A method for online estimating the state of charge of a lithium battery under a working condition

Abstract

The present invention discloses a method for online estimating the state of charge of a lithium battery under working conditions, which is applicable to the method for estimating the state of charge of a lithium battery by obtaining the OCV-SOC characteristics based on a fractional-order equivalent circuit model and a co-evolutionary particle swarm algorithm, and belongs to the field of battery equivalent parameter identification. This method is applicable to obtaining an accurate open-circuit voltage to improve the SOC estimation accuracy of the look-up table method. It solves the problems that the identification result of the battery open-circuit voltage is often related to the working conditions, resulting in inconsistent and fluctuating identification results, and the low accuracy of the SOC estimation result due to the hysteresis characteristic of the original OCV-SOC characteristic. According to this method, the stability of the open-circuit voltage identification result is enhanced, and in addition, the hysteresis characteristic of the SOC generated by integrating the dynamic current is linearized. An updatable OCV-SOC mapping relationship is established based on the processing results of the open-circuit voltage and the SOC, and the accuracy of this mapping relationship for estimating the SOC of the lithium battery is verified under different working conditions.

Description

Technical Field

[0001] The present invention relates to lithium battery power estimation, and in particular to a method for online lithium battery charge state estimation under working conditions. Background Art

[0002] Lithium-ion batteries have the advantages of high energy density, high power density and long cycle life, and are widely used in the new energy vehicle industry. Among them, open circuit voltage (OCV) is one of the key characteristic parameters of lithium-ion batteries. The relationship between open circuit voltage and state of charge (SOC) plays an important role in the estimation of the state of charge of lithium-ion batteries. By establishing an equivalent circuit model (ECM) to simulate the dynamic and static characteristics of the battery, the state estimation algorithm can be used to estimate the SOC. Using the model-based SOC estimation method, it is first necessary to identify the model parameters to obtain the model parameters. As an important parameter in the ECM, researchers have found through a large number of experiments that there is a strong linear relationship between the open circuit voltage of lithium-ion batteries and the SOC of the battery. The OCV-SOC mapping curve is used to establish a connection between the measured electrical parameters and the SOC, so that the current open circuit voltage value can be measured to map the current SOC of the battery. This method is also called the open circuit voltage method. However, due to the polarization phenomenon inside the lithium-ion battery, the battery needs to be left alone for a long enough time to eliminate the polarization phenomenon, so that the open circuit voltage of the battery can be accurately obtained. The open circuit voltage method requires obtaining the open circuit voltage value of a certain SOC state by statically placing the vehicle, so it cannot be used for online SOC estimation, and there is a problem that the research model has poor accuracy within the local SOC range. Therefore, obtaining an accurate OCV-SOC curve is crucial to improving the accuracy of SOC estimation.

[0003] When using open circuit voltage to estimate SOC, the accuracy of open circuit voltage parameter identification will directly affect the accuracy of SOC estimation. There are two types of SOC parameter identification algorithms: offline and online, corresponding to offline modeling and online modeling respectively. Offline parameter identification does not require the identification and calculation of model parameters during the SOC estimation process of the battery, and is simple and convenient. However, the model accuracy of this method will decrease under complex working conditions and after battery aging. The main offline parameter identification algorithms are the least squares algorithm. The online parameter identification method can estimate the model parameters in real time according to the specific situation of the model, avoid the shortcomings of the offline parameter identification method, and obtain higher model accuracy. However, this method has large computational complexity and high cost. The main online parameter identification algorithms are the recursive least squares algorithm, the extended Kalman filter algorithm, the particle swarm algorithm, the genetic algorithm, etc.

[0004] Due to its simple calculation and easy calibration of the initial SOC, the method combining the ampere-hour integration method with a relatively high reliability OCV has been successfully applied to SOC estimation in BMS. Filtering techniques using state space models are used to enhance SOC estimation, and various filtering algorithms applicable to nonlinear systems have been developed based on the Kalman filter algorithm, including the Extended Kalman Filter (EKF), Unscented Kalman Filter (UKF), Cubature Kalman Filter (CKF), etc. Some scholars have proposed a new method for parameterizing the OCV curve polynomial. As the lithium-ion battery ages, the parameters are adjusted to ensure the accurate fitting of the polynomial to the OCV curve. Using an OCV-SOC model based on fractional calculus improves the accuracy of the model, which can well represent the strong nonlinear relationship between the OCV and SOC of lithium-ion batteries. However, the hysteresis characteristics of the OCV-SOC curve affect SOC estimation. Therefore, it is necessary to systematically study the hysteresis characteristics of the lithium-ion battery OCV-SOC to improve the accuracy of SOC estimation. Summary of the Invention

[0005] Aiming at the technical problems existing in the prior art, the present invention proposes a method for online estimating the state of charge of a lithium battery under working conditions, with reasonable design, overcoming the deficiencies of the prior art and having good effects.

[0006] The present invention adopts the following technical solutions:

[0007] A method for online estimating the state of charge of a lithium battery under working conditions, comprising the following steps:

[0008] Step 1: Use an experimental platform to conduct characteristic tests on a ternary lithium battery under different working conditions to obtain test data.

