A SOC estimation method for ternary-lithium iron hybrid battery pack

Abstract

The present invention provides a SOC estimation method for a ternary-lithium iron hybrid battery pack. The method first adopts an extended Kalman filter algorithm to obtain an estimated SOC value of a ternary lithium battery. On this basis, based on the characteristic that the amount of charge change of each cell in the same time period of a series battery pack is the same, an estimated SOC value of a lithium iron phosphate battery is calculated. Since the SOC estimation value of a lithium iron phosphate battery depends only on the amount of charge change, the algorithm complexity is low and hardly increases the computational burden of the BMS. During its use, due to the particularity of the hybrid battery pack configuration, under the condition of full charge of the lithium iron phosphate battery cell, an LFP self-correction mechanism is adopted. At the same time, a variable correction coefficient K is introduced for possible self-discharge of the lithium iron phosphate battery. i , between two full charge conditions, continuous feedback correction can be performed to avoid the situation where the error gradually increases. Compared with the existing technology, the estimation method of the present invention has the advantages of simple calculation, high accuracy and strong adaptability.

Description

Technical Field

[0001] The present invention belongs to the technical field of battery management systems, and in particular relates to a SOC estimation method for a ternary-lithium iron hybrid battery pack. Background Art

[0002] In the past, due to automakers' high demands for electric vehicle power performance, the high-energy-density ternary lithium-ion batteries were the primary choice. However, recent improvements in battery assembly technology have narrowed the energy density gap between lithium iron phosphate (LFP) batteries and ternary lithium-ion batteries. As the focus shifts away from energy density and range, more and more automakers are choosing the safer and more affordable LFP batteries. As of July 2021, LFP battery installations surpassed ternary lithium-ion batteries.

[0003] It's worth noting that lithium iron phosphate batteries have a significant drawback: the presence of a voltage plateau in their open-circuit voltage curve, making SOC estimation difficult and inaccurate. This severely impacts user experience. However, the industry's newly proposed ternary-lithium iron phosphate hybrid battery pack leverages the more accurate SOC estimate of the ternary lithium battery to indirectly estimate the SOC of the lithium iron phosphate battery. This not only reduces the difficulty of estimation but also avoids the open-circuit voltage plateau issue.

[0004] However, it is worth noting that the battery cells of the two different systems have different characteristics such as self-discharge rate and aging rate. If corresponding corrective measures are not taken, the SOC estimation accuracy of the lithium iron phosphate battery will become increasingly poor during the long-term use of the battery pack. The current hybrid battery pack SOC estimation patent only proposes different estimation methods, and does not consider how to avoid the situation where the SOC estimation error of the lithium iron phosphate battery cells gradually increases due to the different battery characteristics of the two systems. Summary of the Invention

[0005] To solve the above problems, a simple and accurate method for estimating the SOC of lithium iron phosphate batteries in a ternary-lithium iron phosphate hybrid battery pack is provided. The present invention adopts the following technical solutions:

[0006] The present invention provides a SOC estimation method for a ternary-iron-lithium hybrid battery pack, which is used to accurately estimate the state of charge (SOC) value of a lithium iron phosphate battery in a ternary-iron-lithium hybrid battery pack. The method is characterized in that it comprises: step S1, constructing a ternary-iron-lithium hybrid battery pack to be tested, and based on a first-order RC equivalent circuit model, using a predetermined algorithm to estimate the state of charge (SOC) value of a ternary lithium battery (NCM) in the hybrid battery pack in real time; step S2, based on the characteristic that each cell in a series battery pack has the same charge and discharge amount at the same time, calculating the SOC change of a lithium iron phosphate battery (LFP) according to the SOC value of the ternary lithium battery; step S3, judging whether the cell voltage U2 of the lithium iron phosphate battery reaches 3.65V, that is, whether the lithium iron phosphate battery reaches a fully charged state; if not, then Calculate the estimated SOC values ​​of the lithium iron phosphate battery without feedback correction and after feedback correction respectively. If the battery reaches the fully charged state, jump to step S4; in step S4, determine whether the corrected and uncorrected estimated SOC values ​​of the lithium iron phosphate battery are both equal to 100% at the time of full charge. If they are equal, directly output the current value. If they are not equal, perform self-correction to reduce subsequent estimation errors; in step S5, determine whether the lithium iron phosphate battery is fully charged for the first time. If it is the first time, only correct the estimated SOC value at the time of full charge. If it is not the first time, jump to step S6; in step S6, calculate the estimated error per second based on the estimated SOC value of the lithium iron phosphate battery without feedback correction at the time of full charge, and use this as the input of the fuzzy inference system to output the self-discharge ratio, that is, the feedback correction coefficient K i , and then obtain the correction amount of the SOC estimate per second, which is used to continuously correct the SOC estimate of the lithium iron phosphate battery before the next full charge. Among them, the ternary-lithium iron hybrid battery pack is a series battery pack, and the lithium iron phosphate monomer controls the full charge state of the battery pack, and the ternary lithium monomer controls the empty state of the battery pack.

