CN105203963A

CN105203963A - Charge state estimation method based on open-circuit voltage hysteretic characteristics

Info

Publication number: CN105203963A
Application number: CN201510578190.XA
Authority: CN
Inventors: 戴海峰; 魏学哲; 朱乐涛; 孙泽昌
Original assignee: Tongji University
Current assignee: Tongji University
Priority date: 2015-09-11
Filing date: 2015-09-11
Publication date: 2015-12-30
Anticipated expiration: 2035-09-11
Also published as: CN105203963B

Abstract

The invention relates to a charge state estimation method based on open-circuit voltage hysteretic characteristics. The method is used for estimating the charge state of a lithium ion battery online. The method comprises the following steps that 1, a hysteretic characteristic curve of open-circuit voltage and the charge state of the lithium ion battery is obtained offline; 2, initial parameters of an open-circuit voltage hysteretic characteristic self-adaptive model based on a Preisach operator is determined through training according to the hysteretic characteristic curve, and the open-circuit voltage hysteretic characteristic self-adaptive model based on the Preisach operator is established; 3, the charge state of the lithium ion battery is estimated online according to the open-circuit voltage hysteretic characteristic self-adaptive model based on the Preisach operator, and the charge state of the lithium ion battery at the present moment is obtained. Compared with the prior art, the charge state estimation method based on the open-circuit voltage hysteretic characteristics has the advantages of being accurate in modeling, capable of improving the accuracy, accurate in estimation and the like.

Description

An Estimation Method of State of Charge Based on Hysteresis Characteristics of Open Circuit Voltage

技术领域technical field

本发明涉及一种荷电状态的估计方法，尤其是涉及一种基于开路电压滞回特性的荷电状态的估计方法。The invention relates to a method for estimating the state of charge, in particular to a method for estimating the state of charge based on the hysteresis characteristic of the open circuit voltage.

背景技术Background technique

动力电池系统作为关键的零部件在电动汽车和电力储能等领域得到越来越多的应用。在应用过程中，需要电池管理系统(BatteryManagementSystem，BMS)对电池状态进行监控，防止过充过放延长电池使用寿命。在这其中，SOC(荷电状态)的准确估计尤为关键。大多数的SOC估计方法是利用SOC与开路电压OCV(OpenCircuitVoltage，开路电压)的对应关系得到，如开路电压法，基于模型的SOC估计方法等。对OCV与SOC对应关系的描述是这些SOC估计方法的核心基础。锂离子电池中开路电压和OCV并不完全一一对应，而是存在滞回关系(同一SOC下，充电过程的OCV大于放电过程的OCV)。As a key component, the power battery system has been used more and more in the fields of electric vehicles and electric energy storage. In the application process, a battery management system (Battery Management System, BMS) is required to monitor the battery status to prevent overcharging and over-discharging to prolong the service life of the battery. Among them, the accurate estimation of SOC (state of charge) is particularly critical. Most of the SOC estimation methods are obtained by using the corresponding relationship between the SOC and the open circuit voltage OCV (Open Circuit Voltage, open circuit voltage), such as the open circuit voltage method, the SOC estimation method based on the model, and the like. The description of the corresponding relationship between OCV and SOC is the core basis of these SOC estimation methods. In lithium-ion batteries, the open circuit voltage and OCV are not exactly one-to-one correspondence, but there is a hysteresis relationship (under the same SOC, the OCV in the charging process is greater than the OCV in the discharging process).

传统的锂离子电池开路电压滞回特性的建模方法中引入了较多的简化，使得滞回模型的建模精度低，从而影响SOC估计。本文采用的基于Preisach算子的滞回模型建模方法，在镍氢电池(NiMH)的开路电压滞回建模中已经应用，但由于锂离子电池相比NiMH电池开路电压滞回关系并不显著并且也无对称性，使得离散Preisach模型无法较好应用到锂离子电池中。More simplifications are introduced in the traditional modeling method of the hysteresis characteristics of the open circuit voltage of lithium-ion batteries, which makes the modeling accuracy of the hysteresis model low, thus affecting the SOC estimation. The hysteresis model modeling method based on the Preisach operator used in this paper has been applied in the hysteresis modeling of the open circuit voltage of NiMH batteries, but the hysteresis relationship of the open circuit voltage of lithium-ion batteries is not significant compared with that of NiMH batteries. And there is no symmetry, so that the discrete Preisach model cannot be well applied to lithium-ion batteries.

