KR-20260065627-A - METHOD FOR DIAGNOSING STATE OF HEALTH OF ELECTRIC VEHICLE BATTERY PACK

Abstract

A method for diagnosing the health status of an electric vehicle battery pack is disclosed. The method for diagnosing the health status of an electric vehicle battery according to the present invention comprises: (a) a step of measuring and storing data including the charging current, charging voltage, and charging time of the electric vehicle battery during charging at a regular interval; (b) a step of calculating the cumulative charging current capacity (C a ) expressed by Equation 1 using the data; (c) a step of calculating the incremental capacity, which is the difference between the maximum and minimum values of the cumulative charging current capacity (C a ) for each predetermined charging voltage interval; (d) a step of calculating an incremental capacity curve representing the relationship between the charging voltage and the incremental capacity, with the charging voltage as the x-axis and the incremental capacity as the y-axis; and (e) a step of calculating the in-boundary capacity, which is the area (Ap) of the peak having the largest area among the peaks of the incremental capacity curve. and (f) a step of diagnosing the state of health (SOH) of the electric vehicle battery using the above-mentioned in-boundary capacity; and the method for diagnosing the state of health of an electric vehicle battery according to the present invention has the effect of diagnosing the state of health of the electric vehicle battery during actual charging with a low error rate by using the in-boundary capacity value as a learning factor of an LSTM model.

Inventors

• 홍영선
• 강민준
• 강병수
• 김우중
• 박상준

Assignees

• 한국생산기술연구원

Dates

Publication Date

20260511

Application Date

20241023

Claims (20)

