A SOC estimation method for energy storage management systems based on offline battery model parameter identification

Abstract

The application relates to an energy storage management system SOC estimation method based on offline identification of battery model parameters, comprising the following steps: S1, extracting current characteristics of a direct current side as input parameters required for dynamic testing for obtaining battery model parameters in the application scenarios of peak regulation, frequency regulation and peak-valley regulation of an energy storage system; S2, performing dynamic testing on a battery cell by using dynamic testing cases with different current characteristics, and obtaining dynamic response test data of input current and output voltage of the battery cell; S3, identifying a parameter set of a battery model by using an offline identification method according to the dynamic response test data of the input current and the output voltage of the battery cell; S4, selecting battery model parameters according to system working modes, and the system working modes include peak regulation, frequency regulation and peak-valley regulation; and performing SOC estimation by using the selected battery model parameters. The battery model parameter set required by the application is obtained through offline identification, and an online identification method is not needed, so that the requirements of operation complex calculation and data storage are reduced, and the applicability is wider.

Description

Technical Field

[0001] This invention relates to the field of energy storage battery technology, and more specifically to a method for estimating the State of Charge (SOC) of an energy storage management system based on offline identification of battery model parameters. Background Technology

[0002] Kalman filters offer advantages such as high interference resistance and estimation accuracy, leading to their widespread application in battery management systems. Kalman filters construct state equations using a battery model and employ a recursive method to achieve optimal state estimation; their estimation accuracy is highly dependent on the accuracy of the battery model. In peak shaving, frequency regulation, and peak-valley operation scenarios of energy storage systems, the strong nonlinear characteristics of batteries make using single model parameters for state estimation impractical. Furthermore, real-time parameter estimation methods cannot guarantee the convergence of SOC estimation results; parameter misalignment can cause the entire system to collapse. Summary of the Invention

[0003] The present invention aims to provide a method for estimating the State of Charge (SOC) of an energy storage management system based on offline identification of battery model parameters.

[0004] The present invention is as follows:

[0005] The SOC estimation method for energy storage management systems based on offline battery model parameter identification includes the following steps:

[0006] S1 extracts the DC-side current characteristics as input parameters for dynamic testing to obtain battery model parameters in the application scenarios of peak shaving, frequency regulation, and peak-valley in energy storage systems.

[0007] S2 uses dynamic test cases with different current characteristics to perform dynamic testing on the battery cell and obtain dynamic response test data of the battery cell's input current and output voltage.

[0008] S3. Based on the dynamic response test data of the cell input current and output voltage, the parameter set of the battery model is identified using an offline identification method.

[0009] S4. Select battery model parameters according to the system operating mode, which includes peak shaving, frequency modulation, and peak-valley; use the selected battery model parameters to estimate SOC.

[0010] A preferred embodiment of the SOC estimation method for an energy storage management system based on offline identification of battery model parameters of the present invention is that the dynamic test in step S2 specifically includes the following:

[0011] S201, let stand for 2 hours, charge at 1C constant current to the cutoff voltage, and then charge at constant voltage until the battery is fully charged;

[0012] S202. After resting for 2 hours, continuously discharge the cell with different current rates until the SOC drops from 100% to 5%; S203. After resting for 2 hours, discharge at 0.05C to the cutoff voltage.

[0013] S204. Charge the battery at a constant current of 1C to the cutoff voltage, and then charge it at a constant voltage until the battery is fully charged. Repeat steps S201 to S204 to perform tests for different test cases.

[0014] A preferred embodiment of the energy storage management system SOC estimation method based on offline identification of battery model parameters of the present invention is that, in step S3, the battery model used refers to a first-order RC equivalent circuit model with hysteresis characteristics, and the parameters to be identified include the gain and deviation of the hysteresis voltage hyst, internal resistance, first-order RC circuit resistance, and first-order RC circuit resistance.

[0015] A preferred embodiment of the energy storage management system SOC estimation method based on offline identification of battery model parameters of the present invention is that, in step S3, the offline identification method adopts the least squares method.