[0009] Step 2: Establish a second-order fractional equivalent circuit model, input the test data into the second-order fractional equivalent circuit model, and use the co-evolutionary particle swarm optimization algorithm for online identification of the open-circuit voltage and parameter Δu oc of the second-order fractional equivalent circuit model.

[0010] Step 3: By adjusting the length of the parameter identification window in the co-evolutionary particle swarm optimization algorithm, obtain the parameter Δu oc The length of the parameter identification window with a value lower than 0.1 mv and no fluctuations is the optimal parameter identification window length. Under the condition of the optimal parameter identification window length, identify the working condition test data to obtain open-circuit voltage data, and obtain the state of charge data through the ampere-hour integration method.

[0011] Step 4: Process the open-circuit voltage data using the mean value and filtering method to eliminate errors; in order to avoid random errors caused by the identification results of the open-circuit voltage data, the open-circuit voltage data obtained by identifying 5 times under the condition of the optimal parameter identification window length are averaged and then filtered using the recursive average filtering method, and the state of charge data obtained by the ampere-hour integration method is filtered using the recursive average filtering method.

[0012] Step 5: The open-circuit voltage data and the state of charge data processed in Step 4 form an OCV-SOC curve, and the state of charge data corresponding to the open-circuit voltage is obtained by the look-up table method under the working condition, completing the estimation of the lithium battery power.

[0013] Preferably, the working conditions are the FUDS working condition and the DST working condition, the FUDS working condition is the Federal Urban Driving Test working condition, and the DST working condition is the Dynamic Stress Test working condition.

[0014] Preferably, the second-order fractional-order equivalent circuit model replaces the capacitor in the integer-order equivalent circuit model with two constant phase elements.

[0015] Preferably, the parameter Δu oc is the change in the open-circuit voltage between two adjacent sampling points in the parameter identification window, and the numerical change of the parameter Δu oc can directly reflect the fluctuation of the open-circuit voltage identification result.

[0016] Preferably, the optimal parameter identification window length is 1200.

[0017] The beneficial effects of the present invention are as follows:

[0018] Based on the co-evolutionary particle swarm optimization algorithm, the present invention studies a method for online estimating the lithium battery power under working conditions, determines the selection criteria for the parameter identification window that can resist the influence of working conditions according to this method, enhances the smoothness of the open-circuit voltage identification result, and additionally smooths the hysteresis effect of the OCV-SOC characteristics caused by the dynamic working condition current fluctuation, increasing the engineering application value. It is applicable to obtaining accurate open-circuit voltage to improve the SOC estimation accuracy of the look-up table method. It solves the problems that the identification result of the battery open-circuit voltage is often related to the working condition, resulting in inconsistent and fluctuating identification results, and the accuracy of the SOC estimation result is not high due to the hysteresis characteristic of the original OCV-SOC characteristic. An updatable OCV-SOC mapping relationship is established according to the processing results of the open-circuit voltage and SOC, and the accuracy of this mapping relationship for the lithium battery SOC estimation is verified under different conditions, that is, the accuracy of the lithium battery power estimation. The results show that the SOC estimated values of the processed OCV-SOC curve can accurately follow the true value of SOC under different working conditions, with good accuracy and stability, and the SOC estimated values can quickly converge to the vicinity of the true value in any interval, with good consistency. Description of the Drawings

[0019] Figure 1 is the fractional-order equivalent circuit model.

[0020] Figure 2 is the schematic diagram of the CPSO algorithm.

[0021] Figure 3 is the calibration of the OCV-SOC curve by the small current charge and discharge method; where (a) is the OCV-SOC curve at different temperatures, and (b) is the OCV-SOC curve at 25 degrees Celsius.

[0022] Figure 4 In (a) is the load current diagram under the DST condition, and (b) is the load current under the FUDS condition.

[0023] Figure 5 In (a) are the test data and OCV identification results under the DST condition, and (b) are the test data and OCV identification results under the FUDS condition.

[0024] Figure 6 In (a) are the original OCV-SOC curves under two working conditions, and (b) is the hysteresis characteristic in the original OCV-SOC mapping relationship.

[0025] Figure 7 is the influence of load current fluctuation on the original SOC.

[0026] Figure 8 In (a) are the OCV identification results under different window lengths, and (b) is the partial enlarged view of the OCV identification results under different window lengths.

[0027] Figure 9 In (a) is the Δu oc identification result under different identification window lengths, and (b) is the Δu oc partial enlarged view of the identification result under different window lengths.

[0028] Figure 10 In (a) are the Δu oc identification results corresponding to the high SOC interval, (b) are the Δu oc identification results corresponding to the medium SOC interval, and (c) are the Δu oc identification results corresponding to the low SOC interval.

[0029] Figure 11 is the process of the recursive average filtering method.

[0030] Figure 12 In (a) are the 5 OCV identification results and the average value, and (b) is the SOC filtering result.

[0031] Figure 13 In (a), it is the OCV average filtering result, and in (b), it is the filtering error.