1. (a) A step of measuring and storing data including the charging current, charging voltage, and charging time of the electric vehicle battery at regular intervals during charging; (b) A step of calculating the cumulative charging current capacity (C a ) expressed by Equation 1 using the above data; (c) A step of calculating the incremental capacity, which is the difference between the maximum and minimum values of the accumulated charging current capacity (C a ) for each predetermined charging voltage interval; (d) A step of obtaining an incremental capacity curve that represents the relationship between the charging voltage and the incremental capacity, with the charging voltage on the x-axis and the incremental capacity on the y-axis; (e) a step of determining the in-boundary capacity, which is the area (Ap) of the peak having the largest area among the peaks of the incremental capacity curve; and (f) a step of diagnosing the state of health (SOH) of the electric vehicle battery using the above boundary-inclusive capacity; Method for diagnosing the health status of an electric vehicle battery including: [Equation 1] In the above Equation 1, C a is the cumulative charging current capacity, and I c is the charging current of the electric vehicle battery, and t c is the charging time of the electric vehicle battery.
2. In paragraph 1, A method for diagnosing the health status of an electric vehicle battery, characterized in that the electric vehicle battery comprises one or more types selected from the group consisting of battery cells, battery modules, and battery packs.
3. In paragraph 1, A method for diagnosing the health status of an electric vehicle battery, characterized in that the constant period of step (a) is 0.1 to 5 Hz.
4. In paragraph 1, A method for diagnosing the health status of an electric vehicle battery, characterized in that the measurement precision of the charging voltage in step (c) is 0.001 to 0.1 V in the battery cell and 0.1 to 10 V in the battery pack.
5. In paragraph 1, A method for diagnosing the health status of a battery, characterized in that the area (A P ) of the peak (P) in step (e) is the area of the peak (A P) from the first lowest point in the direction of a charging voltage smaller than the charging voltage at the peak (P) to the first lowest point in the direction of a charging voltage larger than the charging voltage at the peak ( P ).
6. In paragraph 1, The above step (f) (f-1) A step of obtaining a training model for diagnosing battery health status by training a machine learning model using training data including the above-mentioned boundary-in-capacity and health status; (f-2) A step of verifying whether the learning model has reached the target diagnostic performance using verification data including the boundary-inclusive capacity and health status to obtain a diagnostic model that has reached the target battery health status diagnostic performance; and (f-3) A method for diagnosing the health status of an electric vehicle battery from the boundary-inclusive capacity using the above diagnostic model; characterized by including the step of diagnosing the health status of the electric vehicle battery.
7. In paragraph 6, A method for diagnosing the health status of an electric vehicle battery, characterized in that the machine learning model of step (f-1) comprises one or more selected from the group consisting of Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), Autoencoder (AE), Long Short-Term Memory Autoencoder (LSTM-AE), Long Short-Term Memory Variational Autoencoder (LSTM-VAE), Gated Recurrent Unit Autoencoder (GRU-AE), and Gated Recurrent Unit Variational Autoencoder (GRU-VAE).
8. In paragraph 6, A method for diagnosing the health status of an electric vehicle battery, characterized in that the machine learning model of step (f-1) above is a Long Short-Term Memory (LSTM) model, and the LSTM model adds a dropout layer after the LSTM layer to prevent the machine learning model from overfitting.
9. In paragraph 6, A method for diagnosing the health status of an electric vehicle battery, characterized in that the ratio of the training data to the verification data is 6.5:3.5 to 9.0:1.0.
10. In paragraph 6, A method for diagnosing the health status of an electric vehicle battery, characterized in that the above training data and verification data each include one or more selected from the group consisting of charging time, incremental capacity corresponding to the maximum peak, SOC (state of charge), cumulative charging current, cumulative discharging current, battery cell voltage, temperature, and driving distance.
11. In paragraph 6, A method for diagnosing the health status of an electric vehicle battery, characterized by obtaining a diagnostic model that has reached a target diagnostic performance by training the learning model of step (f-2) using validation data including the bounded-in capacity and health status, until at least one selected from a group consisting of the mean squared error (MSE) of the learning model represented by Equation 4 below, the mean absolute error (MAE) of the learning model represented by Equation 5 below, and the root mean square error (RMSE) of the learning model represented by Equation 6 below, each having a value less than or equal to a predetermined value. [Equation 3] [Equation 4] [Equation 5] In Equations 3 to 5, y is the actual value, and is the predicted value, and n is the number of data.
12. A computer for diagnosing the health status of an electric vehicle battery (a) A step of measuring and storing data including the charging current, charging voltage, and charging time of the electric vehicle battery at regular intervals during charging; (b) A step of calculating the cumulative charging current capacity (C a ) expressed by Equation 1 using the above data; (c) A step of calculating the incremental capacity, which is the difference between the maximum and minimum values of the accumulated charging current capacity (C a ) for each predetermined charging voltage interval; (d) A step of obtaining an incremental capacity curve that represents the relationship between the charging voltage and the incremental capacity, with the charging voltage on the x-axis and the incremental capacity on the y-axis; (e) a step of determining the in-boundary capacity, which is the area (Ap) of the peak having the largest area among the peaks of the incremental capacity curve; and (f) a step of diagnosing the state of health (SOH) of the electric vehicle battery using the above boundary-inclusive capacity; Computer-readable medium storing a software program to be executed: [Equation 1] In the above Equation 1, C a is the cumulative charging current capacity, and I c is the charging current of the electric vehicle battery, and t c is the charging time of the electric vehicle battery.
13. In Paragraph 12, A computer-readable medium storing a software program characterized in that the electric vehicle battery comprises one or more types selected from the group consisting of battery cells, battery modules, and battery packs.
14. In Paragraph 12, A computer-readable medium storing a software program characterized in that the constant period of step (a) is 0.1 to 5 Hz.
15. In Paragraph 12, A computer-readable medium having a software program stored thereon, characterized in that the measurement precision of the charging voltage in step (c) is 0.001 to 0.1 V in the battery cell and 0.1 to 10 V in the battery pack.
16. In Paragraph 12, A computer-readable medium storing a software program, characterized in that the area (A P ) of the peak (P) in step (e) is the area of the peak (A P) from the first lowest point in the direction of a charging voltage smaller than the charging voltage at the peak (P) to the first lowest point in the direction of a charging voltage larger than the charging voltage at the peak ( P ).
17. In Paragraph 12, The above step (f) (f-1) A step of obtaining a training model for diagnosing battery health status by training a machine learning model using training data including the above-mentioned boundary-in-capacity and health status; (f-2) A step of verifying whether the learning model has reached the target diagnostic performance using verification data including the boundary-inclusive capacity and health status to obtain a diagnostic model that has reached the target battery health status diagnostic performance; and (f-3) A computer-readable medium having a software program stored therein, characterized by including the step of diagnosing the health status of an electric vehicle battery from the boundary-inclusive capacity using the above diagnostic model.
18. In Paragraph 17, A computer-readable medium storing a software program characterized in that the machine learning model of step (f-1) comprises one or more selected from the group consisting of Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), Autoencoder (AE), Long Short-Term Memory Autoencoder (LSTM-AE), Long Short-Term Memory Variational Autoencoder (LSTM-VAE), Gated Recurrent Unit Autoencoder (GRU-AE), and Gated Recurrent Unit Variational Autoencoder (GRU-VAE).
19. A communication unit that receives data including the charging current, charging voltage, and charging time of an electric vehicle battery during charging; A processor for diagnosing the state of health (SOH) of the electric vehicle battery using in-boundary capacity; and A storage unit that provides storage space necessary for the processor to diagnose the health status of the electric vehicle battery; A diagnostic system for the health status of an electric vehicle battery that includes.
20. In Paragraph 19, The above processor (a) A step of measuring and storing data including the charging current, charging voltage, and charging time of the electric vehicle battery at regular intervals during charging; (b) A step of calculating the cumulative charging current capacity (C a ) expressed by Equation 1 using the above data; (c) A step of calculating the incremental capacity, which is the difference between the maximum and minimum values of the accumulated charging current capacity (C a ) for each predetermined charging voltage interval; (d) A step of obtaining an incremental capacity curve that represents the relationship between the charging voltage and the incremental capacity, with the charging voltage on the x-axis and the incremental capacity on the y-axis; (e) a step of determining the in-boundary capacity, which is the area (Ap) of the peak having the largest area among the peaks of the incremental capacity curve; and (f) a step of diagnosing the state of health (SOH) of the electric vehicle battery using the above boundary-inclusive capacity; Electric vehicle battery health diagnosis system including: [Equation 1] In the above Equation 1, C a is the cumulative charging current capacity, and I c is the charging current of the electric vehicle battery, and t c is the charging time of the electric vehicle battery.

