A method and system for rapid diagnosis of internal short circuit of a battery

Abstract

The application discloses a battery internal short circuit rapid diagnosis method and system, and relates to the field of battery internal short circuit diagnosis; a second-order RC equivalent circuit model and a charging segment CCCV correction circuit model of a battery are constructed; the terminal voltage and the current of the battery are measured; whether the battery is in a constant current charging working condition is determined based on the terminal voltage and the current, and a first judgment result is obtained; if the judgment result indicates no, the second-order RC equivalent circuit model is observed based on an extended Kalman filtering algorithm, and the short circuit current of the battery is obtained; if the judgment result indicates yes, the charging segment CCCV correction circuit model is observed based on the extended Kalman filtering algorithm, and the short circuit current of the battery is obtained; and the short circuit resistance of the battery is identified based on the short circuit current by using a recursive least square method; the application solves the problem of large error in the constant current segment, directly takes the short circuit current as a state variable, observes by using Kalman filtering, estimates the short circuit information within a few sampling periods of fault occurrence, and improves the speed of internal short circuit diagnosis.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of battery internal short circuit diagnosis, in particular to a battery internal short circuit rapid diagnosis method and system. BACKGROUND

[0002] The electric vehicle technology is increasingly mature, and consumers' demand for electric vehicle performance is also getting higher and higher, especially the endurance mileage. However, while the energy density of lithium ion battery is increasing, the risk of thermal runaway of lithium ion battery under abnormal conditions will also increase significantly. As one of the core components of electric vehicles, although the current research focuses on developing high-energy-density lithium ion batteries, the safety of lithium ion batteries also needs to be paid enough attention. The safety of lithium ion battery is the prerequisite for improving the energy density of lithium ion battery, and the safety research of lithium ion battery is the inexhaustible power for the long-term sustainable development of electric vehicles.

[0003] The thermal runaway of lithium ion battery is mainly caused by two aspects: one is the problem of lithium ion battery materials and production process, and the other is the problem in the use process of lithium ion battery. There are many reasons for the thermal runaway of lithium ion battery in the use process, such as internal and external short circuit of lithium ion battery, overcharge and discharge, high-rate charge and discharge, high and low temperature environment, cycle aging, extrusion deformation, etc. Among them, the internal short circuit of lithium ion battery is the most common cause of thermal runaway.

[0004] In order to maximize the performance of lithium ion battery, improve its safety and prolong its service life, it is necessary to monitor the lithium ion battery on-line and diagnose the internal short circuit of lithium ion battery. The internal short circuit diagnosis of lithium ion battery is to monitor the temperature, voltage, current and other state information of lithium ion battery in real time, and to realize the early warning of lithium ion battery internal short circuit through certain model algorithm.

[0005] At present, some researchers have proposed some methods for detecting and diagnosing the internal short circuit of lithium ion battery from different angles. (1) Model-based diagnosis method. (2) Consistency of battery pack using statistical methods, etc. (3) Method based on thermal and electrical characteristic threshold diagnosis (4) Based on external auxiliary measurement circuit method. The equivalent circuit model is simple and has high precision, which has been widely used in engineering. Considering the economy and efficiency, most of the diagnosis methods in engineering are based on circuit model for diagnosis.

[0006] Most of the existing methods have problems: (1) The method of detecting the battery state of charge SOC generally needs a long time to find the existence of internal short circuit. (2) The method based on the consistency of the battery pack requires the consistency of the battery pack itself to be high, but it is difficult to maintain this condition during the use of the battery, and other disturbances such as battery aging may affect the detection. (3) The method using external auxiliary means or instruments is usually offline or requires relatively high-precision and expensive equipment, such as infrared detectors, X-ray diffraction, etc., which has little value in engineering.

[0007] The method of diagnosing short circuit condition based on model through state observation generally takes the battery SOC as the observation quantity, and the related short circuit information is observed. Due to the slow change of the SOC of the battery, and due to the model accuracy and other disturbances, the SOC estimation error is generally 2%-3%, so a relatively long observation interval is generally selected for SOC estimation, and the short circuit information is further observed through the abnormal change of SOC. The method based on SOC generally cannot quickly diagnose the fault within a few sampling periods after the fault occurs. SUMMARY

[0008] In view of the above problems, the present application provides a battery internal short circuit rapid diagnosis method and system, which solves the problem of large error in the constant current section and improves the speed of internal short circuit diagnosis.

[0009] To achieve the above object, the present application provides the following scheme:

[0010] A battery internal short circuit rapid diagnosis method, the battery internal short circuit rapid diagnosis method comprising the following steps:

[0011] Constructing a second-order RC equivalent circuit model of the battery and a charging section CCCV correction circuit model;

[0012] Measuring the terminal voltage and current of the battery;

[0013] Determining whether the battery is in constant current charging condition based on the terminal voltage and current, and obtaining a first judgment result;

[0014] If the judgment result is no, observing the second-order RC equivalent circuit model based on the extended Kalman filtering algorithm, and obtaining the short circuit current of the battery;

[0015] If the judgment result is yes, observing the charging section CCCV correction circuit model based on the extended Kalman filtering algorithm, and obtaining the short circuit current of the battery;

[0016] Identifying the short circuit resistance of the battery based on the short circuit current using the recursive least squares method.

[0017] Optionally, a second-order RC equivalent circuit model of the battery and a charging section CCCV correction circuit model are constructed, and specifically include:

[0018] The internal resistance of the battery is equivalent to two resistance-capacitance networks and a resistance in series, and a first equivalent circuit of the battery is constructed;

[0019] The parameters in the first equivalent circuit are identified by using the HPPC method, and a second-order RC equivalent circuit model is obtained;

[0020] The internal resistance of the battery is equivalent to a resistance, and a second equivalent circuit of the battery is constructed;

[0021] The parameters of the second equivalent circuit are identified based on an OCV-SOC curve of the battery and different rate CCCV curves, and a charging section CCCV correction circuit model is constructed; the OCV-SOC curve is obtained by pulse discharge testing of the battery, and the different rate CCCV curves are obtained by CCCV charging testing of the battery at different rates.

