A Data-Driven Battery Health Status Analysis Method

Abstract

This invention proposes a data-driven battery health status analysis method. This method addresses the practical problem of the continuous capacity decay of power lithium-ion batteries during use, requiring effective estimation of their health status. It proposes a health factor feature extraction method and a corresponding training network, and then trains the network based on this to achieve accurate analysis of battery health status.

Description

Technical Field

[0001] This invention relates to a data-driven method for analyzing battery health status, belonging to the field of battery analysis. Background Technology

[0002] Due to their excellent performance, lithium-ion batteries are widely used in electric vehicles, electrical equipment, and energy storage systems. However, as the number of charge-discharge cycles of power lithium-ion batteries increases, the battery capacity irreversibly decreases, thus affecting the overall performance of the battery. Summary of the Invention

[0003] This invention provides a data-driven method for analyzing battery health status, which solves the problems disclosed in the background art.

[0004] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: a data-driven battery health status analysis method.

[0005] Extract the battery's health factors, which include: voltage signal energy, temperature signal energy, interval capacity, and peak voltage;

[0006] Input the voltage signal energy, temperature signal energy, interval capacity, and peak voltage into a pre-constructed BP neural network;

[0007] The BP neural network outputs the battery health status (SOH).

[0008] Furthermore: the formula for extracting voltage signal energy is:

[0009]

[0010] In the formula: v(t) represents the battery voltage value, t0 represents the start time of the measurement, and t1 represents the end time of the measurement.

[0011] Furthermore: the formula for extracting temperature signal energy is:

[0012]

[0013] In the formula: T(t) represents the battery temperature value, t0 represents the start time of the measurement, and t1 represents the end time of the measurement.

[0014] Furthermore, the method for extracting the interval capacity is as follows:

[0015] 1) Calculate the battery IC curve using equation (3);

[0016]

[0017] In the formula: Q represents battery capacity, V represents battery voltage, t represents time, and I represents current;

[0018] 2) Calculate the voltage V1 at the left end of the peak and the voltage V2 at the right end of the peak using equations (4) and (5); where: V peak Indicates the peak voltage value of the IC curve; ΔV reg Indicates voltage interval;

[0019]

[0020]

[0021] 3) Calculate the interval capacity Q using equation (6). reg In the formula: Q reg Q1(V1) represents the capacity within the peak range of the IC curve, which is also the peak area corresponding to the peak range; Q2(V2) represents the discharge capacity corresponding to the voltage V1 to the left of the peak; Q2(V2) represents the discharge capacity corresponding to the voltage V2 to the right of the peak.

[0022] Q reg =Q2(V2)-Q1(V1) (12).

[0023] Furthermore: the peak voltage is the voltage value V at the peak of the battery charging IC curve. peak .

[0024] Furthermore: the construction process of the BP neural network is as follows:

[0025] 1) Determine the topology of the BP neural network and calculate the initial weights and thresholds of the BP neural network;

[0026] 2) Population initialization: Randomly generate initial population individuals as the parent population, encode the chromosomes of the parent population, and assign the individual encoding to the weights and thresholds of the BP neural network.

[0027] 3) Calculate the fitness function. The fitness of an individual is a key indicator for judging the quality of an individual. The mean square error function is selected as the fitness function to solve for the fitness of the individuals in the parent population.

[0028] 4) Selecting, crossovering, and mutation operations are performed on parent individuals to generate offspring populations. The selection operation selects a fixed number of individuals from the old population with a certain probability to form a new population. The probability of an individual being selected is related to its fitness value. Individuals with higher fitness have a greater probability of being selected, while individuals with low fitness are eliminated. The crossover operation selects two individuals to exchange information by swapping their chromosome positions, thereby generating new individuals as offspring. The mutation operation causes the position of chromosomes in the offspring to mutate with a certain low probability.

[0029] 5) Calculate the fitness of the offspring population, iterate the population until the iteration objective is met, and output the optimal individual;

[0030] 6) Decode the optimal individual, solve for the optimal initial weights and thresholds, and assign them to the BP neural network;

[0031] 7) Backpropagation (BP) neural networks rapidly adjust and update the weights and thresholds of the hidden and output layers through error backpropagation and algorithms.

[0032] 8) Repeatedly train the BP neural network until the output error meets the accuracy requirements and the predicted SOH is output.

[0033] Accordingly, a computer-readable storage medium storing one or more programs: the one or more programs include instructions that, when executed by a computing device, cause the computing device to perform any of the methods described above.

[0034] Accordingly, a computing device includes:

[0035] One or more processors, one or more memories, and one or more programs, wherein the one or more programs are stored in the one or more memories and configured to be executed by the one or more processors, and the one or more programs include instructions for performing any of the methods described above.