发明内容Contents of the invention

本发明的目的就是为了克服上述现有技术存在的缺陷而提供一种建模准确、提高精度、估计准确的基于开路电压滞回特性的荷电状态的估计方法。The object of the present invention is to provide a method for estimating the state of charge based on the hysteresis characteristic of the open circuit voltage with accurate modeling, improved precision and accurate estimation in order to overcome the above-mentioned defects in the prior art.

本发明的目的可以通过以下技术方案来实现：The purpose of the present invention can be achieved through the following technical solutions:

一种基于开路电压滞回特性的荷电状态的估计方法，用于在线估计锂离子电池的荷电状态，包括以下步骤：A method for estimating the state of charge based on the hysteresis characteristic of the open circuit voltage is used to estimate the state of charge of the lithium-ion battery online, comprising the following steps:

1)离线获取锂离子电池开路电压和荷电状态的滞回特性曲线；1) Obtain the hysteresis characteristic curve of the open circuit voltage and state of charge of the lithium-ion battery offline;

2)根据滞回特性曲线训练确定基于Preisach算子的开路电压滞回特性自适应模型的初始参数，并建立基于Preisach算子的开路电压滞回特性自适应模型；2) Determine the initial parameters of the hysteresis characteristic adaptive model of the open circuit voltage based on the Preisach operator according to the hysteresis characteristic curve training, and establish the adaptive model of the hysteresis characteristic of the open circuit voltage based on the Preisach operator;

3)根据基于Preisach算子的开路电压滞回特性自适应模型在线估计锂离子电池的荷电状态，得到当前时刻锂离子电池的荷电状态。3) The state of charge of the lithium-ion battery is estimated online according to the adaptive model of the open circuit voltage hysteresis characteristic based on the Preisach operator, and the state of charge of the lithium-ion battery at the current moment is obtained.

所述的步骤2)中的基于Preisach算子的开路电压滞回特性自适应模型的初始参数包括Preisach三角形的网格划分数目N和权重向量的初始值μ。The initial parameters of the Preisach operator-based open-circuit voltage hysteresis characteristic adaptive model in step 2) include the grid division number N of the Preisach triangle and the initial value μ of the weight vector.

所述的步骤2)中基于Preisach算子的开路电压滞回特性自适应模型为：In the described step 2), the hysteresis characteristic adaptive model of the open circuit voltage based on the Preisach operator is:

${SOC SOC}_{{t t}_{k k}} (({t t}_{k k})) = = {ω ω}^{T T} (({t t}_{k k})) μ μ (({t t}_{k k}))$

其中，为t_k时刻开路电压对应的荷电状态值，ω(t_k)为t_k时刻开路电压值在Preisach三角形中对应的滞回状态向量，μ(t_k)为t_k时刻Preisach三角形中所有网格的滞回权重向量。in, is the state of charge value corresponding to the open circuit voltage at time t _k , ω(t _k ) is the hysteresis state vector corresponding to the open circuit voltage value at time t _k in the Preisach triangle, μ(t _k ) is all nets in the Preisach triangle at time t _k The hysteretic weight vector for the lattice.

所述的步骤3)具体包括以下步骤：Described step 3) specifically comprises the following steps:

31)在线获取上一时刻t_k-1的Preisach三角形网格中的权重向量μ(t_k-1)和当前时刻t_k的滞回状态值ω(t_k)，计算得到当前时刻的先验荷电状态值为：31) Obtain online the weight vector μ(t _k-1 ) in the Preisach triangle grid at the previous moment t _k-1 and the hysteresis state value ω(t _k ) at the current moment t _k , and calculate the prior State of charge for:

${SOC SOC}_{{t t}_{k k - - 11}} (({t t}_{k k})) = = {ω ω}^{T T} (({t t}_{k k})) μ μ (({t t}_{k k - - 11}));;$

32)根据电池容量Q₀、上一时刻t_k-1的荷电状态值和当前时刻t_k的先验荷电状态值计算得到当前时刻的电流估计值I_cal(t_k)为：32) According to the battery capacity Q ₀ , the state of charge value at the last moment t _k-1 and the prior state of charge value at the current moment t _k Calculate the current estimated value I _cal (t _k ) at the current moment as:

${I I}_{c c a a l l} (({t t}_{k k})) = = {Q Q}_{00} [[{SOC SOC}_{{t t}_{k k - - 11}} (({t t}_{k k})) - - {SOC SOC}_{{t t}_{k k - - 11}} (({t t}_{k k - - 11}))]] / / Δ Δ t t$

