Error estimation improvement for capacity estimators estimating events

Abstract

This invention relates to a method for reducing the estimation error of the state of charge (SOC) of a battery, particularly a traction battery for an electric vehicle. The method comprises at least the following steps: a) introducing a SOC error (118) for calculating the SOC, wherein the SOC error (118) is caused by an erroneous SOC value and is generated due to a capacity estimation error; b) introducing a weighting factor for calculating the SOC; and c) introducing a residual error based on neglecting open-circuit voltage hysteresis. The SOC error (118) corresponds to the maximum positive sum and / or SOC deviation between a starting SOC value (108) and a ending SOC value (110), wherein the starting SOC value (108) is the SOC value at a starting measurement point (112), and the ending SOC value (110) is the SOC value at an ending measurement point (114).

Description

Technical Field

[0001] This invention relates to a method for reducing estimation errors in the state of capacity of batteries, particularly vehicle batteries. Furthermore, this invention relates to the use of said method for calculating the state of capacity of batteries, particularly vehicle batteries. Background Technology

[0002] Accurate determination of the state of charge (SOC) is crucial for lithium-ion batteries. SOC is an input variable for many important battery functions, such as state of health, cell balancing, and SOC-based current limiting. Currently, SOC is determined by a set of open-circuit voltages derived from measured cell voltages. In the laboratory, the charging open-circuit voltage (e.g., measured during 1C charging) and discharging open-circuit voltage (e.g., measured during 1C discharging) are determined. The average open-circuit voltage is then stored in the battery management system and used for subsequent calculations. The difference between the charging and discharging open-circuit voltages produces a state of health hysteresis, which is primarily influenced by anode and cathode capacities. The key is the intensity (in terms of amplitude and charge) at which charging or discharging is performed.

[0003] BR 10 2020 0262 26 A2 relates to a method and system for predicting the end of life of a secondary battery, applicable to all types of secondary batteries. The method identifies end-of-life by monitoring voltage hysteresis cycles occurring between the battery's charging and discharging phases. This method is non-invasive, low-cost, field-applicable, and requires no special equipment.

[0004] US 2011 / 0264381 A1 relates to a system comprising an electrochemical cell, monitoring hardware, and a computer system. The monitoring hardware captures performance characteristics of the electrochemical cell. The computer system derives information about the cell from these performance characteristics. Furthermore, the computer system uses a modified Butler-Volmer (BV) expression to analyze cell data of the electrochemical cell to determine the exchange current density of the electrochemical cell, taking into account cell kinetic information regarding pulse time dependence, electrode area availability information, or a combination of both. To determine the kinetic performance as a function of pulse time, a series of sigmoid-based expressions can be incorporated into the modified BV expression. The obtained exchange current density can be used to analyze other characteristics of the electrochemical cell using the modified BV with or without sigmoid terms. Model parameters can be defined for the aging of the cell. Therefore, the overall kinetic model can predict the kinetic performance of the cell over an aging cycle.

[0005] According to conventional methods, capacity estimation is calculated using the following formula: At time t start and t end The calculated charge integral is divided by the SOC span (SOC-Hub). The SOC value (state of charge) is obtained using the aging-related OCV curve (open-circuit voltage) and the measured OCV resting voltage (U). cell (t) start ) and U cell (t) end ))calculate.

[0006] The calculated estimation error is used to evaluate the accuracy of capacity measurement. If the known accuracy is better than a threshold, the result is used; otherwise, it is discarded. Furthermore, the capacity estimation result is always partially incorporated into the existing capacity value. The higher the accuracy, the greater the impact of the newly calculated capacity value on the final result when weighted.

[0007] To estimate the error, according to the current method, a value derived from the starting point of the measurement (ΔSC) is used. start ) and endpoint (ΔSC) end The estimated SOC error is caused by voltage errors (voltage measurement error (sensor), incomplete relaxation). In addition, integration error is also included in the results.

