Lithium battery residual capacity estimation method and system

Abstract

The invention provides a lithium battery residual capacity estimation method and system, and belongs to the technical field of lithium battery management, and the method comprises the steps: S1, obtaining a relation curve between an open-circuit voltage and residual capacity based on a lithium battery equivalent circuit model when it is detected that the state of a lithium battery meets a preset condition; s2, establishing a state equation and a measurement equation of the lithium battery by taking the residual capacity as a state quantity according to the equivalent circuit model, the relation curve and the capacity calculation formula of the lithium battery; s3, on the basis of a multi-time-scale recursive least square method with a forgetting factor, identifying open-circuit voltage under the urban road circulation working condition on line, and correcting the relation curve according to an identified open-circuit circuit to obtain a corrected relation curve; and S4, establishing a weighted multi-innovation volume Kalman filter based on the state equation and the measurement equation, and estimating the residual capacity. The method has the beneficial effects that the state equation and the measurement equation of the lithium battery are established by taking the residual capacity as the state quantity, and the residual capacity can be directly estimated.

Description

Technical Field

[0001] The present invention relates to the technical field of lithium battery management, and in particular to a method and system for estimating the remaining capacity of a lithium battery. Background Art

[0002] Lithium-ion batteries (lithium ion cells and batteries) are rechargeable batteries that use lithium-ion embedded compounds as positive electrode materials. They mainly rely on the movement of lithium ions between the positive and negative electrodes to work. As an efficient and recyclable energy storage device, lithium batteries have been widely used in many fields such as electric vehicles, portable electronic devices, and energy storage systems.

[0003] Lithium battery capacity is one of the important indicators for evaluating battery performance. Lithium battery capacity will gradually decay with the increase of cycle times and the passage of time, thus affecting the battery life, energy efficiency and user experience. Regular and accurate lithium battery capacity estimation can be used to help users monitor the health of the battery, understand how much power the battery can store, and guide users in using the battery. For example, timely discovering the trend of capacity decline can guide users to take measures such as replacing the battery or reducing the charge and discharge rate to extend the battery life, thereby achieving better user experience and environmental protection benefits.

[0004] Existing lithium battery capacity estimation usually refers to estimating the available capacity of the lithium battery in the current cycle. The result obtained by dividing the available capacity of the lithium battery in the current cycle by its nominal capacity is the lithium battery's state of health (SOH). The state of charge (SOC) of a lithium battery refers to the current remaining capacity of the lithium battery divided by the available capacity in the current cycle. When estimating the state of charge SOC, it is generally necessary to estimate the available capacity in the current cycle in order to obtain a more accurate state of charge SOC value.

[0005] At present, the capacity estimation methods of lithium batteries mainly include data-driven methods, capacity estimation methods based on state of charge (SOC) estimation values, and capacity estimation methods based on capacity increment method or differential voltage method. The data-driven method establishes a mapping relationship between features and capacity through feature selection and machine learning algorithms. This method requires a large amount of historical data for training, and the cost of data acquisition and model training is high. The capacity estimation method based on the state of charge (SOC) estimation value relies on high-precision state of charge (SOC) prediction results, and its accuracy is limited by the accuracy of the state of charge (SOC) estimation. The capacity estimation method based on the capacity increment method or differential voltage method is more used in offline research in laboratory environments to analyze the mechanism and law of battery capacity attenuation. Summary of the invention

[0006] In order to solve the above technical problems, the present invention provides a method and system for estimating the remaining capacity of a lithium battery.

[0007] The technical problem solved by the present invention can be achieved by adopting the following technical solutions:

[0008] A method for estimating the remaining capacity of a lithium battery, comprising:

[0009] Step S1, when it is detected that the lithium battery state meets a preset condition, a relationship curve between the open circuit voltage and the remaining capacity is obtained based on a lithium battery equivalent circuit model;

[0010] Step S2, establishing a state equation and a measurement equation of the lithium battery with the remaining capacity as the state quantity according to the lithium battery equivalent circuit model, the relationship curve and the capacity calculation formula;

[0011] Step S3, online identifying the open circuit voltage under the urban road cycle condition based on the recursive least square method with a forgetting factor on multiple time scales, and correcting the relationship curve according to the identified open circuit voltage to obtain a corrected relationship curve;

[0012] Step S4, establishing a weighted multi-innovation volumetric Kalman filter based on the state equation and the measurement equation to estimate the remaining capacity.

