A simulation verification method and system of a battery SOC algorithm

Abstract

The application discloses a kind of simulation verification method and system of battery SOC algorithm, method includes: using SOC algorithm obtains fΔt SOC estimated value under Δt time range;Based on battery simulation model, fΔt SOC reference value under Δt is obtained;Bootstrap method is used to fΔt SOC estimated value is sampled with replacement n times, and the confidence interval of the preset confidence level of SOC estimated value at t time is constructed;The moving average of SOC reference value at t time is calculated, if the moving average is located in the confidence interval at t time, then consider that SOC estimated value at t time meets the accuracy requirement under the preset confidence level.This application can simulate long-term working condition using battery simulation model, and quickly evaluate the accuracy of SOC algorithm results, make up the defects that various algorithms must follow battery pack long-term test before evaluating accuracy, and ensure the accuracy and correctness of fast simulation evaluation.

Description

Technical Field

[0001] This invention relates to the field of battery management technology, and more specifically to a simulation verification method and system for a battery SOC algorithm. Background Technology

[0002] Currently, several methods exist for verifying the accuracy of SOC estimation algorithms. QC / T897, "Technical Conditions for Battery Management Systems for Electric Vehicles," proposes a method for testing the SOC accuracy of a battery management system. This method can test the accuracy of the SOC estimation algorithm using various methods within different SOC ranges. However, the lengthy testing process limits its application, and this method cannot verify accuracy when external conditions change. Invention patent CN107422289A discloses a method and system for verifying the SOC estimation accuracy of a battery pack in electric vehicles. However, this method obtains the true SOC value by charging the battery pack to full capacity after discharging and recording the charging amount to calculate the true SOC value. Since the battery pack can still charge to a certain extent after reaching the cutoff voltage, there is a lack of a standard for determining whether the battery pack is fully charged, making it difficult to obtain the true SOC value. Furthermore, this method can only verify SOC accuracy during discharge, lacking the requirement to verify SOC accuracy during charging. Invention patent CN103713264A discloses a battery management system SOC estimation accuracy testing system and method. This method uses a single-cell voltage simulation source, a battery pack total voltage and current simulation source, and a battery box temperature simulation source to simulate the battery pack. The simulated calculation data is then transmitted to the battery management system for SOC estimation accuracy testing. However, this patent is vague about the working principle and process of the battery simulator; the estimation data comes from the battery simulator, and the simulator can only partially simulate the battery's operating state. CN109061477B discloses a verification and evaluation method and device for battery SOC algorithms, capable of evaluating various SOC estimation algorithms based on memory space occupied, computation time, and algorithm accuracy. However, this method still requires evaluation during actual battery use, which is time-consuming.

[0003] The methods mentioned above cannot quickly evaluate the SOC estimation algorithm. Therefore, how to provide a method and system that can calculate the SOC estimate of a single lithium battery cell or battery pack and quickly verify the accuracy of the estimation algorithm based on the estimation results is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0004] In view of this, the present invention provides a simulation verification method and system for battery SOC algorithm, which can simulate the battery's operating conditions and reliability, evaluate the accuracy of various SOC estimation algorithms based on the simulation platform, make up for the shortcomings of existing algorithms in that they cannot perform rapid evaluation, ensure the correctness of rapid and accurate evaluation, and point out the parts of the algorithm that need to be optimized.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] A simulation verification method for a battery SOC algorithm includes the following steps:

[0007] S1. Input the SOC algorithm, which can estimate the SOC of the battery at a frequency of f within the simulation time range from t-Δt to t, and obtain fΔt SOC estimates.

[0008] S2. Within the simulation time range from t-Δt to t, the voltage, current and temperature information of the battery are input into the battery simulation model at a frequency of f. Based on the input battery voltage, current and temperature, and combined with the charge and discharge cycle number and initial capacity information, the battery simulation model calculates fΔt SOC values ​​within the simulation time range from t-Δt to t, which are used as SOC reference values.

[0009] S3. Use the Bootstrap method to perform n resampling with replacement on fΔt SOC estimates to construct the confidence interval of the SOC estimate at time t within the preset confidence level.

