A method for estimating the state of charge of a lithium-ion battery

Abstract

The present invention provides a method for estimating the state of charge of a lithium-ion battery, which is used to estimate the state of charge when the battery is started. Before the battery leaves the factory, the hybrid pulse power characteristics test is carried out under different temperature and different state of charge combinations, and the circuit element parameter values under different test combinations are identified based on the equivalent circuit model, and a data table is formed and stored; during the service process of the battery, the terminal voltage value and temperature value when the battery starts but has not started charging and discharging are measured and recorded, and the current value and terminal voltage value are synchronously sampled when the battery starts charging and discharging, forming a current measurement value sequence and a terminal voltage measurement value sequence; the terminal voltage prediction value sequence with the largest determination coefficient r<supgt;2< / supgt> is determined by comparison, and the state of charge corresponding to the circuit element parameter value of the equivalent circuit model corresponding to it is found, which is the estimated value of the state of charge when the battery starts. The method for estimating the state of charge of the lithium-ion battery of the present invention is scientific and reasonable, has clear logic, and has low hardware cost for implementation.

Description

Technical Field

[0001] The present invention relates to the field of batteries, and particularly to a method for estimating the state of charge of a lithium-ion battery. Background Art

[0002] Currently, lithium-ion batteries are widely used due to their high energy density, power density, green environmental protection and other advantages. During the operation of a lithium-ion battery, it is necessary to accurately obtain its state of charge (SOC). Generally, it is necessary to know the initial state of charge when the battery starts to work, and then online monitor the magnitude of the charge and discharge current during the battery operation, and calculate the change in the amount of electricity by integrating the current over time (ampere-hour integration method) and convert it into the change in the state of charge, so as to obtain the state of charge during the battery operation in real time.

[0003] Due to the limitations of the sampling frequency and sampling accuracy of ampere-hour integration, errors will inevitably accumulate over time and lead to a decrease in the measurement accuracy of the state of charge. Therefore, it is necessary to calibrate the state of charge of the battery at regular intervals, and for convenience, it is best to perform online calibration, so that the battery does not need to be sent back to the factory for processing. The currently widely used online calibration idea is based on the equivalent circuit model of the battery, that is, the relationship between the internal resistance and other component parameters in the equivalent circuit model and the state of charge of the battery is obtained through pre-factory measurement and calculation and constitutes a data table, and then during the battery operation, the online identification of the equivalent circuit model parameters is performed through the sampled values of the current and voltage, and finally the state of charge of the battery is obtained by looking up the table based on the identification results of the model parameters. This method of calibrating the state of charge based on the equivalent circuit model is particularly suitable for lithium iron phosphate batteries, because the open-circuit voltage change curve of lithium iron phosphate batteries with the state of charge has a large flat stage, and even if the state of charge changes significantly during this flat stage, the change in the open-circuit voltage of the battery is very small; however, if an equivalent circuit model is established for lithium iron phosphate batteries, the change law of relevant circuit component parameters such as ohmic internal resistance with the state of charge is very obvious. Therefore, compared with the calibration method based solely on the open-circuit voltage, the online calibration of the state of charge based on the equivalent circuit model has more significant effects.

[0004] The identification of the equivalent circuit model parameters depends on the pulse charge and discharge data of the battery. The most typical method is to conduct a Hybrid Pulse Power Characteristic (HPPC) test on the battery. The higher the sampling frequency during the test, the higher the parameter identification accuracy. In the currently known technical solutions, most rely on the sampled current and voltage data during the battery operation process for online parameter identification of the equivalent circuit model. This method has two deficiencies: First, due to cost factors, it is difficult to deploy high-precision, high-frequency current and voltage sampling circuits and corresponding data communication and storage modules in battery products to meet the needs of high-precision online identification. Second, the charge and discharge conditions during the actual operation process of battery products vary widely and are not controlled by the manufacturer. That is, it is not possible to artificially stipulate in advance that the battery product needs to execute a specific charge and discharge process for the need of equivalent circuit model parameter identification. For example, to accurately obtain the ohmic internal resistance value in the RC equivalent circuit, it is usually necessary to capture the voltage drop at the moment of the battery's constant current pulse discharge. This requires the battery to execute a constant current pulse discharge and perform high-frequency voltage value sampling. For most actual battery products, it is difficult to fully meet this requirement.

[0005] Therefore, it is necessary to develop new technical means to improve the existing state-of-charge estimation method based on the equivalent circuit model, which can estimate the state-of-charge of the battery at low hardware cost and without affecting the normal operation of the battery to provide its calibration basis. Summary of the Invention

[0006] To solve the above technical problems, the present invention provides a scientific, reasonable, logically clear, low-execution hardware cost and non-battery-normal-operation-affecting method for estimating the state-of-charge of a lithium-ion battery. The technical solution is a method for estimating the state-of-charge of a lithium-ion battery, which is used to estimate the state-of-charge when the battery is started.