[0016] A preferred embodiment of the SOC estimation method for an energy storage management system based on offline identification of battery model parameters of the present invention is as follows: The specific process of the least squares method is as follows:

[0017] S301, according to the equivalent circuit model, the battery terminal voltage equals the sum of the open-circuit voltage and the hysteresis voltage minus the voltage division caused by the first-order RC circuit and the ohmic internal resistance, that is:

[0018] v(t)=OCV(z(t))+hyst-R1i R1 (t)-R0i(t)

[0019] Furthermore, the discrete representation has:

[0020] v[k]=OCV(z[k])+M0s[k]+Mh[k]-R1i R1 [k]-R0i[k]

[0021] Where M and M0 are the gain and deviation of the hysteresis voltage hyst, respectively, and s[k] and h[k] are the sign value of the current direction and the normalized state value of the hysteresis voltage at time k.

[0022] S302, calculate the difference between the lower terminal voltage and the open-circuit voltage at each moment, i.e.:

[0023]

[0024] S303, calculate the normalized state value of the hysteresis voltage, the sign value of the current direction, and the first-order RC circuit current at each time step, and construct the following linear equation, which can be fitted by the least squares method to obtain M, M0, R0, R1;

[0025]

[0026] A preferred embodiment of the SOC estimation method for an energy storage management system based on offline identification of battery model parameters of the present invention further includes the following steps:

[0027] Before step S4, determine whether the communication with the energy management system is normal. If yes, proceed to S4; otherwise, proceed to S5.

[0028] Step S5: Use the battery model parameters from the previous time step to estimate the SOC.

[0029] A preferred embodiment of the energy storage management system SOC estimation method based on offline identification of battery model parameters of the present invention is that, in step S4 or step S5, the Kalman filter algorithm is used to estimate the SOC of each cell.

[0030] A preferred embodiment of the energy storage management system SOC estimation method based on offline identification of battery model parameters of the present invention is that the system operating mode is set by the energy management system.

[0031] The beneficial effects of this invention are as follows: This invention aims to coordinate with the energy management system (EMS) to adjust the parameters of the battery model according to the real-time working mode of the energy storage battery for SOC estimation using the Kalman filter. The battery model parameter set can be identified offline under different operating conditions, thereby overcoming the limitations of the prior art and improving the adaptability of the algorithm in different scenarios.

[0032] Compared with existing technologies, this invention has the advantages of easy implementation, stable performance, and simple architecture. The battery model parameter set required by this invention is obtained through offline identification, eliminating the need for online identification methods and reducing the complexity of computation and data storage requirements for the battery management system. Simultaneously, the offline identified parameter set requires further verification testing of SOC estimation accuracy. This involves conducting charge-discharge tests under peak-shaving, frequency regulation, and peak-valley application scenarios in the energy storage system, recording changes in cell SOC, observing whether there are any jumps in the state value, and comparing it with the SOC corresponding to the open-circuit voltage after resting to confirm the accuracy of the SOC estimation result. This ensures the stability of the estimation algorithm under different parameter inputs. Furthermore, this invention, through collaboration with the energy management system (EMS), can adjust parameters in real time according to the operating mode of the energy storage system, eliminating the need to construct a cloud-edge collaborative hierarchical architecture for parameter adjustment, thereby reducing the complexity of collaborative control. Attached Figure Description

[0033] Figure 1 This is a flowchart of the SOC estimation method for an energy storage management system based on offline identification of battery model parameters, as described in this invention.

[0034] Figure 2 This is a diagram of the energy storage system architecture involved in this invention.

[0035] Figure 3 This is a diagram of the first-order RC equivalent circuit involved in this invention. Detailed Implementation

[0036] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described below are only for explaining the present invention and do not limit the scope of protection of the present invention.

[0037] The terms "first," "second," etc., used in the specification, claims, and embodiments of this application are used to distinguish similar objects, rather than to describe a specific order or sequence.

[0038] The present invention will be further described in detail below through preferred embodiments:

[0039] As attached Figure 1 As shown: A method for estimating the State of Charge (SOC) of an energy storage management system based on offline identification of battery model parameters, including the following steps:

[0040] S1, extract the DC-side current characteristics in the application scenarios of peak shaving, frequency regulation, and peak-valley in the energy storage system as input parameters for dynamic testing required to obtain battery model parameters; the DC-side current characteristics are different in different application scenarios, and the current characteristics refer to the statistical characteristics of the current at different rates and their duration during the charging and discharging process of the battery.

