Method and device for estimating dynamic thermal resistance of battery module based on machine learning

Abstract

The invention discloses a method and a device for estimating dynamic thermal resistance of a battery module based on machine learning, relates to the technical field of battery thermal resistance testing, and solves the problem of high manpower and material resource cost of lithium battery thermal resistance testing in the prior art. Constructing a machine learning neural network model; training and verifying the machine learning neural network model by using the test data to obtain a dynamic thermal resistance training model; the dynamic thermal resistance of the battery module is estimated by using the dynamic thermal resistance training model, and the thermal resistance of the battery pack can be predicted by using the trained neural network machine learning model, so that the thermal resistance prediction can be quickly realized without actual thermal resistance test, and the method can be used for evaluating the thermal balance dynamic thermal resistance of the battery pack under different working conditions. And the manpower and material resource cost of thermal resistance testing is saved.

Description

Technical Field

[0001] The present application relates to the technical field of battery thermal resistance testing, and in particular to a method and device for estimating the dynamic thermal resistance of a battery module based on machine learning. Background Art

[0002] The dynamic thermal resistance of a battery module is a parameter that describes how the difficulty of heat transfer changes with time and working conditions during the charging, discharging or working process of the battery module. It reflects the resistance to heat flow inside the battery module and has an important impact on the temperature management, thermal safety performance and overall performance of the battery module.

[0003] At present, the patents for analyzing the thermal resistance between lithium batteries and their surrounding thermal management environment are all based on thermal test data and related mathematical and physical models. If you want to examine the thermal resistance between lithium batteries and their surrounding thermal management environment at different coolant flow rates, you can usually only do more thermal group tests at different coolant flow rates. However, the thermal testing process is often complicated, the test cycle is relatively long, and the test manpower and material costs are high. Summary of the invention

[0004] The purpose of this application is to overcome the problem of high manpower and material costs in the thermal resistance test of lithium batteries in the prior art, and to provide a method and device for estimating the dynamic thermal resistance of a battery module based on machine learning.

[0005] In a first aspect, a method for estimating a dynamic thermal resistance of a battery module based on machine learning is provided, comprising:

[0006] Obtain test data of battery modules;

[0007] Build machine learning neural network models;

[0008] Using the test data to train and verify the machine learning neural network model to obtain a dynamic thermal resistance training model;

[0009] The dynamic thermal resistance training model is used to estimate the dynamic thermal resistance of the battery module.

[0010] In some possible implementations, the test data includes coolant temperature, test environment temperature, test flow rate, and dynamic thermal resistance.

[0011] In some possible implementations, the dynamic thermal resistance includes a maximum thermal resistance, an average thermal resistance, and a minimum thermal resistance.

[0012] In some possible implementations, the dynamic thermal resistance is measured by measuring different positions of the battery module under thermal equilibrium by changing test parameters, wherein the test parameters include at least one of the following: coolant temperature, test environment temperature, and test flow rate.

[0013] In some possible implementations, the machine learning neural network model includes an input layer, a hidden layer, and an output layer, wherein the coolant temperature, test environment temperature, and test flow rate serve as input data sets of the input layer, and the maximum thermal resistance, average thermal resistance, and minimum thermal resistance serve as output data sets of the output layer.

[0014] In some possible implementations, Batch Normalization is used to accelerate the training of the neural network model, Dropout is used to prevent the overfitting of the neural network model, Leaky ReLu is used as the activation function, MSE is used as the loss function, Adam is used as the solution algorithm of the neural network model, and the initial learning rate of the neural network model is set to 0.00005-0.00015.

[0015] In a second aspect, a device for estimating dynamic thermal resistance of a battery module based on machine learning is provided, comprising:

[0016] An acquisition module is used to acquire test data of a battery module;

[0017] Building blocks for constructing machine learning neural network models;

[0018] A training module, used to train and verify the machine learning neural network model using the test data to obtain a dynamic thermal resistance training model;

[0019] The prediction module is used to estimate the dynamic thermal resistance of the battery module using the dynamic thermal resistance training model.