[0013] Preferably, the preset condition is that the state of charge of the lithium battery reaches a preset state of charge value;

[0014] The preset state of charge value is 60%.

[0015] Preferably, the lithium battery equivalent circuit model is a first-order equivalent circuit model.

[0016] Preferably, the relationship curve between the open circuit voltage and the remaining capacity is:

[0017]

[0018] Wherein, Cap represents the remaining capacity; a, b, c, d, e, and f represent the fitting parameters of the open circuit voltage and the remaining capacity respectively; the subscripts h, j, and l represent the number of general terms; n h 、n j 、n l represents the total of all general terms; OCV(Cap) represents the open circuit voltage.

[0019] Preferably, the state equation and the measurement equation are respectively:

[0020]

[0021] Among them, Cap represents the remaining capacity; U1 represents the polarization voltage; T represents the sampling interval; τ1 represents the time constant; R1 represents the polarization internal resistance; I represents the current; U represents the terminal voltage; OCV represents the open circuit voltage; R0 represents the ohmic internal resistance; and the subscript k represents the k moment.

[0022] Preferably, the multi-time scale recursive least squares method with forgetting factor is:

[0023]

[0024] Where G represents gain; H represents covariance matrix; represents system data; λ represents the forgetting factor; θ represents the parameter to be identified; U is the terminal voltage; I represents the unit matrix; the subscript k represents the kth moment; and the superscript T represents the matrix transpose.

[0025] Preferably, step S4 comprises:

[0026] Step S41, initializing state values ​​and covariance matrices;

[0027] Step S42, performing a prediction step, includes:

[0028] Step S421, calculate volume points

[0029]

[0030] Among them, S k-1 Denotes the covariance matrix P k-1 The matrix obtained after decomposition; ξ i is the vector of the i-th column of the matrix ξ; represents the estimated value of the state at the k-1th moment;

[0031] Step S422, calculate the volume point state prediction value

[0032]

[0033] Among them, i=0,...,m; u k Represents system input; represents the nonlinear state transfer function;

[0034] Step S423, calculate the state prior value

[0035]

[0036] Step S424, calculate the state covariance prior value P k / k-1 :

[0037]

[0038] Where Q represents the process noise covariance matrix;

[0039] Step S43, performing an update step, includes:

[0040] Step S431, calculate volume points

[0041]

[0042] Among them, S k / k-1 Represents the state covariance prior value P k / k-1 The square root of; Step S432, calculate the volume point observation prediction value

[0043] in, represents the nonlinear state transfer function;

[0044] Step S433, calculate the observation estimate

[0045]

[0046] Step S434, calculate the observation covariance P zz,k :

[0047]

[0048] Where R represents the measurement noise covariance;

[0049] Step S435, calculate the covariance P between the state quantity and the observation quantity xz,k :

[0050]

[0051] Step S436, calculate the Kalman gain K k :

[0052] K k =P xz,k / P zz,k

[0053] Step S437, construct a weighted multi-innovation matrix:

[0054]

[0055] K L =[K k K k-1 …K k-L+1 ]

[0056]

[0057] Among them, z k represents the terminal voltage measurement value at time k; e k represents the error information at time k; E L represents the error matrix; L represents the weighted multi-information window size; K L represents the Kalman gain matrix; W e,L represents the error weight matrix, w e,i Denotes the error weight matrix W e,L Elements in; W t,L represents the time weight matrix, w t,i Represents the time weight matrix W t,L Elements in

[0058] Step S438, state update, state covariance update:

[0059] After exceeding the preset time period, the weighted multi-innovation volumetric Kalman filter is enabled to update the state and the state covariance update formula is:

[0060]

[0061] in, represents the state posterior estimate; P k represents the posterior estimate of the state covariance.

[0062] Preferably, in step S438, within the preset time period, a cubature Kalman filter is used to perform state update and state covariance update:

[0063]

[0064]

[0065] Preferably, the weighted multi-innovation window size is 2.

[0066] The present invention also provides a lithium battery remaining capacity estimation system, which is used to implement the lithium battery remaining capacity estimation method as described above, comprising:

[0067] A curve acquisition module is used to obtain a relationship curve between open circuit voltage and remaining capacity based on a lithium battery equivalent circuit model when it is detected that the lithium battery state meets a preset condition;

[0068] an equation establishing module, connected to the curve acquiring module, for establishing a state equation and a measurement equation of the lithium battery with the remaining capacity as the state quantity according to the lithium battery equivalent circuit model, the relationship curve and the capacity calculation formula;

[0069] An identification module, connected to the curve acquisition module, is used to online identify the open circuit voltage under the urban road cycle condition based on the recursive least square method with a multi-time scale and a forgetting factor, and to correct the relationship curve according to the open circuit voltage obtained by the identification to obtain a corrected relationship curve;

[0070] The remaining capacity estimation module is connected to the equation establishment module and the identification module respectively, and is used to establish a weighted multi-information volumetric Kalman filter based on the state equation and the measurement equation to estimate the remaining capacity.