[0032] Figure 14 It is the processed OCV-SOC curve.

[0033] Figure 15 In (a), it is the comparison of the estimated values of SOC under the DST working condition, and in (b), it is the estimation error of SOC.

[0034] Figure 16 In (a), it is the comparison of the estimated values of SOC under the FUDS working condition, and in (b), it is the estimation error of SOC.

[0035] Figure 17 In (a), it is the comparison of the SOC estimations under five FUDS working conditions, and in (b), it is the enlarged view of the local part of the comparison of the SOC estimations under five FUDS working conditions.

[0036] Figure 18 It is the convergence of the SOC estimation under different initial values. Among them, in Figure (a), it is the SOC estimation result under the condition that the initial value is 80%, and in Figure (b), it is the SOC estimation result under the condition that the initial value is 60%.

[0037] Figure 19 In (a), it is the comparison of the estimated values of SOC with the small current method, and in (b), it is the estimation error of SOC. Specific Embodiments

[0038] The following further describes the specific embodiments of the present invention in conjunction with the accompanying drawings and specific embodiments:

[0039] The electrochemical impedance spectrum of the ternary lithium battery has obvious fractional-order characteristics. Therefore, elements with fractional-order characteristics are used to replace the capacitors in the integer-order circuit, and a fractional-order equivalent circuit model is constructed to describe the impedance characteristics of the battery, so that the state and parameters of the battery can be accurately extracted.

[0040] Second-order fractional-order equivalent circuit model: In the present invention, two constant phase elements CPE1 and CPE2 are used to replace the two integer-order capacitors in the equivalent circuit model, and a fractional-order equivalent circuit model (FECM) is established as Figure 1 shown.

[0041] In Figure 1 , u oc is the open-circuit voltage of the lithium battery, u t is the terminal voltage, R0 is the ohmic internal resistance, CPE1 and CPE2 are two fractional-order capacitors with values of C1 and C2, R1 and R2 are the parallel resistors of the fractional-order capacitors, and i is the load current. To reduce the calculation consumption and the subsequent identification time, the fractional-order orders α and β of the constant phase elements CPE1 and CPE2 are determined as fixed values.

[0042] After a large number of experiments and data analyses, the fractional orders α and β are respectively set to fixed values of 0.7 and 0.9 to reduce the computational consumption and shorten the duration in the subsequent identification process.

[0043] The Co-evolutionary Particle Swarm Optimization (CPSO) algorithm is characterized by combining the Particle Swarm Optimization algorithm and the Parameter Identification Window (PIW) to address the issue of low real-time performance in bionic optimization algorithms. Through the advancement of the parameter identification window, online identification of the open-circuit voltage and other equivalent parameters is achieved.

[0044] The length of the Parameter Identification Window (PIW), as a key parameter of CPSO, affects the stability and consistency of the equivalent parameter identification results. By setting different values for it and testing the algorithm, the influence of the working condition on the open-circuit voltage is weakened by adjusting the length of the parameter identification window. It is verified that under the DST working condition standard, the time period for each working condition test is 360 seconds. Therefore, the number of data points sampled within a 360s time length is 720. The length of the data points identified during the identification of the parameter identification window length affects the optimization result of the particle swarm and thus the identification result. The longer the window length, the more data information is stored, tending to represent the steady-state characteristics of the battery. Therefore, a longer parameter identification window is beneficial for obtaining stable parameter identification results. To better analyze the relationship between the length of the parameter identification window and the fluctuation of the open-circuit voltage, Δu oc parameter is added to the parameter set identified by the CPSO algorithm. Δu oc is the change in the open-circuit voltage between two adjacent sampling points in the parameter identification window. The numerical change of the parameter Δu oc can directly reflect the fluctuation of the OCV identification result. Through analysis and summary, it can be seen that when the window length is longer than the data length within one DST working condition time period, the algorithm has a better inhibitory effect on the fluctuation of the open-circuit voltage identification result caused by the working condition. So the window length should exceed 720. However, as the window length increases, the identification time also increases, and when the window is too long, the inhibitory effect of the algorithm on the fluctuation is no longer obvious when the open-circuit voltage identification value is relatively smooth. It is verified that when the parameter identification window length is between 900 and 1200, the algorithm has a more obvious inhibitory effect on the fluctuation and the open-circuit voltage identification result is relatively smooth.

[0045] CPSO uses a PIW with a window length of w to save a small segment of data before the current moment of the battery, and identifies the equivalent parameters of the battery within the PIW. The PIW moves with the current sampling data point, and CPSO completes one evolutionary step each time the PIW moves. Within each evolutionary step, CPSO only identifies and updates one parameter in the parameter set P. As the sampling points progress, the PIW stores the latest data and discards the oldest data. CPSO sequentially identifies and updates each parameter in P. When the parameters in P are all identified, the identification starts from OCV in a loop until the PIW moves to the last original data point. Figure 2 Figure 1 shows the schematic diagram of the CPSO principle. In the figure, PIW(n) represents the current identification window, and PIW(n - 1) represents the previous identification window; L S represents the predicted step length of the window advancement; ut(k - L S ) and ut(L S ) represent the voltage data loaded within the predicted step length respectively.