Δt＝t_k-t_k-1；Δt ₌ tk _-tk-1 ;

33)根据实测的当前时刻的电流实际值I_m(t_k)和电流估计值I_cal(t_k)得到当前时刻的电流误差值η_current(t_k)为：33) According to the measured current actual value I _m (t _k ) and current estimated value I _cal (t _k ), the current error value η_current(t _k ) at the current moment is obtained as:

η_current(t_k)＝I_m(t_k)-I_cal(t_k)；η_current(t _k )=I _m (t _k )-I _cal (t _k );

34)根据当前时刻的电流误差值η_current(t_k)和当前时刻的先验荷电状态值并采用最小均方误差法得到当前时刻权重向量增量并计算当前时刻的权重值μ(t_k)为：34) According to the current error value η_current(t _k ) at the current moment and the prior state of charge value at the current moment And use the minimum mean square error method to get the weight vector increment at the current moment And calculate the weight value μ(t _k ) at the current moment as:

μ(t_k)＝μ(t_k-1)+λη(t_k)ω(t_k)μ(t _k )=μ(t _k-1 )+λη(t _k )ω(t _k )

$η η (({t t}_{k k})) = = \frac{Δ Δ t t}{{Q Q}_{00} η η__c c u u r r r r e e n no t t (({t t}_{k k}))}$

其中λ为步长因子，且λ∈[0,1]；Where λ is the step factor, and λ∈[0,1];

35)根据当前时刻的权重向量μ(t_k-1)和滞回状态值ω(t_k)通过开路电压滞回特性自适应模型得到当前时刻的后验荷电状态值即当前时刻的锂离子电池的荷电状态，并返回步骤31)进行下一时刻的锂离子电池的荷电状态估计。35) According to the weight vector μ(t _k-1 ) and the hysteresis state value ω(t _k ) at the current moment, the posterior state of charge value at the current moment is obtained through the adaptive model of the hysteresis characteristic of the open circuit voltage That is, the state of charge of the lithium-ion battery at the current moment, and return to step 31) to estimate the state of charge of the lithium-ion battery at the next moment.

所述的步骤34)中，当前时刻权重向量增量的计算式为：In the described step 34), the weight vector increment at the current moment The calculation formula is:

${Δμ Δμ}_{{t t}_{k k}} = = λ λ η η (({t t}_{k k})) ω ω (({t t}_{k k}))$

其中，λ为步长因子，且λ∈[0,1]，η(t_k)为t_k时刻计算电流和测量电流的误差值，ω(t_k)为t_k时刻Preisach三角形中滞回状态向量。Among them, λ is the step size factor, and λ∈[0,1], η(t _k ) is the error value between the calculated current and the measured current at time t _k , ω(t _k ) is the hysteresis state in the Preisach triangle at time t _k vector.

与现有技术相比，本发明具有以下优点：Compared with the prior art, the present invention has the following advantages:

一、建模准确：本发明通过引入测量电流和计算电流的误差，在每一时刻对Preisach三角形中网格对应的权重值进行自适应调节，通过先验荷电状态值和滞回状态值与当前电流实际值通过迭代计算，使得锂离子电池开路电压滞回特性准确进行建模。1. Accurate modeling: the present invention adjusts the weight value corresponding to the grid in the Preisach triangle at each moment by introducing the error of the measured current and the calculated current, and through the prior state of charge value and hysteresis state value and The current actual value is iteratively calculated to accurately model the hysteresis characteristics of the open circuit voltage of the lithium-ion battery.

二、提高精度：本发明是通过算法的改进提高建模的精度，算法中起核心作用的测量电流可通过电池管理系统中原有的电流传感器直接获取，在精度提高的同时并没有增加硬件成本。2. Improving accuracy: The present invention improves the accuracy of modeling through the improvement of the algorithm. The measurement current that plays a core role in the algorithm can be directly obtained by the original current sensor in the battery management system, and the hardware cost is not increased while the accuracy is improved.

三、估计准确：本发明是通过对锂离子电池滞回特性进行荷电状态的估计，针对滞回特性严重不能忽略的锂离子电池(如磷酸铁锂电池)的荷电状态估计更加准确可靠。3. Accurate estimation: the present invention estimates the state of charge of the hysteresis characteristic of the lithium-ion battery, and is more accurate and reliable for the estimation of the state of charge of the lithium-ion battery (such as lithium iron phosphate battery) whose hysteresis characteristic is serious and cannot be ignored.

