A method, device and electronic device for estimating battery state of charge

Abstract

The present invention discloses a method, device and electronic device for estimating the state of charge of a battery, wherein the internal temperature of the battery and the state of charge of the battery output at the current moment are fed back as the input state quantity in the process of estimating the state of charge of the battery and the input state quantity in the process of estimating the internal temperature of the battery at the next moment, respectively, and the estimation calculation is performed after updating the respective models. The internal temperature of the battery and the state of charge of the battery are mutually influential and interrelated, and the influence of the internal temperature of the battery on the estimation of the state of charge is combined, and the state of charge of the battery is estimated by using an adaptive extended Kalman filter algorithm, so as to realize the coupled estimation of the state of charge of the battery and the internal temperature, thereby improving the accuracy of the state of charge estimation result.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of battery technology, and in particular to a method, device and electronic device for estimating the state of charge of a battery. Background Art

[0002] With the development of battery technology, electric vehicles have become popular. As the energy source of electric vehicles, the state of charge (SOC) of the battery can reflect the remaining power of the lithium battery. Accurate estimation of the battery SOC value is of great significance to improving battery safety and energy utilization.

[0003] However, the existing estimation methods still have certain shortcomings and cannot reflect the SOC of the battery well. The estimation effect is poor, which is not conducive to the further development of battery technology. Summary of the invention

[0004] The present invention provides a method and device for estimating the state of charge of a battery and electronic equipment, which realize coupled estimation of the battery SOC and internal temperature and improve the accuracy of the SOC estimation result.

[0005] In a first aspect, an embodiment of the present invention provides a method for estimating a state of charge of a battery, comprising:

[0006] Establishing an equivalent model; wherein the equivalent model includes a battery equivalent circuit model and a battery equivalent thermal circuit model;

[0007] The internal temperature of the battery at the kth moment is estimated according to the battery equivalent thermal circuit model and the first adaptive extended Kalman filter algorithm; the state of charge of the battery at the kth moment is estimated according to the battery equivalent circuit model and the second adaptive extended Kalman filter algorithm;

[0008] The battery internal temperature at the kth moment is used as the input state quantity at the k+1th moment in the battery state of charge estimation process, and the battery state of charge at the k+1th moment is estimated based on the second adaptive extended Kalman filter algorithm; the battery state of charge at the kth moment is used as the input state quantity in the battery internal temperature estimation process, and the battery internal temperature at the k+1th moment is estimated based on the first adaptive extended Kalman filter algorithm; wherein k is a positive integer greater than 0;

[0009] The estimated results of the battery internal temperature at the k+1th moment and the battery state of charge at the k+1th moment are fed back again as the input state quantity in the subsequent corresponding battery state of charge estimation process and the input state quantity in the battery internal temperature estimation process, and are calculated based on the second adaptive extended Kalman filter algorithm, so that the battery state of charge at each moment can be estimated.

[0010] Optionally, estimating the internal temperature of the battery at the kth moment according to the battery equivalent thermal circuit model and the first adaptive extended Kalman filter algorithm includes:

[0011] According to the discretized state equation of the battery equivalent thermal circuit model, the predicted battery internal temperature at the kth moment is obtained based on the state quantity prediction of the battery internal temperature at the k-1th moment;

[0012] Obtaining a first Kalman gain at the kth moment according to the first error covariance and the first system noise covariance;

[0013] The predicted internal battery temperature at the kth moment is corrected according to the first gain factor, the first Kalman gain at the kth moment and the first observation quantity innovation rate to obtain the internal battery temperature at the kth moment, wherein the first gain factor is positively correlated with the first observation quantity innovation rate.

[0014] Optionally, using the internal temperature of the battery at the kth moment as the input state quantity at the k+1th moment in the battery state of charge estimation process, estimating the battery state of charge at the k+1th moment based on the second adaptive extended Kalman filter algorithm, includes:

[0015] updating the capacitance and resistance parameters in the battery equivalent circuit model according to the battery internal temperature and capacitance and resistance parameter identification conditions at the kth moment;

[0016] The battery state of charge at the k+1th moment is estimated according to the updated battery equivalent circuit model and the second adaptive extended Kalman filter algorithm.