[0007] Before the battery leaves the factory, a hybrid pulse power characteristic test is carried out in the laboratory under different temperature and different state-of-charge combinations, and the circuit element parameter values under different test combinations are identified based on the equivalent circuit model, forming a data table and storing it: The equivalent circuit model is composed of a constant voltage source, an ohmic resistor, and an RC parallel circuit connected in series in sequence. The RC parallel circuit is composed of a polarization resistor and a capacitor connected in parallel with each other. The circuit element parameter values include the constant voltage source voltage Voc, the ohmic internal resistance Ro, the polarization internal resistance R p and the capacitor C; the data table is composed of a series of sub-tables, and each sub-table corresponds to a temperature value; each row of data in each sub-table corresponds to a certain state-of-charge and the circuit element parameter values under that state-of-charge;

[0008] During the battery's service process, measure and record the terminal voltage value U at the moment when the battery is started but before charging and discharging 0and the temperature value T 0 ; and starting from the moment when the battery is started and charging and discharging begin, at every fixed time interval Δt, the charging and discharging current value and the terminal voltage value of the battery are synchronously sampled n times to form a current measurement value sequence [I 1 , I 2 , …, I i , …, I n and a terminal voltage measurement value sequence [U 1 , U 2 , …, Ui, …, U n , where the time interval Δt is between 0.1 s and 2 s, and n is between 10 and 100; in the current measurement value sequence, when the battery is in the discharging state, the current value is positive, and when the battery is in the charging state, the current value is negative;

[0009] Estimate the state of charge of the battery according to the following steps:

[0010] Step S1, find the sub-table in the data table whose temperature value is closest to the temperature value T 0 before the battery starts to work;

[0011] Step S2, in the sub-table obtained in Step S1, find the data rows where the constant voltage source voltage Voc is between (1 - m)U 0 and (1 + m)U 0 , where the parameter m is between 0.005 and 0.01, and U 0 is the terminal voltage value when the battery starts but has not started charging and discharging;

[0012] Step S3, for each data row obtained in Step S2, respectively construct an equivalent circuit model corresponding to it based on the circuit element parameter value information it contains;

[0013] Step S4, for each equivalent circuit model constructed in Step S3, respectively input the current measurement value sequence and calculate to obtain its corresponding terminal voltage prediction value sequence [V 1 , V 2 , …, V i , …, V n :

[0014] For any equivalent circuit model constructed in Step S3, let the current value passing through its ohmic resistance at each moment be the current measurement value corresponding to that moment in the current measurement value sequence, and thus calculate the terminal voltage value at each moment and form the terminal voltage prediction value sequence [V 1 , V 2 , …, V i , …, V n under this equivalent circuit model;

[0015] Step S5. For each sequence of predicted terminal voltage values [V 1 , V 2 , …, V i , …, V n obtained in Step S4, compare and calculate it with the sequence of measured terminal voltage values [U 1 , U 2 , …, U i , …, U n respectively, and obtain the corresponding determination coefficient r 2 :

[0016]

[0017] where is the arithmetic mean of all elements in the sequence of measured terminal voltage values [U 1 , U 2 , …, U i , …, U n ;

[0018] Step S6. Find the sequence of predicted terminal voltage values with the largest determination coefficient r 2 in Step S5, thereby finding the corresponding equivalent circuit model, and looking up the state of charge corresponding to the circuit element parameter values of this circuit model from the sub-table in Step 2, which is the estimated value of the state of charge at battery startup.

[0019] In the above method for estimating the state of charge of a lithium-ion battery, the positive active material of the lithium-ion battery is lithium iron phosphate.

[0020] In the above method for estimating the state of charge of a lithium-ion battery, the hybrid pulse power characteristic test under different temperature and state of charge combinations in the laboratory is carried out with the temperature values taken at intervals of ΔT within the range of -20°C to 50°C, and the state of charge values taken at intervals of ΔSOC within the range of 0% to 100%, where the temperature value interval ΔT is between 1°C and 5°C, and the state of charge value interval ΔSOC is between 1% and 5%.

[0021] In the above method for estimating the state of charge of a lithium-ion battery, the hybrid pulse power characteristic test under a certain temperature and state of charge combination is carried out according to the following steps:

[0022] Step A1. Place the battery in a temperature environment of 20 ± 2°C, let it stand for 1 h, then charge it to 100% state of charge according to the standard charging system specified by the manufacturer and let it stand for 1 h again;

[0023] Step A2. Discharge the battery at a constant current of 1C rate to the state of charge to be measured;

[0024] Step A3: Transfer the battery to the temperature environment to be measured and leave it standing for 1 h;

[0025] Step A4: Discharge the battery with a constant current pulse at a C rate for 10 s;

[0026] Step A5: Leave the battery standing for 40 s;

[0027] Step A6: Charge the battery with a constant current pulse at a C rate for 10 s;

[0028] The value of the parameter a is between 1 and 10, and the battery terminal voltage is continuously sampled at a frequency not lower than 10 Hz during the process of steps A4 to A6.

