A self-learning method for the voltage curve of a battery in an energy storage system

Abstract

The present invention belongs to the technical field of battery management. A method for self-learning the voltage curve of a battery in an energy storage system comprises the following steps in sequence: collecting the real-time voltage, current and temperature values of the energy storage system; respectively performing mean value processing on the collected real-time voltage, current and temperature values; respectively storing the voltage, current and temperature values; calculating and storing the constant current parameter RF; extracting the stored voltage, current and temperature values and generating a curve. The advantage of this patent is that it self-learns based on the actual voltage sampling value, solving the problem of inconsistent battery voltage characteristics.

Description

Technical Field

[0001] The present invention belongs to the technical field of battery management, and particularly relates to a method for self-learning the voltage curve of batteries in an energy storage system. Background Art

[0002] The rapid development of new energy fields such as electric vehicles and energy storage power stations has led to the application of a large number of electrochemical batteries such as lithium-ion power batteries or lead-acid batteries in environments such as electric vehicles and energy storage power stations. As a representative lithium battery, it has received more attention and has been applied in many fields. It has been applied on a large scale in the energy storage field, providing an effective tool for power grid frequency modulation, peak shaving, and valley filling, and improving the stability and reliability of the power grid. With the rapid development of new energy vehicles, lithium batteries also have very good prospects in the transportation field.

[0003] Environments such as electric vehicles and energy storage power stations all require a large number of batteries to form a high voltage. To adapt to the safe operation of the battery pack, it is essential to equip a corresponding battery management system. The battery management system needs to analyze and calculate relevant battery information, such as the state of charge, that is, SOC. No matter what method is used, basically the relevant analysis and calculation of the battery management system for this will generally use the built-in relationship curve between voltage and SOC for analysis. Due to reasons such as production processes of single batteries, and differences between each battery and the initial state due to long-term operation later, in the case of using the built-in relationship curve mode and being unable to update it, as the operation time increases, it becomes more and more unable to adapt to the current situation of the battery. Summary of the Invention

[0004] The purpose of the present invention is to solve the above problems and provide a method and device that can enable the voltage curve of the batteries in the energy storage system to perform self-learning.

[0005] For the purpose of the present invention, the following technical solutions are adopted to achieve it:

[0006] A method for self-learning the voltage curve of batteries in an energy storage system, which successively passes through the following steps:

[0007] S1, collect the real-time voltage, current, and temperature values of the energy storage system;

[0008] S2, according to the actual capacity of the battery, when the accumulated SOC change amount by the ampere-hour integration method reaches the set value, perform mean processing on the real-time voltage, current, and temperature values collected within this change amount, where SOC is a value from 0 to 100;

[0009] Preferably, the set value at which the accumulated SOC change amount by the ampere-hour integration method reaches is set to 1%.

[0010] S3. After averaging the voltage, current, and temperature values processed by cumulative SOC change using the ampere-hour integration method in S2, store them as V, I, T respectively, and the cumulative SOC change dSOC for each ampere-hour integration method;

[0011] V[len] = {V1...V n ...V len} (1)

[0012] I[len] = {I1...I n ...I len} (2)

[0013] T[len] = {T1...T n ...T len} (3)

[0014] dSOC[len] = {dSOC1...dSOC n ...dSOC len} (4)

[0015] where len is the length of each data.

[0016] S4. Calculate the constant current parameter RF. Calculate RF from the I[len] stored in S3 using formula (5), and store RF, where RF0 = 0 and Q is the actual capacity of the battery.

[0017] When and

[0018] RF[len] = {RF1...RF n ...RF len} (6)

[0019] When the open-circuit duration in the cumulative SOC change for each ampere-hour integration method is greater than the open-circuit duration standard time, RF = 0.

[0020] Preferably, err1 is 0.1 and err2 is 50%.

[0021] Preferably, the open-circuit duration standard time is 1 hour.

[0022] S5. When the charge and discharge are completed and the set conditions are met, extract the voltage, current, and temperature values stored in S3 and generate curves.

[0023] S51. When the number of times the voltage during the charging process is continuously greater than or equal to the self-learning judgment point Vup of the charging voltage curve accumulates to N times, it is recorded as reaching the set condition time. During the period from reaching the set condition time to the end of charging, the SOC corresponding to the charging SOC curve is found through the voltage during the charging process, and the maximum SOC is recorded, denoted as SOCVup.

[0024] When the number of times the voltage during the discharging process is continuously less than or equal to the self-learning judgment point Vdown of the discharging voltage curve accumulates to N times, it is recorded as reaching the set condition time. During the period from reaching the set condition time to the end of discharging, the SOC corresponding to the discharging SOC curve is found through the voltage during the discharging process, and the minimum SOC is recorded, denoted as SOCVdown.