[0017] Optionally, estimating the battery state of charge at the k+1th moment according to the updated battery equivalent circuit model and the second adaptive extended Kalman filter algorithm includes:

[0018] According to the updated state equation of the battery equivalent circuit model, based on the state quantity prediction of the battery state of charge at the kth moment, a predicted battery state of charge at the k+1th moment is obtained;

[0019] Updating the second error covariance and the second system noise covariance, and obtaining the second Kalman gain at the k+1th moment according to the updated second error covariance and the second system noise covariance; wherein the update calculation of the second system noise covariance does not include a subtraction operation of the second error covariance;

[0020] The predicted battery state of charge at the k+1th moment is corrected according to the second gain factor, the second Kalman gain at the k+1th moment and the second observation value innovation rate to obtain the battery state of charge at the k+1th moment, wherein the second gain factor is positively correlated with the second observation value innovation rate.

[0021] Optionally, estimating the battery state of charge at the kth moment according to the battery equivalent circuit model and the second adaptive extended Kalman filter algorithm includes:

[0022] According to the state equation of the battery equivalent circuit model, based on the state quantity of the battery state of charge at the k-1th moment, a predicted battery state of charge at the kth moment is obtained;

[0023] Obtaining a second Kalman gain at the kth moment according to the second error covariance and the second system noise covariance;

[0024] The predicted battery state of charge at the kth moment is corrected according to the second gain factor, the second Kalman gain at the kth moment and the second observation value innovation rate to obtain the battery state of charge at the kth moment, wherein the second gain factor is positively correlated with the second observation value innovation rate.

[0025] Optionally, the battery state of charge at the kth moment is used as an input state quantity in the battery internal temperature estimation process, and the battery internal temperature at the k+1th moment is estimated based on the first adaptive extended Kalman filter algorithm, including:

[0026] Updating the heat generation power in the battery equivalent thermal circuit model according to the battery state of charge at the kth moment;

[0027] The internal temperature of the battery at the k+1th moment is estimated according to the updated battery equivalent thermal circuit model and the first adaptive extended Kalman filter algorithm.

[0028] Optionally, estimating the internal temperature of the battery at the k+1th moment according to the updated battery equivalent thermal circuit model and the first adaptive extended Kalman filter algorithm includes:

[0029] According to the discretized state equation of the battery equivalent thermal circuit model, the predicted battery internal temperature at the k+1th moment is obtained based on the state quantity of the battery internal temperature at the kth moment;

[0030] Updating a first error covariance and a first system noise covariance, and obtaining a first Kalman gain at the k+1th moment according to the updated first error covariance and the second system noise covariance; wherein the update calculation of the first system noise covariance does not include a subtraction operation of the first error covariance;

[0031] The predicted internal temperature of the battery at the k+1th moment is corrected according to the first gain factor, the first Kalman gain at the k+1th moment and the first observation value innovation rate to obtain the internal temperature of the battery at the k+1th moment, wherein the first gain factor is positively correlated with the first observation value innovation rate.

[0032] In a second aspect, an embodiment of the present invention provides a battery state of charge estimation device, comprising:

[0033] A model unit, used to establish an equivalent model; wherein the equivalent model includes a battery equivalent circuit model and a battery equivalent thermal circuit model;

[0034] a first estimation unit, configured to estimate the internal temperature of the battery at the kth moment according to the battery equivalent thermal circuit model and the first adaptive extended Kalman filter algorithm; and to estimate the state of charge of the battery at the kth moment according to the battery equivalent circuit model and the second adaptive extended Kalman filter algorithm;