[0029] The above method for estimating the state of charge of the lithium-ion battery, for the test results of the hybrid pulse power characteristics under a certain temperature and state of charge combination, the identification of the circuit element parameter values under this combination based on the equivalent circuit model is divided into the following steps:

[0030] Step B1: Obtain the arithmetic mean of all battery terminal voltage sampling values within 1 s before the start of the constant current pulse discharge, which is the constant voltage source voltage Voc;

[0031] Step B2: Divide the difference between the constant voltage source voltage Voc and the first battery terminal voltage sampling value at the moment of the start of the constant current pulse discharge by the current value during the constant current pulse discharge process, which is the ohmic internal resistance Ro;

[0032] Step B3: Use the battery terminal voltage sampling time series during the standing stage between the constant current pulse discharge and the constant current pulse charge, and adopt the least squares method to fit the parameter β based on the following function:

[0033] U(t) = Voc - (Voc - U r0 )e -βt (2)

[0034] In the formula, t is the time from the start moment of the standing, U(t) is the terminal voltage sampling value at time t, Voc is the constant voltage source voltage obtained in step B1, U r0 is the first battery terminal voltage sampling value during the standing stage, and e is the natural constant;

[0035] Step B4: Calculate the polarization internal resistance R p :

[0036]

[0037] In the formula, Voc is the constant voltage source voltage obtained in step B1, U r0 is the first battery terminal voltage sampling value during the standing stage, e is the natural constant, β is the parameter value obtained in step B3, and t d is the total duration of the constant current pulse discharge, and Id is the magnitude of the current value for constant current pulse discharge;

[0038] Step B5. Calculate the capacitance C according to the following formula:

[0039] C = 1 / (βR p ) (4)

[0040] where β is the parameter value obtained in step B3, and R p is the polarization internal resistance obtained in step B4.

[0041] In the above method for estimating the state of charge of a lithium-ion battery, in step S4, for a constructed equivalent circuit model, the input current measurement value sequence [I 1 , I 2 , …, I i , …, I n and the specific method for calculating the predicted terminal voltage value sequence [V 1 , V 2 , …, Vi, …, V n are as follows. The values of each element in the predicted terminal voltage value sequence are calculated sequentially by performing the following sub-steps:

[0042] Sub-step S4.1. Let k = 0, W 0 = 0, I 0 = 0;

[0043] Sub-step S4.2. Calculate the value of W k+1 :

[0044]

[0045] Sub-step S4.3. Calculate the value of V k+1 :

[0046] V k+1 = Voc - W k+1 - RoI k+1 (6)

[0047] Sub-step S4.4. Increase the value of k by 1 on the original basis;

[0048] Sub-step S4.5. If k + 1 > n, end the calculation; otherwise, return to sub-step S4.2.

[0049] In the above technical solutions, for the detailed information on the composition of the equivalent circuit model, parameter identification, and calculation of the predicted terminal voltage value sequence from the current measurement value sequence based on the equivalent circuit model, reference can be made to the article "Modelling of Li-ion batteries using equivalent circuit diagrams" by Ahmad Rahmoun.

[0050] This technical solution estimates the state of charge based on the correspondence between the circuit element parameters and the state of charge of the battery in the equivalent circuit model. Compared with the prior art, the improvement effect is that this technical solution does not require online identification of the parameters of the battery's equivalent circuit model. Therefore, in the operation of the battery product, there is no need for high-precision and high-frequency current and voltage sampling equipment, no need to perform a specific charge and discharge operation on the battery deliberately, and no need to perform relatively complex least squares parameter fitting online. This technical solution does the opposite. Before the battery leaves the factory, the battery manufacturer conducts a large number of tests and equivalent circuit model parameter identifications, and stores the relevant results. In this way, in the operation of the battery product, only need to run each equivalent circuit model with parameter identification results within the possible range one by one, calculate the voltage prediction value sequence under the corresponding equivalent circuit model from the current measurement value sequence and compare it with the real voltage measurement value sequence, so as to find the equivalent circuit model that most conforms to the actual situation and finally estimate the state of charge of the battery. Therefore, this estimation method is not only scientific and reasonable, with clear logic, but also for the operation of the battery product, the amount of calculation for performing relevant operations is small, and there is no need for high-cost sampling and calculation hardware costs. Description of the Drawings

[0051] Figure 1 is the schematic diagram of the equivalent circuit model in the embodiment of the present invention. Figure 1 The two terminal blocks at the rightmost end are used to connect the load, and the voltage difference between them is the battery terminal voltage.

[0052] Figure 2 is the flowchart of sub-steps S4.1 to S4.5 in the embodiment of the present invention.

[0053] Figure 3 is the voltage and current curve diagram of the partial process of the hybrid power characteristic test in the embodiment of the present invention. Detailed Embodiment

[0054] The present invention will be further described below with reference to the drawings and embodiments.

[0055] An estimation method for the state of charge of a lithium-ion battery is used to estimate the state of charge when the battery is started.

[0056] Before the battery leaves the factory, a hybrid pulse power characteristic test is carried out in the laboratory under different temperature and state of charge combinations, and the circuit element parameter values under different test combinations are obtained based on equivalent circuit model identification, forming a data table and storing it: The equivalent circuit model is shown in the appendix Figure 1 , which is composed of a constant voltage source, an ohmic resistor and an RC parallel circuit connected in series in sequence. The RC parallel circuit is composed of a polarization resistor and a capacitor connected in parallel with each other. The circuit element parameter values include the constant voltage source voltage Voc, the ohmic internal resistance Ro, the polarization internal resistance Rp and a capacitor C; the data table consists of a series of sub - tables, and each sub - table corresponds to a temperature value; each row of data in each sub - table corresponds to a certain state of charge and the parameter values of circuit elements in that state of charge;