[0046] Example 1: The battery used for testing is a ternary lithium battery, and the model of the tested lithium battery is NCR18650BE. The test is carried out at a constant temperature of 25°C, and the detailed parameters are shown in Table 1. The experimental platform uses a main controller with the model of STM32F407ZGT6. The DST load condition is defined through the 12-bit DA converter integrated in the MCU, and the programmable LoadProfile is provided to the battery by controlling the charge and discharge current through a power amplifier. The ADS1274 is used to collect the terminal voltage and current of the battery, and the collected data is sent to the PC side (Intel Core i7 - 7500U CPU, 2.70GHz, 12GB, 64bit, Windows 10) for saving the experimental data after being preprocessed by the MCU. After the original experimental data is tested and stored, the MATLAB (R2020b) is used to verify the parameter identification algorithm.

[0047] Table 1 Technical parameters of the ternary lithium battery

[0048]

[0049] Lithium battery characteristic test: (1) Small current charge and discharge test. The small current charge and discharge is to charge and discharge the battery with a current of C / 20. At a very low discharge rate, the polarization effect of the battery is very weak, and then the average voltage during the discharge and charge processes is recorded as OCV, and the average voltage can weaken the influence of the hysteresis effect and ohmic internal resistance of the battery.

[0050] At a fixed temperature, using the terminal voltage curves measured during charging and discharging, the OCV - SOC curves at different temperatures are as shown in Figure 3 (a). The OCV - SOC curve under the temperature condition of 25 degrees is obtained by interpolating the median value as shown inFigure 3 as shown in (b) of

[0051] (2) Dynamic condition test. According to the Federal Urban Driving Schedule (FUDS) and Dynamic Stress Test (DST) standards in the "USABC Battery Test Method Manual", the dynamic condition test of lithium batteries is carried out, and the test data is saved in the database. The data points stored in the database are sorted according to the sampling time. The actual data sampling is 0.5 s, which is used for algorithm analysis and verification; The FUDS condition is the time–speed curve of standard urban driving vehicles in the automotive industry, and the time period for completing one cycle is 1372 seconds. The DST condition is a simplified version based on the power–time demand in FUDS, and the time period for completing one cycle is 360 seconds. The load currents under the two dynamic electrical conditions are as Figure 4 shown.

[0052] Figure 5 represent the battery discharge test results under the two dynamic conditions, Figure 5 (a) in Figure 5 is the test curve of the DST condition,

[0053] Under dynamic condition, the identified OCV-SOC curve has a hysteresis characteristic, which affects engineering applications. The commonly used OCV-SOC calibration methods in engineering, including the HPPC method and the small current charge and discharge method, usually result in a monotonic OCV-SOC mapping relationship. However, the CPSO algorithm is improved based on the standard particle swarm algorithm. When conducting algorithm tests, the part with large working condition fluctuations will show a hysteresis characteristic, Figure 5 represents the identification results of OCV under the DST condition and the FUDS condition, using the second-order fractional-order equivalent circuit model and the typical CPSO algorithm structure characteristics. Since the discharge times of the two conditions are different, it is difficult to compare their characteristics with time as the abscissa. Therefore, the OCV-SOC curves of the two conditions are used for comparative analysis. Among them, Figure 6 (a) in Figure 6 represents the OCV-SOC identification results of the two conditions.

[0054] The open-circuit voltage is related to the operating conditions and algorithm conditions: (1) The operating conditions affect the identification result of the open-circuit voltage. Analysis Figure 5 shows that the OCV can vary with the fluctuation of the terminal voltage u t , but under different operating conditions, the OCV identification result will change. This is because in the Figure 1 circuit shown, due to the existence of the discharge current and the battery internal resistance, the OCV is converted into u t . Therefore, the open-circuit voltage is affected by the operating conditions and has a certain relationship with the terminal voltage conditions. In dynamic operating condition tests, the battery open-circuit voltage data cannot be directly obtained, so the influence of the battery open-circuit voltage by the operating conditions can be reflected by the change of the terminal voltage.

[0055] The fluctuation characteristics of the operating conditions are the main reason for affecting the local hysteresis characteristics of the OCV-SOC curve. Under a typical algorithm test scheme, two operating conditions obtained relatively consistent estimation results. Due to the different cycle periods and current distribution intensities of the DST operating condition and the FUDS operating condition, therefore Figure 6 the OCV-SOC identification results in (a) also show corresponding fluctuation periods. Figure 6 In (b), a part where the OCV-SOC of the two operating conditions differs greatly is found. The blue frame represents the open-circuit voltage difference at the same SOC, with a maximum of about 7 mV. The yellow frame represents the SOC difference at the same OCV, with a maximum difference of about 0.7. The differences at both places are within a reasonable range. Therefore, it can be considered that the OCV-SOC curves of the DST operating condition and the FUDS operating condition have high consistency. In addition, in Figure 6 (b), both OCV-SOC curves show obvious hysteresis characteristics. On the one hand, because the ampere-hour integration method is used to calculate the SOC of the battery during the test, due to the dynamic characteristics of the load current, when the current is negative, the SOC decreases; when the current is positive, the SOC increases. Therefore, the final SOC is related to the load current, that is, while decreasing as a whole, it shows periodic fluctuation characteristics locally. The influence of the load current fluctuation on the SOC is as Figure 7 shown. On the other hand, the OCV identification results also show fluctuation characteristics related to the algorithm and operating conditions. The above two aspects are the main reasons for the local hysteresis characteristics of the OCV-SOC curve under dynamic operating conditions.