Description

Method for Diagnosing State of Health of Electric Vehicle Battery Pack The present invention relates to a method for diagnosing the health status of an electric vehicle battery pack. Lithium batteries are rechargeable batteries used as power sources in various technologies, such as electric vehicles, portable devices, and energy storage systems, due to their high power density, low self-discharge rate, and light weight. However, as accidents caused by explosions during use of lithium batteries occur frequently, research on battery management systems is necessary. The aforementioned battery management system improves stability by monitoring the battery's voltage, current, and temperature. Furthermore, it estimates the state of charge (SOC) and state of health (SOH) through the battery's internal parameters. Battery State of Charge (SOC) and State of Health (SOH) cannot be measured directly and must be determined using algorithms that utilize internal or external parameters such as voltage, current, and temperature. Conventionally, various studies have been conducted to estimate the health status of batteries. However, since these studies focus on laboratory-scale battery cell experiments conducted under full charge/discharge conditions and constant current and temperature, they often fail to account for actual driving environments in automobiles. Therefore, to address these issues, there is a need to develop a method for diagnosing battery health using actual electric vehicle battery charging data, and there is an increasing trend of research on methods to diagnose battery health and improve the accuracy of such diagnosis. These drawings are for reference to explain exemplary embodiments of the present invention, and therefore, the technical concept of the present invention should not be interpreted as being limited to the attached drawings. FIG. 1 is a flowchart showing the sequence of a method for diagnosing the health status of an electric vehicle battery pack according to Test Examples 4 to 6 after calculating the incremental capacity of an electric vehicle battery cell and a battery pack according to Examples 1 and 2 of the present invention. Figure 2 is a graph showing the incremental capacity according to the voltage of the battery cell of Example 1 of the present invention. Figure 3 is a graph showing the incremental capacity according to the voltage of the battery pack of Example 2 of the present invention. Figure 4 illustrates a method for deriving an in-boundary capacity value from an incremental capacity curve according to the voltage of a battery cell according to Example 1 of the present invention. Figure 5 illustrates a method for deriving an in-boundary capacity value from an incremental capacity curve according to the voltage of a battery pack according to Example 2 of the present invention. Figure 6 is a graph showing the in-boundary capacity value according to the cycle of the battery cell of Test Example 1 of the present invention. Figure 7 is a graph showing the In-boundary Capacity value according to the cumulative driving distance of the battery pack of Test Example 2 of the present invention. Figure 8 is a graph showing the In-boundary Capacity and SOH values according to the cycle of the battery cell of Test Example 3 of the present invention. Figure 9 shows a heatmap of the correlation between the In-boundary Capacity and SOH of a battery cell according to Test Example 3 of the present invention. FIG. 10 shows the configuration of the LSTM models of Test Examples 4 to 6 of the present invention. Figure 11 shows a correlation heatmap between input parameters of an LSTM model of a battery pack according to Test Examples 4 to 6 of the present invention. Figure 12 is a graph showing the estimation of the in-boundary capacity of the LSTM model of the battery pack according to Test Example 4 of the present invention at a training rate of 50%. Figure 13 is a graph showing the estimation of the in-boundary capacity of the LSTM model of the battery pack according to Test Example 5 of the present invention at a training rate of 60%. Figure 14 is a graph showing the estimation of the in-boundary capacity of the LSTM model of the battery pack according to Test Example 6 of the present invention at a training rate of 70%. Figure 15 is a graph showing the error in the estimation of LSTM-based In-boundary Capacity according to the cumulative driving distance for each training ratio of the battery pack of the present invention. Hereinafter, embodiments of the present invention will be described in detail with reference to the attached drawings so that those skilled in the art can easily implement the present invention. However, the following description is not intended to limit the present invention to specific embodiments, and detailed descriptions of related prior art are omitted if it is determined that such detailed descriptions could obscure the essence of the present invention. The terms used