[0007] The SOC estimation method for a ternary-iron-lithium hybrid battery pack provided by the present invention may also have the following technical features: in step S1, the predetermined algorithm is an extended Kalman filter algorithm.

[0008] The SOC estimation method of a ternary-lithium iron phosphate hybrid battery pack provided by the present invention may also have the following technical features, wherein the calculation formula of the SOC change of the lithium iron phosphate battery is:

[0009]

[0010] Where, Cap NCM Cap is the actual capacity of the ternary lithium battery cell; LFP is the actual capacity of the lithium iron phosphate battery cell; ΔSOC NCM(k-1) and ΔSOC LFP(k-1)They are the SOC changes of the ternary lithium battery and the lithium iron phosphate battery from time k-1 to time k.

[0011] The SOC estimation method of a ternary-lithium iron hybrid battery pack provided by the present invention may also have such a technical feature, wherein, in step S3, the calculation formula of the SOC estimation value of the lithium iron phosphate battery without feedback correction is:

[0012]

[0013] Where, is the estimated SOC value without feedback correction of LFP at time k; is the estimated SOC value without feedback correction of LFP at time k-1;

[0014] The calculation formula for the feedback-corrected SOC estimate of the lithium iron phosphate battery is:

[0015]

[0016] Where, is the SOC estimate after feedback correction of LFP at time k; is the SOC estimate after feedback correction of LFP at time k-1; SOC corr(i) is the feedback correction amount per second during the SOC estimation process between the i-th full charge and the i+1-th full charge.

[0017] The SOC estimation method of a ternary-lithium iron hybrid battery pack provided by the present invention may also have the following technical features, wherein the self-correction is: when the lithium iron phosphate battery cell is fully charged, its SOC value should be 100%, but due to the estimated existence of various influencing factors, its SOC estimated value may cause an error and is not 100%. In order to avoid increasing the error in the subsequent estimation process, the SOC estimated value of the lithium iron phosphate battery at the time of full charge is changed to 100% to conform to the actual situation.

[0018] The SOC estimation method of a ternary-iron-lithium hybrid battery pack provided by the present invention may also have the following technical features: wherein, in step S6, the calculation formula of the estimation error per second is:

[0019]

[0020] Where, SOC err / s(i) is the estimated error at the time of the i-th full charge in the lithium iron phosphate battery without feedback correction method; ΔT is the time difference between the i-1-th full charge and the i-th full charge, where the number of full charges is greater than 1; 10000 is a conversion factor created because the SOC estimation error per second is extremely small.

[0021] The present invention provides a method for estimating the SOC of a ternary-iron-lithium hybrid battery pack, which may also have the following technical features: in step S6, the fuzzy inference system adopts the Mamdani model and defuzzifies it using the centroid method, the fuzzy rules are formulated using the expert experience method, the membership functions of the fuzzy input and output are mainly triangular membership functions, and both include five fuzzy sets, and the fuzzy inference system is based on SOC err / s(i) As input, the correction coefficient K acting on the i-th full charge and the i+1-th full charge is output through the predetermined fuzzy rules. i , where i is greater than 1, when i≤1, K i =0, correction coefficient K i The domain of SOC is [0,1], where err / s(i) The domain is [0,0.5], SOC err / s(i) When it is greater than 0.5, the degree of self-discharge is already huge. Combined with the battery pack fault diagnosis method, it is believed that the battery management system has issued a warning and taken corresponding treatment measures before this situation occurs.

[0022] The SOC estimation method of a ternary-iron-lithium hybrid battery pack provided by the present invention may also have such a technical feature, wherein, in step S6, in order to avoid adding K i The correction will interfere with the judgment of the true self-discharge of the single cells in the battery pack. The correction amount of the SOC value per second is calculated based on the self-discharge correction amount of the uncorrected SOC estimated value of the lithium iron phosphate battery at the i-th full charge moment:

[0023]

[0024] Where, SOC corr(i) is the SOC estimation error per second caused by self-discharge between the i-1th full charge and the i-th full charge, which will be used for feedback correction of the SOC estimation between the i-th and i+1th full charges.