[0056] The OCV-SOC curve will change with the change of temperature. As Figure 3 shows, as the SOC increases, the OCV of the battery monotonically increases from 2.7 V to 4.2 V. In the 20%-80% SOC range, the influence of temperature on the OCV-SOC curve is not obvious. At both ends of the SOC region, the influence of low temperature on the ternary lithium battery is more obvious. The OCV-SOC curve at 15 °C is significantly shifted compared with those at 25 °C and 35 °C.

[0057] (2) The open-circuit voltage identification result is related to the algorithm conditions, and the parameter identification window length affects the smoothness of the OCV identification result. The OCV identification result is related to the algorithm conditions. The length of the parameter identification window (PIW), as a key parameter of CPSO, affects the stability and consistency of the equivalent parameter identification result. Set the PIW with a length of 300 - 1200 to test the algorithm and discuss the influence of the PIW weakening condition on the open-circuit voltage. The identification results of the open-circuit voltage under different PIWs are as Figure 8 shown.

[0058] The longer the parameter identification window, the more it can suppress the open-circuit voltage fluctuation and make the open-circuit voltage result smoother. Figure 8 (a) in shows the open-circuit voltage identification results corresponding to different PIW lengths. When the PIW is 300 - 1200, the overall consistency of the open-circuit voltage is relatively high. It can be seen from the enlarged view that the longer the PIW, the smoother the open-circuit voltage identification result. Especially at the end of battery discharge, the shorter the PIW, the more obvious the open-circuit voltage fluctuation. Figure 8 (b) in is a partial enlarged view. To explore the influencing factors of the open-circuit voltage fluctuation, the terminal voltage u t is added. The results show that the fluctuation period of the open-circuit voltage is basically consistent with the operating cycle of the working condition. Especially when the PIW lengths are 300 and 600, at the places where the working condition changes violently, the open-circuit voltage fluctuation increases significantly. However, as the PIW becomes longer, the identification result of the open-circuit voltage almost becomes a uniform decrease, and the correlation with the working condition fluctuation weakens. That is, the longer the PIW, the more it can suppress the open-circuit voltage fluctuation caused by the change of the working condition.

[0059] To better analyze the relationship between the PIW length and the open-circuit voltage fluctuation, the Δu oc parameter is added to the parameter set identified by the CPSO algorithm. Δu oc is the change amount of the OCV between two adjacent sampling points in the parameter identification window.

[0060] Set the PIW with a length of 300 - 1200 and test the algorithm under the condition that other algorithm parameters remain the same to obtain the Δu oc parameter as Figure 9 shown. Figure 9 (a) in is the Δu oc parameter identified from the complete working condition test data under the PIW with a length of 300 - 1200. It can be seen from Figure 9 (a) that during the entire working condition test, the Δu oc value identified under the condition of PIW = 300 is basically the largest, which reflects that the OCV value identified under this condition fluctuates most obviously. Figure 9Figure (b) is a partial enlarged view during the mid-identification stage when the data identification result is relatively stable. Figure 9 In Figure (b), as the identification window length gradually increases, it can be seen that Δu oc The value gradually becomes smaller, indicating that as the identification window length increases, the change in OCV between two adjacent sampling points becomes smaller, that is, the fluctuation of the OCV identification result decreases. This is also consistent with Figure 8 The characteristic that the longer the window length, the smoother the OCV identification result. From Figure 9 Figure (b), it can also be seen that when PIW = 900, the value of Δu oc Is already relatively small, stably distributed around 0.1 mV. Compared with the Δu oc Values when PIW = 300 and 600, there has been a significant improvement. When PIW = 1200, the value of Δu oc Is stably lower than 0.1 mV and basically has no fluctuation, corresponding to Figure 8 At this time, the OCV identification result is already very smooth, meeting the requirements for constructing a smooth OCV-SOC characteristic.