[0057] During the service process of the battery, measure and record the terminal voltage value U at the moment when the battery starts but has not yet started charging or discharging. 0 and the temperature value T 0 ; and starting from the moment when the battery starts and begins to charge or discharge, synchronously sample the charging and discharging current value and the terminal voltage value of the battery n times at fixed time intervals Δt to form a current measurement value sequence [I 1 , I 2 , …, I i , …, I n and a terminal voltage measurement value sequence [U 1 , U 2 , …, Ui, …, U n , where the time interval Δt is between 0.1 s and 2 s, and n is between 10 and 100; in the current measurement value sequence, the current value is positive when the battery is in the discharge state and negative when the battery is in the charge state;

[0058] Estimate the state of charge of the battery according to the following steps:

[0059] Step S1: Find the sub - table in the data table whose temperature value is closest to the temperature value T before the battery starts to work. 0 ;

[0060] Step S2: In the sub - table obtained in step S1, find the data rows where the constant - voltage source voltage Voc is between (1 - m)U 0 and (1 + m)U 0 , where the parameter m is between 0.005 and 0.01, and U 0 is the terminal voltage value at the moment when the battery starts but has not yet started charging or discharging;

[0061] Step S3: For each data row obtained in step S2, construct an equivalent circuit model corresponding to it based on the circuit element parameter value information it contains;

[0062] Step S4: For each equivalent circuit model constructed in step S3, input the current measurement value sequence and calculate to obtain its corresponding terminal voltage prediction value sequence [V 1 , V 2 , …, V i , …, V n :

[0063] For any equivalent circuit model constructed in step S3, let the current value passing through its ohmic resistance at each moment be the current measurement value corresponding to that moment in the current measurement value sequence, and thus calculate the terminal voltage value at each moment and form a terminal voltage prediction value sequence [V 1 ,V 2 ,…,V i ,…,V n ;

[0064] Step S5. For each terminal voltage prediction value sequence [V 1 ,V 2 ,…,V i ,…,V n obtained in step S4, respectively compare and calculate it with the terminal voltage measurement value sequence [U 1 ,U 2 ,…,U i ,…,U n and obtain its corresponding determination coefficient r 2 :

[0065]

[0066] In the formula is the arithmetic mean of all elements in the terminal voltage measurement value sequence [U 1 ,U 2 ,…,U i ,…,U n ;

[0067] Step S6. Find the terminal voltage prediction value sequence with the largest determination coefficient r 2 in step S5, thereby finding the corresponding equivalent circuit model, and checking the state of charge corresponding to the circuit element parameter value of this circuit model from the sub-table in step 2, which is the estimated value of the state of charge at battery startup.

[0068] In the above method for estimating the state of charge of a lithium-ion battery, the positive active material of the lithium-ion battery is lithium iron phosphate.

[0069] In the above method for estimating the state of charge of a lithium-ion battery, the hybrid pulse power characteristic test under different temperature and state of charge combinations in the laboratory, the temperature values for different temperatures are temperature values in the range of -20°C to 50°C at intervals of ΔT, and the state of charge values for different states of charge are state of charge values in the range of 0% to 100% at intervals of ΔSOC, where the temperature value interval ΔT is between 1°C and 5°C, and the state of charge value interval ΔSOC is between 1% and 5%.

[0070] In the above method for estimating the state of charge of a lithium-ion battery, the hybrid pulse power characteristic test under a certain temperature and state of charge combination is carried out according to the following steps:

[0071] Step A1: Place the battery in a temperature environment of 20±2°C, let it stand for 1 h, then charge it to 100% state of charge according to the standard charging system specified by the manufacturer, and then let it stand for another 1 h;

[0072] Step A2: Discharge the battery at a constant current of 1C to the state of charge to be measured;

[0073] Step A3: Transfer the battery to the temperature environment to be measured and let it stand for 1 h;

[0074] Step A4: Discharge the battery with a constant current pulse of aC for 10 s;

[0075] Step A5: Let the battery stand for 40 s;

[0076] Step A6: Charge the battery with a constant current pulse of aC for 10 s;

[0077] The value of the parameter a above ranges from 1 to 10, and during the process of steps A4 to A6, the battery terminal voltage is continuously sampled at a frequency not lower than 10 Hz.

[0078] For the above method for estimating the state of charge of a lithium-ion battery, for the test results of the hybrid pulse power characteristics under a certain temperature and state of charge combination, the identification of the circuit element parameter values under this combination based on the equivalent circuit model is divided into the following steps:

[0079] Step B1: Obtain the arithmetic mean of all battery terminal voltage sampling values within 1 s before the start of the constant current pulse discharge, which is the constant voltage source voltage Voc;

[0080] Step B2: Divide the difference between the constant voltage source voltage Voc and the first battery terminal voltage sampling value at the moment of the start of the constant current pulse discharge by the current value during the constant current pulse discharge process, which is the ohmic internal resistance Ro;

[0081] Step B3: Use the battery terminal voltage sampling time series during the standstill stage between the constant current pulse discharge and the constant current pulse charge, and adopt the least squares method to fit the parameter β based on the following function:

[0082] U(t) = Voc - (Voc - U r0 )e -βt (2)

[0083] where t is the time from the start moment of the standstill, U(t) is the terminal voltage sampling value at time t, Voc is the constant voltage source voltage obtained in step B1, U r0 is the first battery terminal voltage sampling value during the standstill stage, and e is the natural constant;

[0084] Step B4: Calculate the polarization internal resistance R p :

[0085]

[0086] Wherein, Voc is the constant voltage source voltage obtained in step B1, and U r0 is the first battery terminal voltage sampling value during the rest period, e is the natural constant, β is the parameter value obtained in step B3, and t d is the total duration of the constant current pulse discharge, and I d is the magnitude of the current value of the constant current pulse discharge;

[0087] Step B5. Calculate the capacitance C according to the following formula:

[0088] C = 1 / (βR p ) (4)

[0089] Wherein, β is the parameter value obtained in step B3, and R p is the polarization internal resistance obtained in step B4.