[0025] Functions and effects of the invention

[0026] According to a SOC estimation method for a ternary-lithium iron hybrid battery pack of the present invention, the method first adopts an extended Kalman filter algorithm to obtain an estimated SOC value of the ternary lithium battery. Since this method is relatively mature, and since the open circuit voltage curve of the ternary lithium battery has a larger slope, the calculated SOC estimation accuracy of the ternary lithium battery is relatively high. On this basis, based on the characteristic that the amount of charge change of each single cell in the same time period of the series battery pack is the same, the estimated SOC value of the lithium iron phosphate battery is calculated. Since the SOC estimation value of the lithium iron phosphate battery only depends on the amount of charge change, the algorithm complexity is relatively low and it hardly increases the computational burden of the BMS. In its use process, due to the particularity of the hybrid battery pack configuration, under the condition of full charge of the lithium iron phosphate battery cell, the present invention also adopts the self-correction mechanism of LFP. At the same time, a variable correction coefficient K is introduced for the possible self-discharge of the lithium iron phosphate battery. i , between two full charging conditions, continuous feedback correction can be performed to avoid the gradual increase of errors.

[0027] In addition, since the SOC estimation of the lithium iron phosphate battery in the present invention does not blindly rely on the ternary lithium battery, compared with the existing technology, the SOC estimation method of the ternary-lithium iron hybrid battery pack of the present invention is not only simple to calculate and highly accurate, but also has the advantage of strong adaptability. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 is a flow chart of SOC estimation of a ternary-lithium iron hybrid battery pack in an embodiment of the present invention;

[0029] Figure 2 This is a schematic diagram of a circuit of a ternary-iron-lithium hybrid battery pack used in an experiment in an embodiment of the present invention;

[0030] Figure 3 Schematic diagram of charge and discharge current of a ternary-lithium iron hybrid battery pack according to an embodiment of the present invention;

[0031] Figure 4 1 is the discharge open circuit voltage curve of two battery cells used in the experiment in the embodiment of the present invention;

[0032] Figure 5 2 is a schematic diagram of a first-order RC equivalent circuit model of a ternary lithium battery in an embodiment of the present invention;

[0033] Figure 6 1. It is a schematic diagram of the SOC value estimation result and absolute error of a ternary lithium battery cell in an embodiment of the present invention;

[0034] Figure 7 Schematic diagram of membership functions and fuzzy rules corresponding to fuzzy input and output quantities of the fuzzy inference system according to an embodiment of the present invention;

[0035] Figure 8 1. is a schematic diagram of the SOC estimation result and absolute error of a lithium iron phosphate battery cell in an embodiment of the present invention;

[0036] Figure 9 This is a comparison chart of the absolute error obtained by estimating the SOC value of a lithium iron phosphate battery using the method of the present invention in an embodiment of the present invention and the absolute error results obtained by the other two methods (no correction at all and no feedback correction). DETAILED DESCRIPTION

[0037] In order to make the technical means, creative features, objectives and effects achieved by the present invention easier to understand, the SOC estimation method of a ternary-iron-lithium hybrid battery pack of the present invention is specifically described below in conjunction with embodiments and drawings.

[0038] <Example>

[0039] Figure 1 4 is a flow chart of SOC estimation of a ternary-lithium iron hybrid battery pack in an embodiment of the present invention.

[0040] like Figure 1 As shown, a SOC estimation method for a ternary-lithium iron hybrid battery pack specifically includes the following steps:

[0041] In step S1, a ternary lithium iron hybrid battery pack to be tested is constructed, and based on a first-order RC equivalent circuit model, an extended Kalman filter (EKF) algorithm is used to estimate the state of charge (SOC) value of the ternary lithium battery (NCM) in the hybrid battery pack in real time.

[0042] Figure 2 This is a circuit diagram of a ternary-iron-lithium hybrid battery pack used in an experiment in an embodiment of the present invention.

[0043] In this embodiment, two ternary lithium batteries and lithium iron phosphate batteries with a nominal capacity of 40Ah are arbitrarily selected and connected in series to form a group (the voltage cutoff range of the ternary lithium battery is 2.5V~4V, and the voltage cutoff range of the lithium iron phosphate battery is 2V~3.65V), and the lithium iron phosphate monomer controls the full charge state of the battery pack, and the ternary lithium monomer controls the empty state of the battery pack.