[0022] Optionally, the second-order RC equivalent circuit model is:

[0023] x2(k+1)=A2x2(k)+B2u2(k)+w(k)

[0024] y2(k)=g2(x2(k),u2(k))+v(k);

[0025]

[0026] g2(x2(k),u k (t))=OCV(ξ(k))+U1(k)+U2(k)+(I L (k)-I SC (k))R0;

[0027] wherein x2(k+1) represents a state variable of the second-order RC equivalent circuit model at the k+1th iteration, A2 represents a state variable matrix of the second-order RC equivalent circuit model acting on x2(k+1), x2(k) represents a state variable of the second-order RC equivalent circuit model at the kth iteration, x2(k)=[ξ(k) U1(k) U2(k) Isc(k)] T ,ξ(k) represents a state of charge of the battery at the kth iteration, U1(k) represents a voltage value of the first resistance-capacitance network of the second-order RC equivalent circuit model at the kth iteration, U2(k) represents a voltage value of the second resistance-capacitance network of the second-order RC equivalent circuit model at the kth iteration, I sc(k) represents the short circuit current value at the kth iteration, B2 represents the input control matrix of the second order RC equivalent circuit model acting on u2(k), u2(k) represents the input variable of the second order RC equivalent circuit model at the kth iteration, u2(k) = I L (k), I L (k) represents the current of the battery, w(k) represents the system noise at the kth iteration, y2(k) represents the observation variable of the second order RC equivalent circuit model at the kth iteration, y2(k) = U t (k), U t (k) represents the terminal voltage of the battery, g2() represents the observation function of the second order RC equivalent circuit model, v(k) represents the measurement noise at the kth iteration, η represents the charging efficiency of the battery, ΔT represents the sampling time, Q0 represents the battery capacity, τ1 represents the time constant of the first resistance-capacitance network of the second order RC equivalent circuit model, τ1 = R1C1, R1 represents the resistance value of the first resistance-capacitance network of the second order RC equivalent circuit model, C1 represents the capacitance value of the first resistance-capacitance network of the second order RC equivalent circuit model, τ2 represents the time constant of the second resistance-capacitance network of the second order RC equivalent circuit model, τ2 = R2C2, R2 represents the resistance value of the second resistance-capacitance network of the second order RC equivalent circuit model, C2 represents the capacitance value of the second resistance-capacitance network of the second order RC equivalent circuit model, OCV() represents the open circuit voltage function of the battery, R0 represents the resistance value of the second order RC equivalent circuit model;

[0028] The charging section CCCV correction circuit model is:

[0029] x1(k+1) = A1x1(k) + B1u1(k) + w(k)

[0030] y1(k) = g1(x1(k), u1(k)) + v(k);

[0031]

[0032] g1(x1(k), u1(k)) = OCV(ξ(k)) + (I L (k) - I SC (k))R s ;

[0033] wherein x1(k+1) represents the state variable of the charging section CCCV correction circuit model at the k+1th iteration, A1 represents the state variable matrix of the charging section CCCV correction circuit model acting on x1(k+1), x1(k) represents the state variable of the charging section CCCV correction circuit model at the kth iteration, x1(k) = [ξ(k) Isc(k)] TB1 represents an input control matrix of the charging section CCCV correction circuit model acting on u1(k), u1(k) represents an input variable of the charging section CCCV correction circuit model at the kth iteration, u1(k) = I L (k), y1(k) represents an observation variable of the charging section CCCV correction circuit model at the kth iteration, y1(k) = U t (k), g1() represents an observation function of the charging section CCCV correction circuit model, R s represents a resistance value of the charging section CCCV correction circuit model.

[0034] Optionally, the extended Kalman filter EKF specifically comprises:

[0035] Let the value of the iteration number k be 0;

[0036] Set the initial condition as:

[0037] wherein, represents the posterior state variable at the 0th iteration, x0 represents an initial random setting value, E[] represents taking the mean, represents the posterior error covariance matrix at the 0th iteration;

[0038] Let the value of k increase by 1;

[0039] Based on the posterior state estimation at the (k-1)th iteration, determine the prior state estimation at the kth iteration as:

[0040]

[0041] wherein, represents the prior state variable of the target model at the kth iteration, f() represents a mapping of the target model, represents the posterior state variable of the target model at the (k-1)th iteration, u k-1 represents the input variable of the target model at the (k-1)th iteration;

[0042] The target model is a second-order RC equivalent circuit model or a charging section CCCV correction circuit model;

[0043] Based on the prior state estimation at the kth iteration and the posterior error covariance matrix at the (k-1)th iteration, determine the prior error covariance matrix at the kth iteration as:

[0044]

[0045] wherein, represents the prior error covariance matrix of the target model at the kth iteration, represents the state variable matrix of the target model at the (k-1)th iteration, represents a posteriori error covariance matrix of the target model at the k-1th iteration, represents a transpose of a state variable matrix of the target model at the k-1th iteration, Q k represents a process error of the target model at the kth iteration;

[0046] determines a gain matrix of the kth iteration based on the a priori error covariance matrix of the kth iteration, as:

[0047]

[0048] wherein, K k represents a Kalman gain of the target model at the kth iteration that minimizes the estimation and measurement variance, represents a transpose of an observation matrix of the target model at the kth iteration, R k represents an observation error matrix of the target model at the kth iteration, represents an observation matrix of the target model at the kth iteration;

[0049] determines a posteriori state estimation of the kth iteration based on the a priori state estimation of the kth iteration and the gain matrix of the kth iteration, as:

[0050]

[0051] wherein, represents a posteriori state variable of the target model at the kth iteration, y k represents a sampled voltage value of the target model at the kth iteration, g() represents an observation function of the target model;

[0052] determines a posteriori error covariance matrix of the kth iteration based on the posteriori state estimation of the kth iteration, the a priori error covariance matrix of the kth iteration and the gain matrix of the kth iteration, as:

[0053]

[0054] wherein, represents a posteriori error covariance matrix of the target model at the kth iteration, I represents a unit matrix;

[0055] determines whether a preset iteration number is reached, to obtain a second determination result;

[0056] if the second determination result indicates no, the value of k is increased by 1, and the step of determining the a priori state estimation of the kth iteration based on the posteriori state estimation of the k-1th iteration is returned to;

[0057] If the second determination result indicates yes, output a posterior state estimation of the kth iteration; the posterior state estimation of the kth iteration contains a short-circuit current of the battery.