[0035] a second estimation unit, configured to use the battery internal temperature at the kth moment as an input state quantity at the k+1th moment in the battery state of charge estimation process, and estimate the battery state of charge at the k+1th moment based on the second adaptive extended Kalman filter algorithm; use the battery state of charge at the kth moment as an input state quantity in the battery internal temperature estimation process, and estimate the battery internal temperature at the k+1th moment based on the first adaptive extended Kalman filter algorithm; wherein k is a positive integer greater than 0;

[0036] The circulation unit is used to feed back the results of the estimated internal temperature of the battery at the k+1th moment and the state of charge of the battery at the k+1th moment as the input state quantity in the subsequent corresponding battery state of charge estimation process and the input state quantity in the battery internal temperature estimation process, and calculate based on the second adaptive extended Kalman filter algorithm, so as to estimate the battery state of charge at each moment.

[0037] In a third aspect, an embodiment of the present invention provides an electronic device, the electronic device comprising:

[0038] at least one processor; and,

[0039] a memory communicatively connected to the at least one processor; wherein,

[0040] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the battery state of charge estimation method described in any embodiment of the present invention.

[0041] The technical solution provided by the embodiment of the present invention is to calculate the internal temperature of the battery at the kth moment (current moment) in real time based on the equivalent thermal circuit model and the first adaptive extended Kalman filter algorithm, and to calculate the battery state of charge at the kth moment (current moment) in real time based on the equivalent circuit model and the second adaptive extended Kalman filter algorithm, and to feed back the internal temperature of the battery and the battery state of charge output at the current moment as the input state quantity in the battery state of charge estimation process and the input state quantity in the battery internal temperature estimation process at the next moment, respectively, and perform estimation calculation after updating the respective models. The internal temperature of the battery and the battery state of charge are mutually influential and interrelated, and the influence of the internal temperature of the battery on the state of charge estimation is combined, and the battery state of charge is estimated by using the adaptive extended Kalman filter algorithm, so as to realize the coupled estimation of the battery state of charge and the internal temperature, and improve the accuracy of the state of charge estimation result. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 A schematic flow chart of a method for estimating the state of charge of a battery is provided for an embodiment of the present invention;

[0043] Figure 2 A schematic diagram of the structure of a second-order RC equivalent circuit model provided in an embodiment of the present invention;

[0044] Figure 3 A schematic diagram of the structure of a battery equivalent thermal circuit model provided in an embodiment of the present invention;

[0045] Figure 4 A schematic flow chart of another method for estimating the state of charge of a battery is provided for an embodiment of the present invention;

[0046] Figure 5 A flowchart of coupled estimation of battery SOC and battery internal temperature based on an adaptive extended Kalman filter algorithm is provided for an embodiment of the present invention

[0047] Figure 6 A comparison chart of three SOC changes, namely, SOC(Tin), SOC(Ts) and measured SOC, provided in an embodiment of the present invention;

[0048] Figure 7 for Figure 6 A partial enlarged view of the dotted frame;

[0049] Figure 8 The error variation diagram between SOC(Tin), SOC(Ts) and the measured SOC provided in the embodiment of the present invention;

[0050] Fig. 9 A schematic diagram of the structure of a battery SOC estimation device provided by an embodiment of the present invention;

[0051] Fig.10 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0052] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are 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.

[0053] Due to the nonlinear characteristics of lithium batteries, SOC cannot be directly obtained through sensors. It must be estimated by measuring physical quantities such as lithium battery voltage, operating current, and lithium battery internal resistance and using certain mathematical methods. At present, the adaptive Kalman filter algorithm is used in related technologies to estimate the battery SOC, but the current Kalman filter algorithm estimation process uses the battery surface temperature as the measurement data. Since the battery surface temperature can be directly measured, but the internal temperature cannot be directly obtained, only the impact of the battery surface temperature on the SOC estimation result is considered, which is equivalent to estimating the single state of SOC. However, the surface temperature of the battery lags behind the internal temperature change during the charging and discharging process, so the internal temperature of the battery can better reflect the actual situation of the battery. Therefore, if only the single state of SOC is estimated separately, it will inevitably affect the accuracy of the final result.