[0041] S2, using dynamic test cases with different current characteristics to perform dynamic testing on the battery cell, and obtaining dynamic response test data of the battery cell's input current and output voltage; the dynamic testing in step S2 specifically includes the following:

[0042] S201, let stand for 2 hours, charge at 1C constant current to the cutoff voltage, and then charge at constant voltage until the battery is fully charged;

[0043] S202. After resting for 2 hours, continuously discharge the cell with different current rates until the SOC drops from 100% to 5%; S203. After resting for 2 hours, discharge at 0.05C to the cutoff voltage.

[0044] S204. Charge the battery at a constant current of 1C to the cutoff voltage, and then charge it at a constant voltage until the battery is fully charged. Repeat steps S201 to S204 to perform tests for different test cases.

[0045] S3, based on the dynamic response test data of cell input current and output voltage, uses an offline identification method to identify the parameter set of the battery model, and integrates the battery model parameters identified under different working conditions into a set, which is then integrated into the energy storage battery management system for real-time parameter adjustment;

[0046] In step S3, the battery model used refers to a first-order RC equivalent circuit model with hysteresis characteristics, as shown in the attached figure. Figure 3 As shown, the parameters to be identified include the gain and bias of the hysteresis voltage hyst, internal resistance, first-order RC circuit resistance, and first-order RC circuit resistance.

[0047] In step S3, the offline identification method employs the least squares method. The specific process of the least squares method is as follows:

[0048] S301, according to the equivalent circuit model, the battery terminal voltage equals the sum of the open-circuit voltage and the hysteresis voltage minus the voltage division caused by the first-order RC circuit and the ohmic internal resistance, that is:

[0049] v(t)=OCV(z(t))+hyst-R1i R1 (t)-R0i(t)

[0050] Furthermore, the discrete representation has:

[0051] v[k]=OCV(z[k])+M0s[k]+Mh[k]-R1i R1 [k]-R0i[k]

[0052] Where M and M0 are the gain and deviation of the hysteresis voltage hyst, respectively, and s[k] and h[k] are the sign value of the current direction and the normalized state value of the hysteresis voltage at time k.

[0053] S302, calculate the difference between the lower terminal voltage and the open-circuit voltage at each moment, i.e.:

[0054]

[0055] S303, calculate the normalized state value of the hysteresis voltage, the sign value of the current direction, and the first-order RC circuit current at each time step, and construct the following linear equation, which can be fitted by the least squares method to obtain M, M0, R0, R1;

[0056]

[0057] S4. Select battery model parameters according to the system operating mode, which includes peak shaving, frequency modulation, and peak-valley operation. Use the selected battery model to estimate the SOC of each cell. The parameters of each cell are the same in the algorithm, but because the voltage changes of each cell are inconsistent during the charging and discharging process, the SOC changes are inconsistent. Therefore, it is necessary to estimate the SOC of each cell.

[0058] This embodiment also includes the following steps:

[0059] Before step S4, determine if communication with the energy management system is normal. If yes, proceed to S4; otherwise, proceed to S5; as shown in the appendix. Figure 2 As shown, the system operating mode is set by the energy management system (EMS) and can be sent to the battery management system via a communication protocol.

[0060] Step S5: Use the battery model parameters from the previous moment to estimate the State of Charge (SOC). It's important to note that in cases of communication failure, the energy storage battery management system will identify the fault and request the energy storage converter to stop charging and discharging within a certain timeframe. However, this process requires response time. To minimize the impact on the SOC estimation algorithm, the battery model parameters from the previous moment must still be used for estimation. This is because if the response time to stop charging and discharging is too long, current will still exist in the system, and using incorrect battery model parameters will result in inaccurate SOC estimation.

[0061] In this embodiment, in either step S4 or step S5, the Kalman filter algorithm is used to estimate the SOC of each cell.

[0062] This invention extracts current features based on various application scenarios of energy storage systems and uses them for parameter identification of battery models, forming multiple parameter sets to solve the problems of poor adaptability of single parameters and high requirements for dynamic parameter identification algorithms.