[0071] The advantages or beneficial effects of the technical solution of the present invention are:

[0072] The present invention uses the remaining capacity as the state quantity to establish the state equation and measurement equation of the lithium battery, which can directly estimate the remaining capacity without inferring the remaining capacity by estimating the state of charge SOC; online parameter identification is based on the recursive least squares method with a multi-time scale and a forgetting factor, which reduces the number of calculations and has parameter updates compared to offline parameter identification, making it more accurate; and the utilization rate of the new information is improved based on the weighted multi-new information volumetric Kalman filter. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] Figure 1 A schematic diagram of a flow chart of a method for estimating the remaining capacity of a lithium battery in a preferred embodiment of the present invention;

[0074] Figure 2 A schematic diagram of an OCV-remaining capacity curve under different aging conditions in a preferred embodiment of the present invention;

[0075] Figure 3 A schematic diagram of an OCV-state of charge SOC curve under different aging conditions in a preferred embodiment of the present invention;

[0076] Figure 4 It is a partial enlarged schematic diagram of the OCV-remaining capacity curve under different aging conditions in a preferred embodiment of the present invention;

[0077] Figure 5 It is a partial enlarged schematic diagram of the OCV-state of charge SOC curve under different aging conditions in a preferred embodiment of the present invention;

[0078] Figure 6 A schematic diagram of a first-order equivalent circuit model in a preferred embodiment of the present invention;

[0079] Figure 7a-7i A schematic diagram of the remaining capacity estimation in different health states in a preferred embodiment of the present invention;

[0080] Figure 8 FIG. 4 is a structural block diagram of a lithium battery remaining capacity estimation system in a preferred embodiment of the present invention. DETAILED DESCRIPTION

[0081] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0082] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0083] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments, but they are not intended to limit the present invention.

[0084] In this embodiment, the terms are explained as follows:

[0085] State of health (SOH) refers to the ratio of the available capacity of a lithium battery in the current cycle divided by its nominal capacity.

[0086] State of charge (SOC) refers to the ratio of the current remaining capacity of a lithium battery divided by the available capacity in the current cycle.

[0087] Open circuit voltage (OCV): OCV refers to the terminal voltage of a battery in an open circuit state. In this embodiment, the voltage under a low rate discharge can be considered as the open circuit voltage.

[0088] See also Figure 1 In a preferred embodiment of the present invention, based on the above problems existing in the prior art, a method for estimating the remaining capacity of a lithium battery is provided, comprising:

[0089] Step S1, when it is detected that the lithium battery state meets a preset condition, a relationship curve between the open circuit voltage and the remaining capacity is obtained based on a lithium battery equivalent circuit model;

[0090] In this embodiment, the lithium battery includes a lithium iron phosphate (LFP) battery, which may be referred to as an LFP battery hereinafter.

[0091] An aging experiment was carried out on the lithium iron phosphate battery to obtain the current and voltage data of the LFP battery under low-rate current discharge in different health states and a constant temperature environment of 25°C, as well as the current and voltage data under urban road cycle conditions (Urban Dynamometer Driving Schedule, UDDS).

[0092] The open circuit voltage (OCV) is the voltage at a small rate of discharge. The remaining capacity and state of charge SOC are calculated based on the current and time. The OCV-remaining capacity curve under different aging conditions can be drawn, such as Figure 2 As shown; draw the OCV-state of charge SOC curve under different aging conditions as shown Figure 3 . Figure 2 and Figure 3 The lines of different colors in the figure represent the nth battery test. Due to testing errors, the 3rd and 10th results are missing.

[0093] Depend on Figure 2 It can be seen that with the increase in the number of aging cycles, the maximum remaining capacity of the lithium battery obtained in each aging test gradually decreases, and the remaining capacity cannot be deduced simply through an OCV-remaining capacity curve and OCV value.