附图说明Description of drawings

图1为锂离子电池OCV-SOC滞回特性曲线示意图。Figure 1 is a schematic diagram of the OCV-SOC hysteresis characteristic curve of a lithium-ion battery.

图2为基本Preisach算子示意图。Figure 2 is a schematic diagram of the basic Preisach operator.

图3为Preisach三角形及阶梯形记忆曲线示意图。Figure 3 is a schematic diagram of a Preisach triangle and a ladder-shaped memory curve.

图4为离散Preisach三角形及网格示意图。Figure 4 is a schematic diagram of discrete Preisach triangles and meshes.

图5为锂离子电池OCV-SOC滞回模型训练流程图。Figure 5 is a flowchart of the training of the OCV-SOC hysteresis model for lithium-ion batteries.

图6为锂离子电池OCV-SOC自适应离散Preisach模型应用流程图。Figure 6 is a flow chart of the application of the OCV-SOC adaptive discrete Preisach model for lithium-ion batteries.

具体实施方式Detailed ways

下面结合附图和具体实施例对本发明进行详细说明。The present invention will be described in detail below in conjunction with the accompanying drawings and specific embodiments.

实施例：Example:

本发明的主要目的是建立一种锂离子电池开路电压滞回特性的准确建模方法，从而最终能够用于锂离子电池荷电状态SOC的准确估计，本发明的另一目的是通过提出的建模方法由已知的准确SOC得到滞回存在下的OCV值，从而能进一步用于分析电池的极化电压以及阻抗等，为电池管理系统提供更多的信息。The main purpose of the present invention is to establish an accurate modeling method for the hysteresis characteristic of the open circuit voltage of a lithium-ion battery, thereby finally being able to be used for accurate estimation of the SOC of the state of charge of the lithium-ion battery. The model method obtains the OCV value in the presence of hysteresis from the known accurate SOC, which can be further used to analyze the polarization voltage and impedance of the battery, and provide more information for the battery management system.

为了实现本发明如上所述的目标和其他优点，如这里具体地和广泛地描述，提供一种基于电流调节和离散Preisach算子的锂离子电池开路电压滞回特性的建模方法，传统的开路电压滞回特性建模中存在误差较大精度难以保证的缺点，并不适用于本发明，本发明在离散Preisach算子对滞回特性建模的基础上，将Preisach三角形划分的网格所对应的权重值视为时变量，通过求取每一时刻下定义的先验SOC以及电池中的电流值，得到电流的偏差值，该值结合最小均方误差理论(LMS)得到每一时刻下权重值变化的增量，与上一时刻的权重值求和得到当前时刻的权重值，进一步结合当前时刻的滞回状态，得到后验SOC即为当前时刻的SOC输出，其中模型中网格划分的数目和权重向量的初始值是通过锂离子电池OCV-SOC滞回曲线离线确定。In order to achieve the objectives and other advantages of the present invention as described above, as specifically and broadly described herein, a modeling method for the hysteresis characteristics of the open circuit voltage of lithium-ion batteries based on current regulation and discrete Preisach operators is provided. The traditional open circuit In the modeling of voltage hysteresis characteristics, there is a shortcoming that the error is large and the accuracy is difficult to guarantee, which is not applicable to the present invention. On the basis of the discrete Preisach operator modeling the hysteresis characteristics, the present invention divides the grid corresponding to the Preisach triangle The weight value of the current is regarded as a time variable. By obtaining the prior SOC defined at each moment and the current value in the battery, the deviation value of the current is obtained. This value is combined with the minimum mean square error theory (LMS) to obtain the weight at each moment. The increment of the value change is summed with the weight value of the previous moment to obtain the weight value of the current moment, and further combined with the hysteresis state of the current moment, the posterior SOC is the SOC output of the current moment, where the grid division in the model The initial values of the number and weight vectors are determined off-line through the Li-ion battery OCV-SOC hysteresis curve.