[0044] Since the capacity attenuation of lithium iron phosphate battery packs at low temperatures greatly affects the user experience, in this embodiment, in order to minimize the impact of low temperature on the available capacity of the hybrid battery pack, the battery cells of the two systems are pre-treated separately before the hybrid battery pack is assembled: the ternary lithium battery is discharged at a constant current of 1 / 3C, and the lithium iron phosphate battery is discharged at a constant current of 1 / 3C and then charged at a constant current of 1 / 3C to 10% SOC. After the battery pack is assembled, a 500Ω resistor is connected in parallel to the lithium iron phosphate battery cell, as shown in the figure. Figure 2 As shown, this simulates the situation where there is self-discharge.

[0045] Figure 3 Schematic diagram of the charge and discharge current of the ternary-lithium iron hybrid battery pack in an embodiment of the present invention.

[0046] The battery pack is subjected to Figure 3 The five groups of charge and discharge cycle experimental currents are shown, where each group of charge and discharge cycles consists of the following operating conditions: (1) the lithium iron phosphate battery is fully charged in a two-stage charging mode; (2) the lithium iron phosphate battery is discharged to 3.196V under NEDC conditions; (3) the lithium iron phosphate battery is charged to 3.395V; (4) the lithium iron phosphate battery is discharged to 3.166V under NEDC conditions; (5) the lithium iron phosphate battery is charged to 3.389V; (6) the lithium iron phosphate battery is discharged to 3.215V under NEDC conditions; (7) there is a 3-hour pause between each charge and discharge condition. In order to simulate the actual operating conditions where few users will use the low SOC power range, and to prevent the discharge to the lithium iron phosphate battery voltage plateau period from increasing the difficulty of calculating the reference SOC, the lithium iron phosphate battery cell is used to control the battery pack discharge cutoff.

[0047] Among the many SOC estimation methods, the extended Kalman filter algorithm has been widely used due to its good robustness, moderate computational complexity and applicability to nonlinear systems.

[0048] Figure 4 1 is the discharge open circuit voltage curve of two battery cells used in the experiment in the embodiment of the present invention.

[0049] like Figure 4 From the discharge open circuit voltage curves of the ternary lithium battery cell and the lithium iron phosphate battery cell shown, it can be seen that the curve slope of the ternary lithium battery is larger and its SOC estimation accuracy is more accurate. Therefore, this embodiment directly adopts the extended Kalman filter algorithm to perform SOC estimation of the ternary lithium battery.

[0050] Figure 5 Schematic diagram of the first-order RC equivalent circuit model of the ternary lithium battery in an embodiment of the present invention.

[0051] The first-order RC model of a battery is a commonly used equivalent circuit model. The RC order is a key factor affecting model accuracy. Too low an order can lead to low model accuracy; too high an order increases the computational complexity and may result in overfitting.

[0052] In this embodiment, the Figure 5 The first-order RC model shown can accurately simulate the characteristics of ternary lithium batteries with moderate computational complexity. Based on basic circuit principles, the external characteristic equation of the first-order RC model is as follows:

[0053] U1=IR1×[1-exp(-t / τ1)]

[0054] U t =U OCV -IR0-U1

[0055] Where I is the charging current; R1 and C1 are the polarization internal resistance and polarization capacitance respectively; τ1 = R1C1, which is the time constant; U1 represents the voltage across the RC loop; U OCV represents the open circuit voltage; R0 is the ohmic internal resistance; U t is the model output voltage.

[0056] In this embodiment, the particle swarm algorithm is used in combination with the standard capacity test of the ternary lithium battery monomer, the HPPC experiment and the NEDC operating condition data to calculate the parameters (R 0-cha R 0-dsc R1τ1) for offline identification. Specifically:

[0057] In the parameter identification process, it is believed that the closer the model terminal voltage is to the measured terminal voltage, the more accurate the parameters are. Therefore, the root mean square error (RMSE) between the two can be used as the fitness function to evaluate the quality of the identification results:

[0058]

[0059] Where n represents the data length in the selected optimization interval; U t,k and They represent the terminal voltage measured by the voltage sensor at time k and the terminal voltage estimated by the model respectively.

[0060] The system state, input and output of EKF are as follows:

[0061] x k =(SOC k ,U 1,k )

[0062] u k =I k

[0063] y k =U t,k

[0064] Where x k is the system state quantity, including the state of charge SOC k and polarization voltage U 1,k ;u k is the system input, specifically the current I k ;y k is the system output, specifically the terminal voltage U t,k .