[0054] In view of this, Figure 1 A flow chart of a method for estimating the state of charge of a battery is provided for an embodiment of the present invention. This embodiment is applicable to the case of estimating the SOC of a battery. The method can be executed by a device for estimating the state of charge of a battery, and the device can be implemented in hardware and / or software. The method specifically includes the following steps:

[0055] S110, establishing an equivalent model; wherein the equivalent model includes a battery equivalent circuit model and a battery equivalent thermal circuit model;

[0056] Among them, the battery equivalent circuit model adopts the second-order RC equivalent circuit model. Figure 2 FIG. 1 is a schematic diagram of a second-order RC equivalent circuit model provided in an embodiment of the present invention. Figure 2 , R0 represents the ohmic internal resistance of the battery, R1 and R2 represent the electrochemical polarization resistance and concentration polarization resistance of the battery respectively, C1 and C2 represent the electrochemical polarization capacitance and concentration polarization capacitance of the battery respectively. U1 and U2 represent the voltage across the electrochemical polarization capacitance and the voltage across the concentration polarization capacitance respectively, I is the current, Ut is the battery output voltage, U OCVIt represents the open circuit voltage of the battery. Figure 3 This is a schematic diagram of the structure of a battery equivalent thermal circuit model provided in an embodiment of the present invention, see Figure 3 , where H is the equivalent concentrated heat source inside the battery (heat generating power), Cc and Cs are the internal heat capacity and surface environment heat capacity, Rc and Rs are the internal thermal resistance and external thermal resistance, respectively, and T in and T s They are the internal temperature and surface temperature of the battery respectively.

[0057] S120, estimating the internal temperature of the battery at the kth moment according to the battery equivalent thermal circuit model and the first adaptive extended Kalman filter algorithm;

[0058] Specifically, the parameters of the internal heat capacity Cc, the surface environment heat capacity Cs, the internal thermal resistance Rc and the external thermal resistance Rs in the battery equivalent thermal circuit model change according to the ambient temperature or the internal temperature, and the parameters corresponding to the internal heat capacity Cc, the surface environment heat capacity Cs, the internal thermal resistance Rc and the external thermal resistance Rs can be obtained according to known identification means. Since the parameters of the internal heat capacity Cc, the surface environment heat capacity Cs, the internal thermal resistance Rc and the external thermal resistance Rs are mainly related to the material properties of the battery itself, and the external thermal resistance Rs is also related to the external environment in which the battery is located, the identification results of the internal heat capacity Cc, the surface environment heat capacity Cs and the internal thermal resistance Rc can be used as fixed values, and the identification results of the external thermal resistance Rs can be used only as the initial value of the external thermal resistance Rs. The parameters of the external thermal resistance Rs can be used as a state variable and estimated simultaneously with the internal temperature of the battery in the subsequent process to improve the adaptability of the model. The internal heat capacity Cc, the surface environment heat capacity Cs, the internal thermal resistance Rc and the external thermal resistance Rs are used together with other variables as input conditions of the first adaptive extended Kalman filter algorithm, and the battery equivalent thermal circuit model is discretized to obtain a discretized state equation.

[0059] According to the discretized state equation, the difference between the internal temperature of the battery and the ambient temperature, and the external thermal resistance Rs of the battery are taken as the state variables of the system, the difference between the surface temperature of the battery and the ambient temperature is taken as the observation variable of the system, and the heat generation power is taken as the external excitation. The estimated values ​​of the internal temperature of the battery and the external thermal resistance Rs of the battery are updated using the first adaptive extended Kalman filter algorithm to estimate the internal temperature of the battery at the kth moment.

[0060] The Kalman filter algorithm is a technology that iteratively updates the system state variables based on the principle of optimal filtering. It is a recursive estimator, that is, it uses the predicted state of the previous time step and the current measurement value to calculate the optimal estimate of the current state. Through a one-step prediction and one-step correction cyclic iteration method, the estimated value is constantly approaching the true value. The discretized state equation is used as the system state equation of the adaptive extended Kalman filter algorithm. By giving the data of the previous moment, the estimated value at the kth moment (that is, the internal temperature of the battery at the current moment) is calculated, and the estimated value is corrected by combining the error covariance and the Kalman gain.