[0058] The method for rapidly diagnosing internal short circuit of the battery under the constant current charging condition comprises the following steps:

[0059] A charging section CCCV correction circuit model of the battery is constructed.

[0060] The charging section CCCV correction circuit model is observed based on an extended Kalman filtering algorithm, so as to obtain the short-circuit current of the battery.

[0061] The short-circuit resistance of the battery is identified based on the short-circuit current by using a recursive least square method.

[0062] Optionally, the charging section CCCV correction circuit model of the battery comprises the following steps:

[0063] An internal resistance of the battery is equivalent to a resistor, and a second equivalent circuit of the battery is constructed.

[0064] Parameters of the second equivalent circuit are identified based on an OCV-SOC curve of the battery and different rate CCCV curves, so as to construct the charging section CCCV correction circuit model; the OCV-SOC curve is obtained by performing pulse discharge test on the battery, and the different rate CCCV curves are obtained by performing CCCV charging test on the battery under different rates.

[0065] Optionally, the charging section CCCV correction circuit model is as follows:

[0066] x1(k+1) = A1x1(k) + B1u1(k) + w(k)

[0067] y1(k) = g1(x1(k), u1(k)) + v(k);

[0068]

[0069] g1(x1(k), u1(k)) = OCV(ξ(k)) + (I L (k) - I SC (k))R s ;

[0070] Wherein, x1(k+1) represents a state variable of the charging section CCCV correction circuit model at the k+1th iteration, A1 represents a state variable matrix of the charging section CCCV correction circuit model acting on x1(k+1), x1(k) represents a state variable of the charging section CCCV correction circuit model at the kth iteration, x1(k) = [ξ(k) Isc(k)]T, ξ(k) represents a state of charge of the battery at the kth iteration, Isc (k) represents the short-circuit current value at the kth iteration, B1 represents the input control matrix of the charging section CCCV correction circuit model acting on u1(k), u1(k) represents the input variable of the charging section CCCV correction circuit model at the kth iteration, u1(k) = I L (k), I L (k) represents the current of the battery, w(k) represents the system noise at the kth iteration, y1(k) represents the observation variable of the charging section CCCV correction circuit model at the kth iteration, y1(k) = U t (k), U t (k) represents the terminal voltage of the battery, g1() represents the observation function of the charging section CCCV correction circuit model, v(k) represents the measurement noise at the kth iteration, η represents the charging efficiency of the battery, ΔT represents the sampling time, Q0 represents the battery capacity, OCV() represents the open-circuit voltage function of the battery, R s represents the resistance value of the charging section CCCV correction circuit model.

[0071] The battery internal short-circuit rapid diagnosis system is applied to any one of the battery internal short-circuit rapid diagnosis methods described above, and the system comprises:

[0072] A building module is configured to build a second-order RC equivalent circuit model of a battery and a charging section CCCV correction circuit model.

[0073] A measuring module is configured to measure the terminal voltage and current of the battery.

[0074] A judging module is configured to determine whether the battery is in a constant-current charging working condition based on the terminal voltage and current, and obtain a first judgment result.

[0075] A calculating module is configured to obtain the short-circuit current of the battery based on an extended Kalman filter, and identify the short-circuit resistance of the battery based on the short-circuit current by using a recursive least squares method.

[0076] The calculating module specifically comprises:

[0077] If the judgment result indicates no, the second-order RC equivalent circuit model is observed based on an extended Kalman filter algorithm to obtain the short-circuit current of the battery.

[0078] If the judgment result indicates yes, the charging section CCCV correction circuit model is observed based on an extended Kalman filter algorithm to obtain the short-circuit current of the battery.

[0079] The short-circuit resistance of the battery is identified based on the short-circuit current by using a recursive least squares method.

[0080] An electronic device includes a memory, a processor, and a computer program stored on the memory and executable on the processor, and the processor implements the battery internal short circuit rapid diagnosis method according to any one of the above when executing the computer program.

[0081] A computer readable storage medium, the storage medium stores a computer program, and the computer program implements the battery internal short circuit rapid diagnosis method according to any one of the above when executed.

[0082] According to the specific embodiments of the present application, the following technical effects are disclosed.

[0083] The application discloses a battery internal short circuit rapid diagnosis method and system, a second-order RC equivalent circuit model and a charging segment CCCV correction circuit model of a battery are constructed; the terminal voltage and the current of the battery are measured; whether the battery is in a constant current charging working condition is determined based on the terminal voltage and the current, and a first judgment result is obtained; if the judgment result is no, the second-order RC equivalent circuit model is observed based on an extended Kalman filtering algorithm, and a short circuit current of the battery is obtained; if the judgment result is yes, the charging segment CCCV correction circuit model is observed based on the extended Kalman filtering algorithm, and the short circuit current of the battery is obtained; and the short circuit resistance of the battery is identified based on the short circuit current by using a recursive least square method, compared with a traditional HPPC method for identifying an RC model based on the charging segment CCCV correction circuit model, the problem that the traditional model has a large error in a constant current segment is solved, the short circuit current I SC Directly as a state variable, directly observed by using Kalman filtering EKF, short circuit information can be estimated within a few sampling periods after the fault occurs, compared with a traditional method for estimating abnormal changes of SOC, the speed of internal short circuit diagnosis is greatly improved. BRIEF DESCRIPTION OF DRAWINGS

[0084] In order to more clearly illustrate the technical solutions of the embodiments of the present application or the prior art, the drawings needed in the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0085] Figure 1 The equivalent circuit of the internal short circuit of the battery used in the present application;

[0086] Figure 2 The charging test schematic diagram for identifying the parameters of the charging segment CCCV correction circuit model of different rates used in the present application;

[0087] Figure 3 The charging curve under different rates of CCCV;

[0088] Figure 4 The CCCV correction circuit model is a second-order RC equivalent circuit model.

[0089] Figure 5 The short circuit information is identified based on an observer and a recursive least square (RLS) method.

[0090] Figure 6 A flowchart of the entire diagnosis algorithm. DETAILED DESCRIPTION

[0091] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative efforts fall within the scope of the present application.