[0025] S52. Judge the validity of RF[len]. When there are more than M consecutive cases where RF[len] is 0, this data is judged as invalid.

[0026] S53. When RF[len] is judged to be valid, in the case of charging data, with SOCVup as the end point, a new charging SOC curve is formed in sequence from len to 1 minus dSOC[len], that is, the V[len], I[len], and T[len] corresponding to the SOC.

[0027] When RF[len] is judged to be valid, in the case of discharging data, first reverse the saved V[len], I[len], T[len], dSOC[len], and RF[len], that is, the data in each array is swapped from the head and tail to the middle. With SOCVdown as the starting point, a new discharging SOC curve is formed in sequence from 1 to len plus dSOC[len], that is, the V[len], I[len], and T[len] corresponding to the SOC.

[0028] S54. When there is a 0 in RF[len], the V[len], I[len], and T[len] are interpolations of the V[len], I[len], and T[len] when RF[len] is 1 in the data before or after it.

[0029] S55. Judge the validity of the newly formed charge-discharge SOC curve. It is valid when there is no decreasing data in V[len].

[0030] Preferably, the set condition in S5 is that the corresponding SOC during discharging ≤ 5% and the corresponding SOC during charging ≥ 95%.

[0031] Preferably, the self-learning judgment point Vup of the charging voltage curve in S51 is the voltage point corresponding to SOC = 95 in the charging SOC curve.

[0032] Preferably, the self-learning judgment point Vdown of the discharge voltage curve in S51 is the voltage point corresponding to the SOC of 5 in the discharge SOC curve.

[0033] Preferably, in S51, when the number of times that the voltage is continuously greater than or equal to the self-learning judgment point Vup of the charging voltage curve during the charging process accumulates to N times, and when the number of times that the voltage is continuously less than or equal to the self-learning judgment point Vdown of the discharge voltage curve during the discharge process accumulates to N times, this N is 5.

[0034] Preferably, in S52, when there are M consecutive cases where RF[len] is 0, this M is 5.

[0035] Preferably, the interpolation method in S54 is linear interpolation.

[0036] In summary, the advantages of this patent are to solve the problems existing in the above-mentioned prior art, and to provide a method for self-learning the voltage curve of a battery in an energy storage system. This method is based on self-learning of actual voltage sampling values, solves the problem of inconsistent battery voltage characteristics, is easy to implement, and is convenient for engineering applications. Description of the Drawings

[0037] Figure 1 : Schematic diagram of the process of the method for self-learning the voltage curve of a battery in an energy storage system of this patent and the relationship between each module. Detailed Embodiment

[0038] The following will give a detailed description of the specific embodiments of the invention with reference to the drawings.

[0039] The present invention provides a device for self-learning the voltage curve of a battery in an energy storage system, which is composed of a battery detection module, a battery data processing module, a battery data storage module, a battery constant current judgment module, and a battery self-learning module.

[0040] A method for self-learning the voltage curve of a battery in an energy storage system includes the following steps in sequence:

[0041] S1, collecting the real-time voltage, current, and temperature values of the energy storage system through the battery detection module;

[0042] S2, through the battery data processing module, according to the actual capacity of the battery, when the change amount of the SOC accumulated by the ampere-hour integration method reaches the set value, the real-time voltage, current, and temperature values collected within this change amount are respectively subjected to mean processing, where the SOC is a value from 0 to 100;

[0043] Preferably, the value at which the change amount of the SOC accumulated by the ampere-hour integration method reaches the set value is set to 1%.

[0044] Example 1: When the actual battery capacity is 100 Ah, 1% of each SOC change is 1 Ah, where the ampere-hour integration method is the integration data of current and time.

[0045] S3. Through the battery data storage module, the voltage, current, and temperature values after mean processing of each ampere-hour integration method cumulative SOC change in S2 are respectively stored as V, I, T, and each ampere-hour integration method cumulative SOC change dSOC.

[0046] V[len] = {V1...V n ...V len} (1)

[0047] I[len] = {I1...I n ...I len} (2)

[0048] T[len] = {T1...T n ...T len} (3)

[0049] dSOC[len] = {dSOC1...dSOC n ...dSOC len} (4)

[0050] Among them, len is the length of each data.