[0008] Systems and methods for accurately determining the aging state of lithium-sulfur batteries for corresponding modules or cells, or systems and methods for estimating the state of charge of batteries, are known from US2020 / 082631A1 and CN110286324A, respectively. Summary of the Invention

[0009] According to the present invention, a method for reducing the estimation error of the state of capacity of batteries, particularly vehicle batteries, is proposed, wherein at least the following method steps are performed: a) Introducing a state-of-charge (POC) error for calculating the capacity state, wherein the POC error is caused by an inaccurate capacity state value, and the POC error is generated due to capacity estimation error. b) Introduce weighting factors for calculating capacity state, and c) Based on neglecting the open-circuit voltage hysteresis, a residual error is introduced. Wherein, the state of charge error corresponds to the maximum positive and / or negative state of charge deviation between the initial state of charge value and the final state of charge value, and the initial state of charge value is the state of charge value at the starting measurement point, and the final state of charge value is the state of charge value at the ending measurement point.

[0010] The method proposed according to the present invention can achieve significant improvements in the calculation of capacity estimation errors. Furthermore, erroneous capacity estimation errors can be better filtered out, or alternatively, erroneous capacity estimation errors have a smaller impact on determining the final result.

[0011] In an advantageous extension of the method proposed according to the invention, the capacity estimation error is 5%, preferably 3%, and particularly advantageously 2%.

[0012] Furthermore, the method proposed according to the invention advantageously incorporates a scaling parameter to increase or decrease the impact of capacity estimation error on state of charge error. This scaling parameter is used relative to other error effects (e.g., the state of charge (SOC) at the start of the charging process due to insufficient voltage relaxation or voltage sensor error). start ) and the state of charge at the end (SOC) end This can be used to enhance or reduce the impact of capacity estimation error on state-of-charge error. If the scaling parameter is set to zero, the standard for capacity estimation error is not included in the state-of-charge error.

[0013] Advantageously, according to the method proposed in this invention, the value of the scaling parameter is between 0 and 1.

[0014] Furthermore, in the method proposed according to the present invention, the initial state of charge value and the final state of charge value are obtained from the aging open-circuit voltage curve.

[0015] In an extension of the method proposed according to the present invention, the aging open-circuit voltage curve represents the change process of open-circuit voltage and state of charge within a time window. Furthermore, the aging open-circuit voltage curve is provided for calculating the state of charge error. By referencing the aging open-circuit voltage curve, the accuracy of open-circuit voltage and state of charge can be improved within a defined time window.

[0016] Furthermore, the method proposed according to the present invention is characterized in that the weighting factor decreases as the open-circuit voltage change due to battery, especially vehicle battery, aging increases and as the capacity estimation error increases.

[0017] Furthermore, the method proposed according to the present invention specifies that the increased capacity estimation error is limited to a maximum capacity estimation error, which is calculated by referencing the difference between the initial state of charge and the final state of charge of the battery, particularly the vehicle traction battery.

[0018] Furthermore, in the method proposed according to the invention, it is advantageously stipulated that the residual error is assumed to be a percentage share of the open-circuit voltage hysteresis. In the present case, residual error should be understood as a performance deviation of the battery cell due to manufacturing tolerances. This residual error cannot be improved or can only be eliminated at a very high cost.

[0019] Furthermore, the present invention relates to the use of this method for calculating the state of capacity of a battery, particularly a vehicle traction battery of an electric drive vehicle.

[0020] Advantages of the present invention The method proposed according to the present invention can prevent state-of-health drifts in batteries and significantly improve the robustness of the obtained estimation results compared to those obtained by methods used to date for estimating battery remaining capacity. Among these methods applied to date, SOHC bias leading to EOL (end of life) (i.e., SOH < 80%) has been detected.

[0021] By using the method proposed in this invention, a significant improvement can be achieved in the calculation of capacity estimation errors. This results in erroneous capacity estimation errors being filtered out from the beginning and their impact being reduced in subsequent calculations, thus reducing their influence on the final result.

[0022] To prevent drift in capacity estimates, the error calculation method was extended. Based on... The new calculation specification reveals that errors in SOHC estimation can lead to variations in SOC. start and SOC end The maximum positive and negative SOC deviation between them. The maximum SOC difference represents ΔSOC. SOH error.

[0023] The greater the change in open-circuit voltage (OCV) caused by aging, the greater the capacity estimation error (maximum conservative case SOHC). err =SOH BOL –SOH EOL The lower the weight of the calculated capacity value in the final result, the smaller its weight. The impact of capacity estimation error can be increased or decreased using an optional scaling parameter. This scaling parameter can be advantageously selected between 0 (off) and 1 (on). Attached Figure Description

[0024] Embodiments of the invention will be explained in more detail with reference to the accompanying drawings and the following description.