[0036] The beneficial effects achieved by this invention are as follows:

[0037] This invention proposes a data-driven method for analyzing battery health status. Based on battery voltage and temperature data, a capacity increment curve is obtained, health factors characterizing battery health status are extracted, and a GA-BP neural network is used to estimate the battery health status. Attached Figure Description

[0038] Figure 1 This is a schematic diagram of the BP neural network analysis process in this invention;

[0039] Figure 2 This is a schematic diagram of the experimental results of the present invention;

[0040] Figure 3 This is a schematic diagram of the experimental results of the present invention;

[0041] Figure 4 This is a schematic diagram of the experimental results of the present invention;

[0042] Figure 5 This is a schematic diagram of the experimental results of the present invention. Detailed Implementation

[0043] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and should not be used to limit the scope of protection of the present invention.

[0044] The battery health status analysis method of the present invention includes a health factor feature extraction method.

[0045] Health factor feature extraction method: used to extract health factors that can reflect the health status of the battery, and then realize battery health status analysis through estimation methods.

[0046] Health factor feature extraction methods include voltage-signal-energy (VSE) extraction, temperature-signal-energy (TSE) extraction, and interval capacity Q. reg Extraction method, peak voltage V peak Extraction method components.

[0047] The voltage signal energy extraction method is shown in equation (1).

[0048]

[0049] In the formula: v(t) represents the battery voltage value, t0 represents the start time of the measurement, and t1 represents the end time of the measurement.

[0050] The method for extracting energy from temperature signals is shown in equation (2).

[0051]

[0052] In the formula: T(t) represents the battery temperature value, t0 represents the start time of the measurement, and t1 represents the end time of the measurement.

[0053] Peak voltage V peak It is determined by the voltage value at the peak of the battery charging IC curve.

[0054] Interval capacity Q reg The extraction methods are shown in equations (3), (4), (5), and (6). Where Q represents battery capacity, V represents battery voltage, t represents time, and I represents current. peak Indicates the peak voltage value of the IC curve; ΔV reg Indicates the voltage interval; V1 is the voltage at the left end of the peak; V2 is the voltage at the right end of the peak. Q reg Q1(V1) represents the capacity within the peak range of the IC curve, which is also the peak area corresponding to the peak range; Q2(V2) represents the discharge capacity corresponding to the voltage V1 to the left of the peak; Q2(V2) represents the discharge capacity corresponding to the voltage V2 to the right of the peak.

[0055]

[0056]

[0057]

[0058] Q reg =Q2(V2)-Q1(V1) (18)

[0059] like Figure 1 As shown, the battery health status analysis method of the present invention is implemented by the following steps:

[0060] (1) Extracting neural network feature parameters based on battery aging data. In this paper, voltage signal energy, temperature signal energy, interval capacity, and peak voltage are extracted as input features of the BP neural network, and the battery health state (SOH) is used as the output feature.

[0061] (2) Determine the topology of the BP neural network and calculate the initial weights and thresholds of the BP neural network.

[0062] (3) Population initialization. Randomly generate initial population individuals as the parent population, encode the chromosomes of the parent population, and assign the individual encodings to the weights and thresholds of the BP neural network.

[0063] (4) Calculate the fitness function. The fitness of an individual is a key indicator for judging the quality of an individual. This paper chooses the mean square error function as the fitness function to solve for the fitness of individuals in the parent population.

[0064] (5) Selection, crossover, and mutation operations are performed on parent individuals to generate offspring populations. Selection involves choosing a fixed number of individuals from the old population with a certain probability to form a new population. The probability of an individual being selected is related to its fitness value; individuals with higher fitness have a greater probability of being selected, while individuals with low fitness are eliminated. Crossover involves selecting two individuals and exchanging chromosome positions to generate new individuals as offspring. Mutation involves causing a mutation in the position of chromosomes in the offspring with a certain low probability.

[0065] (6) Calculate the fitness of the offspring population, iterate the population until the iteration target is met, and output the optimal individual.

[0066] (7) Decode the optimal individual, solve for the optimal initial weights and thresholds, and assign them to the BP neural network.

[0067] (8) The BP neural network quickly adjusts and updates the weights and thresholds of the hidden layer and the output layer through the backpropagation algorithm.

[0068] (9) Repeatedly train the BP neural network until the output error meets the accuracy requirements and output the predicted SOH.

[0069] like Figure 2-4 As shown, this invention addresses the health status problem of lithium-ion batteries during use by researching a lithium-ion battery health status analysis method based on data feature mining. The health status estimation of lithium-ion batteries includes the following steps. The specific implementation of this invention is described below using the NASA battery dataset as an example.

[0070] (1) Processing of lithium-ion battery aging data.

[0071] (2) Obtain battery voltage and temperature data, and extract four different health factors: voltage signal energy (VSE), temperature signal energy (TSE), interval capacity Qreg, and peak voltage Vpeak.

[0072] (3) Extract the feature parameters of the GA-BP neural network. Based on the discharge voltage signal, temperature signal and capacity increment curve of the NASA battery dataset, solve [VSE; TSE; Qreg; Vpeak] as the input parameters of the GA-BP neural network and the battery SOH as the output parameter.