[0010] S4. Calculate the moving average of the SOC reference value at time t in the simulation. If the moving average is within the confidence interval at time t, then the estimated SOC value at time t in the simulation is considered to meet the accuracy requirements under the preset confidence level.

[0011] S5. As the simulation progresses, confidence intervals of the SOC estimates at different times are continuously constructed according to S1-S4, based on the preset confidence levels. The online determination is made as to whether the SOC estimates at different times meet the accuracy requirements under the preset confidence levels.

[0012] Furthermore, the simulation verification method for the aforementioned battery SOC algorithm also includes:

[0013] S6. In at least five charge-discharge cycles, if the moving average of the SOC reference value at a certain moment is outside the confidence interval at that moment, record the battery state, including at least SOC, current and voltage; according to the SOC algorithm, the corresponding battery state information needs to be optimized, find the overlapping area of ​​battery state, and determine in which battery state (full charge, low charge or random) the SOC algorithm is not accurate enough and needs to be optimized.

[0014] Preferably, in the above-mentioned method for verifying the simulation accuracy of a battery SOC algorithm, step S3 includes:

[0015] The Bootstrap method is used to perform n resamplings with replacement on fΔt SOC estimates to obtain n SOC estimate samples; where n≥1000.

[0016] Calculate the mean of each SOC estimate sample and arrange the means in ascending order;

[0017] Within the sorted mean, based on the preset confidence level 1-α, the nα / 2th number and the n(1-α / 2th)th number are calculated and used as the upper and lower limits of the interval, respectively. This interval is then used as the confidence interval for the preset confidence level 1-α.

[0018] Optionally, in the above-mentioned simulation verification method for battery SOC algorithm, the preset confidence level is not less than 80%.

[0019] Preferably, in the above-mentioned method for verifying the simulation accuracy of a battery SOC algorithm, the preset confidence level is 90%.

[0020] Preferably, in the above-mentioned method for verifying the simulation accuracy of a battery SOC algorithm, the calculation length of the moving average of the SOC reference value is Δt.

[0021] This invention also provides a simulation verification system for a battery SOC algorithm, comprising:

[0022] The SOC algorithm module is used to input the SOC algorithm and estimate the SOC of the battery at a frequency of f within the simulation time range from t-Δt to t, obtaining fΔt SOC estimates.

[0023] The simulation module is used to collect battery voltage, current and temperature data at a sampling frequency of f within the simulation time range from t-Δt to t and input them into the battery simulation model. Based on the input battery voltage, current and temperature, and combined with the number of charge and discharge cycles and initial capacity information, the battery simulation model calculates fΔt SOC values ​​within the simulation time range from t-Δt to t, which are used as SOC reference values.

[0024] The confidence interval construction module is used to construct the confidence interval of the SOC estimate at time t by performing n resamplings with replacement on fΔt SOC estimates using the Bootstrap method; and

[0025] The accuracy judgment module is used to calculate the moving average of the SOC reference value at time t. If the moving average is within the confidence interval of the simulation at time t, the estimated SOC value at the simulation at time t is considered to meet the accuracy requirements under the preset confidence level.

[0026] Preferably, the simulation accuracy verification system for the above-mentioned battery SOC algorithm further includes:

[0027] The algorithm optimization and judgment module is used to record the battery state, including at least SOC, current and voltage, when the moving average of the SOC reference value at a certain moment is outside the confidence interval at that moment during at least five charge-discharge cycles. Based on the battery state information that the SOC algorithm needs to optimize, it finds the overlapping area of ​​the battery state and determines the SOC algorithm's accuracy is insufficient and needs optimization in which battery state (full charge, low charge or random).