[0020] In a third aspect, a computer program product comprising instructions is provided, which, when executed on a computer, enables the computer to execute the method described in the first aspect.

[0021] In a fourth aspect, a computer-readable storage medium is provided, wherein the computer-readable medium stores a program code for execution by a device, wherein the program code includes steps for executing the method in any one of the implementations of the first aspect described above.

[0022] In a fifth aspect, an electronic device is provided, comprising a processor, a memory, and a program or instruction stored in the memory and executable on the processor, wherein the program or instruction, when executed by the processor, implements a method as in any one of the implementations in the first aspect above.

[0023] The present application has the following beneficial effects: the present application utilizes historical thermal resistance test data to train a neural network machine learning model, and utilizes the trained neural network machine learning model to predict the thermal resistance of the battery pack, thereby enabling rapid prediction of thermal resistance without the need for actual thermal resistance testing. This can not only be used to evaluate the thermal equilibrium dynamic thermal resistance of the battery pack under different operating conditions, saving the manpower and material costs of thermal resistance testing, but can also evaluate non-thermal equilibrium dynamic thermal resistance and static thermal resistance conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application.

[0025] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments 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 creative work.

[0026] Figure 1 is a flow chart of a method for estimating dynamic thermal resistance of a battery module based on machine learning according to Example 1 of the present application;

[0027] Figure 2 is a schematic diagram of a machine learning neural network model in the method for estimating dynamic thermal resistance of a battery module based on machine learning in Example 1 of the present application;

[0028] FIG3( a ) is a comparison diagram of the prediction and test results of the highest thermal resistance in the method for estimating the dynamic thermal resistance of a battery module based on machine learning in Example 1 of the present application;

[0029] FIG3( b ) is a comparison diagram of the prediction and test results of the average thermal resistance in the method for estimating the dynamic thermal resistance of a battery module based on machine learning in Example 1 of the present application;

[0030] FIG3( c ) is a comparison chart of the prediction and test results of the minimum thermal resistance in the method for estimating the dynamic thermal resistance of a battery module based on machine learning in Example 1 of the present application;

[0031] FIG4( a ) is a comparison diagram of the prediction and test results of the maximum thermal resistance at different coolant flow rates in the method for estimating the dynamic thermal resistance of a battery module based on machine learning in Example 1 of the present application;

[0032] FIG4( b ) is a comparison chart of the prediction and test results of the average thermal resistance at different coolant flow rates in the method for estimating the dynamic thermal resistance of a battery module based on machine learning in Example 1 of the present application;

[0033] FIG4( c ) is a comparison chart of the prediction and test results of the minimum thermal resistance at different coolant flow rates in the method for estimating the dynamic thermal resistance of a battery module based on machine learning in Example 1 of the present application;

[0034] FIG5( a ) is a line graph of the absolute deviation between the maximum thermal resistance prediction and the test result at a flow rate of 0.5-15 L / min in the method for estimating the dynamic thermal resistance of a battery module based on machine learning in Example 1 of the present application;

[0035] FIG5( b ) is a line graph of the absolute deviation between the average thermal resistance prediction and the test result at a flow rate of 0.5-15 L / min in the method for estimating the dynamic thermal resistance of a battery module based on machine learning in Example 1 of the present application;

[0036] FIG5( c ) is a line graph of the absolute deviation between the minimum thermal resistance prediction and the test result at a flow rate of 0.5-15 L / min in the method for estimating the dynamic thermal resistance of a battery module based on machine learning in Example 1 of the present application;

[0037] Figure 6 is a structural block diagram of a device for estimating dynamic thermal resistance of a battery module based on machine learning according to Example 2 of the present application;

[0038] Figure 7 It is a schematic diagram of the internal structure of the electronic device of Example 5 of the present application.