[0061] In some embodiments, when correcting the state variable (internal battery temperature), a first gain factor can also be introduced. The first gain factor is positively correlated with the first observation value innovation rate. The first gain factor is used to change the weights of the state estimate and the measurement value to prevent the increase in the error of the estimation result due to the sudden change of the first observation value innovation, thereby ensuring the stability of the final estimation result. Among them, an exemplary method for setting the first gain factor is: when the first observation value innovation rate is in the third preset range, the size of the gain factor should be set to the third preset value, thereby weakening the correction of the actual measurement value to the expected result; conversely, when the first observation value innovation rate is in the fourth preset range, it should be set to the fourth preset value to strengthen the correction of the actual measurement value to the estimation result, wherein the fourth preset value is greater than the third preset value, and when the left side of the numerical axis is the positive direction, the third preset range is within the right range of the fourth preset value. Among them, the first gain factor can also add the influence of temperature on it. The specific size setting can be debugged according to the simulation test results until the final result meets the requirements.

[0062] S130, using the internal temperature of the battery at the kth moment as the input state quantity at the k+1th moment in the battery state of charge estimation process, and estimating the battery state of charge at the k+1th moment based on the second adaptive extended Kalman filter algorithm; wherein k is a positive integer greater than 0.

[0063] Specifically, the internal temperature of the battery at the kth moment is used as the input state quantity at the k+1th moment in the battery SOC estimation process, and the parameters of the battery equivalent circuit model at the internal temperature of the battery are identified. According to the identification relationship, the corresponding identification parameters that need to be updated are obtained, for example, the corresponding parameters of the variables such as the electrochemical polarization resistance R1 and the concentration polarization resistance R2 of the battery, the actual capacity Ca of the battery, the electrochemical polarization capacitance C1 and the concentration polarization capacitance C2 of the battery, so as to update the parameters in the battery equivalent circuit model. In the SOC estimation process, the SOC, the voltage U1 across the electrochemical polarization capacitor, and the voltage U2 across the concentration polarization capacitor are set as state variables, the output terminal voltage Ut of the battery is set as an observed quantity, and the current I is used as an external excitation. The estimated value of the SOC of the battery is updated using the second adaptive extended Kalman filter algorithm, and the battery SOC at the k+1th moment is estimated, so as to realize the coupled estimation of the battery SOC and the internal temperature.

[0064] In some embodiments, when correcting the state variable (battery SOC), a second gain factor can also be introduced. The second gain factor is positively correlated with the second observation value innovation rate. The second gain factor is used to change the weight of the state estimate and the measurement value to prevent the error of the estimation result from increasing due to the sudden change of the observation value innovation, thereby ensuring the stability of the final estimation result. Among them, an exemplary second gain factor setting method is: when the second observation value innovation rate is in the first preset range, the size of the second gain factor should be set to the first preset value, thereby weakening the correction of the actual measurement value to the expected result; conversely, when the second measurement innovation rate is in the second preset range, it should be set to the second preset value to strengthen the correction of the actual measurement value to the estimation result, wherein the second preset value is greater than the first preset value, and when the left side of the numerical axis is the positive direction, the first preset range is within the right range of the second preset value. Among them, the second gain factor can also add the influence of temperature on it. It should be noted that the first adaptive extended Kalman filter algorithm in the above embodiment refers to the Kalman filter algorithm used in the process of estimating the internal temperature of the battery, and the second adaptive extended Kalman filter algorithm refers to the Kalman filter algorithm used in the process of estimating the battery state of charge.