[0051] Example 2: The data storage situation of a certain section on a certain occasion (where len = 5) is as follows:

[0052] V[5] = {3.245 3.246 3.247 3.248 3.249}

[0053] I[5] = {30 30 30 30 30}

[0054] T[5] = {25 25 25 25 25}

[0055] dSOC[5] = {1 1 1 1 1}

[0056] S4. Through the battery constant current judgment module, calculate the constant current parameter RF. Calculate RF from the I[len] stored in S3 through formula (5), and store RF through the battery data storage module, where RF0 = 0 and Q is the actual battery capacity.

[0057] When And

[0058] RF[len] = {RF1...RF n ...RFlen} (6)

[0059] When the open - circuit duration in each ampere - hour integration method for accumulating the SOC change amount is greater than the open - circuit duration standard time, RF = 0.

[0060] Preferably, err1 is 0.1 and err2 is 50%.

[0061] Preferably, the open - circuit duration standard time is 1 hour.

[0062] Example 3: When the actual battery capacity Q is 100 Ah, in step S3, when the current I this time is 50 and the previous current is 45,

[0063] S5. After the charge - discharge ends and meets the set conditions, the battery self - learning module extracts the voltage, current, and temperature values stored in S3 and generates and obtains a new curve.

[0064] S51. When the number of times the voltage during the charging process is continuously greater than or equal to the self - learning judgment point Vup of the charging voltage curve accumulates to N times, it is recorded as the time when the set conditions are met. During the period from the time when the set conditions are met to the end of charging, the SOC corresponding to the charging SOC curve is found through the voltage during the charging process, and the maximum SOC is recorded, denoted as SOCVup;

[0065] When the number of times the voltage during the discharging process is continuously less than or equal to the self - learning judgment point Vdown of the discharging voltage curve accumulates to N times, it is recorded as the time when the set conditions are met. During the period from the time when the set conditions are met to the end of discharging, the SOC corresponding to the discharging SOC curve is found through the voltage during the discharging process, and the minimum SOC is recorded, denoted as SOCVdown.

[0066] Example 4: Suppose the SOC is greater than 95 and a certain charging SOC curve is as follows.

[0067] SOC 95 96 97 98 99 100 Voltage 3.450 3.460 3.475 3.495 3.520 3.650

[0068] The self - learning judgment point Vup of the charging voltage curve is 3.450.

[0069] When the charging voltage is greater than Vup until the end of charging, the SOC data in the above table is found through the voltage, and the maximum SOC is recorded, denoted as SOCVup. For example, when searching through 3.495, the SOC is found to be 98, and when searching through 3.475, the SOC is found to be 97, and SOCVup is 98.

[0070] S52. Judge the validity of RF[len]. When there are M consecutive cases where RF[len] is 0, this data is judged to be invalid.

[0071] Example 5: When RF[9] = {1 1 0 0 0 0 0 0 1}, the data is judged as invalid this time.

[0072] S53. When RF[len] is judged as valid, in the case of charging data, with SOCVup as the end point, a new charging SOC curve is formed in sequence from len to 1 minus dSOC[len], that is, V[len], I[len], T[len] corresponding to SOC.

[0073] When RF[len] is judged as valid, in the case of discharging data, first, the saved V[len], I[len], T[len], dSOC[len], and RF[len] are reversed, that is, the data in each array is swapped from the head and tail to the middle. With SOCVdown as the starting point, a new discharging SOC curve is formed in sequence from 1 to len plus dSOC[len], that is, V[len], I[len], T[len] corresponding to SOC.

[0074] Example 6: In the case of discharging data, first, the saved V[len], I[len], T[len], dSOC[len], and RF[len] are reversed, that is, the data in each array is swapped from the head and tail to the middle. The following is the method:

[0075] When in the case of discharging data, if the saved V[len] and I[len] data are as follows:

[0076] V[5] = {3.020 2.990 2.960 2.930 2.900}

[0077] I[5] = {48 49 50 51 52}

[0078] The reversal is as follows:

[0079] V[5] = {2.900 2.930 2.960 2.990 3.020}

[0080] I[5] = {52 51 50 49 48}

[0081] S54. When there is a 0 in RF[len], V[len], I[len], and T[len] are the interpolations of V[len], I[len], and T[len] when RF[len] is 1 in the previous or subsequent data.

[0082] Example 7: The latter part of a certain charging SOC curve is as follows.

[0083] SOC 91 92 93 94 95 96 97 98 99 100 Voltage 3.410 3.420 3.430 3.440 3.450 3.460 3.475 3.495 3.520 3.650

[0084] When the charging voltage is greater than Vup until the end of charging, search for the SOC data in the above table by voltage, record the maximum SOC among them, denoted as SOCVup. For example, when searching with 3.520, the SOC is found to be 99, and when searching with 3.475, the SOC is found to be 97. In fact, SOCVup is 99.