[0025] The image shows: Figure 1 : The various changes in the open-circuit voltage curves of batteries at different aging stages, along with capacity estimation parameters, and Figure 2 : Illustrations of the SOC hysteresis change process of batteries in different aging states. Detailed Implementation

[0026] In the following description of embodiments of the invention, the same or similar elements are denoted by the same reference numerals, and repeated descriptions of these elements are omitted in certain cases. These figures are only schematic representations of the subject matter of the invention.

[0027] according to Figure 1 The diagram illustrates the various changes in the open-circuit voltage (OCV) curve for different capacity estimation parameters. For example... Figure 1 As shown, in the aging open-circuit voltage diagram 100, the open-circuit voltage 104 is plotted with respect to the state of charge 102 (SOC). The aging open-circuit voltage curve 106 depicts the aging open-circuit voltage curve from BOL (start of life) to EOL (end of life). From... Figure 1 As shown in the diagram, the various open-circuit voltage curves 106 in the aging open-circuit voltage curve series 106 exhibit significantly different changes within the range of 40% SOC to approximately 80% SOC in the aging open-circuit voltage diagram 100. To calculate the capacity of a battery cell, the corresponding open-circuit voltage curve 106 is used to determine the battery capacity. The calculated capacity estimate typically carries an error of + / - 5%. This capacity estimation error, i.e., state-of-charge error 118, leads to the use of an open-circuit voltage curve 106 with an incorrect aging state in subsequent estimations. This additional accumulated error can result in further, more inaccurate capacity estimates, a phenomenon known as drift. This primarily occurs in regions of the various open-circuit voltage curves 106 where significant changes occur during battery aging, particularly in traction batteries.

[0028] From the basis Figure 1 The diagram shows that the state-of-charge integral 116 can be calculated from the initial state-of-charge value 108 to the final state-of-charge value 110. From the initial measurement point 112 to the final measurement point 114, which is approximately the same for all aging open-circuit voltage curves 106, the aforementioned state-of-charge error 118 arises depending on the location of the final measurement point 114. This is because different aging open-circuit voltage curves 106 may apply to the final measurement point 114, resulting in this error. Figure 1 The state-of-charge error 118 is shown. The various aging open-circuit voltage curves 106 plotted in the aging open-circuit voltage diagram 100 are summarized in the capacity table 120 of the aging open-circuit voltage curves 106. For different capacities of 71%, 80%, 73%, 90%, and 100%, the aforementioned series of aging open-circuit voltage curves 106 were obtained, as they are plotted in the aging open-circuit voltage diagram 100.

[0029] To prevent the drift of the aforementioned capacity estimates, the error calculation for capacity estimates has been extended according to the method proposed in this invention. In the new calculation specification, the SOC error (ΔSOC) caused by the erroneous SOHC value is included. SOH ) is included and added to ΔSC start and ΔSC end Above. This can be seen from the following relationship: The newly added addend ΔSOC in the above formula SOH The error can be determined according to the following relationship: Ideally, the capacity estimation error is known, for example, SOH. C,err =+ / -5%. This value is included in the calculation. Therefore, the error in SOHC estimation due to SOHC estimation is known in the SOC calculation. start and SOC end The maximum positive and negative OC deviation between them. This maximum SOC difference represents ΔSOC. SOH error.

[0030] The greater the change in open-circuit voltage caused by aging, the greater the capacity estimation error (maximum conservative case SOH). C,err =SOH BOL –SOH EOL The smaller the weight of the calculated capacity value in the final result, the more significant the scaling parameter becomes. This can be seen from the relationship above. The scaling parameter can optionally increase or decrease the impact of capacity estimation errors. Its value ranges from 0 (off) to 1 (on).

[0031] Advantageously, the uncorrectable residual error of open-circuit voltage hysteresis (OCV hysteresis) can also be included in the error calculation of capacity estimation error. This is based on... Figure 2 It is schematically represented in the diagram.

[0032] From the basis Figure 2 The diagram shows the state of charge (SOC) hysteresis plot 200. The percentage change in SOC 204 is plotted with respect to the percentage change in SOC 202. The changes in the battery's state of charge 206 and—which are essentially mirror images of these states—the end-of-discharge state 208 are obtained. For each SOC state, refer to Table 210.