[0092] The present application aims to provide a battery internal short circuit rapid diagnosis method and system, which solves the problem of large error of a traditional model in a constant current segment by using a CCCV correction circuit model, and estimates short circuit information in a few sampling periods after a fault occurs by using a short circuit current I SC as a state variable, directly observing by using an extended Kalman filter (EKF), so as to estimate the short circuit information in a few sampling periods after a fault occurs, and greatly improve the speed of internal short circuit diagnosis compared with a traditional method of estimating abnormal changes of SOC.

[0093] In order to make the above-mentioned purposes, features and advantages of the present application more apparent and easy to understand, the present application will be further described in detail below with reference to the drawings and specific embodiments.

[0094] As shown in FIG. 1, the internal short circuit equivalent circuit model of a lithium ion battery is first established in the present application. Figure 1 Figure 1 As shown in FIG. 1, the internal short circuit equivalent circuit model of a lithium ion battery is first established in the present application.The entire model includes the OCV-SOC voltage of the battery, which is related to the charge state in the positive and negative active materials of the battery, R0 represents Ohmic polarization, and the two RC links usually represent the concentration polarization in the battery and the electrochemical polarization of the electrode particles. The OCV represents the open circuit voltage of the battery, which is the equilibrium potential of the battery in the equilibrium state without external excitation, and is related to the positive and negative material properties of the battery, and is a function of the charge state SOC of the battery. R0 represents resistance, C1, R1, C2 and R2 represent the polarization capacitance and resistance of the polarization link, and all the parameters can be measured by the battery HPPC standard test method. The present application selects a second-order RC circuit, so that the polarization link of the battery has two RC inertia links, and U t is the terminal voltage of the battery at the external port.

[0095] The whole model is simple, and the parameters of the whole model can be usually obtained through pulse discharge test.

[0096]

[0097] wherein ξ represents the SOC of the battery, Q0 is the capacity of the battery, η is the charging efficiency, I L is the external excitation current of the battery, and the inflow is positive.

[0098] The specific connection mode of the equivalent circuit of the internal short circuit of the battery used in the application is as follows:

[0099] As shown in Figure 1 , the open circuit voltage OCV(ξ) is connected in series with the resistance R0 and two resistance-capacitance networks to obtain a series branch, and the series branch is connected in parallel with the short circuit resistance R ISC . The first resistance-capacitance network is obtained by connecting the first resistance R1 and the first capacitance C1 in series, the second resistance-capacitance network is obtained by connecting the second resistance R2 and the second capacitance C2 in series, and the first resistance-capacitance network is connected in series with the second resistance-capacitance network.

[0100] When the internal short circuit occurs in the battery, the model of the battery can be represented by II in Figure 1 , and the resistance R ISC represents the internal short circuit of the battery, at this time, due to the existence of the short circuit resistance R ISC , there is an electron channel from the positive electrode to the negative electrode of the battery, so there is a part of internal leakage current I SC . At this time, the real electrochemical current I in flowing into the battery is shunted, and the relationship between the electrochemical current I in and the external excitation current I L can be represented as:

[0101] I L =I in +I Sc .

[0102] Embodiment 1

[0103] As shown in Figure 6 , the battery internal short circuit rapid diagnosis method provided by the embodiment of the application comprises the following steps:

[0104] A second-order RC equivalent circuit model and a charging segment CCCV correction circuit model of the battery are constructed.

[0105] The terminal voltage U t and the current I L of the battery are measured.

[0106] Based on the terminal voltage U t and the current I L , it is determined whether the battery is in a constant current charging working condition, and a first judgment result is obtained.

[0107] If the judgment result is negative, then the second-order RC equivalent circuit model is observed based on the extended Kalman filter algorithm to obtain the battery's short-circuit current I. SC .

[0108] If the judgment result indicates yes, then the CCCV correction circuit model of the charging section is observed based on the extended Kalman filter algorithm to obtain the short-circuit current I of the battery. SC .

[0109] The short-circuit resistance of the battery is identified using the recursive least squares method based on the short-circuit current.

[0110] In actual implementation, a second-order RC equivalent circuit model of the battery and a CCCV correction circuit model for the charging section are constructed, specifically including:

[0111] The battery's internal resistance is represented by two RC networks and a resistor R0 connected in series, thus constructing the first equivalent circuit of the battery.

[0112] The parameters in the first equivalent circuit are identified using the HPPC method to obtain a second-order RC equivalent circuit model.

[0113] like Figure 4 As shown in model.II, the obtained second-order RC equivalent circuit model is obtained by connecting the open-circuit voltage OCV(ξ) with the resistor R0 and two RC networks in series. The first RC network is obtained by connecting the first resistor R1 and the first capacitor C1 in series, and the second RC network is obtained by connecting the second resistor R2 and the second capacitor C2 in series. The first RC network and the second RC network are connected in series.

[0114] The internal resistance of the battery is equivalent to a resistor R. S Construct the second equivalent circuit of the battery.

[0115] The parameters of the second equivalent circuit are identified based on the battery's OCV-SOC curve and CCCV curves at different rates, and a CCCV correction circuit model for the charging segment is constructed. The OCV-SOC curve is obtained by performing pulse discharge tests on the battery, and the CCCV curves at different rates are obtained by performing CCCV charging tests on the battery at different rates.

[0116] like Figure 4 As shown in model.II, the obtained CCCV correction circuit model for the charging section is composed of the open-circuit voltage OCV(ξ) and the resistance R. S Obtained by connecting multiple series.

[0117] The equivalent circuit model of the second-order RC circuit is:

[0118] x2(k+1)=A2x2(k)+B2u2(k)+w(k)

[0119] y2(k) = g2(x2(k), u2(k)) + v(k);

[0120]

[0121] g2(x2(k), u2(k)) = OCV(ξ(k)) + U1(k) + U2(k) + (I L (k) - I SC (k))R0;