[0085] V[9] = {3.415 3.425 3.435 3.445 3.455 3.465 3.480 3.500 3.530}

[0086] I[9] = {30 30 30 29 30 28 30 30 30}

[0087] T[9] = {25 25 26 25 25 25 25 25 25}

[0088] dSOC[9] = {1 1 1 1 1 1 1 1 1}

[0089] RF[9] = {1 1 1 1 1 1 1 1 1}

[0090] That is, starting from SOC = 99 and reversing, the new charging SOC curve is as follows:

[0091] SOC 91 92 93 94 95 96 97 98 99 Voltage 3.415 3.425 3.435 3.445 3.455 3.465 3.480 3.500 3.530

[0092] Example 8: Suppose the latter part of a certain charging SOC curve is as shown below.

[0093] SOC 91 92 93 94 95 96 97 98 99 100 Voltage 3.410 3.420 3.430 3.440 3.450 3.460 3.475 3.495 3.520 3.650

[0094] When the charging voltage is greater than Vup until the end of charging, search for the SOC data in the above table by voltage, record the maximum SOC among them, denoted as SOCVup. For example, when searching with 3.520, the SOC is found to be 99, and when searching with 3.475, the SOC is found to be 97. SOCVup is 99.

[0095] V[9] = {3.415 3.425 3.405 3.445 3.455 3.465 3.450 3.500 3.530}

[0096] RF[9] = {1 1 0 1 1 1 0 1 1}

[0097] That is, starting from SOC = 99 and reversing, the new charging SOC curve is as follows:

[0098] SOC 91 92 93 94 95 96 97 98 99 Voltage 3.415 3.425 - 3.445 3.455 3.465 - 3.500 3.530

[0099] After the redundant data is linearly interpolated before and after, the new charging SOC curve is as follows:

[0100] SOC 91 92 93 94 95 96 97 98 99 Voltage 3.415 3.425 3.435 3.445 3.455 3.465 3.4825 3.500 3.530

[0101] S55. To determine the validity of the newly formed charge-discharge SOC curve, if there is no decreasing data in V[len], it is valid.

[0102] Example 9: Suppose the new charging SOC curve is as follows:

[0103] SOC 91 92 93 94 95 96 97 98 99 Voltage 3.415 3.425 3.420 3.445 3.455 3.465 3.4825 3.500 3.530

[0104] Among them, the data 3.420 of SOC93 is less than the data 3.425 of SOC92, which is invalid.

[0105] Preferably, the set condition in S5 is that for discharge, the corresponding SOC <= 5%, and for charging, the corresponding SOC >= 95%.

[0106] Preferably, the self-learning judgment point Vup of the charging voltage curve in S51 is the voltage point corresponding to SOC = 95 in the charging SOC curve.

[0107] Preferably, the self-learning judgment point Vdown of the discharge voltage curve in S51 is the voltage point corresponding to SOC = 5 in the discharge SOC curve.

[0108] Preferably, in S51, when the number of times the voltage continuously is greater than or equal to the self-learning judgment point Vup of the charging voltage curve accumulates to N times during the charging process, and when the number of times the voltage continuously is less than or equal to the self-learning judgment point Vdown of the discharge voltage curve accumulates to N times during the discharge process, this N is 5.

[0109] Preferably, in S52, when there are M consecutive cases where RF[len] is 0, this M is 5. The interpolation method in S54 is linear interpolation.

[0110] For the above voltage curve self-learning device of the energy storage system battery, the battery detection module is used to collect the real-time voltage, current, and temperature values of the energy storage system; the battery data processing module is used to perform mean processing on the collected real-time voltage, current, and temperature values respectively; the battery data storage module is used to store the mean-processed voltage, current, and temperature values respectively, and is also used to store the constant current parameters; the battery constant current judgment module is used to calculate the constant current parameters; the battery self-learning module is used to extract the data from the battery data storage module and generate a new curve.