[0028] As can be seen from the above technical solution, compared with the prior art, this invention discloses a simulation verification method and system for battery SOC algorithm. For SOC, a quantity that cannot be directly measured, it is calculated through a mathematical model, and the accuracy of this calculation also needs to be evaluated using mathematical methods. This invention can quickly verify the estimation accuracy of the battery SOC algorithm, and when there is a deviation in the SOC estimate, an optimization algorithm can be used to correct the SOC estimate. The SOC estimation verification method is computationally simple and does not involve complex calculations such as differentiation and integration. Attached Figure Description

[0029] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0030] Figure 1 The attached figure is a flowchart of a simulation verification method for a battery SOC algorithm provided by the present invention;

[0031] Figure 2 The attached figure is a structural block diagram of a simulation verification system for a battery SOC algorithm provided by the present invention. Detailed Implementation

[0032] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0033] like Figure 1 As shown in the figure, this invention discloses a simulation verification method for a battery SOC algorithm, including the following steps:

[0034] S1. Input the SOC algorithm. Within the simulation time range from t-Δt to t, estimate the SOC of the battery at a frequency of f to obtain fΔt estimated SOC values.

[0035] S2. Within the simulation time range from t-Δt to t, the voltage, current and temperature of the battery are input into the battery simulation model at a frequency of f. Based on the input battery voltage, current and temperature, and combined with information such as the number of charge and discharge cycles and initial capacity, the battery simulation model calculates fΔt SOC values ​​within the simulation time range from t-Δt to t, which are used as SOC reference values.

[0036] S3. Use the Bootstrap method to perform n resampling with replacement on fΔt SOC estimates to construct the confidence interval of the SOC estimate at time t within the preset confidence level.

[0037] S4. Calculate the moving average of the SOC reference value at time t in the simulation. If the moving average is within the confidence interval at time t, then the estimated SOC value at time t in the simulation is considered to meet the accuracy requirements under the preset confidence level.

[0038] S5. As the simulation progresses, confidence intervals of the SOC estimates at different times are continuously constructed according to S1-S4, based on the preset confidence levels. The online determination is made as to whether the SOC estimates at different times meet the accuracy requirements under the preset confidence levels.

[0039] In other embodiments, it also includes:

[0040] S6. In at least five charge-discharge cycles, if the moving average of the SOC reference value at a certain moment is outside the confidence interval at that moment, record the battery state, including at least SOC, current and voltage. Based on the SOC algorithm, optimize the corresponding battery state information, find the overlapping area of ​​battery state, and determine in which battery state (full charge, low charge or random) the SOC algorithm is not accurate enough and needs to be optimized.

[0041] The bootstrap method is a simulated sampling statistical inference method based on original data. It is commonly used for interval estimation and hypothesis testing problems where parameters are difficult to calculate using conventional methods. The basic procedure involves resampling with replacement within the range of the original data, maintaining the same sample size, and ensuring that each observation in the original data is selected with equal probability each time. The resulting sample is called the bootstrap sample, and the statistic α' is calculated from the bootstrap sample. The number of resampling iterations, m, is typically greater than 1000, resulting in m bootstrap samples and m corresponding statistics α1', α2', ... α mThe distribution constructed using these m statistics approximates the distribution of the original sample statistic α, thereby achieving interval estimation of the statistic.

[0042] In one embodiment, S2 includes:

[0043] The battery simulation model simulates the battery's operating conditions based on the battery voltage, current, and temperature data input from S2, combined with the battery's current charge-discharge cycle count and initial capacity at the start of operation.

[0044] During the simulation of battery usage conditions, the SOC value is calculated at the same frequency f within the simulation time range from t-Δt to t, and output as the SOC reference value.

[0045] In one embodiment, S3 includes:

[0046] The Bootstrap method is used to perform n resamplings with replacement on fΔt SOC estimates to obtain n SOC estimate samples; where n≥1000.

[0047] Calculate the mean of each SOC estimate sample and arrange the means in ascending order;

[0048] Within the sorted mean, based on the pre-set confidence level 1-α, calculate the n(nα / 2)th number and the n(n(1-α / 2)th number), which will be used as the upper and lower limits of the interval, respectively. This interval will then be used as the confidence interval for the pre-set confidence level 1-α. The confidence level 1-α typically needs to be at least 80%.

[0049] The above method will be further illustrated below with a specific embodiment.

[0050] The battery management system collects battery voltage, current, and temperature data at a preset sampling frequency and inputs them into the battery simulation model. An algorithm for calculating the battery's State of Charge (SOC) is used to estimate the SOC at a frequency of 1Hz within a simulation time range of 0 to 5 seconds, yielding five estimated SOC values. Based on the voltage, current, and temperature information, the battery simulation model calculates five reference SOC values, as shown in Table 1.