[0094] However, it can be found that in the part of the remaining capacity below 12Ah (i.e. the second voltage plateau), the OCV-remaining capacity curves under different aging conditions have a high degree of overlap. Figure 4 As shown, for example, when the OCV is 3.25 V, the remaining capacity corresponding to the OCV-remaining capacity curve under the first aging test is substantially the same as the remaining capacity corresponding to the OCV-remaining capacity curve under the eleventh aging test. Based on this characteristic, the embodiment of the present invention can estimate the remaining capacity within a certain range.

[0095] Depend on Figure 3 It can be seen that as the lithium battery ages, the OCV-state of charge SOC curve will also change, and the curve overlap is high only during the voltage plateau period. Therefore, the OCV-state of charge SOC curve used in the model-based state of charge SOC estimation method needs to be continuously updated according to the aging state. However, the present invention directly uses the remaining capacity as the state quantity, and for batteries in different aging states, only the OCV-remaining capacity curve measured at the beginning needs to be used.

[0096] Enlarge the OCV-state of charge SOC curve as shown Figure 5 As shown, it can be seen that under the same OCV, the SOC corresponding to different curves has a large difference. For example, under the same OCV of 3.25V, the SOC corresponding to the OCV-SOC curve in the first aging test is about 28%, and the SOC corresponding to the OCV-SOC curve in the eleventh aging test is about 36%, which has a large error. Figure 5 It can be seen that the test battery at different aging stages is in the second voltage plateau at 60% SOC.

[0097] Therefore, the present invention estimates the remaining capacity starting from 60% SOC.

[0098] As a preferred embodiment, the preset condition is that the state of charge of the lithium battery reaches a preset state of charge value;

[0099] The default state of charge value is 60%.

[0100] In this embodiment, the function of OCV with respect to the remaining capacity is obtained based on the equivalent potential model fitting. The equivalent potential model is used to fit the OCV-remaining capacity curve obtained from the first battery test. The fitting formula is as follows:

[0101] As a preferred embodiment, the relationship curve between the open circuit voltage and the remaining capacity is:

[0102]

[0103] Where Cap represents the remaining capacity; a, b, c, d, e, and f represent the fitting parameters of the open circuit voltage and the remaining capacity, respectively; the subscripts h, j, and l represent the number of general terms; n h 、n j 、n l represents the total number of general terms, i=0,...,n i ,j=0,...,n j ,l=0,...,n l ; OCV(Cap) stands for open circuit voltage.

[0104] Since the OCV-remaining capacity curves under different aging conditions have a high degree of overlap starting from 60% SOC, the remaining capacities corresponding to the OCV-remaining capacity curves under different aging tests are basically the same. Therefore, the present invention only needs to fit one OCV-remaining capacity curve to be used for the remaining capacity estimation under different health conditions.

[0105] Step S2, establishing a state equation and a measurement equation of the lithium battery with the remaining capacity as the state quantity according to the lithium battery equivalent circuit model, the relationship curve and the capacity calculation formula;

[0106] As a preferred embodiment, the lithium battery equivalent circuit model is a first-order equivalent circuit model.

[0107] Specifically, in this embodiment, the first-order equivalent circuit model is selected as the lithium battery equivalent circuit model, such as Figure 6 As shown in the figure, this model uses an ideal voltage source to describe OCV, an ideal internal resistance to describe the ohmic internal resistance R0, and an RC unit to describe the polarization voltage. R1 represents the polarization internal resistance, C1 represents the polarization capacitance, and U1 represents the polarization voltage.

[0108] In this embodiment, the system state equation and measurement equation are established by combining the capacity calculation formula, the first-order equivalent circuit model and the OCV-remaining capacity curve. Different from other commonly used system equations, the system state equation and measurement equation established in the present invention do not use the state of charge SOC as the state quantity, but use the remaining capacity as the state quantity.

[0109] As a preferred embodiment, the state equation and the measurement equation are respectively:

[0110]

[0111] Among them, Cap k represents the remaining capacity at time k; U 1,k represents the polarization voltage at time k; T represents the sampling interval, which is used as an example but not as a limitation. In the experiment of this embodiment, T is 1s; τ 1,k Cap represents the time constant at time k; k-1 represents the remaining capacity at time k-1; U 1,k-1 represents the polarization voltage at time k-1; R 1,k represents the polarization internal resistance at time k; I k represents the current at time k; U k represents the terminal voltage at time k; OCV (Cap k ) represents the remaining capacity Cap at time k k The corresponding open circuit voltage; R 0,k Represents the ohmic internal resistance at time k; the subscript k represents time k.