[0073] (4) First, set the basic parameters of the BP neural network. The number of nodes in the input layer, hidden layer, and output layer of the BP neural network are set to 4, 1, and 9, respectively. The linear function "purelin" is selected as the output layer transfer function, and the LM algorithm function "trainlm" with a fast convergence speed is selected as the training function. The training times of the neural network are set to 1000, the training objective is 0.01, and the learning rate is 0.01. Next, determine the basic parameters of the genetic algorithm. The number of individuals in the population is 40, the maximum number of generations is 50, the crossover probability is 0.7, and the mutation probability is 0.01.

[0074] (5) Conduct training and verification.

[0075] (6) Based on the SOH trajectory of lithium-ion batteries and the relative error, root mean square error, mean absolute error, and mean relative error, the prediction accuracy and adaptability of the battery health state estimation method based on GA-BP neural network are analyzed. The verification data are as follows.

[0076] Table 1 Evaluation metrics of GA-BP neural networks based on NASA datasets

[0077]

[0078] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

[0079] A computer-readable storage medium storing one or more programs, the one or more programs including instructions that, when executed by a computing device, cause the computing device to perform a data-driven battery health state analysis method.

[0080] A computing device includes one or more processors, one or more memories, and one or more programs, wherein the one or more programs are stored in the one or more memories and configured to be executed by the one or more processors, and the one or more programs include instructions for performing a data-driven battery health state analysis method.

[0081] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0082] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0083] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0084] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0085] The above are merely embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention are included within the scope of the claims of the present invention pending approval.

Claims

1. A data-driven method for analyzing battery health status, characterized in that: Extract the battery's health factors, which include: voltage signal energy, temperature signal energy, interval capacity, and peak voltage; Input the voltage signal energy, temperature signal energy, interval capacity, and peak voltage into a pre-constructed BP neural network; The BP neural network outputs the battery health status (SOH). The method for extracting the interval capacity is as follows: 1) Calculate the battery IC curve using equation (3); （3）； In the formula: Q represents battery capacity, V represents battery voltage, t represents time, and I represents current; 2) Calculate the voltage V1 at the left end of the peak and the voltage V2 at the right end of the peak using equations (4) and (5); where: V peak Indicates the peak voltage value of the IC curve; ΔV reg Indicates voltage interval; （4）； （5）； 3) Calculate the interval capacity Q using equation (6). reg In the formula: Q reg Q1(V1) represents the capacity within the peak range of the IC curve, which is also the peak area corresponding to the peak range; Q2(V2) represents the discharge capacity corresponding to the voltage V1 to the left of the peak; Q2(V2) represents the discharge capacity corresponding to the voltage V2 to the right of the peak. （6）； The construction process of the BP neural network is as follows: 1) Determine the topology of the BP neural network and calculate the initial weights and thresholds of the BP neural network; 2) Population initialization: Randomly generate initial population individuals as the parent population, encode the chromosomes of the parent population, and assign the individual encoding to the weights and thresholds of the BP neural network; 3) Calculate the fitness function. The fitness of an individual is a key indicator for judging the quality of an individual. The mean square error function is selected as the fitness function to solve for the fitness of the individuals in the parent population. 4) Selecting, crossovering, and mutation operations are performed on parent individuals to generate offspring populations. The selection operation selects a fixed number of individuals from the old population with a certain probability to form a new population. The probability of an individual being selected is related to its fitness value; individuals with higher fitness have a greater probability of being selected, while individuals with low fitness are eliminated. The crossover operation selects two individuals to exchange information by swapping their chromosome positions, thereby generating new individuals as offspring. The mutation operation causes the position of chromosomes in the offspring to mutate with a certain low probability. 5) Calculate the fitness of the offspring population, iterate the population until the iteration objective is met, and output the optimal individual; 6) Decode the optimal individual, solve for the optimal initial weights and thresholds, and assign them to the BP neural network; 7) Backpropagation (BP) neural networks rapidly adjust and update the weights and thresholds of the hidden and output layers through error backpropagation and algorithms. 8) Repeatedly train the BP neural network until the output error meets the accuracy requirements and the predicted SOH is output.

2. The data-driven battery health status analysis method according to claim 1, characterized in that: The formula for extracting energy from the voltage signal VSE is: （1）； In the formula: v(t) represents the battery voltage value, t0 represents the start time of the measurement, and t1 represents the end time of the measurement.

3. The data-driven battery health status analysis method according to claim 1, characterized in that: The formula for extracting the energy of the temperature signal TSE is: （2）； In the formula: T(t) represents the battery temperature value, t0 represents the start time of the measurement, and t1 represents the end time of the measurement.

4. The data-driven battery health status analysis method according to claim 1, characterized in that: Peak voltage is the voltage value V at the peak of the battery charging IC curve. peak .

5. A computer-readable storage medium for storing one or more programs, characterized in that: The one or more programs include instructions that, when executed by a computing device, cause the computing device to perform any of the methods according to claims 1 to 4.

6. A computing device, characterized in that, include: One or more processors, one or more memories, and one or more programs, wherein the one or more programs are stored in the one or more memories and configured to be executed by the one or more processors, the one or more programs including instructions for performing any of the methods according to claims 1 to 4.

Images