[0063] By working in conjunction with the Energy Management System (EMS), parameters can be adjusted in real time according to the working mode of the energy storage system, avoiding the problem of complex dynamic parameter adjustment architecture in cloud-edge collaboration.

[0064] The preferred embodiments of this application have been described in detail above with reference to the accompanying drawings. Typical known structures and common knowledge techniques in the preferred embodiments have not been described in detail here. Those skilled in the art can improve and implement the technical solutions of this invention based on the guidance provided in these embodiments and their own capabilities. Some typical known structures, known methods or common knowledge techniques should not be obstacles for those skilled in the art to implement this application.

[0065] The scope of protection claimed in this application shall be determined by the contents of its claims, and the contents described in the invention description, specific embodiments and drawings shall be used to interpret the claims.

[0066] Within the scope of the technical concept of this application, several modifications can be made to the specific implementation of this application, and these modified implementations should also be considered within the protection scope of this application.

Claims

1. A method for state of charge (SOC) estimation of an energy storage management system based on offline identification of battery model parameters, characterized in that, The method comprises the following steps: S1, extracting the current characteristics of the DC side in the application scenarios of energy storage system peak regulation, frequency regulation and peak-valley regulation as input parameters required for dynamic test for obtaining battery model parameters; S2, performing dynamic test on the battery cell using dynamic test cases with different current characteristics to obtain dynamic response test data of the input current and output voltage of the battery cell; S3, identifying the parameter set of the battery model using an offline identification method according to the dynamic response test data of the input current and output voltage of the battery cell; S4, selecting the battery model parameters according to the system working mode, wherein the system working mode includes peak regulation, frequency regulation and peak-valley regulation; and performing SOC estimation using the selected battery model parameters.

2. The method of claim 1, wherein, The dynamic test in step S2 specifically includes the following contents: S201, resting for 2 hours, 1C constant current charging to the cut-off voltage, and then constant voltage charging until the battery reaches the full charge state; S202, resting for 2 hours, continuously discharging the battery cell at different current rates to reduce the SOC from 100% to 5%; S203, resting for 2 hours, discharging at 0.05C to the cut-off voltage; S204, 1C constant current charging to the cut-off voltage, and then constant voltage charging until the battery reaches the full charge state; repeating steps S201 to S204 to test different test cases.

3. The method of claim 1, wherein, In step S3, the battery model refers to a first-order RC equivalent circuit model with hysteresis characteristics, and the parameters to be identified include the gain and offset of the hysteresis voltage hyst, the internal resistance, the first-order RC circuit resistance and the first-order RC circuit resistance.

4. The method of claim 3, wherein, In step S3, the offline identification method uses the least square method.

5. The method of claim 4, wherein, The specific process of the least square method is as follows: S301, according to the equivalent circuit model, the battery terminal voltage is equal to the sum of the open circuit voltage and the hysteresis voltage minus the voltage division of the first-order RC circuit and the ohmic internal resistance, that is: v(t) = OCV(z(t)) + hyst - R1i R1 (t) - R0i(t) Further, the discrete representation is: v[k] = OCV(z[k]) + M0s[k] + Mh[k] - R1i R1 [k] - R0i[k] Wherein, M and M0 are the gain and offset of the hysteresis voltage hyst, s[k] and h[k] are the current direction symbol value and the normalized state value of the hysteresis voltage at time k; S302, calculating the difference between the terminal voltage and the open circuit voltage at each time, that is: S303, calculating the normalized state value of the hysteresis voltage, the current direction symbol value and the first-order RC circuit current at each time, and constructing the following linear equation, that is, M, M0, R0 and R1 can be obtained by fitting using the least square method; 6. The method of claim 1, wherein, Further comprising the following steps: Before step S4, it is judged whether the communication with the energy management system is normal, if yes, jumping to S4; otherwise, jumping to S5; Step S5, using the battery model parameters of the previous time to perform SOC estimation.

7. The method of claim 6, wherein, In step S4 or step S5, the Kalman filter algorithm is used to estimate the SOC of each battery cell.

8. The method of claim 1, wherein: The system working mode is set by the energy management system.

Images