[0122] where x2(k + 1) represents the state variable of the second-order RC equivalent circuit model at the k+1th iteration, A2 represents the state variable matrix of the second-order RC equivalent circuit model acting on x2(k + 1), x2(k) represents the state variable of the second-order RC equivalent circuit model at the kth iteration, x2(k) = [ξ(k) U1(k) U2(k) Isc(k)] T , ξ(k) represents the state of charge of the battery at the kth iteration, U1(k) represents the voltage value of the first RC network of the second-order RC equivalent circuit model at the kth iteration, U2(k) represents the voltage value of the second RC network of the second-order RC equivalent circuit model at the kth iteration, I sc (k) represents the short-circuit current value at the kth iteration, B2 represents the input control matrix of the second-order RC equivalent circuit model acting on u2(k), u2(k) represents the input variable of the second-order RC equivalent circuit model at the kth iteration, u2(k) = I L (k), I L (k) represents the current of the battery, w(k) represents the system noise at the kth iteration, y2(k) represents the observation variable of the second-order RC equivalent circuit model at the kth iteration, y2(k) = U t (k), U t (k) represents the terminal voltage of the battery, g2() represents the observation function of the second-order RC equivalent circuit model, v(k) represents the measurement noise at the kth iteration, η represents the sampling time, Q0 represents the capacity of the battery, τ1 represents the time constant of the first RC network of the second-order RC equivalent circuit model, τ1 = R1C1, R1 represents the resistance value of the first RC network of the second-order RC equivalent circuit model, C1 represents the capacitance value of the first RC network of the second-order RC equivalent circuit model, τ2 represents the time constant of the second RC network of the second-order RC equivalent circuit model, τ2 = R2C2, R2 represents the resistance value of the second RC network of the second-order RC equivalent circuit model, C2 represents the capacitance value of the second RC network of the second-order RC equivalent circuit model, OCV() represents the open-circuit voltage function of the battery, and R0 represents the resistance value of the second-order RC equivalent circuit model.

[0123] In the specific implementation process, each iteration is generally spaced 1 s.

[0124] In the specific implementation, the application proposes a charging segment CCCV correction circuit model as shown in Figure 4 The specific model establishment steps are as follows: referring to Figure 2 , perform CCCV charging under four different rates, set the current of the constant current segment to 0.5 / 0.75 / 1 / 1.25 C, set the initial SOC to 0, i.e., discharge the battery to its empty state, the external voltage is generally 3 V, and then set the cutoff voltage of the CV segment to 4.2 V, i.e., the open circuit voltage of the battery when the SOC is calibrated to 1. Perform charging test on the battery to measure the terminal voltage of the battery under different charging rates. Four groups of voltage curves obtained by experiment are as shown in Figure 3 , plus the OCV-SOC voltage curve of HPPC test, a total of five voltage curves.

[0125] When the battery works in the constant current segment, the battery is in a relatively stable polarization state, showing stable characteristics, and the dynamic polarization characteristics are hidden, which can be represented by a pure resistor R S .

[0126] U t = OCV + I L R0+U1+U2≈OCV+I L (R0+R1+R2)≈OCV+I L R s

[0127] Referring to Figure 4 , R s is obtained by experimental test on the charging segment CCCV as prior information of the entire model, and the model is switched to the prior charging model when the battery works in the CCCV mode charging segment. The voltage on the entire R s is the difference between U t and OCV(ξ), i.e.:

[0128] ΔU s (ξ)=U t (ξ)-OCV(ξ)

[0129] R s (ξ, I C )=ΔU s (ξ) / I C

[0130] I c is the battery current, and note that C means constant, so the current is represented as I c .

[0131] The specific identification method is to convert the voltage quantity into a function of ξ, which only requires the integration of current with respect to time.

[0132] Observation Figure 3 where OCV is the lowermost curve, U t The four curves above are at four different charge rates, and they are all functions of the horizontal coordinate SOC, ΔU s is the voltage difference between two curves at the same horizontal coordinate. At this time, R s The value is specifically expressed as this voltage difference ΔU s Divided by the current I c at this time.

[0133] At this time, the model of the entire battery can be represented in a discrete system as when the battery is in the CCCV working condition, the model uses Figure 4 The CCCV correction circuit model for the charging section in model I.b is:

[0134] x1(k+1) = A1x1(k) + B1u1(k) + w(k)

[0135] y1(k) = g1(x1(k), u1(k)) + v(k);

[0136]

[0137] g1(x1(k), u1(k)) = OCV(ξ(k)) + (I L (k) - I SC (k))R s ;

[0138] where x1(k+1) represents the state variable of the CCCV correction circuit model for the charging section at the k+1th iteration, A1 represents the state variable matrix of the CCCV correction circuit model for the charging section acting on x1(k+1), x1(k) represents the state variable of the CCCV correction circuit model for the charging section at the kth iteration, x1(k) = [ξ(k) Isc(k)] T , B1 represents the input control matrix of the CCCV correction circuit model for the charging section acting on u1(k), u1(k) represents the input variable of the CCCV correction circuit model for the charging section at the kth iteration, u1(k) = I L (k), y1(k) represents the observation variable of the CCCV correction circuit model for the charging section at the kth iteration, y1(k) = U t (k), g1() represents the observation function of the CCCV correction circuit model for the charging section, R s represents the resistance value of the CCCV correction circuit model for the charging section.

[0139] Where ΔT is the sampling time of the discrete system (1 second), and R s (ξ,I c ) for battery and constant current I C The function is derived from the prior information identification of CCCV segment tests.

[0140] After establishing the CCCV correction model for the charging section, refer to Figure 5 As shown, only the battery voltage and current signals need to be sampled, and the short-circuit current I in the direct observation model can be obtained based on the Kalman filter (EKF). SC The short-circuit resistance is identified online using the Recursive Least Squares (RLS) method. The Kalman Filter (KF) algorithm is a minimum variance estimation algorithm applied to linear systems, capable of estimating multidimensional state variables. However, lithium-ion batteries are nonlinear systems, making the standard Kalman Filter algorithm unsuitable. Therefore, the Extended Kalman Filter (EKF) algorithm is used for state estimation of short-circuit information, which involves linearizing the system's state-space expression using Taylor series expansion. The specific recursive process of EKF is as follows:

[0141] Set the iteration number k to 0;

[0142] The initial conditions are set as follows:

[0143] in, Let x0 represent the posterior state variable in iteration 0, x0 represent the initial random setting value, and E[] represent taking the mean. This represents the posterior error covariance matrix at iteration 0;

[0144] Increment the value of k by 1;

[0145] Based on the posterior state estimate of the (k-1)th iteration, the prior state estimate of the kth iteration is determined as follows:

[0146]

[0147] in, Let f represent the prior state variables of the target model in the k-th iteration, and let f() represent a mapping of the target model. Let u represent the posterior state variable of the target model in the (k-1)th iteration. k-1 This represents the input variables of the target model in the (k-1)th iteration.