[0090] For the above method for estimating the state of charge of a lithium-ion battery, in step S4, for a constructed equivalent circuit model, input the sequence of current measurement values [I 1 , I 2 , …, I i , …, I n and calculate to obtain the sequence of predicted terminal voltage values [V 1 , V 2 , …, Vi, …, V n The specific method is as shown in the appendix Figure 2 . The values of each element in the sequence of predicted terminal voltage values are calculated sequentially by performing the following sub-steps:

[0091] Sub-step S4.1. Let k = 0, W 0 = 0, I 0 = 0;

[0092] Sub-step S4.2. Calculate the value of W k+1 :

[0093]

[0094] Sub-step S4.3. Calculate the value of V k+1 :

[0095] V k+1 = Voc - W k+1 - RoI k+1 (6)

[0096] Sub-step S4.4. Let the value of k be increased by 1 on the original basis;

[0097] Sub-step S4.5: If k + 1 > n, end the calculation; otherwise, return to sub-step S4.2.

[0098] Embodiment

[0099] Please refer to Appendix Figure 1 to Appendix Figure 3 to understand this embodiment.

[0100] A certain lithium iron phosphate battery has a rated capacity of 70 Ah. Before leaving the factory, the battery is tested for its hybrid pulse power characteristics under different temperature and state of charge combinations in the laboratory. The temperature values are taken at intervals of ΔT = 5 °C within the range of -20 °C to 50 °C, and the state of charge values are taken at intervals of ΔSOC = 5% within the range of 0% to 100%.

[0101] Taking the hybrid pulse power characteristic test of this battery at 25 °C and SOC = 90% as an example, the process of identifying the parameters of the equivalent circuit model is introduced. The principle of the equivalent circuit composition can be seen in Figure 1 , Figure 1 The two circles at the rightmost end in [reference] represent the terminals connected to the load, and the voltage difference between them is the battery terminal voltage. The voltage and current sampling frequency during the above-mentioned hybrid pulse power characteristic test is 100 Hz. The voltage and current variation during the local test process can be seen in Figure 3 , Figure 3 The stage where the current value is 70 A represents the constant current discharge stage, the stage where the current value is -70 A represents the constant current charge stage, and the stage where the current value is 0 represents the standby stage.

[0102] The identification process in the above example is as follows:

[0103] Step B1: Obtain the arithmetic mean of all battery terminal voltage sampling values within 1 s before the start of the constant current pulse discharge, which is the constant voltage source voltage Voc = 3.3295 V;

[0104] Step B2: Subtract the first battery terminal voltage sampling value 3.2410 V during the constant current pulse discharge stage from the constant voltage source voltage Voc = 3.3294 V, and divide the difference 0.0884 V by the current value 70 A during the constant current pulse discharge process, which is the ohmic internal resistance Ro = 1.263 mΩ;

[0105] Step B3: Using the battery terminal voltage sampling time series during the standby stage between the constant current pulse discharge and the constant current pulse charge, the least squares method is used to fit the parameter β based on the following function:

[0106] U(t) = Voc - (Voc - U r0 )e -βt (2)

[0107] where t is the time from the start of the rest period, U(t) is the sampled terminal voltage at time t, Voc is the constant voltage source voltage Voc = 3.3295 V obtained in step B1, and U r0 = 3.3075 V is the first sampled battery terminal voltage during the rest period, and e is the natural constant 2.7183. The fitting gives β = 0.0458 s -1 .

[0108] Step B4. Calculate the polarization internal resistance R according to the following formula p :

[0109]

[0110] where Voc is the constant voltage source voltage Voc = 3.3295 V obtained in step B1, U r0 = 3.3075 V is the first sampled battery terminal voltage during the rest period, e = 2.7183 is the natural constant, and β = 0.0458 s -1 is the parameter value obtained in step B3, t d = 10 s is the total duration of the constant current pulse discharge, and I d = 70 A is the magnitude of the constant current pulse discharge current. The calculated polarization internal resistance R p = 0.855 mΩ.

[0111] Step B5. Calculate the capacitance C according to the following formula:

[0112] C = 1 / (βR p ) (4)

[0113] where β = 0.0458 s -1 is the parameter value obtained in step B3, and R p = 0.855 mΩ is the polarization internal resistance obtained in step B4. The calculated capacitance C = 25.528 kF.