[0112] Step S3, online identifying the open circuit voltage under the urban road cycle (UDDS) condition based on the multi-time scale recursive least square method with forgetting factor (MFFRLS), and correcting the relationship curve according to the identified open circuit voltage to obtain a corrected relationship curve;

[0113] Specifically, the recursive least square method with forgetting factor is a method for dynamic system parameter estimation. The algorithm can realize real-time identification of lithium battery parameters and real-time identification of OCV. Since the OCV curve will be different under different working conditions, the embodiment of the present invention uses the recursive least square method to online identify the OCV under the first UDDS working condition test, and corrects the above-obtained remaining capacity-OCV curve under dynamic working conditions.

[0114] Considering that the change of lithium battery parameters is a slow process, it is not necessary to identify the parameters every moment. Therefore, the embodiment of the present invention adopts the recursive least square method with forgetting factor in multiple time scales to identify parameters, and a large time scale Ls is given to perform parameter identification and update parameters, that is, the parameters are calculated once every time scale, otherwise the parameters calculated last time are kept.

[0115] The recursive least squares method with forgetting factor in multiple time scales can reduce the number of calculations and maintain relatively stable parameter values. Compared with offline parameter identification, it has the function of parameter updating and is more accurate.

[0116] against For this system, the algorithm flow is as follows. As a preferred implementation, the recursive least square method with multiple time scales and forgetting factors is:

[0117]

[0118] Where G represents gain; H represents covariance matrix; represents system data; λ represents the forgetting factor; θ represents the parameter to be identified; U represents the system output, that is, the terminal voltage; I represents the unit matrix; the subscript k represents the kth moment; and the superscript T represents the matrix transpose.

[0119] In this embodiment, the identification time is 100 seconds.

[0120] The embodiment of the present invention adopts the recursive least square method with multiple time scales and forgetting factors to identify parameters online, which has a small amount of calculation and is more accurate than offline identification.

[0121] Step S4: A weighted multi-innovation cubature Kalman filter (WMICKF) is established based on the state equation and the measurement equation, and the remaining capacity is estimated using the weighted multi-innovation cubature Kalman filter.

[0122] The main idea of ​​CKF is to approximate the propagation and observation of the probability density function by selecting a set of representative points (called "volume points") in the state space. These volume points are usually obtained by numerically integrating the Gaussian distribution to ensure that the approximation is Gaussian. Compared with the standard Kalman filter, CKF has better approximation performance on nonlinear problems because it avoids linearizing the Jacobian matrix, thereby improving the accuracy of the prediction and update steps.

[0123] The basic steps of CKF are similar to those of standard Kalman filtering, including two steps: prediction and update. The prediction step predicts the state at the next moment through the state transfer function and calculates the mean and covariance matrix of the predicted state. The update step corrects the predicted state using the observed value and updates the mean and covariance matrix of the state.

[0124] The weighted multi-information cubature Kalman filter is an improvement on the CKF. Compared with the CKF, the multi-information cubature Kalman filter introduces the innovation vector and improves the estimation accuracy by making more use of the innovation. The weighted multi-information cubature Kalman filter further improves the utilization rate of the innovation, assigns different weights to the innovation at different times, and further improves the estimation accuracy.

[0125] Specifically, the above-mentioned online identified model parameters and the relationship curve between the open circuit voltage and the remaining capacity are updated in the state equation and the measurement equation.

[0126] As a preferred embodiment, step S4 includes:

[0127] Step S41, initializing state values ​​and covariance matrices;

[0128] Step S42, performing a prediction step, includes:

[0129] Step S421, calculate volume points

[0130]

[0131] ξ=[EE]

[0132] Among them, S k-1 Denotes the covariance matrix P k-1 The matrix obtained after decomposition; ξ i is the vector of the i-th column of the matrix ξ, i=0,...,m, m=2n, n represents the dimension of the state vector; represents the state estimate at the k-1th moment; E represents the identity matrix, whose order is the same as the state vector;

[0133] Step S422, calculate the volume point state prediction value

[0134]

[0135] Among them, u k Represents system input; represents the nonlinear state transfer function;

[0136] Step S423, calculate the state prior value

[0137]

[0138] Step S424, calculate the state covariance prior value P k / k-1 :

[0139]

[0140] Where Q represents the process noise covariance matrix;

[0141] Step S43, performing an update step, includes:

[0142] Step S431, calculate volume points

[0143]

[0144] Among them, S k / k-1 Represents the state covariance prior value P k / k-1 The square root of