[0039] Reference numerals:

[0040] 100, acquisition module; 200, construction module; 300, training module; 400, prediction module. DETAILED DESCRIPTION

[0041] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0042] Example 1

[0043] like Figure 1 As shown, a method for estimating the dynamic thermal resistance of a battery module based on machine learning according to Example 1 of the present application includes:

[0044] S100, obtain training and test data sets, i.e., test data of the battery module: according to the existing data, the test environment temperature, coolant temperature, test flow rate and dynamic thermal resistance value are summarized in the same file, wherein the dynamic thermal resistance includes the highest thermal resistance, the average thermal resistance and the lowest thermal resistance. The dynamic thermal resistance of different positions of the battery cell under thermal equilibrium is measured by changing the coolant temperature, the ambient temperature and the coolant flow rate, as shown in Table 1 below, which summarizes the highest thermal resistance, the average thermal resistance and the lowest thermal resistance of a battery cell under different coolant flow rates. The data in Table 1 are used as the data set for training the neural network model.

[0045] Table 1: Dataset examples

[0046] Coolant temperature Ambient temperature flow Maximum thermal resistance Average thermal resistance Minimum thermal resistance 23℃ 23℃ 0.5L / min 2.38K / W 1.46K / W 0.67K / W 23℃ 23℃ 3L / min 1.97K / W 1.14K / W 0.41K / W 23℃ 23℃ 6L / min 1.85K / W 1.03K / W 0.32K / W 23℃ 23℃ 10L / min 1.81K / W 0.99K / W 0.28K / W 23℃ 23℃ 15L / min 1.78K / W 0.96K / W 0.26K / W

[0047] S200, constructing a machine learning neural network model: In this embodiment, a neural network (NN) is used for data fitting. The neural network includes structures such as an input layer, a hidden layer, and an output layer, and has parameters such as an activation function, a learning rate, a loss function, and a solution algorithm.

[0048] First, the input layer of the machine learning neural network model is determined. In this embodiment, the test environment temperature, coolant temperature and coolant flow rate are used as three variables as the input layer.

[0049] Then, the output layer of the machine learning neural network model is determined, and the corresponding highest thermal resistance, average thermal resistance, and lowest thermal resistance values ​​obtained from the test are used as the output layer data set.

[0050] In addition, it is necessary to determine the hidden layer of the machine learning neural network model, wherein the number of hidden layers can be modified according to actual usage. In this embodiment, the hidden layer is set to 10 as an example. In theory, the more hidden layers there are, the more accurate the trained model will be and the smaller the error will be.

[0051] S300, using the test data to train and verify the machine learning neural network model to obtain a dynamic thermal resistance training model.

[0052] First, you need to select an algorithm for training the machine learning neural network model. The neural network algorithms are divided into: Levenberg-Marquardt\Bayesian Regularization\Scaled Conjugate Gradient. In this embodiment, the Levenberg-Marquardt algorithm is used. In actual application, the neural network algorithm with the smallest error can be selected based on the error analysis after training.

[0053] In the machine learning neural network model proposed in this embodiment, the three variables of the test environment temperature, coolant temperature and coolant flow obtained in step S100 are input into the input layer of the machine learning neural network model. Then, the number of neurons in each layer and the number of hidden layers are set. For example, in this embodiment, 10 hidden layers are set, each layer has 10 neurons, and Figure 2 In the equation, wi and bi are the weight parameters when mapping each neuron in the upper layer to a neuron in the lower layer. That is, assuming that the second neuron in the second layer (valued as θ) is mapped to the fourth neuron in the third layer, the mapping is: 22→34 =w 22→34 ×θ+b 22→34 . We can know that each neuron in the second layer has a mapping relationship with the fourth neuron in the third layer, so the value of the fourth neuron in the third layer is: y 34 =∑y 2i→34 , which is the sum of all the mappings from the second layer to the fourth neuron in the third layer. Then, Batch Normalization is used to accelerate model training, Dropout is used to prevent model overfitting, and Leaky ReLu is used as the activation function to increase the nonlinearity of the model. In the output layer, the final output result is set to 3 parameters, that is, the 10 neurons in the 10th layer of the hidden layer output 3 parameters after the activation function. Then, the output parameters are compared with the highest thermal resistance, average thermal resistance and lowest thermal resistance values ​​obtained from the test, MSE is used as the loss function, Adam is used as the solution algorithm, the initial learning rate is set to 0.00005-0.00015, and the optimal initial learning rate is set to 0.0001. The neural network is trained to finally obtain the corresponding dynamic thermal resistance training model.