[0065] The iterative estimation equation of EKF is as follows:

[0066] State vector time update:

[0067] Error covariance matrix time update:

[0068] Kalman gain update:

[0069] State vector measurement update:

[0070] Error covariance matrix measurement update:

[0071] Where A k and C k are the first-order Taylor expansion coefficients of the state equation and the output equation respectively; ∑ ω and ∑ υ are ω k and υ k The covariance matrix of .

[0072] Figure 6 Schematic diagram of the SOC value estimation result and absolute error of the ternary lithium battery cell in an embodiment of the present invention.

[0073] Figure 6 In the figure, (a) is the estimated result of the SOC value of the ternary lithium battery cell, and (b) is the absolute error, where the reference SOC is calculated by the ampere-hour integration method. Figure 6 It can be seen that the estimation accuracy of the SOC value of the ternary lithium battery cell is good, with an error within 0.5%.

[0074] Step S2, based on the characteristic that the charge and discharge amount of each cell in the series battery pack is the same at the same time, the SOC change of the lithium iron phosphate battery (LFP) is calculated according to the SOC value of the ternary lithium battery:

[0075]

[0076] Where, Cap NCM Cap is the actual capacity of the ternary lithium battery cell; LFP is the actual capacity of the lithium iron phosphate battery cell; ΔSOC NCM(k-1) and ΔSOC LFP(k-1) They are the SOC changes of the ternary lithium battery NCM and the lithium iron phosphate battery LFP from time k-1 to time k.

[0077] Step S3, determine whether the single cell voltage U2 of the lithium iron phosphate battery reaches 3.65V, that is, whether the lithium iron phosphate battery reaches a fully charged state. If it does not reach a fully charged state, calculate the SOC value of the lithium iron phosphate battery before and after feedback correction respectively. If it reaches a fully charged state, jump to step S4.

[0078] In a ternary-lithium iron hybrid battery pack, the voltage limits of the two battery systems differ, so it is necessary to monitor the voltage of each cell separately. Let's take the voltage of the ternary lithium battery cell as U1 and the voltage of the lithium iron phosphate battery cell as U2. Based on the charge and discharge characteristics of the series battery pack, it can be seen that in the battery pack configuration of this embodiment, the ternary lithium battery and the lithium iron phosphate battery respectively control the empty and full states of the battery pack.

[0079] Since the SOC of ternary lithium batteries can be estimated in real time through EKF, the SOC value of lithium iron phosphate batteries is more difficult to estimate and has poor accuracy due to the existence of the plateau period of the open circuit voltage curve. Therefore, the SOC of the lithium iron phosphate battery cell can only be confirmed to be 100% at the full charge moment. When the full charge state is not reached, the SOC of the lithium iron phosphate battery is calculated in two ways: without feedback correction and with feedback correction. Among them, the SOC without feedback correction is calculated and recorded in order to calculate the degree of SOC error generated by the LFP cell per second, so as to facilitate the selection of the appropriate correction coefficient K later. i . Specifically:

[0080] The calculation formula for the SOC value of lithium iron phosphate battery without feedback correction is:

[0081]

[0082] Where, is the estimated SOC value without feedback correction of LFP at time k; It is the estimated SOC value without feedback correction of LFP at time k-1.

[0083] The calculation formula of the SOC value of the feedback-corrected lithium iron phosphate battery is:

[0084]

[0085] Where, is the SOC estimate after feedback correction of LFP at time k; is the SOC estimate after feedback correction of LFP at time k-1; SOC corr(i) is the feedback correction amount per second during the SOC estimation process after the i-th full charge and before the i+1-th full charge, and when i=0,1, SOC corr(i) Equal to 0, when i>1, SOC corr(i)The calculation method of is shown in step S6.

[0086] Step S4, determining whether the corrected and uncorrected SOC estimated values ​​of the lithium iron phosphate battery are both equal to 100% at the time of full charge, if so, directly outputting the current value, if not, performing self-correction to reduce subsequent estimation errors.

[0087] Theoretically, when the hybrid battery pack reaches a full charge, that is, when the LFP cell is fully charged, its SOC estimate should be 100%. However, in practice, the charge changes of each cell in the series battery pack are not exactly the same, and there is also some estimation error when the EKF estimates the SOC of the ternary lithium battery. Therefore, the SOC of the lithium iron phosphate battery at the time of full charge calculated based on the SOC estimate of the ternary lithium battery is not necessarily 100%. At this time, in order to avoid increasing errors in the subsequent estimation process, the SOC estimate of the lithium iron phosphate battery needs to be self-calibrated, that is, the SOC estimate of the lithium iron phosphate battery at the time of full charge is changed to 100% to conform to the actual situation.