[0061] According to the previous analysis of the original OCV identification result, at the end of discharge, the open-circuit voltage fluctuates most violently in the low SOC section. Therefore, it is necessary to analyze the change of Δu oc In different SOC intervals. Figure 10 Shows the change of Δu oc Corresponding to different SOC intervals. From Figure 10 Figure (a), it can be seen that in the low SOC section, when PIW = 300, the value of Δu oc Is relatively large, and the maximum value exceeds 0.5 mV, which is consistent with the characteristic that the OCV identification result fluctuates violently in the low SOC section at the end of discharge. The obvious fluctuation error in the identification process under this condition significantly reduces the accuracy of the open-circuit voltage value. However, as the window length increases, when PIW = 1200, the value of Δu oc In the low SOC section can still be stably distributed within 0.1 mV. This proves that even in the low SOC interval, the increase in the PIW length still has an obvious effect on suppressing the open-circuit voltage fluctuation caused by working condition changes. Figure 10 Figures (b) and (c) respectively show the change trends of Δu oc Corresponding to the middle and high SOC sections. Figure 10 It can be seen from Figures (b) and (c) that in the middle and high SOC sections, although the value of Δu oc Is generally lower than that in the low SOC section, the data value obtained when PIW = 300 is still much larger than the values obtained under other window conditions. And in the three SOC regions, the value of Δu oc Obtained under the condition of PIW = 1200 is not only always lower than 0.1 mV but also has been stably fluctuating.

[0062] By analyzing the identification results of OCV and Δu oc For the window length PIW, it can be seen that when its value is 900, the fluctuations of the identified OCV and Δu oc values have significantly decreased; when its value is 1200, the Δu oc value has been basically stable, and the identified OCV value is basically smooth as a whole at this time. The reason why a longer window length can suppress the open-circuit voltage fluctuation caused by the change of the working condition load current is that there are more identified data points in the longer window, the data information is more sufficient, and it is easier for the particles in the algorithm to maintain the consistency of the optimal particle position gbest during optimization. The optimal particle position affects the update of the identified model parameter values, so a longer window is more likely to obtain a parameter identification result with slower fluctuations and higher smoothness.

[0063] Use methods such as mean value and filtering to process the OCV data to eliminate errors and increase the smoothness of the OCV data. The CPSO algorithm is improved based on the standard particle swarm algorithm. Bionic optimization algorithms similar to PSO search for their optimal solutions among many parameter values. Restricted by the nonlinear characteristics of lithium batteries and the fitness function of the CPSO algorithm itself, the optimization results of the parameters are not completely consistent, but fluctuate within a reasonable range. The experimental tests were carried out 5 times on the identification results of the equivalent parameters of lithium batteries by the CPSO algorithm. There are certain errors in the OCV identification results of the 5 experiments, and the maximum difference is about within 0.5 mV. The identified results of OCV under the DST working condition show a continuous decrease.

[0064] Since the actual working state of the battery is very complex and will be affected by the surrounding environment. The design of the battery's SOC estimator needs to ensure the SOC estimation accuracy under more complex working conditions. Therefore, the accuracy of the OCV-SOC mapping relationship is crucial for the SOC estimation based on the look-up table method.

[0065] After averaging the open-circuit voltage data identified 5 times under the condition of the optimal parameter identification window length. Figure 12 In (a) is the partial enlarged view of the average value of OCV and the original value of the five tests. When the original data was tested, the CPSO algorithm used the terminal voltage fitting error as the fitness function. When each particle approaches the optimal solution, there is a probability of obtaining similar solutions. Therefore, the OCV identification results will have the same values. When calibrating the OCV-SOC mapping relationship, if there are the same OCV values, there will be a situation where one OCV value corresponds to multiple SOC values in the look-up table, which will increase the error. Therefore, it is necessary to further process the SOC data and OCV data to avoid the appearance of the same values in the constructed OCV-SOC mapping relationship while retaining the original data trend.

[0066] When processing the SOC data and the OCV data after averaging, the recursive average filtering method is adopted. The CPSO algorithm has high stability, and the identified OCV results rarely have large deviations. Therefore, the recursive average filtering method is used to process the OCV data. As Figure 11 shown, this method follows the first-in-first-out principle when reading data. A queue with a length of N is used as the OCV data memory. As the queue moves forward, the oldest data at the head of the queue is discarded, and the latest data is placed at the tail of the queue. In this way, there is always "the latest" data in the queue. When calculating the arithmetic mean, only the N OCV data in the queue need to be arithmetically averaged to obtain a new arithmetic mean value.

[0067] The load current in the DST working condition changes frequently. In order to avoid the hysteresis characteristic caused by current integration when calibrating the OCV-SOC mapping relationship, the SOC obtained by the ampere-hour integration method is filtered. When the battery discharges, the SOC of the battery continuously decreases. Since the tested dynamic working conditions all have periodic characteristics, it can be considered that the SOC in the DST working condition continuously and uniformly decreases with a certain fluctuation. When conducting dynamic working condition tests, the sampling frequency of the battery data is 100Hz. Since the sampling frequency is small, although there are fluctuations in the original SOC data, the overall trend is stable and the fluctuation value is small. Therefore, the recursive average filtering method, which can effectively suppress periodic interference, make the signal smoother while maintaining the basic shape characteristics of the signal, is used to process the OCV data and the SOC data. The selection of the filtering window length N is determined by the working condition characteristics. The cycle period of the DST working condition is 360 seconds, and the number of sampling data points in one cycle under the sampling condition of 0.5s is 720. Therefore, the number of data points in the filtering window should be selected as 720, that is, the same as the number of data points in a single cycle of the working condition. Figure 12 Figure (b) in shows the filtered SOC data.

