a lithium battery residual capacity model using physical knowledge according to claim 4, characterized in that: The step 3.5: using Kirchhoff's law to construct a time function of the terminal voltage under the zero-state response in the equivalent circuit model using the terminal voltage and the discharge current, includes: Using the formula: Construct a time function of the zero-state response lower terminal voltage; wherein U″(t) represents the time function of the zero-state response lower terminal voltage, and R0 represents the internal resistance of the battery.
6. According to claim 5, a lithium battery remaining capacity model training method using physical knowledge is characterized in that: The step 3.6: obtaining the electrochemical parameters of the lithium battery according to the time function of the lower terminal voltage of the zero-state response, and constructing a data set using the electrochemical parameters of the lithium battery, comprises: Step 3.6.1: Perform parameter identification on the time function of the lower terminal voltage of the zero-state response to obtain the electrochemical polarization resistance and the electrochemical concentration polarization resistance; Step 3.6.2: Calculate the electrochemical polarization capacitance and the electrochemical concentration polarization capacitance using the electrochemical polarization resistance and the electrochemical concentration polarization resistance; wherein the calculation formula of the electrochemical polarization capacitance is: The calculation formula of electrochemical concentration polarization capacitance is: Step 3.6.3: Construct a data set using the electrochemical polarization resistance, electrochemical concentration polarization resistance, electrochemical polarization capacitance, electrochemical concentration polarization capacitance, battery internal resistance, and remaining capacity of the lithium battery.
7. According to claim 6, a lithium battery remaining capacity model training method using physical knowledge is characterized in that: The step 4: inputting the data set into a neural network for training to obtain a lithium battery remaining power prediction model, including: Multiple loss functions are used to measure the difference between the predicted value and the true value of the lithium battery remaining power prediction model. When the difference value is within the preset range, the training is completed; the loss function is: Among them, MAE represents the difference value output by the first loss function, MRE represents the difference value output by the second loss function, RMSE represents the difference value output by the third loss function, and n represents the number of sample points. Represents the predicted value of the lithium battery remaining power prediction model, y i Represents the true value.
8. A lithium battery remaining capacity model training system using physical knowledge, characterized in that: include: An equivalent circuit model building module is used to apply a pulse current to a lithium battery and then establish an equivalent circuit model for the discharge condition of the lithium battery; A power parameter acquisition module is used to obtain the power parameters of the lithium battery during the discharge process; A parameter processing module, used for processing the power parameters based on physical knowledge to form a data set; A model training module, used to input the data set into a neural network for training to obtain a lithium battery remaining power prediction model; The remaining power prediction module is used to use the lithium battery remaining power prediction model to complete the prediction of the remaining power of the target lithium battery.
9. An electronic device, comprising a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the transceiver, the memory, and the processor are connected via the bus, wherein: When the computer program is executed by the processor, the steps in the method for training a lithium battery remaining power model using physical knowledge are implemented as described in any one of claims 1-7.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, the steps in the method for training a lithium battery remaining capacity model using physical knowledge are implemented as described in any one of claims 1-7.