[0114] During a certain start of the battery for charge and discharge operations, the terminal voltage value U 0 = 3.2620 V and the temperature value T 0 = 21 °C at the moment when the battery starts but before charge and discharge begins are measured and recorded; and starting from the moment when the battery starts and charge and discharge begin, the charge and discharge current values and terminal voltage values of the battery are synchronously sampled 20 times at fixed time intervals Δt = 0.5 s, forming a current measurement value sequence [I 1 , I 2 , …, I i , …, I 20= [65A, 73A, 76A, 60A, 54A, 32A, 12A, -25A, -36A, -40A, -12A, 2A, 15A, 28A, 34A, 45A, 58A, 63A, 71A, 75A], and the sequence of terminal voltage measurement values [U 1 , U 2 , …, Ui, …, U 20 = [3.1046V, 3.0854V, 3.0657V, 3.1314V, 3.1500V, 3.2053V, 3.2481V, 3.3416V, 3.3671V, 3.3568V, 3.3129V, 3.2673V, 3.2182V, 3.2056V, 3.1979V, 3.1677V, 3.1334V, 3.1267V, 3.0715V, 3.0676V].

[0115] Estimate the state of charge of the battery according to the following steps:

[0116] Step S1: Find the sub-table in the data table whose temperature value is closest to the temperature value T 0 = 21°C before the battery starts to work, which is the sub-table with a temperature value of 20°C.

[0117] Step S2: Take m = 0.007, and find the data rows where the constant voltage source voltage Voc is between (1 - m)U 0 = 3.239V and (1 + m)U 0 = 3.285V. There are a total of three data rows involved. The state of charge and the circuit element parameter values under the corresponding state of charge are shown in Table 1.

[0118] Table 1 State of Charge and Circuit Element Parameter Table

[0119]

[0120]

[0121] Step S3: For the three data rows obtained in Step S2, respectively construct the corresponding 1#, 2#, and 3# equivalent circuit models based on the circuit element parameter value information they contain. The circuit element parameters included in these three equivalent circuit models are shown in the 2nd, 3rd, and 4th rows of Table 1 respectively.

[0122] Step S4: For each equivalent circuit model constructed in Step S3, respectively input the sequence of current measurement values and calculate the corresponding sequence of terminal voltage prediction values [V 1 , V 2 , …, V i , …, V n :

[0123] For any equivalent circuit model constructed in step S3, let the current value passing through its ohmic resistance at each moment be the current measurement value corresponding to that moment in the current measurement value sequence, and thus calculate the terminal voltage value at each moment and form a terminal voltage prediction value sequence [V 1 , V 2 , …, V i , …, V n .

[0124] For each equivalent circuit model in 1#, 2#, and 3#, calculate the terminal voltage prediction values at different k values as shown in Table 2.

[0125] Table 2 Terminal voltage prediction values of different equivalent circuit models / V

[0126] k Model 1 Model 2 Model 3 1 3.0393 3.0543 3.1144 2 3.0126 3.0277 3.0929 3 3.0019 3.0170 3.0840 4 3.0521 3.0673 3.1239 5 3.0705 3.0857 3.1383 6 3.1403 3.1557 3.1940 7 3.2041 3.2195 3.2448 8 3.3228 3.3382 3.3395 9 3.3586 3.3740 3.3682 10 3.3721 3.3874 3.3792 11 3.2828 3.2981 3.3081 12 3.2381 3.2534 3.2725 13 3.1964 3.2117 3.2392 14 3.1545 3.1697 3.2057 15 3.1348 3.1501 3.1899 16 3.0990 3.1144 3.1613 17 3.0567 3.0720 3.1273 18 3.0398 3.0552 3.1137 19 3.0132 3.0287 3.0923 20 2.9994 3.0149 3.0810

[0127] Taking the 1# model as an example, first let k = 0, W 0 = 0, I 0 = 0, and then calculate:

[0128]

[0129] Therefore, V k+1 = V 1 = Voc - W 1 - RoI 1 = 3.248 - 0 - 3.211E - 3 * 65 = 3.0393V.

[0130] Then let the value of k increase by 1 on the original basis, that is, k = 1. At this time, k + 1 = 2 < 20, and then calculate:

[0131]

[0132] Therefore, V k+1 = V 2 = Voc - W 2 - RoI 2 = 3.248 - 1.001E - 3 - 3.211E - 3 * 73 = 3.0126V.

[0133] Then let the value of k increase by 1 on the original basis, that is, k = 2. At this time, k + 1 = 3 < 20, and then return to calculate W 3 and V 3 . Loop like this until the values of W 20 and V 20 are calculated when k = 19. After that, let the value of k increase by 1 on the original basis, that is, k = 20. At this time, k + 1 = 21 > 20, so the calculation ends.

[0134] Step S5. For each sequence of predicted terminal voltage values [V 1 , V 2 , …, V i , …, V n obtained in Step S4, compare and calculate it with the sequence of measured terminal voltage values [U 1 , U 2 , …, U i , …, U n respectively, and obtain the corresponding coefficient of determination r 2 :

[0135]

[0136] where is the arithmetic mean of all elements in the sequence of measured terminal voltage values [U 1 , U 2 , …, U i , …, U n .

[0137] 1# equivalent circuit model is compared with the sequence of measured terminal voltage values, and the corresponding coefficient of determination r 2 is 0.6286;

[0138] 2# equivalent circuit model is compared with the sequence of measured terminal voltage values, and the corresponding coefficient of determination r 2 is 0.3966;

[0139] 3# equivalent circuit model is compared with the sequence of measured terminal voltage values, and the corresponding coefficient of determination r 2 is 0.9334.