根据本发明的优选实施例，完整的实施步骤如下：1)实验离线得到锂离子电池OCV-SOC滞回特性曲线；2)离线确定Preisach三角形网格划分的数目和对应的权重向量的初始值；3)在线应用时Preisach三角形网格对应的权重值视为时变，当前值由上一时刻权重值加上当前时刻变化量得到；4)根据当前时刻OCV值更改Preisach三角形每一网格的滞回状态值，并根据上一时刻的权重值得到当前时刻的先验SOC值；5)由先验SOC值、上一时刻的SOC值、电池容量得到当前时刻的电流估计值，与实际电流比较得到误差值；6)当前时刻的电流误差值结合最小均方误差(LMS)理论得到当前时刻权重值的增量和当前时刻的权重值；7)当前时刻的权重值结合当前时刻的滞回状态值得到当前时刻的后验SOC值即为当前时刻的SOC值，下一时刻重复上述过程。According to a preferred embodiment of the present invention, the complete implementation steps are as follows: 1) experiment off-line to obtain the lithium-ion battery OCV-SOC hysteresis characteristic curve; 2) off-line determine the number of Preisach triangular mesh division and the initial value of the corresponding weight vector; 3) The weight value corresponding to the Preisach triangle grid is regarded as time-varying during online application, and the current value is obtained by adding the weight value of the previous moment to the current moment change; 4) Change the hysteresis of each grid of the Preisach triangle according to the current OCV Return the state value, and obtain the prior SOC value at the current moment according to the weight value at the previous moment; 5) Obtain the current estimated value at the current moment from the prior SOC value, the SOC value at the previous moment, and the battery capacity, and compare it with the actual current Obtain the error value; 6) The current error value at the current moment is combined with the least mean square error (LMS) theory to obtain the increment of the weight value at the current moment and the weight value at the current moment; 7) The weight value at the current moment is combined with the hysteresis state at the current moment The posteriori SOC value obtained at the current moment is the SOC value at the current moment, and the above process is repeated at the next moment.

图1所示为实验测试的锂离子电池OCV-SOC滞回特性曲线(OCV充分静置，时间大于3小时)，该实验曲线用于确定本发明提出的滞回模型的初始参数，为了充分覆盖锂离子电池的滞回特性，实验测定时先从电池满充状态放电，每隔5％SOC变化和静置后记录OCV值直至SOC＝0满放状态；然后进入充电状态每隔5％SOC变化和静置后记录OCV值到SOC＝95％后，转为放电，直到5％SOC后又改为充电；该过程循环进行并保证每一充电状态比上一充电状态的SOC值减少5％，每一放电状态比上一放电状态的SOC值增加5％，直到SOC＝50％实验结束；在整个过程中，OCV均是SOC每变化5％后静置并记录。Fig. 1 shows the lithium-ion battery OCV-SOC hysteresis characteristic curve (OCV fully stands still, the time is greater than 3 hours) of experimental test, and this experimental curve is used for determining the initial parameter of the hysteresis model that the present invention proposes, in order to fully cover The hysteresis characteristics of lithium-ion batteries are firstly discharged from the fully charged state during the experimental measurement, and the OCV value is recorded every 5% SOC change and after standing still until SOC=0 full discharge state; then enter the charging state every 5% SOC change And record the OCV value after standing still until SOC=95%, then switch to discharge, until 5% SOC and then change to charging; this process is cyclical and ensures that each charging state is 5% less than the SOC value of the previous charging state, The SOC value of each discharge state is increased by 5% compared with the previous discharge state, until the end of the SOC=50% experiment; in the whole process, the OCV is the SOC after every 5% change, and then stand still and record.

如图2-4所示，Preisach算子(α,β)描述了输入输出的一种关系，即当输入大于阈值α时输出为1，当输入小于阈值β(β<α)时输出为-1，当输入介于阈值α和β之间时输出值不变，由于滞回关系中输入值幅值有限，如图3所示，所有的Preisach算子(α,β)构成了二维平面中一直角三角形，滞回特性的历史信息可通过该三角形中一记忆曲线进行表示，记忆曲线是一条阶梯形曲线，其形成规则是：当输入随时间增加时，阶梯形曲线反映为三角形内水平上升的线段；当输入随时间减小时，阶梯形曲线反映为三角形内垂直左移的线段，三角形内部阶梯形曲线左下区域滞回状态值为1，右上区域滞回状态值为-1，同时三角形内部每一点对应一个权重值，输出值由三角形内所有点的滞回状态值以及权重值乘积积分得到。As shown in Figure 2-4, the Preisach operator (α, β) describes a relationship between input and output, that is, when the input is greater than the threshold α, the output is 1, and when the input is smaller than the threshold β (β<α), the output is - 1. When the input is between the threshold α and β, the output value remains unchanged. Due to the limited amplitude of the input value in the hysteresis relationship, as shown in Figure 3, all Preisach operators (α, β) constitute a two-dimensional plane The right-angled triangle in the middle, the historical information of the hysteresis characteristic can be expressed by a memory curve in the triangle, the memory curve is a ladder-shaped curve, and its formation rule is: when the input increases with time, the ladder-shaped curve is reflected as the level in the triangle Rising line segment; when the input decreases with time, the step-shaped curve is reflected as a line segment moving vertically to the left in the triangle. Each internal point corresponds to a weight value, and the output value is obtained by multiplying and integrating the hysteresis state values of all points in the triangle and the weight value.