[0145] Step S432, calculate the volume point observation prediction value

[0146]

[0147] in, represents the nonlinear state transfer function;

[0148] Step S433, calculate the observation estimate

[0149]

[0150] Step S434, calculate the observation covariance P zz,k :

[0151]

[0152] Where R represents the measurement noise covariance;

[0153] Step S435, calculate the covariance P between the state quantity and the observation quantity xz,k :

[0154]

[0155] Step S436, calculate the Kalman gain K k , K k is the covariance P between the state quantity and the observation quantity xz,k and the observed covariance P zz,k ratio:

[0156] K k =P xz,k / P zz,k

[0157] Step S437, construct a weighted multi-innovation matrix:

[0158]

[0159] K L =[K k K k-1 …K k-L+1 ]

[0160]

[0161] Among them, zk represents the terminal voltage measurement value at time k; e k represents the error information at time k; E L represents the error matrix; L represents the weighted multi-information window size; K L represents the Kalman gain matrix; W e,L represents the error weight matrix, w e,i Denotes the error weight matrix W e,L Elements in; W t,L represents the time weight matrix, w t,i Represents the time weight matrix W t,L Elements in

[0162] In this embodiment, the error weight matrix W e,L , time weight matrix W t,L Different weights are assigned to new information based on error distribution and time distribution. The larger the error, the more corrections are needed, and the larger the weight is assigned; the closer the time is, the more meaningful the new information is, and the larger the weight is assigned.

[0163] Step S438, state update, state covariance update:

[0164] During the preset time period, ordinary cubature Kalman filtering is used to update the state and state covariance:

[0165]

[0166] in, represents the Kalman gain K k The transpose of

[0167] Specifically, at the beginning of the algorithm, ordinary CKF is used to estimate the capacity, and the state update and state covariance update formulas are as above. After the algorithm has been running for a period of time, WMICKF is enabled, that is:

[0168] After exceeding the preset time period, the weighted multi-innovation volumetric Kalman filter is enabled to update the state and the state covariance update formula is:

[0169]

[0170] in, represents the state posterior estimate; P k represents the posterior estimate of the state covariance; where, represents the Kalman gain matrix K L The transpose of ;.

[0171] The weighted multi-innovation cubature Kalman filter proposed in the embodiment of the present invention can make better use of the innovations compared to the cubature Kalman filter.

[0172] As a preferred implementation, the weighted multi-innovation window size is 2.

[0173] In this embodiment, the algorithm sets the initial capacity value to 8Ah, and the initial values ​​of the P, Q, and R matrices are all the same.

[0174] In the above preferred embodiment, the method of the present invention starts the algorithm from 60% SOC, firstly performs parameter identification, and the identification time is 100s. After the parameters are stable, the weighted multi-innovation CKF is used to estimate the remaining capacity.

[0175] Different aging states are also different health states. Generally, the percentage of the ratio of the current fully charged battery capacity to the initial nominal battery capacity is used as the battery SOH value. The remaining capacity estimation results under different health states are as follows: Figure 7a-7i As shown, the values ​​above the chart, such as 21.36Ah, 20.47Ah, 19.38Ah, 18.88Ah, 18.49Ah, 17.79Ah, 17.37Ah, 16.91Ah, and 16.03Ah, respectively represent the capacity of the battery when it is fully charged in different health states. In the nine tests, the remaining capacity estimation successfully converged to the reference value, and good estimation results were achieved.

[0176] Since the OCV-remaining capacity curves have certain similarities starting from 60% SOC, the embodiment of the present invention only needs to obtain one OCV-remaining capacity curve to complete the remaining capacity estimation of the LFP battery under different health states.

[0177] See also Figure 8 The present invention also provides a lithium battery remaining capacity estimation system for implementing the above-mentioned lithium battery remaining capacity estimation method, comprising:

[0178] The curve acquisition module 1 is used to obtain the relationship curve between the open circuit voltage and the remaining capacity based on the lithium battery equivalent circuit model when it is detected that the lithium battery state meets the preset conditions;

[0179] The equation establishing module 2 is connected to the curve acquiring module 1 and is used to establish the state equation and measurement equation of the lithium battery with the remaining capacity as the state quantity according to the lithium battery equivalent circuit model, the relationship curve and the capacity calculation formula;

[0180] The identification module 3 is connected to the curve acquisition module 1, and is used to online identify the open circuit voltage under the urban road cycle condition based on the recursive least square method with a forgetting factor on multiple time scales, and to correct the relationship curve according to the open circuit voltage obtained by the identification to obtain a corrected relationship curve;

[0181] The remaining capacity estimation module 4 is connected to the equation establishment module 2 and the identification module 3 respectively, and is used to establish a weighted multi-innovation volumetric Kalman filter based on the state equation and the measurement equation to estimate the remaining capacity.