[0148] The target model is a second-order RC equivalent circuit model or a charging segment CCCV correction circuit model.

[0149] determining a priori error covariance matrix of the kth iteration based on the a priori state estimation of the kth iteration and the a posteriori error covariance matrix of the (k-1)th iteration as:

[0150]

[0151] wherein, represents the a priori error covariance matrix of the target model at the kth iteration, represents the state variable matrix of the target model at the (k-1)th iteration, represents the a posteriori error covariance matrix of the target model at the (k-1)th iteration, represents the transpose of the state variable matrix of the target model at the (k-1)th iteration, Q k represents the process error of the target model at the kth iteration.

[0152] determining a gain matrix of the kth iteration based on the a priori error covariance matrix of the kth iteration as:

[0153]

[0154] wherein, K k represents the Kalman gain of the target model at the kth iteration that minimizes the variance of the estimation and the measurement, represents the transpose of the observation matrix of the target model at the kth iteration, R k represents the observation error matrix of the target model at the kth iteration, represents the observation matrix of the target model at the kth iteration.

[0155] determining a posteriori state estimation of the kth iteration based on the a priori state estimation of the kth iteration and the gain matrix of the kth iteration as:

[0156]

[0157] wherein, represents the a posteriori state variable of the target model at the kth iteration, y k represents the sampled voltage value of the target model at the kth iteration, g() represents the observation function of the target model.

[0158] determining a posteriori error covariance matrix of the kth iteration based on the a posteriori state estimation of the kth iteration, the a priori error covariance matrix of the kth iteration and the gain matrix of the kth iteration as:

[0159]

[0160] wherein, represents the a posteriori error covariance matrix of the target model at the kth iteration, I represents the unit matrix.

[0161] determining whether a preset iteration number is reached to obtain a second determination result.

[0162] If the second determination result indicates no, the value of k is increased by 1, and the step of "determining the prior state estimation of the kth iteration based on the posterior state estimation of the (k-1)th iteration" is returned.

[0163] If the second determination result indicates yes, the posterior state estimation of the kth iteration is output, and the short-circuit current of the battery is contained in the posterior state estimation of the kth iteration.

[0164] The flowchart of the whole internal short circuit identification algorithm is shown in Figure 6 The terminal voltage U of the battery is measured t and the current I L When the battery works in different working conditions, the model is switched, the modified model of the CCCV working condition is used when the battery works in constant current charging, the short-circuit current I SC is directly estimated by EKF, and the short-circuit information is obtained.

[0165] Embodiment 2

[0166] The embodiment of the application provides a battery internal short circuit rapid diagnosis method under constant current charging working condition, which comprises the following steps:

[0167] A CCCV modified circuit model of the battery is constructed.

[0168] The CCCV modified circuit model of the battery is observed based on an extended Kalman filter algorithm, and the short-circuit current I SC of the battery is obtained.

[0169] The short-circuit current I SC is used to identify the short-circuit resistance R s of the battery by using a recursive least square method.

[0170] In the specific implementation, the CCCV modified circuit model of the battery is constructed, and specifically comprises the following steps:

[0171] The internal resistance of the battery is equivalent to a resistance R s , and a second equivalent circuit of the battery is constructed.

[0172] The parameters of the second equivalent circuit are identified based on an OCV-SOC curve of the battery and different rate CCCV curves, the CCCV modified circuit model of the charging section is constructed, the OCV-SOC curve is obtained by pulse discharge testing of the battery, and the different rate CCCV curves are obtained by CCCV charging testing of the battery under different rates.

[0173] The CCCV modified circuit model of the charging section is:

[0174] x1(k+1) = A1x1(k) + B1u1(k) + w(k)

[0175] y1(k) = g1(x1(k), u1(k)) + v(k);

[0176]

[0177] g1(x1(k), u1(k)) = OCV(ξ(k)) + (I L (k) - I SC (k))R s ;

[0178] wherein x1(k+1) represents the state variable of the charging section CCCV correction circuit model at the k+1th iteration, A1 represents the state variable matrix of the charging section CCCV correction circuit model acting on x1(k+1), x1(k) represents the state variable of the charging section CCCV correction circuit model at the kth iteration, x1(k) = [ξ(k) Isc(k)] T , ξ(k) represents the state of charge of the battery at the kth iteration, I sc (k) represents the short circuit current value at the kth iteration, B1 represents the input control matrix of the charging section CCCV correction circuit model acting on u1(k), u1(k) represents the input variable of the charging section CCCV correction circuit model at the kth iteration, u1(k) = I L (k), I L (k) represents the current of the battery, w(k) represents the system noise at the kth iteration, y1(k) represents the observation variable of the charging section CCCV correction circuit model at the kth iteration, y1(k) = U t (k), U t (k) represents the terminal voltage of the battery, g1() represents the observation function of the charging section CCCV correction circuit model, v(k) represents the measurement noise at the kth iteration, η represents, ΔT represents the sampling time, Q0 represents the battery capacity, OCV() represents the open circuit voltage function of the battery, R s represents the resistance value of the charging section CCCV correction circuit model.

[0179] Embodiment 3

[0180] The battery internal short circuit rapid diagnosis system provided by the embodiment of the application is applied to the method in embodiment 1, and the system comprises:

[0181] A building module is configured to build a second-order RC equivalent circuit model of the battery and a charging section CCCV correction circuit model.

[0182] A measuring module is configured to measure the terminal voltage Ut and current I L .

[0183] The judgment module is used to determine the terminal voltage U. t and current I L Determine whether the battery is in a constant current charging condition to obtain a first judgment result.

[0184] The calculation module is used to obtain the battery's short-circuit current I based on the extended Kalman filter. SC And used based on the short-circuit current I SC Identifying the short-circuit resistance R of the battery using the recursive least squares method S .

[0185] The calculation module specifically includes:

[0186] If the judgment result is negative, then the second-order RC equivalent circuit model is observed based on the extended Kalman filter algorithm to obtain the battery's short-circuit current I. SC .

[0187] If the judgment result indicates yes, then the CCCV correction circuit model of the charging section is observed based on the extended Kalman filter algorithm to obtain the short-circuit current I of the battery. SC .

[0188] The short-circuit resistance R of the battery is identified using the recursive least squares method based on the short-circuit current. S .