[0065] The technical solution provided by the embodiment of the present invention calculates the internal temperature of the battery at the kth moment (the current moment) in real time based on the equivalent thermal circuit model and the adaptive extended Kalman filter algorithm, and uses the output internal temperature of the battery as the temperature input condition of the SOC estimation process at the k+1th moment (the next moment), adjusts the model parameters, combines the influence of the internal temperature of the battery on the SOC estimation, and uses the adaptive extended Kalman filter algorithm to estimate the battery SOC, realizes the coupled estimation of the battery SOC and the internal temperature, and improves the accuracy of the SOC estimation result.

[0066] Based on the above embodiments, several forms of parameter identification relationships of the battery equivalent circuit model are listed, which are not specific limitations. Specifically, the actual capacity Ca can be obtained by charging and discharging the battery at a certain rate under different temperature conditions, and the actual capacity of the battery under different temperature conditions can be fitted with the temperature to obtain the relationship between the two: Ca(T) = f(T);

[0067] Battery SOC-U OCV The relationship can be obtained by placing the battery at a certain temperature and SOC point for a long time, and recording the value when the battery voltage reaches a stable state as the open circuit voltage of the battery under the current conditions. By changing the temperature conditions and SOC point, the open circuit voltage U corresponding to different temperatures T and different SOC can be obtained. OCV , and fit the relationship between the three, and the open circuit voltage U OCV As the dependent variable, taking temperature T and SOC as independent variables, the relationship between the three can be obtained: OCV (SOC, T) = G(SOC, T);

[0068] In order to identify the capacitance and resistance parameters of the battery equivalent circuit model, it is necessary to perform a hybrid power pulse characteristic (HPPC) test on the experimental battery, and then process the output voltage and current data, and use the least squares method to identify the ohmic internal resistance R0, electrochemical reaction polarization internal resistance R1, concentration polarization internal resistance R2, electrochemical reaction polarization capacitance C1, and concentration polarization capacitance C2. After testing under different temperatures and SOC conditions, the capacitance and resistance parameters under different temperatures and SOC conditions can be obtained: R0=f1(SOC, T); R1=f2(SOC, T); R2=f3(SOC, T); C1=f4(SOC, T); C2=f5(SOC, T).

[0069] Figure 4 A flowchart of another method for estimating the state of charge of a battery is provided for an embodiment of the present invention. Figure 4 ,include:

[0070] S210, establishing an equivalent model; wherein the equivalent model includes a battery equivalent circuit model and a battery equivalent thermal circuit model;

[0071] S220, estimating the internal temperature of the battery at the kth moment according to the battery equivalent thermal circuit model and the first adaptive extended Kalman filter algorithm; estimating the battery state of charge at the kth moment according to the battery equivalent circuit model and the second adaptive extended Kalman filter algorithm;

[0072] Specifically, the battery internal temperature at the kth moment is estimated based on the equivalent thermal circuit model and the first adaptive extended Kalman filter algorithm, and the battery SOC at the kth moment is calculated in real time based on the equivalent circuit model and the second adaptive extended Kalman filter algorithm.

[0073] S230. Using the internal temperature of the battery at the kth moment as the input state quantity at the k+1th moment in the process of estimating the battery state of charge, and estimating the battery state of charge at the k+1th moment based on the second adaptive extended Kalman filter algorithm; using the battery state of charge at the kth moment as the input state quantity in the process of estimating the internal temperature of the battery, and estimating the battery internal temperature at the k+1th moment based on the first adaptive extended Kalman filter algorithm; wherein k is a positive integer greater than 0.

[0074] Specifically, the battery SOC calculation result output at the kth moment is used to update the heat generation power of the battery internal temperature estimation process at the k+1th moment, and the battery internal temperature at the k+1th moment is estimated based on the first adaptive extended Kalman filter algorithm. The battery internal temperature calculation result output at the kth moment is used as the input condition in the battery SOC estimation process at the k+1th moment, and the capacitance and resistance parameters in the equivalent circuit model are updated and adjusted, and the battery SOC at the k+1th moment is estimated based on the second adaptive extended Kalman filter algorithm.