[0182] More implementation details of the system and the method have been disclosed and will not be repeated in this embodiment.

[0183] Specifically, the present invention establishes a state equation and an observation equation based on the current and voltage data measured in real time, a first-order equivalent circuit model, a capacity calculation formula, and a residual capacity-open circuit voltage curve, and adopts a multi-time scale recursive least squares method with a forgetting factor (MFFRLS) combined with a weighted multi-innovation cubature Kalman filter (WMICKF) to estimate the residual capacity of a lithium-ion battery in real time starting from 60% SOC, thereby realizing real-time estimation of the residual capacity of a lithium-ion battery.

[0184] The system equation of the embodiment of the present invention uses the remaining capacity as a state quantity, which is different from the prior art that uses the state of charge SOC as a state quantity. The present invention can directly estimate the remaining capacity instead of inferring the remaining capacity by estimating the state of charge SOC.

[0185] The lithium-ion battery remaining capacity estimation method proposed in the present invention can estimate the current remaining capacity of the lithium-ion battery in real time. Different from estimating the available capacity of the current cycle, it directly estimates the remaining capacity of the battery. This value can not only reflect the health status of the lithium-ion battery, but also provide users with a reference to the current available power of the lithium-ion battery.

[0186] The advantages or beneficial effects of adopting the above technical solution are:

[0187] (1) The present invention fits the residual capacity and voltage data of the LFP battery under low-rate current discharge through an equivalent potential model as a residual capacity-OCV curve, and online identifies the OCV under UDDS conditions through the recursive least squares method, and corrects the residual capacity-OCV curve under dynamic conditions.

[0188] (2) The present invention establishes a state equation and an observation equation based on a first-order equivalent circuit model, a capacity calculation formula, and a remaining capacity-OCV curve.

[0189] (3) Since the residual capacity-OCV curve has a high overlap below 60% SOC, the present invention adopts multi-time scale recursive least squares method to identify model parameters online starting from 60% SOC, and then uses WMICKF algorithm to estimate the residual capacity of lithium battery at micro scale, with less calculation. WMICKF algorithm improves the utilization rate of new information and has higher estimation accuracy.

[0190] (4) The remaining capacity estimation method proposed in the present invention can be used to estimate the remaining capacity of LFP batteries under different SOHs, and only the initial remaining capacity-OCV curve needs to be obtained. The maximum remaining capacity of the lithium battery can be inferred based on the estimated remaining capacity combined with the state of charge SOC, which can be further used to estimate the health status of the lithium battery.

[0191] The above are only preferred embodiments of the present invention, and are not intended to limit the implementation methods and protection scope of the present invention. Those skilled in the art should be aware that all solutions obtained by equivalent substitutions and obvious changes made using the contents of this specification and illustrations should be included in the protection scope of the present invention.

Claims

1. A method for estimating the remaining capacity of a lithium battery, characterized in that: include: Step S1, when it is detected that the lithium battery state meets a preset condition, a relationship curve between the open circuit voltage and the remaining capacity is obtained based on a lithium battery equivalent circuit model; Step S2, establishing a state equation and a measurement equation of the lithium battery with the remaining capacity as the state quantity according to the lithium battery equivalent circuit model, the relationship curve and the capacity calculation formula; Step S3, online identifying the open circuit voltage under the urban road cycle condition based on the recursive least square method with a forgetting factor on multiple time scales, and correcting the relationship curve according to the identified open circuit voltage to obtain a corrected relationship curve; Step S4, establishing a weighted multi-innovation volumetric Kalman filter based on the state equation and the measurement equation to estimate the remaining capacity.

2. The method for estimating the remaining capacity of a lithium battery according to claim 1, characterized in that: The preset condition is that the state of charge of the lithium battery reaches a preset state of charge value; The preset state of charge value is 60%.

3. The method for estimating the remaining capacity of a lithium battery according to claim 1, wherein: The lithium battery equivalent circuit model is a first-order equivalent circuit model.