[0051] Table 1

[0052] Time (s) SOC estimate SOC reference value 1 60.1％ 60.1％ 2 60.1％ 60.0％ 3 59.8％ 60.0％ 4 59.9％ 59.9％ 5 59.8％ 59.9％

[0053] Using the Bootstrap method, 1000 resamplings with replacement were performed on the 5 SOC estimates to obtain 1000 SOC estimate samples. The mean SOC estimate SOC1, SOC2, ..., SOC2 was calculated for each sample. nThen, arrange the means from smallest to largest to obtain SOC*1≤SOC*2≤...≤SOC* n Within the sorted mean, at a 90% confidence level, the 50th and 950th numbers are calculated as the upper and lower limits of the interval, yielding 59.8% and 60.0%, respectively. [59.84%, 60.04%] represents the 90% approximate confidence interval for the SOC estimate. The moving average of the SOC reference value is then calculated to be 59.98%. Since this moving average falls within the confidence interval, the accuracy of the 59.8% SOC estimate at time 5s is considered satisfactory.

[0054] As the simulation time is shifted forward, following the method described above, the moving average of the SOC reference value with a length of 5 seconds is calculated, as shown in Table 2.

[0055] Table 2

[0056]

[0057] As shown in Table 2, the moving average of the reference SOC value at each time point can be included within the 90% approximate confidence interval of the SOC estimate. Therefore, the SOC estimate has an approximate accuracy of 90%.

[0058] The simulation time can be set to any length, such as several hours, days, or months. The computation time of the above program after multiple runs is far less than the actual simulation time, thus fully meeting the requirements for rapid SOC estimation.

[0059] like Figure 2 As shown, this embodiment of the invention also provides a simulation verification system for a battery SOC algorithm, comprising:

[0060] The SOC algorithm input module is used to input the SOC algorithm, which estimates the SOC of the battery at a frequency of f within the time range from t-Δt to t, and obtains fΔt SOC estimates.

[0061] The simulation module is used to collect battery voltage, current and temperature at a sampling frequency of f within the time range from t-Δt to t and input them into the battery simulation model. Based on the input battery voltage, current and temperature, and combined with information such as the number of charge and discharge cycles and initial capacity, the battery simulation model calculates fΔt SOC values ​​within the simulation time range from t-Δt to t, which are used as SOC reference values.

[0062] The confidence interval construction module is used to construct the confidence interval of the SOC estimate at time t by performing n resamplings with replacement on fΔt SOC estimates using the Bootstrap method;

[0063] The accuracy judgment module is used to calculate the moving average of the SOC reference value at time t. If the moving average is within the confidence interval at time t, the estimated SOC value at time t is considered to meet the accuracy requirements under the preset confidence level.

[0064] In other embodiments, it also includes:

[0065] The algorithm optimization and determination module is used to record the battery state at the time of offset, including information such as SOC, voltage and current, if the moving average of the SOC reference value at a certain moment is outside the confidence interval at that moment during at least five charge-discharge cycles.

[0066] The SOC algorithm needs to optimize the corresponding battery state information, find the overlapping areas of battery states, and determine in which battery states (full charge, low charge, or random) the SOC algorithm's accuracy is insufficient and needs optimization.

[0067] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.