[0068] The processed reference characteristics are of representative value. The OCV processing results are as Figure 13 shown, Figure 13 Figure (a) in shows that the processed OCV results maintain the dynamic trend of the original data. From Figure 13 the enlarged view in Figure (a) in , it can be seen that the recursive average filtering method can effectively suppress the periodic fluctuation of the OCV results caused by the periodic change of the dynamic working condition and increase the smoothness of the OCV data. Figure 13 Figure (b) in shows the error between the original data and the filtered data, which is basically within 0.5mV.

[0069] The processed OCV-SOC curve is as Figure 14As shown, it can be seen that the SOC range is 4% - 98%, and the OCV range is 3.2V - 4.2V, which basically covers the normal use range of ternary lithium batteries and has representative value.

[0070] The processed reference characteristics can improve the SOC estimation accuracy (applicability verification). The SOC estimated values of the processed OCV - SOC curve can accurately follow the true value of SOC under the DST and FUDS working conditions. The SOC estimation errors under the two working conditions are within 0.5% and 1% respectively, showing good accuracy and stability.

[0071] Under the DST working condition, the SOC estimation accuracy based on the reference characteristics is verified by the look - up table method. After verification, the SOC estimation has good stability and high accuracy. The estimation results are as Figure 15 shown. Figure 15 In (a), it represents the true value and estimated value of SOC, Figure 15 and in (b), it represents the error between the SOC estimated value and the true value. The results show that under the DST working condition, the SOC estimated value can accurately follow the true value of SOC. Since the mapping relationship of OCV - SOC in the look - up table is obtained by processing the OCV results identified under the DST working condition, the SOC estimation accuracy under the DST working condition is high, and the error is within 0.5%. In addition, when the SOC drops to about 10%, the error increases significantly. There are mainly two reasons: one is that at the end of the discharge of ternary lithium batteries, the internal polarization of the battery is relatively severe, resulting in a rapid change in the terminal voltage; the other is that during the OCV pre - processing, the error between the filtered OCV and the original value increases significantly, which also leads to an increase in the SOC estimation error at the end of discharge.

[0072] The OCV - SOC calibrated by the CPSO algorithm has good applicability and can be used under various working conditions. After verification, the SOC estimation error under the FUDS working condition is within 1%. Figure 16 It shows the SOC estimation results of the OCV look - up table identified by the CPSO algorithm under the FUDS working condition. Figure 16 In (a), it is the comparison between the SOC estimated value and the true value, Figure 16 and (b) is the SOC estimation error. From Figure 16 (a), it can be seen that under the FUDS working condition, the SOC estimated value can still accurately follow the true value of SOC. Figure 16 In (b), it represents the SOC estimation error under the FUDS working condition. The error is basically within 1% and shows a certain periodic characteristic, which is related to the periodic fluctuation of the true value of SOC obtained by integrating the load current.

[0073] Under FUDS conditions, the average error of SOC estimation in multiple tests does not exceed 0.3%, and the stability is good. To avoid contingency, the equivalent parameters of 5 FUDS conditions were identified by CPSO, and the SOC estimation values of 5 FUDS conditions were obtained according to the look-up table. As Figure 17 shown, Figure 17 (a) in it represents the SOC estimation values and true values of 5 FUDS conditions. It can be seen that the SOC estimation values of 5 FUDS conditions can accurately follow the true value of SOC. Table 2 lists the average error and root mean square error of 5 SOC estimations under FUDS conditions. The average error is less than 0.25%, and the root mean square error is less than 0.35%. Within a reasonable error range, the feasibility of the CPSO algorithm for OCV identification accuracy and consistency and look-up table estimation of SOC is verified.

[0074] Table 2 SOC Estimation Error under FUDS Conditions

[0075]

[0076] The SOC identification results have good convergence and consistency. The SOC estimation can quickly converge to the vicinity of the true value in any interval, and has good consistency and stability. In any discharge test interval, the convergence speed of the SOC estimation result is relatively fast, and the convergence error is within 1% within 10 s. To explore the convergence of SOC estimation under the conditions of algorithm restart or different initial values, three tests were carried out under the DST condition with SOC initial values of about 80% and 60% respectively. The results are as Figure 18 shown. Figure 18 (a) in it represents the SOC estimation result with an initial value of 80% under the DST condition. Figure 18 (b) in it represents the SOC estimation result with an initial value of 60%. It can be seen that the SOC estimation can quickly converge to the vicinity of the true value in any interval, and the consistency and stability of the SOC estimation are good.

[0077] The consistency of the OCV-SOC estimated by the CPSO algorithm and the OCV-SOC mapping relationship measured by the small current method is similar under the DST condition. Through the results of small current tests, the OCV is brought into the OCV-SOC calibrated by the CPSO algorithm for look-up table. The results are as Figure 19 shown. Among them Figure 19 (a) is the test comparison result, Figure 19 (b) is the error between the two.