[0088] Step S5, determining whether the lithium iron phosphate battery is fully charged for the first time, if it is the first time fully charged, only correcting the SOC estimation value at the time of full charge, if it is not the first time fully charged, jumping to step S6.

[0089] In a ternary-lithium iron phosphate hybrid battery pack, the two different battery cell systems have different self-discharge rates, aging rates, and other characteristics. Therefore, in practice, the charge changes of each cell in the series battery pack are not exactly the same, resulting in errors in the estimated SOC of the LFP. If the constant charge relationship in step S2 is relied upon to calculate the SOC of the lithium iron phosphate battery over a long period of time, the error will inevitably increase. Therefore, this embodiment takes into account the possible self-discharge error.

[0090] Since it's impossible to know when self-discharge occurs in practice and calculate the self-discharge rate at each moment based on the full-charge correction error, this step determines whether it's the first full charge. During the first full charge, only the LFP SOC estimate is modified, without incorporating the ongoing correction. Furthermore, the initial self-discharge rate is relatively low, so not correcting it won't significantly impact the battery.

[0091] Step S6, calculate the estimated error per second based on the SOC estimated value of the lithium iron phosphate battery at the time of full charge without feedback correction, and use this as the input of the fuzzy inference system to output the self-discharge ratio, that is, the feedback correction coefficient K i , and then obtain the correction amount of the SOC value per second, which is used to continuously correct the SOC value of the lithium iron phosphate battery before the next full charge.

[0092] In a ternary-lithium iron phosphate hybrid battery pack, the estimated SOC value of the lithium iron phosphate battery is directly calculated from the estimated value of the ternary lithium battery through a simple charge and discharge power relationship. If the lithium iron phosphate battery has self-discharge, then when the lithium iron phosphate battery is fully charged, the estimated SOC value will be greater than 100%. The excess is the self-discharge power. In subsequent estimates, the self-discharge error needs to be taken into account to avoid the LFP SOC estimation error from gradually increasing before the next full charge. The SOC estimation error per second is calculated as follows:

[0093]

[0094] Where, SOC err / s(i) is the estimated error at the time of the i-th full charge in the LFP method without feedback correction; ΔT is the time difference between the i-1-th full charge and the i-th full charge, where the number of full charges is greater than 1; 10000 is a conversion factor created because the error in the SOC estimation per second is extremely small.

[0095] Figure 7 It is a schematic diagram of the membership function and fuzzy rules of the fuzzy input and output quantities corresponding to the fuzzy inference system in an embodiment of the present invention.

[0096] like Figure 7 As shown, the fuzzy inference system in this embodiment uses the Mamdani model and defuzzifies using the centroid method. The membership functions for both the fuzzy input and output are primarily triangular membership functions. The input is empirically divided into five fuzzy sets: VS (very small), S (small), M (medium), B (large), and VB (very large). The fuzzy output is divided into five levels: mf1, mf2, mf3, mf4, and mf5, to correct the SOC of the lithium iron phosphate battery. As the estimation error increases, the correction coefficient increases, and the fuzzy rules are formulated using expert experience.

[0097] The fuzzy inference system is based on SOC err / s(i) and correction factor K i As the input and output of the fuzzy inference system.

[0098] Among them, the input SOC err / s(i) The domain of the estimation error is [0,0.5]. When SOC err / s(i) When it is greater than 0.5, the degree of self-discharge is already huge. Combined with the battery pack fault diagnosis method, it is believed that the battery management system has issued a warning and taken corresponding treatment measures before this situation occurs.

[0099] Output correction coefficient K i The domain of K is [0,1]. Since continuous feedback correction is performed after the first full charge, i is greater than 1. When i≤1, K i =0.

[0100] In order to avoid adding K i The correction will interfere with the judgment of the actual self-discharge of the battery cells in the battery pack. Therefore, the self-discharge correction amount needs to be calculated based on the original SOC estimation value at the i-th full charge moment in the SOC estimation method without feedback correction by LFP (that is, the value without self-correction in step S4):

[0101]

[0102] Where, SOC corr(i) is the SOC estimation error per second caused by self-discharge between the i-1th full charge and the i-th full charge, which will be used for feedback correction of the SOC estimation between the i-th and i+1th full charges.

[0103] Furthermore, by updating the relevant input quantities in step S2 through an algorithm cycle, the SOC estimation value of the lithium iron phosphate battery can be continuously corrected in the subsequent estimation process.