Claims

1. A method for self-learning the voltage curve of a battery in an energy storage system, characterized in that, Follow these steps in order: S1, collects the real-time voltage, current and temperature values of the energy storage system; S2, according to the actual capacity of the battery, when the accumulated SOC change by the ampere-hour integration method reaches the setting, the real-time voltage, current and temperature values collected within the change are averaged respectively, where SOC is a value from 0 to 100%; S3, storing the voltage, current and temperature values after the average processing of each accumulated SOC change amount by the ampere-hour integration method in S2 as V, I, T, and each accumulated SOC change amount by the ampere-hour integration method dSOC; V[len] = {V1...V n ...V len}; I[len] = {I1...I n ...I len}; T[len] = {T1...T n ...T len}; dSOC[len] = {dSOC1...dSOC n ...dSOC len}; Among them, len is the length of each data; S4, calculate the steady current parameter RF, calculate RF by the formula using I[len] stored in S3, and store RF, where RF0=0, and Q is the actual capacity of the battery; When And , RF i = 1; RF[len] = {RF1...RF n ...RF len}; err1 is 0.1, err2 is 50%; When the open circuit duration in each ampere-hour integration method accumulated SOC change is greater than the open circuit duration standard time, RF = 0; S5, when charging and discharging are completed and the set conditions are met, the voltage, current and temperature values stored in S3 are extracted through the battery self-learning module to obtain a new curve.

2. The method for self-learning the voltage curve of the energy storage system battery according to claim 1, wherein According to the ampere-hour integration method, the cumulative SOC change reaches the set value and is set to 1%; the standard open circuit duration is 1 hour.

3. The self-learning method for the voltage curve of the energy storage system battery according to claim 1, characterized in that S5 include: S51, when the voltage during the charging process is continuously greater than or equal to the self-learning judgment point Vup of the charging voltage curve, the number of times accumulated reaches N times is recorded as the time of reaching the set condition. From the time of reaching the set condition to the end of charging, the SOC corresponding to the charging SOC curve is found by the voltage during the charging process, and the maximum SOC is recorded, which is recorded as SOCVup; When the voltage during the discharge process is continuously less than or equal to the self-learning judgment point Vdown of the discharge voltage curve, the number of times accumulated reaches N is recorded as the time of reaching the set condition. From the time of reaching the set condition to the end of discharge, the SOC corresponding to the discharge SOC curve is found by the voltage during the discharge process, and the minimum SOC is recorded, which is recorded as SOCVdown. S52, judging the validity of RF[len]. If RF[len] is 0 and there are more than M consecutive RF[len], the data is judged to be invalid. S53, when RF[len] is determined to be valid, under the charging data, with SOCVup as the end point, dSOC[len] is subtracted from len to 1 in sequence to form a new charging SOC curve, that is, V[len], I[len], T[len] corresponding to the SOC; when RF[len] is determined to be valid, under the discharging data, the saved V[len], I[len], T[len], dSOC[len], RF[len] are first flipped back and forth, that is, the data in each array is swapped from the head to the tail to the middle, and with SOCVdown as the starting point, dSOC[len] is added from 1 to len in sequence to form a new discharging SOC curve, that is, V[len], I[len], T[len] corresponding to the SOC; S54. When there is a 0 in RF[len], V[len], I[len], and T[len] are interpolations of V[len], I[len], and T[len] when RF[len] is 1 in the data before or after it. S55. Determine the validity of the newly formed charge-discharge SOC curve. The curve is valid if there is no decreasing data in V[len].

4. The self-learning method for the voltage curve of the battery of the energy storage system according to claim 3, characterized in that In S51, the self-learning judgment point Vup of the charging voltage curve is the voltage point corresponding to 95% SOC in the charging SOC curve; the self-learning judgment point Vdown of the discharging voltage curve in S51 is the voltage point corresponding to 5% SOC in the discharging SOC curve; in S51, when the number of times the voltage continuously is greater than or equal to the self-learning judgment point Vup of the charging voltage curve during charging accumulates to N times, and when the number of times the voltage continuously is less than or equal to the self-learning judgment point Vdown of the discharging voltage curve during discharging accumulates to N times, this N is 5.

5. The method for self-learning the voltage curve of the battery of the energy storage system according to claim 3, wherein In S52, when there are consecutive M or more 0s in RF[len], this M is 5; the interpolation method in S54 is linear interpolation.

6. A voltage curve self-learning device for a battery of an energy storage system, characterized in that, This method implements the method described in any one of claims 1-5, and includes a battery detection module, a battery data processing module, a battery data storage module, a battery constant current judgment module, and a battery self-learning module.

7. The self-learning device for the voltage curve of the energy storage system battery according to claim 6, characterized in that, The battery detection module is used to collect the real-time voltage, current, and temperature values of the energy storage system; the battery data processing module is used to perform mean value processing on the collected real-time voltage, current, and temperature values respectively. The battery data storage module is used to store the voltage, current, and temperature values after mean value processing respectively, and is also used to store the constant current parameters; the battery constant current judgment module is used to calculate the constant current parameters; the battery self-learning module is used to extract the data from the battery data storage module and generate a new curve.

Images