[0078] Under two different OCV-SOC calibration systems, the maximum error between the two does not exceed 1%. The accuracy of the OCV identified by CPSO and the accuracy of the OCV-SOC curve calibrated using the identified OCV are verified. In addition, as one of the most basic methods for estimating the SOC of lithium batteries, the open-circuit voltage method is affected by the cyclic aging of the battery and requires continuous updating of the OCV-SOC mapping relationship. The CPSO algorithm can identify parameters such as OCV in real time during the operation of the battery, and by filtering the identified OCV results, an updated OCV-SOC curve can be obtained. Therefore, the OCV-SOC mapping relationship can also be understood as being updated in real time, which can effectively improve problems such as the decrease in SOC estimation accuracy caused by the change of the OCV-SOC mapping relationship affected by battery aging.

[0079] The present invention studies a method for online estimating the state of charge of a lithium battery under working conditions, which is applicable to the method for estimating the state of charge of a lithium battery by obtaining OCV-SOC characteristics based on a fractional-order equivalent circuit model and a co-evolutionary particle swarm optimization algorithm, and belongs to the field of battery equivalent parameter identification. This method is applicable to obtaining an accurate open-circuit voltage to improve the SOC estimation accuracy of the look-up table method. It solves the problems that the identification result of the battery open-circuit voltage is often related to the working conditions, resulting in inconsistent and fluctuating identification results, and the low accuracy of the SOC estimation result due to the hysteresis characteristic of the original OCV-SOC characteristic. Based on the CPSO algorithm, the correlation between the identification result of the open-circuit voltage of a lithium battery and the working conditions and the weakening method are studied. According to this method, the smoothness of the open-circuit voltage identification result is enhanced, and in addition, the hysteresis characteristic of the SOC generated by integrating the dynamic current is linearized. An updatable OCV-SOC mapping relationship is established based on the processing results of the open-circuit voltage and SOC, and the accuracy of this mapping relationship for estimating the SOC of a lithium battery is verified under different conditions. The results show that:

[0080] (1) The dynamic working conditions have a clear impact on the open-circuit voltage value and the SOC value. The open-circuit voltage identified by the CPSO algorithm shows obvious fluctuation characteristics related to the working condition cycle period. The longer the parameter identification window length of the CPSO algorithm, the weaker the fluctuation characteristics and the higher the smoothness of the open-circuit voltage identification result.

[0081] (2) The hysteresis characteristic in OCV-SOC is eliminated, and a smooth and updatable OCV-SOC reference characteristic with relatively high accuracy is established. The SOC estimation error under two working conditions is within 1%.

[0082] (3) The convergence of SOC estimation is consistent with the convergence of the OCV identification result. Using the reference characteristic to estimate the SOC by the look-up table method, the estimation results have good consistency, a fast convergence speed, and do not depend on the initial value.

[0083] Certainly, the above description is not a limitation of the present invention, and the present invention is not limited to the above examples. Changes, modifications, additions, or substitutions made by those skilled in the art within the scope of the essence of the present invention should also fall within the protection scope of the present invention.

Claims

1. A method for online estimating the state of charge of a lithium battery under a working condition, characterized in that It includes the following steps: Step 1: Use the experimental platform to conduct characteristic tests on the ternary lithium battery under different working conditions to obtain test data; Step 2: Establish a second-order fractional-order equivalent circuit model, input the test data into the second-order fractional-order equivalent circuit model, and use the co-evolutionary particle swarm optimization algorithm for on-line identification of the open-circuit voltage and parameter Δu of the second-order fractional-order equivalent circuit model oc ; The parameter Δu oc is the change in the open-circuit voltage between two adjacent sampling points in the parameter identification window. The parameter Δu oc 's numerical change can directly reflect the fluctuation of the open-circuit voltage identification result; Step 3: Obtain the parameter Δu by adjusting the length of the parameter identification window in the co-evolutionary particle swarm optimization algorithm oc The length of the parameter identification window with a value lower than 0.1 mv and no fluctuations is the optimal parameter identification window length. The optimal parameter identification window length is 1200. Under the condition of the optimal parameter identification window length, identify the open-circuit voltage data from the working condition test data, and obtain the state of charge data through the ampere-hour integration method; Step 4: Process the open-circuit voltage data using the mean value and filtering method to eliminate errors; in order to avoid the random errors caused by the identification results of the open-circuit voltage data, the open-circuit voltage data identified 5 times under the condition of the optimal parameter identification window length are averaged and then filtered using the recursive average filtering method, and the recursive average filtering method is used to filter the state of charge data obtained by the ampere-hour integration method; Step 5: The open-circuit voltage data and the state of charge data processed in Step 4 form an OCV-SOC curve, and the state of charge data corresponding to the open-circuit voltage is obtained through the look-up table method under the working conditions to complete the estimation of the lithium battery's power.

2. The method for online estimating the state of charge of a lithium battery under a working condition according to claim 1, characterized in that, The working conditions are the FUDS condition and the DST condition. The FUDS condition is the Federal Urban Driving Test condition, and the DST condition is the Dynamic Stress Test condition.

3. The method for online estimating the state of charge of a lithium battery under a working condition according to claim 1, wherein The second-order fractional-order equivalent circuit model replaces the capacitor in the integer-order equivalent circuit model with two constant phase elements.

Images