[0189] Example 4

[0190] This invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements any of the methods in Embodiment 1.

[0191] Example 5

[0192] This invention provides a computer-readable storage medium storing a computer program, which, when executed, implements any of the methods in Embodiment 1.

[0193] The present invention has the following beneficial effects:

[0194] 1. The CCCV correction circuit model for the charging section, compared with the RC model identified by the traditional HPPC method, solves the problem of large error in the constant current section of the traditional model.

[0195] 2. The short-circuit current I SCAs a state variable, the short circuit information can be estimated within several sampling periods after the fault occurs by using the Kalman filter EKF direct observation, which greatly improves the speed of internal short circuit diagnosis compared with the traditional method of estimating abnormal changes of SOC.

[0196] The various embodiments are described in the specification by way of progression, each building on the last to facilitate ease of understanding. The same or similar reference numerals are used in the drawings and description to refer to the same or like parts, components and operations throughout.

[0197] The principles and implementations of the present application are described in the specification by using specific examples, and the above description of the examples is only used to help understand the method and core idea of the present application; at the same time, for those skilled in the art, according to the idea of the present application, the specific implementation and application range will be changed. In view of the above, the content of the specification should not be understood as a limitation of the present application.

Claims

1. A method for rapid diagnosis of internal short circuit in a battery, characterized by, The battery internal short circuit rapid diagnosis method comprises the following steps: a second-order RC equivalent circuit model and a charging segment CCCV correction circuit model of the battery are constructed; the terminal voltage and the current of the battery are measured; it is determined whether the battery is in a constant current charging working condition based on the terminal voltage and the current, and a first judgment result is obtained; if the judgment result indicates no, the second-order RC equivalent circuit model is observed based on an extended Kalman filtering algorithm, and a short circuit current of the battery is obtained; if the judgment result indicates yes, the charging segment CCCV correction circuit model is observed based on the extended Kalman filtering algorithm, and the short circuit current of the battery is obtained; the short circuit resistance of the battery is identified based on the short circuit current by using a recursive least square method.

2. The method of claim 1, wherein the method is characterized by: The construction of the second-order RC equivalent circuit model and the charging segment CCCV correction circuit model of the battery specifically comprises the following steps: the internal resistance of the battery is equivalent to two resistance-capacitance networks and one resistance in series, and a first equivalent circuit of the battery is constructed; parameters in the first equivalent circuit are identified by using an HPPC method, and a second-order RC equivalent circuit model is obtained; the internal resistance of the battery is equivalent to one resistance, and a second equivalent circuit of the battery is constructed; parameters of the second equivalent circuit are identified based on an OCV-SOC curve and different rate CCCV curves of the battery, and a charging segment CCCV correction circuit model is constructed; the OCV-SOC curve is obtained by pulse discharge testing of the battery, and the different rate CCCV curves are obtained by CCCV charging testing of the battery at different rates.

3. The method of claim 1, wherein the method is characterized by: The second-order RC equivalent circuit model is as follows: x2(k+1) = A2x2(k) + B2u2(k) + w(k) y2(k) = g2(x2(k), u2(k)) + v(k) g2(x2(k), u2(k)) = OCV(ξ(k)) + U1(k) + U2(k) + (I L (k) - I SC (k))R0; wherein x2(k+1) represents a state variable of the second-order RC equivalent circuit model at the k+1th iteration, A2 represents a state variable matrix of the second-order RC equivalent circuit model acting on x2(k+1), x2(k) represents a state variable of the second-order RC equivalent circuit model at the kth iteration, x2(k) = [ξ(k) U1(k) U2(k) I sc (k)] T ,ξ(k) represents a state of charge of the battery at the kth iteration, U1(k) represents a voltage value of a first RC network of the second-order RC equivalent circuit model at the kth iteration, U2(k) represents a voltage value of a second RC network of the second-order RC equivalent circuit model at the kth iteration, I sc (k) represents a short-circuit current value at the kth iteration, B2 represents an input control matrix of the second-order RC equivalent circuit model acting on u2(k), u2(k) represents an input variable of the second-order RC equivalent circuit model at the kth iteration, u2(k) = I L (k), I L (k) represents a current of the battery, w(k) represents a system noise at the kth iteration, y2(k) represents an observation variable of the second-order RC equivalent circuit model at the kth iteration, y2(k) = U t (k), U t (k) represents a terminal voltage of the battery, g2() represents an observation function of the second-order RC equivalent circuit model, v(k) represents a measurement noise at the kth iteration, η represents a charging efficiency of the battery, ΔT represents a sampling time, Q0 represents a battery capacity, τ1 represents a time constant of the first RC network of the second-order RC equivalent circuit model, τ1 = R1C1, R1 represents a resistance value of the first RC network of the second-order RC equivalent circuit model, C1 represents a capacitance value of the first RC network of the second-order RC equivalent circuit model, τ2 represents a time constant of the second RC network of the second-order RC equivalent circuit model, τ2 = R2C2, R2 represents a resistance value of the second RC network of the second-order RC equivalent circuit model, C2 represents a capacitance value of the second RC network of the second-order RC equivalent circuit model, OCV() represents an open-circuit voltage function of the battery, and R0 represents a resistance value of the second-order RC equivalent circuit model. The charging segment CCCV correction circuit model is as follows: x1(k+1) = A1x1(k) + B1u1(k) + w(k) y1(k) = g1(x1(k), u1(k)) + v(k) g1(x1(k),u1(k)) = OCV(ξ(k)) + (I L (k) - I SC (k))R s ; wherein x1(k+1) represents a state variable of the charging section CCCV correction circuit model at the (k+1)th iteration, A1 represents a state variable matrix of the charging section CCCV correction circuit model acting on x1(k+1), x1(k) represents a state variable of the charging section CCCV correction circuit model at the kth iteration, x1(k) = [ξ(k) Isc(k)] T B1 represents an input control matrix of the charging section CCCV correction circuit model acting on u1(k), u1(k) represents an input variable of the charging section CCCV correction circuit model at the kth iteration, u1(k) = I L (k), y1(k) represents an observation variable of the charging section CCCV correction circuit model at the kth iteration, y1(k) = U t (k), g1() represents an observation function of the charging section CCCV correction circuit model, R s represents a resistance value of the charging section CCCV correction circuit model.