4. The method for estimating the remaining capacity of a lithium battery according to claim 1, characterized in that: The relationship curve between the open circuit voltage and the remaining capacity is: Wherein, Cap represents the remaining capacity; a, b, c, d, e, and f represent the fitting parameters of the open circuit voltage and the remaining capacity respectively; the subscripts h, j, and l represent the number of general terms; n h 、n j 、n l represents the total of all general terms; OCV(Cap) represents the open circuit voltage.

5. The method for estimating the remaining capacity of a lithium battery according to claim 1, characterized in that: The state equation and measurement equation are: Among them, Cap represents the remaining capacity; U1 represents the polarization voltage; T represents the sampling interval; τ1 represents the time constant; R1 represents the polarization internal resistance; I represents the current; U represents the terminal voltage; OCV represents the open circuit voltage; R0 represents the ohmic internal resistance; and the subscript k represents the k moment.

6. The method for estimating the remaining capacity of a lithium battery according to claim 1, characterized in that: The multi-time scale recursive least square method with forgetting factor is: Where G represents gain; H represents covariance matrix; represents system data; λ represents the forgetting factor; θ represents the parameter to be identified; U represents the terminal voltage; I represents the unit matrix; the subscript k represents the kth moment; and the superscript T represents the matrix transpose.

7. The method for estimating the remaining capacity of a lithium battery according to claim 1, characterized in that: The step S4 comprises: Step S41, initializing state values ​​and covariance matrices; Step S42, performing a prediction step, includes: Step S421, calculate volume points Among them, S k-1 Denotes the covariance matrix P k-1 The matrix obtained after decomposition; ξ i is the vector of the i-th column of the matrix ξ; represents the estimated value of the state at the k-1th moment; Step S422, calculate the volume point state prediction value Among them, i=0,...,m; u k Represents system input; represents the nonlinear state transfer function; Step S423, calculate the state prior value Step S424, calculate the state covariance prior value P k / k-1 : Where Q represents the process noise covariance matrix; Step S43, performing an update step, includes: Step S431, calculate volume points Among them, S k / k-1 Represents the state covariance prior value P k / k-1 The square root of Step S432, calculate the volume point observation prediction value in, represents the nonlinear state transfer function; Step S433, calculate the observation estimate Step S434, calculate the observation covariance P zz,k : Where R represents the measurement noise covariance; Step S435, calculate the covariance P between the state quantity and the observation quantity xz,k : Step S436, calculate the Kalman gain K k : K k =P xz,k / P zz,k Step S437, construct a weighted multi-innovation matrix: K L =[K k K k-1 …K k-L+1 ] Among them, z k represents the terminal voltage measurement value at time k; e k represents the error information at time k; E L represents the error matrix; L represents the weighted multi-information window size; K L represents the Kalman gain matrix; W e,L represents the error weight matrix, w e,i Denotes the error weight matrix W e,L Elements in; W t,L represents the time weight matrix, w t,i Represents the time weight matrix W t,L Elements in; Step S438, state update, state covariance update: After exceeding the preset time period, the weighted multi-innovation volumetric Kalman filter is enabled to update the state and the state covariance update formula is: in, represents the state posterior estimate; P k represents the posterior estimate of the state covariance.

8. The method for estimating the remaining capacity of a lithium battery according to claim 7, characterized in that: In step S438, within the preset time period, a cubature Kalman filter is used to perform state update and state covariance update:

9. The method for estimating the remaining capacity of a lithium battery according to claim 7, characterized in that: The weighted multi-innovation window size is 2.

10. A lithium battery remaining capacity estimation system, characterized in that: The method for estimating the remaining capacity of a lithium battery according to any one of claims 1 to 9 comprises: A curve acquisition module is used to obtain a relationship curve between open circuit voltage and remaining capacity based on a lithium battery equivalent circuit model when it is detected that the lithium battery state meets a preset condition; an equation establishing module, connected to the curve acquiring module, for establishing a state equation and a measurement equation of the lithium battery with the remaining capacity as the state quantity according to the lithium battery equivalent circuit model, the relationship curve and the capacity calculation formula; An identification module, connected to the curve acquisition module, is used to online identify the open circuit voltage under the urban road cycle condition based on the recursive least square method with a multi-time scale and a forgetting factor, and to correct the relationship curve according to the open circuit voltage obtained by the identification to obtain a corrected relationship curve; The remaining capacity estimation module is connected to the equation establishment module and the identification module respectively, and is used to establish a weighted multi-information volumetric Kalman filter based on the state equation and the measurement equation to estimate the remaining capacity.

Images