[0033] From the basis Figure 2As shown in the diagram, the state-of-charge hysteresis plot 200 contains multiple state-of-charge hysteresis curves, representing the state-of-charge values ​​of batteries with different states of charge 206, discharge 208, and state of health 210 (SOH). Along the X-axis, the state of charge 202 varies from 0 to 100% as a percentage, while the Y-axis reflects the state-of-charge change ΔSOC204 from -4% to +4%.

[0034] according to Figure 2 As shown in state-of-charge (SOC) hysteresis plot 200, there are state curves illustrating the changes in SOC across multiple batteries with different states of charge (206) and discharge (208). These state curves illustrate the effect of the SOC change process 202 on different states of charge and discharge (206, 208) and the effect on the respective state of health (SOH) 210 of the batteries. With increasing SOC 202, the SOC curve initially rises sharply, while a gradual decline is observed as SOC 202 decreases. The state curves also show that the SOC response varies depending on the battery's state of health 210. Therefore, newer batteries exhibit faster SOC changes 204, while older batteries tend to respond more slowly. This is due to changes in the electrochemical reactions occurring within the batteries.

[0035] For example, hysteresis can be determined by analyzing the difference between the state of charge (SOC) value during charging state 206 and the SOC value during discharging state 208. In the SOC hysteresis plot 200, these differences are represented as curves showing the changes in SOC 206 during battery charging and SOC 208 during battery discharging.

[0036] according to Figure 2 The SOC hysteresis change process shown can be stored in the battery management system of the traction battery or vehicle battery. As an uncorrectable error, a percentage share of the originally present open-circuit voltage hysteresis can be assumed, for example, 10% or 20%. The residual error can be added separately to the starting point (ΔSOC). start ) and endpoint (ΔSOC) end The relaxation error is related to the open-circuit voltage. For example, SOHC measurements can begin at 30% state of charge (SOC). In this case, the maximum SOC error caused by the hysteresis effect of the open-circuit voltage is approximately 2% SOC.

[0037] Then, for the relaxation error ΔSOC caused by sensor error and incomplete relaxation... start According to the above calculation guidelines, if we assume that the residual error of the original OCV hysteresis error measurement is 10%, we can add an additional 0.2% SOC error.

[0038] This invention is not limited to the embodiments described herein and the aspects highlighted therein. Rather, many modifications are possible within the scope of the claims and fall within the realm of professional expertise.

Claims

1. A method for reducing estimation errors in battery state of capacity, comprising at least the following steps: a) Introducing a state-of-charge error (118) for calculating the capacity state, wherein the state-of-charge error (118) is caused by an erroneous capacity state value, wherein the state-of-charge error (118) is generated due to capacity estimation error. b) Introduce weighting factors for calculating capacity state, and c) Based on neglecting the open-circuit voltage hysteresis, a residual error is introduced. The state of charge error (118) corresponds to the maximum positive and / or negative state of charge deviation between the initial state of charge value (108) and the final state of charge value (110), wherein the initial state of charge value (108) is the state of charge value at the starting measurement point (112), and the final state of charge value (110) is the state of charge value at the ending measurement point (114).

2. The method according to claim 1, wherein, The capacity estimation error is 5%, preferably 3%, and particularly advantageously 2%.

3. The method according to any one of the preceding claims, wherein, The effect of the capacity estimation error on the state of charge error (118) can be increased or decreased by scaling parameters.

4. The method according to claim 3, wherein, The scaling parameter has a value between 0 and 1.

5. The method according to any one of the preceding claims, wherein, The initial state of charge (108) and the final state of charge (110) are obtained from the aging open-circuit voltage curve (106).

6. The method according to claim 5, wherein, The aging open-circuit voltage curve (106) represents the change process of open-circuit voltage (104) and state of charge (102) within a time window, wherein the aging open-circuit voltage curve (106) is provided for calculating the state of charge error (118).

7. The method according to any one of the preceding claims, wherein, The weighting factor decreases as the change in open-circuit voltage (104) due to battery aging increases and as the capacity estimation error increases.

8. The method according to claim 7, wherein, The increased capacity estimation error is limited to a maximum capacity estimation error, which is calculated by referencing the difference between the initial state of charge (108) and the final state of charge (110) of the battery.

9. The method according to any one of the preceding claims, wherein, The residual error is assumed to be the percentage share of open-circuit voltage hysteresis.

10. Use of the method according to any one of the preceding claims for calculating the state of battery capacity.

Images