[0054] It should be noted that in the entire neural network process, coolant temperature, test environment temperature and coolant flow rate are used as three input quantities, and the maximum thermal resistance, average thermal resistance and minimum thermal resistance are used as output quantities, and training and verification are carried out based on this. Finally, the trained neural network model is used to further predict the thermal resistance at coolant flow rates of 20L / min, 25L / min and 30L / min.

[0055] like Figure 3(a) , 3(b) As shown in Figure 3(c), for the comparison between the prediction and test results, it can be seen that the data are almost distributed near the fitting line. Figure 4(a) , 4(b) and 4(c) are comparisons of the prediction and test results at different flow rates. The thermal resistance at a flow rate of 0.5-15 L / min is verified and compared, and the thermal resistance at a flow rate of 20-30 L / min is predicted. Figure 5(a) , 5(b)and 5(c) are the absolute deviations of the two at 0.5-15 L / min, where the deviation is the difference between the predicted value and the tested value. It can be seen that the predicted results obtained by the neural network are very consistent with the actual results, with a deviation within 5%.

[0056] In this embodiment, by performing thermal resistance tests to obtain a data set, conducting machine learning, and establishing a machine learning model, the resources consumed in actual testing can be saved, thereby reducing manpower and material costs; based on the established machine learning model, rapid predictions can be achieved with small prediction errors.

[0057] Example 2

[0058] like Figure 6 As shown, a device for estimating the dynamic thermal resistance of a battery module based on machine learning according to Embodiment 2 of the present application includes:

[0059] An acquisition module 100 is used to acquire test data of a battery module;

[0060] Construction module 200, for constructing a machine learning neural network model;

[0061] A training module 300 is used to train and verify the machine learning neural network model using test data to obtain a dynamic thermal resistance training model;

[0062] The prediction module 400 is used to estimate the dynamic thermal resistance of the battery module using a dynamic thermal resistance training model.

[0063] It should be noted that other specific implementations of the device for estimating the dynamic thermal resistance of the battery module based on machine learning in this embodiment can refer to the specific implementations of the method for estimating the dynamic thermal resistance of the battery module based on machine learning mentioned above. To avoid redundancy, they will not be repeated here.

[0064] Example 3

[0065] A computer program product involved in Example 3 of the present application is used to store a computer program. When the computer program runs on a computer, the method in any one of the implementation modes in Example 1 of the present application is implemented.

[0066] Example 4

[0067] A computer-readable storage medium involved in Embodiment 4 of the present application, the computer-readable storage medium stores a program code for execution by a device, the program code including steps for executing a method in any one of the implementations in Embodiment 1 of the present application;

[0068] Among them, the computer-readable storage medium can be a read-only memory (ROM), a static storage device, a dynamic storage device or a random access memory (RAM); the computer-readable storage medium can store program code, and when the program stored in the computer-readable storage medium is executed by the processor, the processor is used to execute the steps of the method in any one of the implementation methods in Example 1 of the present application.

[0069] Example 5

[0070] like Figure 7 As shown, an electronic device involved in Embodiment 5 of the present application includes a processor, a memory, and a program or instruction stored in the memory and executable on the processor, and when the program or instruction is executed by the processor, a method in any one of the implementation modes in Embodiment 1 of the present application is implemented;

[0071] Among them, the processor can adopt a general-purpose central processing unit (CPU), a microprocessor, an application specific integrated circuit (ASIC), a graphics processing unit (GPU) or one or more integrated circuits to execute relevant programs to implement the method in any one of the implementation methods in Example 1 of the present application.