[0104] Figure 8 is a schematic diagram of the SOC estimation result and absolute error of the lithium iron phosphate battery cell in an embodiment of the present invention, and Figure 9 This is a comparison chart of the absolute error obtained by estimating the SOC value of a lithium iron phosphate battery using the method of the present invention in an embodiment of the present invention and the absolute error results obtained by the other two methods (no correction at all and no feedback correction).

[0105] Figure 8 (a) is a schematic diagram of the SOC estimation results of a lithium iron phosphate battery cell, and (b) is a schematic diagram of the absolute error, where the reference SOC is calculated using the ampere-hour integration method.

[0106] In this embodiment, the absolute errors of the SOC values ​​of lithium iron phosphate battery cells are calculated by using the method of the present invention, namely the SOC estimation method of the ternary-lithium iron hybrid battery pack, the completely uncorrected method, and the fully charged method. The results are compared. Figure 9 shown.

[0107] Among them, the first method is completely uncorrected, which means that the SOC of the lithium iron phosphate battery in the whole process is calculated directly based on the SOC estimation result of the ternary lithium battery according to the characteristic of the same charge and discharge capacity of the series battery pack; the second method only modifies the full charge, which refers to the method in step S2, which only corrects the SOC at the full charge moment and does not contain feedback correction.

[0108] Depend on Figure 9 As can be seen from the figure, due to the existence of self-discharge resistance, in the absence of correction, the SOC estimation error of lithium iron phosphate battery increases almost linearly with the increase of battery pack usage time. In addition, compared with the method of only correcting full charge, the first 2×10 5The overlap between the two methods during the period is due to the fact that only self-calibration is performed before the second full charge, without feedback correction. Later in the period, it becomes clear that the proposed method avoids the increasing error trend between full charges, keeping the estimated error within 0.5%, with a maximum reduction of approximately 0.7%.

[0109] Due to time constraints, only two charge-discharge cycles were used between full charges in this embodiment. Furthermore, this was an accelerated test, and the parallel self-discharge resistance in this embodiment was relatively low. However, in actual vehicle use, the intervals between full charges were longer and the self-discharge resistance was higher. In this experiment, a 500Ω resistance value resulted in a 0.5% SOC error over two days, equivalent to a 2000Ω self-discharge resistance value causing a 2% SOC error over one month. This estimate is highly accurate, demonstrating that the method proposed in this embodiment of the present invention can significantly improve the SOC.

[0110] Example Function and Effect

[0111] According to the SOC estimation method of a ternary-lithium iron hybrid battery pack provided in this embodiment, the method first adopts the extended Kalman filter algorithm to obtain the SOC estimation value of the ternary lithium battery. Since this method is relatively mature and the open circuit voltage curve of the ternary lithium battery has a larger slope, the calculated SOC estimation accuracy of the ternary lithium battery is relatively high. On this basis, based on the characteristic that the change in the amount of charge of each single cell in the same time period of the series battery pack is the same, the SOC estimation value of the lithium iron phosphate battery is calculated. Since the SOC estimation value of the lithium iron phosphate battery only depends on the change in charge, the algorithm complexity is relatively low and it hardly increases the computing burden of the BMS. In its use, due to the particularity of the hybrid battery pack configuration, under the condition of full charge of the lithium iron phosphate battery cell, this embodiment also adopts the LFP self-correction mechanism. At the same time, a variable correction coefficient K is introduced for the possible self-discharge of the lithium iron phosphate battery. i , between two full charging conditions, continuous feedback correction can be performed to avoid the gradual increase of errors.

[0112] In addition, since the SOC estimation of the lithium iron phosphate battery in this embodiment does not blindly rely on the ternary lithium battery, compared with the existing technology, the SOC estimation method of the ternary-lithium iron phosphate hybrid battery pack in this embodiment is not only simple to calculate and highly accurate, but also has the advantage of strong adaptability.

[0113] The above embodiments are only used to illustrate specific implementations of the present invention, and the present invention is not limited to the description scope of the above embodiments.