在实际应用时，需要对该连续Preisach三角形进行离散化，即沿水平方向和垂直方向对该三角形进行划分(一般等长进行划分)，最终形成多个矩形网格，每一网格对应的滞回状态值由该网格包含的所有点的滞回状态值决定，该值位于区间[-1,+1]，每一网格同样对应一权重值，在应用之前需要对该模型进行训练，训练采用实验得到的OCV-SOC滞回关系曲线，得到网格划分的数目以及每一网格对应的权重值，训练过程的流程图如图5所示。In practical applications, it is necessary to discretize the continuous Preisach triangle, that is, to divide the triangle along the horizontal and vertical directions (generally equal-length division), and finally form multiple rectangular grids. The hysteresis state value is determined by the hysteresis state value of all points contained in the grid, and the value is in the interval [-1,+1]. Each grid also corresponds to a weight value, and the model needs to be trained before application. The training uses the OCV-SOC hysteresis relationship curve obtained from the experiment to obtain the number of grid divisions and the corresponding weight value of each grid. The flow chart of the training process is shown in Figure 5.

模型的初始参数主要包括离散Preisach模型中Preisach三角形边长划分数目n和该划分下Preisach三角形中每一网格对应的滞回权重初始值μ。Preisach三角形是α-β平面上一顶点坐标为(u_min,u_max)(其中u_min和u_max分别是荷电状态为0和100％时对应的开路电压值)斜边位于α＝β上的等腰直角三角形。Preisach三角形边长划分数目n的选取由模型最终的精度要求决定，一般取n的初始值大于等于20。竖直边和水平边的划分结果可分别表示为α_i<α_i+1,i＝1,2,…,n(其中α₁＝u_min，α_n+1＝u_max)和β_i<β_i+1,i＝1,2,…,n(其中β₁＝u_min，β_n+1＝u_max)。最终整个Preisach三角形中共产生N＝n(n+1)/2个网格。每一网格可表示为S_i(i-1)/2+j＝{(β,α)|β_j≤β＜β_j+1,α_i≤α＜α_i+1}其中j≤i,i＝1,2,…,n,j＝1,2,…,n。在离散Preisach模型中每一网格均对应一权重值μ(称为滞回权重，需要离线计算获取)。取锂离子电池滞回特性实验曲线中OCV值作为模型输入u，每个输入下每个网格均作更新，产生该输入下对应的滞回状态值ω。ω值的确定需要依据Preisach三角形中的阶梯形曲线。该阶梯形曲线是由输入值u决定(当u增加时在Preisach三角形中产生一水平上升直线，当u减小时在Preisach三角形中产生一竖直左移直线，随着输入u不断增减在Preisach三角形中生成一阶梯形曲线)，在阶梯形曲线以下的网格S对应的滞回状态ω等于1，在阶梯形曲线以上的网格S对应的滞回状态ω等于-1，当阶梯形曲线穿过网格S的内部时，相应的滞回状态ω等于网格中阶梯形曲线下方的部分的面积减去阶梯形曲线下方的部分的面积；定义所有的滞回状态ω为一滞回状态向量ω(ω＝[ω₁,ω₂,…,ω_N]^T)，所有的滞回权重μ为一滞回权重向量μ(μ＝[μ₁,μ₂,…,μ_N]^T)，每一输入下对应的输出SOC等于滞回状态向量ω和滞回权重向量μ的乘积，即最终由所有实验获得的输入输出值得到一未知数为滞回权重向量μ的线性方程组，通过离线数值运算确定μ。The initial parameters of the model mainly include the division number n of the side length of the Preisach triangle in the discrete Preisach model and the initial value μ of the hysteresis weight corresponding to each grid in the Preisach triangle under the division. The Preisach triangle is a vertex on the α-β plane whose coordinates are (u _min , u _max ) (where u _min and u _max are the corresponding open-circuit voltage values when the state of charge is 0 and 100% respectively) and the hypotenuse is located on α=β isosceles right triangle. The selection of the number n of the side length divisions of the Preisach triangle is determined by the final accuracy requirements of the model, and the initial value of n is generally greater than or equal to 20. The division results of vertical and horizontal sides can be expressed as α _i <α _i+1 , i=1,2,...,n (where α ₁ =u _min , α _n+1 =u _max ) and β _i < β _i+1 , i=1,2,...,n (wherein β ₁ =u _min , β _n+1 =u _max ). Finally, a total of N=n(n+1)/2 grids are generated in the entire Preisach triangle. Each grid can be expressed as S _i(i-1)/2+j ={(β,α)|β _j ≤β<β _j+1 ,α _i ≤α<α _i+1 } where j≤i ,i=1,2,...,n,j=1,2,...,n. In the discrete Preisach model, each grid corresponds to a weight value μ (called hysteresis weight, which needs to be obtained by off-line calculation). The OCV value in the experimental curve of the hysteresis characteristic of the lithium-ion battery is taken as the model input u, and each grid is updated under each input to generate the corresponding hysteresis state value ω under the input. The determination of the value of ω needs to be based on the step-shaped curve in the Preisach triangle. The step-shaped curve is determined by the input value u (when u increases, a horizontal rising straight line is generated in the Preisach triangle, when u decreases, a vertical left-moving straight line is generated in the Preisach triangle, and as the input u continues to increase and decrease in the Preisach triangle A ladder-shaped curve is generated in the triangle), the hysteresis state ω corresponding to the grid S below the ladder-shaped curve is equal to 1, and the hysteresis state ω corresponding to the grid S above the ladder-shaped curve is equal to -1, when the ladder-shaped curve When passing through the interior of the grid S, the corresponding hysteresis state ω is equal to the area of the part under the step-shaped curve in the grid minus the area of the part under the step-shaped curve; define all hysteresis states ω as a hysteresis state Vector ω(ω＝[ω ₁ ,ω ₂ ,…,ω _N ] ^T ), all hysteresis weight μ is a hysteresis weight vector μ(μ=[μ ₁ ,μ ₂ ,…,μ _N ] ^T ) , the corresponding output SOC under each input is equal to the product of the hysteresis state vector ω and the hysteresis weight vector μ, namely Finally, a system of linear equations in which the unknown is the hysteresis weight vector μ is obtained from the input and output values obtained from all experiments, and μ is determined by off-line numerical calculation.