[0068] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A simulation verification method of a battery SOC algorithm, characterized in that, The method comprises the following steps: S1, input SOC algorithm, in the simulation time range from t to t, the SOC of the battery is estimated at the frequency of f, and a SOC estimation value is obtained; S2, in In the simulation time range from t to t, the voltage, current and temperature used by the battery are input into the battery simulation model at a frequency of f, and the battery simulation model simulates the use condition of the battery according to the input battery voltage, current and temperature, and in combination with the current charge and discharge cycle number and initial capacity information of the battery; in the simulation process of simulating the use condition of the battery, In the simulation time range from t to t, the voltage, current and temperature used by the battery are input into the battery simulation model at a frequency of f, and the battery simulation model simulates the use condition of the battery according to the input battery voltage, current and temperature, and in combination with the current charge and discharge cycle number and initial capacity information of the battery; in the simulation process of simulating the use condition of the battery, The number of SOC values is calculated as the SOC reference value. S3. Using the Bootstrap method The SOC estimates are resampled n times with replacement to construct a confidence interval for the SOC estimates at time t with a pre-set confidence level. S4, calculating a moving average of the SOC reference value at the simulation time t, and if the moving average is within the confidence interval at the simulation time t, it is considered that the SOC estimation value at the simulation time t meets the accuracy requirement at the preset confidence level; S5, with the lapse of simulation time, the confidence interval of the SOC estimation value at different times at the preset confidence level is continuously constructed according to S1-S4, and whether the SOC estimation value at different times meets the accuracy requirement at the preset confidence level is judged; S6, in at least five charge-discharge cycles, if the moving average of the SOC reference value at a certain time is outside the confidence interval at the time, the battery state is recorded, at least including SOC, current and voltage; According to the SOC algorithm, the corresponding battery state information needs to be optimized to find the overlapping area of the battery state, and it is judged that the accuracy of the SOC algorithm is not enough under the state of the battery and needs to be optimized.

2. The method of claim 1, wherein the SOC algorithm is a battery SOC algorithm. The S3 comprises: Using the Bootstrap method Each SOC estimate is resampled n times with replacement to obtain n SOC estimate samples; where n≥1000. calculating the mean value of each SOC estimation value sample, and arranging the mean values in ascending order; Within the sorted mean, the pre-set confidence level is obtained , the first number and the first number are obtained, respectively as the upper and lower limits of the interval, and the interval is taken as the confidence interval of the pre-set confidence level .

3. The method of claim 1, wherein the SOC algorithm is a battery SOC algorithm. The preset confidence level is not less than 80%.

4. The method of claim 3, wherein the SOC algorithm is verified by: The preset confidence level is 90%.

5. The method of claim 1, wherein the SOC algorithm is a battery SOC algorithm. The calculation length of the moving average of the SOC reference value is .

6. A simulation verification system of a battery SOC algorithm, which is suitable for the simulation verification method of a battery SOC algorithm according to any one of claims 1-5, characterized in that, It comprises: The SOC algorithm module is configured to input an SOC algorithm, and estimate the SOC of the battery at a frequency of f in a simulation time range from t to t, and obtain an SOC estimation value. a simulation module for inputting the battery voltage, the current, and the temperature, which are input at a frequency of f, into the battery simulation model in a simulation time range from t to t+1, and calculating the SOC value as the SOC reference value in the simulation time range from t to t+1, according to the input battery voltage, the current, and the temperature, in combination with the number of charge and discharge cycles and the initial capacity information; and the SOC value calculated as the SOC reference value in the simulation time range from t to t+1, according to the input battery voltage, the current, and the temperature, in combination with the number of charge and discharge cycles and the initial capacity information.​​ The confidence interval construction module is configured to perform n times of resampling with replacement on the SOC estimation value by using a Bootstrap method to construct a confidence interval of the SOC estimation value at the preset confidence level at the t time point. The confidence interval construction module is configured to perform n times of resampling with replacement on the SOC estimation value by using a Bootstrap method to construct a confidence interval of the SOC estimation value at the preset confidence level at the t time point. And The accuracy judgment module is used for calculating the moving average of the SOC reference value at the simulation time t, and if the moving average is within the confidence interval at the simulation time t, it is considered that the SOC estimation value at the simulation time t meets the accuracy requirement at the preset confidence level.

7. The simulation verification system of a battery SOC algorithm according to claim 6, characterized in that, It also comprises: The algorithm optimization judgment module is used for recording the battery state including SOC, current and voltage in at least five charge-discharge cycles if the moving average of the SOC reference value at a certain time is outside the confidence interval at the time, and according to the SOC algorithm, the corresponding battery state information needs to be optimized to find the overlapping area of the battery state, and it is judged that the accuracy of the SOC algorithm is not enough under the state of the battery and needs to be optimized.

Images