[0072] The processor may also be an integrated circuit electronic device with signal processing capability. In the implementation process, each step of the method in any implementation of Embodiment 1 of the present application may be completed by an integrated logic circuit of hardware in the processor or by instructions in software form.

[0073] The above-mentioned processor may also be a general-purpose processor, a digital signal processor, an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application may be implemented or executed. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the method disclosed in the embodiments of the present application may be directly embodied as being executed by a hardware decoding processor, or may be executed by a combination of hardware and software modules in a decoding processor. The software module may be located in a mature storage medium in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in a memory, and the processor reads the information in the memory, and completes the functions required to be performed by the unit included in the data processing device of the embodiment of the present application in combination with its hardware, or executes the method in any one of the implementation modes in the embodiment 1 of the present application.

[0074] The above are only preferred specific implementations of the present application; however, the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical solution and its improved ideas of the present application within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A method for estimating the dynamic thermal resistance of a battery module based on machine learning, characterized in that: include: Obtain test data of battery modules; Build machine learning neural network models; Using the test data to train and verify the machine learning neural network model to obtain a dynamic thermal resistance training model; The dynamic thermal resistance training model is used to estimate the dynamic thermal resistance of the battery module.

2. The method for estimating the dynamic thermal resistance of a battery module based on machine learning according to claim 1, characterized in that: The test data include coolant temperature, test environment temperature, test flow rate and dynamic thermal resistance.

3. The method for estimating the dynamic thermal resistance of a battery module based on machine learning according to claim 1, characterized in that: The dynamic thermal resistance includes a maximum thermal resistance, an average thermal resistance and a minimum thermal resistance.

4. The method for estimating the dynamic thermal resistance of a battery module based on machine learning according to claim 2 or 3, characterized in that: The dynamic thermal resistance is measured by changing the test parameters and measuring different positions of the battery module under thermal equilibrium, wherein the test parameters include at least one of the following: coolant temperature, test environment temperature and test flow rate.

5. The method for estimating the dynamic thermal resistance of a battery module based on machine learning according to claim 3, characterized in that: The machine learning neural network model includes an input layer, a hidden layer and an output layer, wherein the coolant temperature, the test environment temperature and the test flow rate are used as input data sets of the input layer, and the maximum thermal resistance, the average thermal resistance and the minimum thermal resistance are used as output data sets of the output layer.

6. The method for estimating the dynamic thermal resistance of a battery module based on machine learning according to claim 1, characterized in that: Batch Normalization is used to accelerate the training of the neural network model, Dropout is used to prevent the overfitting of the neural network model, Leaky ReLu is used as the activation function, MSE is used as the loss function, Adam is used as the solution algorithm of the neural network model, and the initial learning rate of the neural network model is set to 0.00005-0.00015.

7. A device for estimating dynamic thermal resistance of a battery module based on machine learning, characterized in that: include: An acquisition module is used to acquire test data of a battery module; Building blocks for constructing machine learning neural network models; A training module, used to train and verify the machine learning neural network model using the test data to obtain a dynamic thermal resistance training model; The prediction module is used to estimate the dynamic thermal resistance of the battery module using the dynamic thermal resistance training model.

8. A computer program product, characterized in that The computer program product stores computer instructions, and when the computer instructions are executed by a processor, the method according to any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a program code for execution by a device, wherein the program code includes steps for executing the method according to any one of claims 1 to 6.

10. An electronic device, characterized in that: The electronic device comprises a processor, a memory, and a program or instruction stored in the memory and executable on the processor, wherein the program or instruction implements the method as claimed in any one of claims 1 to 6 when executed by the processor.

Images