[0140] Step S6. Find out the sequence of predicted terminal voltage values with the largest coefficient of determination r 2 in Step S5. Thus, the corresponding equivalent circuit model is found to be the 3# equivalent circuit model, and the state of charge corresponding to the circuit element parameter values of this circuit model is found to be 30% from the sub-table in Step 2, which is the estimated value of the state of charge at battery startup.

[0141] The accuracy of the estimation method of the present invention depends to a large extent on the selection of the intervals of temperature and state of charge in the hybrid pulse power characteristic test before the battery leaves the factory. The smaller the interval, the higher the prediction accuracy.

[0142] Furthermore, on the basis of Step S6, while estimating the state of charge value at the battery startup state, synchronously output the value of the coefficient of determination r 2 corresponding to the sequence of predicted terminal voltage values with the largest coefficient of determination r 2 , and compare this value with the threshold r c 2 : If r2 >r c 2 , it is considered that the estimated value of the state of charge is reliable; otherwise, it is considered that the estimated value of the state of charge this time is unreliable, and wait for the estimation of the state of charge value at the next battery startup. If the estimated values of the state of charge are all unreliable in consecutive h estimations, it indicates that the health state of the battery has deteriorated significantly compared with that at the factory, and the battery needs to be returned to the factory for re - testing of the hybrid pulse power characteristics and identification of the equivalent circuit element parameters. The above r c 2 value ranges from 0.7 to 0.9, and the h value ranges from 5 to 50.

[0143] In the embodiment of the present invention, the state of charge is estimated based on the corresponding relationship between the circuit element parameters in the equivalent circuit model and the state of charge of the battery. Compared with the prior art, the improvement effect is that in this embodiment, it is not necessary to perform online identification of the equivalent circuit model parameters of the battery. Therefore, in the operation link of the battery product, there is no need for high - precision and high - frequency current and voltage sampling equipment, no need to specifically perform a certain specific charge - discharge operation on the battery, and no need to perform relatively complex least - squares method parameter fitting online. The technical solution of this embodiment does the opposite. The battery manufacturer conducts a large number of tests and identification of the equivalent circuit model parameters before the battery leaves the factory and stores the relevant results. In this way, in the operation link of the battery product, only need to run each equivalent circuit model within the possible range with the parameter identification results one by one, calculate the voltage prediction value sequence under the corresponding equivalent circuit model from the current measurement value sequence and compare it with the real voltage measurement value sequence, so as to find the equivalent circuit model that best conforms to the actual situation and finally estimate the state of charge of the battery. Therefore, this estimation method is not only scientific and reasonable, with clear logic; but also for the operation link of the battery product, the amount of calculation for performing relevant operations is small, and there is no need for high - cost sampling and calculation hardware costs.

Claims

1. A method for estimating the state of charge of a lithium-ion battery, which is used to estimate the state of charge when the battery is started, Characterized in that, Before the battery leaves the factory, the hybrid pulse power characteristics are tested in the laboratory under different temperature and state-of-charge combinations, and the circuit element parameter values under different test combinations are obtained based on the equivalent circuit model identification, forming a data table and storing it: The equivalent circuit model consists of a constant voltage source, an ohmic resistance, and an RC parallel circuit connected in series in sequence. The RC parallel circuit consists of a polarization resistance and a capacitor connected in parallel with each other. The circuit element parameter values include the constant voltage source voltage Voc, the ohmic internal resistance Ro, the polarization internal resistance R p and the capacitor C; The data table consists of a series of sub-tables, and each sub-table corresponds to a temperature value; Each row of data in each sub-table corresponds to a certain state-of-charge and the circuit element parameter values under that state-of-charge; During the battery's service life, measure and record the terminal voltage value U at the moment when the battery starts but has not yet started charging or discharging. 0 And the temperature value T 0 ; And starting from the moment when the battery starts and begins to charge or discharge, at every fixed time interval Δt, synchronously sample the charge and discharge current value and the terminal voltage value of the battery n times to form a current measurement value sequence [I 1 , I 2 , …, I i , …, I n and a terminal voltage measurement value sequence [U 1 , U 2 , …, Ui, …, U n , where the time interval Δt is between 0.1 s and 2 s, and n is between 10 and 100; in the current measurement value sequence, when the battery is in the discharge state, the current value is positive, and when the battery is in the charge state, the current value is negative. The state of charge of the battery is estimated according to the following steps: Step S1: Find the sub-table in the data table whose temperature value is closest to the temperature value T before the battery starts working. 0 that is closest; Step S2: From the sub-table obtained in Step S1, find the data rows where the constant voltage source voltage Voc is between (1 - m)U 0 and (1 + m)U 0 , where the parameter m is between 0.005 and 0.01, and U 0 is the terminal voltage value at the moment when the battery starts but has not yet started charging or discharging; Step S3: For each data row obtained in Step S2, an equivalent circuit model corresponding to it is constructed respectively based on the circuit element parameter value information it contains; Step S4: For each equivalent circuit model constructed in step S3, input the current measurement value sequence respectively and calculate to obtain the corresponding terminal voltage prediction value sequence [V 1 , V 2 , …, V i , …, V n : For any equivalent circuit model constructed in step S3, let the current value passing through its ohmic resistance at each moment be the current measurement value corresponding to that moment in the current measurement value sequence, and thus calculate the terminal voltage value at each moment and form a terminal voltage prediction value sequence [V 1 , V 2 , …, V i , …, V n ; Step S5. For each sequence of predicted terminal voltage values [V 1 , V 2 , …, V i , …, V n obtained in Step S4, compare and calculate it with the sequence of measured terminal voltage values [U 1 , U 2 , …, U i , …, U n respectively, and obtain its corresponding coefficient of determination r 2 : where is the arithmetic mean of all elements in the sequence of terminal voltage measurement values [U 1 , U 2 , …, U i , …, U n ; Step S6: Find out the coefficient r determined in step S5 2 The maximum terminal voltage prediction value sequence, thereby finding the corresponding equivalent circuit model, and looking up the state of charge corresponding to the circuit element parameter values of this circuit model from the sub-table in step 2, which is the estimated value of the state of charge at battery startup.