4. The method of claim 3, wherein the step of determining the internal short circuit is performed by measuring a voltage of the battery. The extended Kalman filtering EKF specifically comprises the following steps: the value of iteration times k is set to 0; an initial condition is set as follows: wherein, represents the posterior state variable at 0th iteration, xo represents an initial random set value, E[] represents taking the mean, represents the posterior error covariance matrix at 0th iteration; the value of k is increased by 1; the prior state estimation of the kth iteration is determined based on the posterior state estimation of the (k-1)th iteration as follows: wherein, represents the prior state variable of the target model at the kth iteration, f() represents a mapping of the target model, represents the posterior state variable of the target model at the k-1th iteration, u k-1 represents the input variable of the target model at the k-1th iteration; The target model is the second-order RC equivalent circuit model or the charging segment CCCV correction circuit model; the prior error covariance matrix of the kth iteration is determined based on the prior state estimation of the kth iteration and the posterior error covariance matrix of the (k-1)th iteration as follows: wherein, represents the prior error covariance matrix of the target model at the kth iteration, represents the state variable matrix of the target model at the k-1th iteration, represents the posterior error covariance matrix of the target model at the k-1th iteration, represents the transpose of the state variable matrix of the target model at the k-1th iteration, Q k represents the process error of the target model at the kth iteration; the gain matrix of the kth iteration is determined based on the prior error covariance matrix of the kth iteration as follows: where K k represents the Kalman gain of the target model at the kth iteration that minimizes the estimation variance from the measurement, represents the transpose of the observation matrix of the target model at the kth iteration, R k represents the observation error matrix of the target model at the kth iteration, represents the observation matrix of the target model at the kth iteration; the posterior state estimation of the kth iteration is determined based on the prior state estimation of the kth iteration and the gain matrix of the kth iteration as follows: wherein, represents the posterior state variable of the target model at the kth iteration, y k represents the sampled voltage value of the target model at the kth iteration, g() represents the observation function of the target model; the posterior error covariance matrix of the kth iteration is determined based on the posterior state estimation of the kth iteration, the prior error covariance matrix of the kth iteration and the gain matrix of the kth iteration as follows: wherein, denotes the posterior error covariance matrix of the target model at the kth iteration, and I denotes the identity matrix; it is determined whether the preset iteration times are reached, and a second judgment result is obtained. If the second determination result indicates no, the value of k is increased by 1, and the step of "determining the prior state estimation of the kth iteration based on the posterior state estimation of the (k-1)th iteration" is returned to; If the second determination result indicates yes, the posterior state estimation of the kth iteration is outputted; the short-circuit current of the battery is contained in the posterior state estimation of the kth iteration.

5. A method for diagnosing a short circuit inside a battery in a constant current charging operation, characterized by, The battery internal short-circuit rapid diagnosis method under the constant current charging condition comprises the following steps: A charging section CCCV correction circuit model of the battery is constructed; The charging section CCCV correction circuit model is observed based on an extended Kalman filtering algorithm to obtain the short-circuit current of the battery; The short-circuit resistance of the battery is identified based on the short-circuit current by using a recursive least square method; The charging section CCCV correction circuit model is as follows: x1(k+1)=A1x1(k)+B1u1(k)+w(k) y1(k)=g1(x1(k),u1(k))+v(k); g1(x1(k),u1(k)) = OCV(ξ(k)) + (I L (k) - I SC (k))R s ; wherein x1(k+1) denotes a state variable of the charge phase CCCV correction circuit model at the (k+1)th iteration, A1 denotes a state variable matrix of the charge phase CCCV correction circuit model acting on x1(k+1), x1(k) denotes a state variable of the charge phase CCCV correction circuit model at the kth iteration, x1(k) = [ξ(k) Isc(k)] T , ξ(k) denotes a state of charge of the battery at the kth iteration, Isc(k) denotes a short-circuit current value at the kth iteration, sc (k) denotes a state of charge of the battery at the kth iteration, Isc(k) denotes a short-circuit current value at the kth iteration, L (k), I L (k) denotes a state of charge of the battery at the kth iteration, Isc(k) denotes a short-circuit current value at the kth iteration, t (k), U t (k) denotes a state of charge of the battery at the kth iteration, Isc(k) denotes a short-circuit current value at the kth iteration, s (k) denotes a state of charge of the battery at the kth iteration, Isc(k) denotes a short-circuit current value at the kth iteration, 6. The method of claim 5, wherein the method is characterized by: The construction of the charging section CCCV correction circuit model of the battery specifically comprises: The internal resistance of the battery is equivalent to a resistor to construct a second equivalent circuit of the battery; Parameters of the second equivalent circuit are identified based on an OCV-SOC curve and different rate CCCV curves of the battery to construct the charging section CCCV correction circuit model; the OCV-SOC curve is obtained by pulse discharge testing of the battery, and the different rate CCCV curves are obtained by CCCV charging testing of the battery under different rates.

7. A battery internal short circuit rapid diagnosis system characterized by, The battery internal short-circuit rapid diagnosis system is applied to the method of any one of claims 1-4, and the system comprises: A establishing module is configured to construct a second-order RC equivalent circuit model and a charging section CCCV correction circuit model of the battery; A measuring module is configured to measure the terminal voltage and current of the battery; A judging module is configured to determine whether the battery is in a constant current charging condition based on the terminal voltage and current to obtain a first determination result; A calculating module is configured to obtain the short-circuit current of the battery based on an extended Kalman filtering algorithm and to identify the short-circuit resistance of the battery based on the short-circuit current by using a recursive least square method; The calculating module specifically comprises: If the determination result indicates no, the second-order RC equivalent circuit model is observed based on an extended Kalman filtering algorithm to obtain the short-circuit current of the battery; If the determination result indicates yes, the charging section CCCV correction circuit model is observed based on an extended Kalman filtering algorithm to obtain the short-circuit current of the battery; The short-circuit resistance of the battery is identified based on the short-circuit current by using a recursive least square method.

8. An electronic device, comprising: The storage medium stores a computer program, and the computer program is executed to implement the method of any one of claims 1-4.

9. A computer-readable storage medium, characterized in that, The storage medium stores a computer program, and the computer program is executed to implement the method of any one of claims 1-4.

Images