Claims

1. A method for estimating the state of charge (SOC) of a ternary-lithium iron hybrid battery pack, for accurately estimating the state of charge (SOC) value of a lithium iron phosphate battery in a ternary-lithium iron hybrid battery pack, characterized in that: include: Step S1, constructing a ternary iron-lithium hybrid battery pack to be tested, and using an extended Kalman filter algorithm based on a first-order RC equivalent circuit model to estimate the state of charge (SOC) value of the ternary lithium battery in the hybrid battery pack in real time; Step S2, based on the characteristic that the charge and discharge amounts of each cell in the series battery pack are the same at the same time, the SOC change of the lithium iron phosphate battery is calculated according to the SOC value of the ternary lithium battery; Step S3, determining whether the single cell voltage U2 of the lithium iron phosphate battery reaches 3.65V, that is, whether the lithium iron phosphate battery is fully charged. If not, respectively calculating the estimated SOC values ​​of the lithium iron phosphate battery before and after feedback correction. If it is fully charged, jumping to step S4; Step S4, determining whether the corrected and uncorrected estimated SOC values ​​of the lithium iron phosphate battery are both equal to 100% at the time of full charge; if so, directly outputting the current value; if not, performing self-correction to reduce subsequent estimation errors; Step S5, determining whether the lithium iron phosphate battery is fully charged for the first time, if it is the first time fully charged, only correcting the SOC estimation value at the time of full charge, if it is not the first time fully charged, jumping to step S6; Step S6, calculate the estimated error per second based on the SOC estimated value of the lithium iron phosphate battery at the time of full charge without feedback correction, and use this as the input of the fuzzy inference system to output the self-discharge ratio, that is, the feedback correction coefficient K i , and then obtain the correction amount of the SOC estimated value per second, which is used to continuously correct the SOC estimated value of the lithium iron phosphate battery before the next full charge. The ternary-lithium iron hybrid battery pack is a series-connected battery pack, and the full charge state of the battery pack is controlled by the lithium iron phosphate monomer, while the empty state of the battery pack is controlled by the ternary lithium monomer.

2. The SOC estimation method of a ternary-iron-lithium hybrid battery pack according to claim 1, characterized in that: in, The calculation formula for the SOC change of the lithium iron phosphate battery is: Where, Cap NCM Cap is the actual capacity of the ternary lithium battery cell; LFP is the actual capacity of the lithium iron phosphate battery cell; ΔSOC NCM(k-1) and ΔSOC LFP(k-1) They are the SOC changes of the ternary lithium battery and the lithium iron phosphate battery from time k-1 to time k, k = 1, 2, 3…, n.

3. The SOC estimation method of a ternary-iron-lithium hybrid battery pack according to claim 2, characterized in that: in, In step S3, the calculation formula for the estimated SOC value of the lithium iron phosphate battery without feedback correction is: Where, is the estimated SOC value of the lithium phosphate battery at time k without feedback correction; It is the estimated SOC value of the lithium phosphate battery at time k-1 without feedback correction; The calculation formula for the feedback-corrected SOC estimate of the lithium iron phosphate battery is: Where, The SOC estimate of the lithium phosphate battery at time k after feedback correction; The SOC estimated value of the lithium phosphate battery after feedback correction at time k-1; SOC corr(i) is the feedback correction amount per second during the SOC estimation process between the i-th full charge and the i+1-th full charge, i = 0, 1, 2, ..., n.

4. The SOC estimation method of a ternary-lithium iron hybrid battery pack according to claim 3, characterized in that: in, The self-correction is: When a lithium iron phosphate battery cell is fully charged, its SOC value should be 100%. However, due to various influencing factors, the estimated SOC value may cause errors and is not 100%. In order to avoid increasing this error in the subsequent estimation process, the estimated SOC value of the lithium iron phosphate battery at the time of full charge is changed to 100% to conform to the actual situation.

5. The SOC estimation method of a ternary-lithium iron hybrid battery pack according to claim 3, characterized in that: in, In step S6, the calculation formula of the estimation error per second is: Where, SOC err / s(i) is the estimated error at the time of the i-th full charge in the lithium iron phosphate battery without feedback correction method; ΔT is the time difference between the i-1-th full charge and the i-th full charge, where the number of full charges is greater than 1, i = 2, 3, ..., n; 10000 is a conversion factor created because the SOC estimation error per second is extremely small.

6. The SOC estimation method of a ternary-iron-lithium hybrid battery pack according to claim 5, characterized in that: in, In step S6, in order to avoid adding K i The correction will interfere with the judgment of the true self-discharge of the single cells in the battery pack. The correction amount of the SOC estimate per second is calculated based on the self-discharge correction amount of the uncorrected SOC estimate of the lithium iron phosphate battery at the i-th full charge moment: Where, SOC corr(i) is the SOC estimation error per second caused by self-discharge between the i-1th full charge and the i-th full charge, which will be used for feedback correction of the SOC estimation between the i-th and i+1th full charges, i = 2, 3, …, n.

Images