2. The method for estimating the state of charge of a lithium-ion battery according to claim 1, Characterized in that, The positive active material of the lithium-ion battery is lithium iron phosphate.

3. The method for estimating the state of charge of a lithium-ion battery according to claim 1, Characterized in that, For the hybrid pulse power characteristic test carried out in the laboratory under different temperature and state of charge combinations, the temperature values for different temperatures are temperature values in the range of -20°C to 50°C at intervals of ΔT, and the state of charge values for different states of charge are state of charge values in the range of 0% to 100% at intervals of ΔSOC, where the temperature value interval ΔT is between 1°C and 5°C, and the state of charge value interval ΔSOC is between 1% and 5%.

4. The method for estimating the state of charge of a lithium-ion battery according to claim 1, Characterized in that, The hybrid pulse power characteristic test under a certain temperature and state of charge combination is carried out according to the following steps: Step A1: Place the battery in a temperature environment of 20±2°C, let it stand for 1 h, then charge it to 100% state of charge according to the standard charging system specified by the manufacturer and then let it stand for 1 h; Step A2: Discharge the battery at a constant current of 1C rate to the state of charge to be measured; Step A3: Transfer the battery to the temperature environment to be measured and let it stand for 1 h; Step A4: Discharge the battery at a constant current pulse of aC rate for 10 s; Step A5: Let the battery stand for 40 s; Step A6: Charge the battery at a constant current pulse of aC rate for 10 s; The value of the above parameter a is between 1 and 10, and during the process of Steps A4 to A6, the battery terminal voltage is continuously sampled at a frequency not lower than 10 Hz.

5. The method for estimating the state of charge of a lithium-ion battery according to claim 1, Characterized in that, For the test result of the hybrid pulse power characteristic under a certain temperature and state of charge combination, the identification of the circuit element parameter values under this combination based on the equivalent circuit model is divided into the following steps: Step B1: Obtain the arithmetic mean of all battery terminal voltage sampling values within 1 s before the start of the constant current pulse discharge, which is the constant voltage source voltage Voc; Step B2: Divide the difference between the constant voltage source voltage Voc and the first battery terminal voltage sampling value at the moment of the start of the constant current pulse discharge by the current value during the constant current pulse discharge process, which is the ohmic internal resistance Ro; Step B3: Use the battery terminal voltage sampling time series during the idle stage between the constant current pulse discharge and the constant current pulse charge, and adopt the least squares method to fit the parameter β based on the following function: U(t) = Voc - (Voc - U r0 )e -βt (2) where t is the time from the start of the rest period, U(t) is the sampled terminal voltage at time t, Voc is the constant voltage source voltage obtained in step B1, U r0 is the first sampled battery terminal voltage during the rest period, and e is the natural constant; Step B4. Calculate the polarization internal resistance R according to the following formula p :[[]]END]] Where Voc is the constant voltage source voltage obtained in step B1, and U r0 is the first battery terminal voltage sampling value during the resting stage, e is the natural constant, β is the parameter value obtained in step B3, and t d is the total duration of the constant current pulse discharge, and I d is the magnitude of the current value of the constant current pulse discharge; Step B5: Calculate the capacitance C calculated according to the following formula: C = 1 / (βR p )(4) where β is the parameter value obtained in step B3, and R p is the polarization internal resistance obtained in step B4.

6. The method for estimating the state of charge of a lithium-ion battery according to claim 1, Characterized in that, In step S4, for a certain constructed equivalent circuit model, input the sequence of measured current values [I 1 , I 2 , …, I i , …, I n and calculate to obtain the sequence of predicted terminal voltage values [V 1 , V 2 , …, Vi, …, V n . The specific method is to calculate the values of the elements in the sequence of predicted terminal voltage values by successively executing the following sub-steps: Sub-step S4.1: Let k = 0, W 0 = 0, I 0 = 0; Sub-step S4.2, calculate the value of W k+1 : Sub-step S4.3, calculate the value of V k+1 : V k+1 = Voc - W k+1 - RoI k+1 (6) Sub-step S4.4: Let the value of k increase by 1 on the original basis; Sub-step S4.5: If k + 1 > n, end the calculation, otherwise return to Sub-step S4.2.

Images