如图6所示，为了提高锂离子电池滞回曲线的建模精度，本发明提出了将每一网格对应的权重值视为时变量并能自适应改变的建模方法，在具体应用该方法时模型训练得到的权重值视为权重初始值，每一时刻首先根据OCV输入值更新离散Preisach三角形中所有网格对应的滞回状态值，由滞回状态值和上一时刻得到的权重值加权求和得到SOC值(定义为当前时刻的先验SOC)，并根据已知的电池容量和上一时刻得到的SOC值可以确定当前时刻电流估计值，该估计值和BMS检测到的电流值比较得到误差值，根据最小均方误差理论，该误差值和当前滞回状态值以及一个常数(该常数取值在0和1之间，可通过多次取值选取合适值)的乘积作为当前时刻权重值相比上一时刻权重值的增量，将得到的当前时刻权重值与当前时刻的滞回状态值加权求和，得到当前时刻的后验SOC即为当前OCV输入下对应的SOC输出，下一时刻重复该过程。As shown in Figure 6, in order to improve the modeling accuracy of the lithium-ion battery hysteresis curve, the present invention proposes a modeling method that regards the weight value corresponding to each grid as a time variable and can be adaptively changed. In the method, the weight value obtained by model training is regarded as the initial weight value. At each moment, the hysteresis state value corresponding to all the grids in the discrete Preisach triangle is first updated according to the OCV input value, and the hysteresis state value and the weight value obtained at the previous moment The SOC value (defined as the prior SOC at the current moment) is obtained by weighted summation, and the current estimated value at the current moment can be determined according to the known battery capacity and the SOC value obtained at the previous moment, and the estimated value and the current value detected by the BMS The error value is obtained by comparison. According to the minimum mean square error theory, the product of the error value, the current hysteresis state value and a constant (the value of the constant is between 0 and 1, and an appropriate value can be selected by multiple values) is used as the current Compared with the increment of the weight value at the previous time, the weight value at the current time is weighted and summed with the hysteresis state value at the current time, and the posterior SOC at the current time is the corresponding SOC output under the current OCV input , and repeat the process at the next moment.

Claims

1. A method for estimating the state of charge based on open-circuit voltage hysteresis characteristics, used for online estimation of the state of charge of lithium ion battery, is characterized in that, comprises the following steps:

1) Obtain the hysteresis characteristic curve of the open circuit voltage and state of charge of the lithium-ion battery offline;

2) Determine the initial parameters of the hysteresis characteristic adaptive model of the open circuit voltage based on the Preisach operator according to the hysteresis characteristic curve training, and establish the adaptive model of the hysteresis characteristic of the open circuit voltage based on the Preisach operator;

3) The state of charge of the lithium-ion battery is estimated online according to the adaptive model of the open circuit voltage hysteresis characteristic based on the Preisach operator, and the state of charge of the lithium-ion battery at the current moment is obtained.

2. a kind of estimation method based on the state of charge of open circuit voltage hysteresis characteristic according to claim 1, is characterized in that, described step 2) in the open circuit voltage hysteresis characteristic adaptive model based on Preisach operator The initial parameters of include the mesh division number N of the Preisach triangle and the initial value μ of the weight vector.

3. a kind of estimation method based on the state of charge of open circuit voltage hysteresis characteristic according to claim 1, it is characterized in that, described step 2) based on the open circuit voltage hysteresis characteristic adaptive model of Preisach operator is :

{SOC SOC}_{{t t}_{k k}} (({t t}_{k k})) = = {ω ω}^{T T} (({t t}_{k k})) μ μ (({t t}_{k k}))

in, is the state of charge value corresponding to the open circuit voltage at time t _k , ω(t _k ) is the hysteresis state vector corresponding to the open circuit voltage value at time t _k in the Preisach triangle, μ(t _k ) is all nets in the Preisach triangle at time t _k The hysteretic weight vector for the lattice.

4. a kind of method for estimating state of charge based on open circuit voltage hysteresis characteristic according to claim 3, is characterized in that, described step 3) specifically comprises the following steps:

31) Obtain online the weight vector μ(t _k-1 ) in the Preisach triangle grid at the previous moment t _k-1 and the hysteresis state value ω(t _k ) at the current moment t _k , and calculate the prior State of charge for:

{SOC SOC}_{{t t}_{k k - - 11}} (({t t}_{k k})) = = {ω ω}^{T T} (({t t}_{k k})) μ μ (({t t}_{k k - - 11}));;

32) According to the battery capacity Q ₀ , the state of charge value at the last moment t _k-1 and the prior state of charge value at the current moment t _k Calculate the current estimated value I _cal (t _k ) at the current moment as:

{I I}_{c c a a l l} (({t t}_{k k})) = = {Q Q}_{00} [[{SOC SOC}_{{t t}_{k k - - 11}} (({t t}_{k k})) - - {SOC SOC}_{{t t}_{k k - - 11}} (({t t}_{k k - - 11}))]] / / Δ Δ t t

Δt ₌ tk _-tk-1 ;

33) According to the measured current actual value I _m (t _k ) and current estimated value I _cal (t _k ), the current error value η_current(t _k ) at the current moment is obtained as:

η_current(t _k )=I _m (t _k )-I _cal (t _k );

34) According to the current error value η_current(t _k ) at the current moment and the prior state of charge value at the current moment And use the minimum mean square error method to get the weight vector increment at the current moment And calculate the weight value μ(t _k ) at the current moment as:

μ(t _k )=μ(t _k-1 )+λη(t _k )ω(t _k )

η η (({t t}_{k k})) = = \frac{Δ Δ t t}{{Q Q}_{00} η η__c c u u r r r r e e n no t t (({t t}_{k k}))}

Where λ is the step factor, and λ∈[0,1];

35) According to the weight vector μ(t _k-1 ) and the hysteresis state value ω(t _k ) at the current moment, the posterior state of charge value at the current moment is obtained through the adaptive model of the hysteresis characteristic of the open circuit voltage That is, the state of charge of the lithium-ion battery at the current moment, and return to step 31) to estimate the state of charge of the lithium-ion battery at the next moment.

5. A method for estimating state of charge based on open-circuit voltage hysteresis characteristics according to claim 4, characterized in that, in the step 34), the current moment weight vector increment The calculation formula is:

{Δμ Δμ}_{{t t}_{k k}} = = λ λ η η (({t t}_{k k})) ω ω (({t t}_{k k}))

Among them, λ is the step size factor, and λ∈[0,1], η(t _k ) is the error value between the calculated current and the measured current at time t _k , ω(t _k ) is the hysteresis state in the Preisach triangle at time t _k vector.