[0075] S240. The estimated results of the battery internal temperature at the k+1th moment and the battery state of charge at the k+1th moment are fed back again as the input state quantity in the subsequent corresponding battery state of charge estimation process and the input state quantity in the battery internal temperature estimation process, and are calculated based on the second adaptive extended Kalman filter algorithm, so that the battery state of charge at each moment can be estimated.

[0076] Specifically, the outputted battery internal temperature and battery SOC are fed back as the input state quantity in the battery SOC estimation process and the input state quantity in the battery internal temperature estimation process at the next moment, respectively, and the respective models are updated before performing estimation calculations. The battery internal temperature and the battery SOC are mutually influential and interrelated. Therefore, through the iterative operation of the above process, the battery SOC at each moment is obtained, and the coupled estimation between the battery SOC and the internal temperature is realized, further improving the accuracy of the SOC estimation result.

[0077] Figure 5 A coupled estimation flow chart of battery SOC and battery internal temperature based on an adaptive extended Kalman filter algorithm is provided for an embodiment of the present invention. Figure 5, which includes a battery SOC estimation process 5000 based on a second adaptive extended Kalman filter algorithm and a battery internal temperature estimation process 5001 based on a first adaptive extended Kalman filter algorithm.

[0078] S501, establishing a battery equivalent circuit model;

[0079] Combination Figure 2 ,In the process of battery SOC estimation, according to Kirchhoff's law, the state equation of the second-order RC equivalent circuit can be expressed as:

[0080]

[0081]

[0082] Ut=U OCV -U1-U2-IR0

[0083] The above equation is further discretized, and the discrete calculation formula of SOC is listed:

[0084] U1(k)=exp[-ΔT / (R1×C1)]×U1(k-1)+R1×{1-exp[-ΔT / (R1×C1)]}×I(k-1) (1)

[0085] U2(k)=exp[-ΔT / (R2×C2)]×U2(k-1)+R2×{1-exp[-ΔT / (R2×C2)]}×I(k-1) (2)

[0086] Ut(k)=Ux CV (k)-U1(k)-U2(k)-R0×I(k) (3)

[0087] SOC(k)=SOC(k-1)+I(k-1)×ΔT / Ca (4)

[0088] Among them, ΔT is the sampling time, Ca is the rated capacity of the battery, R0 is the ohmic internal resistance of the battery, R1 and R2 are the electrochemical polarization resistance and concentration polarization resistance of the battery, C1 and C2 are the electrochemical polarization capacitance and concentration polarization capacitance of the battery, respectively. U1 and U2 are the voltage across the electrochemical polarization capacitance and the voltage across the concentration polarization capacitance, respectively, I is the current, Ut is the battery output voltage, and U OCV is the open circuit voltage of the battery, where the subscripts (k), (k-1) and (k+1) represent the discrete parameters of the corresponding parameters at the kth moment, the k-1th moment and the k+1th moment respectively.

[0089] In the SOC estimation process, let the state variable X(k) = [SOC(k), U1(k), U2(k)]T , the observation quantity Y(k) is the terminal voltage Ut(k) of the battery, the external excitation is I(k), according to the established second-order RC equivalent circuit model and the basic form of the Kalman filter algorithm, the discrete state equation and observation equation of the battery system in the SOC estimation process are listed in general form:

[0090] X(k)=A×X(k-1)+B×I(k-1)+ω(k); (5)

[0091] Ut(k)=C×X(k)+D×I(k)+ν(k)=U OCV -U1(k)-U2(k)-R0×I(k)+ν(k) (6)

[0092] ω(k) and ν(k) are the second system noise and the second observation noise of the system respectively.

[0093] From the above discrete equation for battery SOC calculation and the discrete state equations (1), (2), (3), and (4) of the battery second-order RC model, we can further obtain the SOC estimation coefficient matrix:

[0094]

[0095] C=[d(U OCV ) / d(SOC) -1 -1] D=-R0

[0096] Among them, τ1=R1×C1, τ2=R2×C2.

[0097] S502, initial state setting of SOC estimation;

[0098] When starting to estimate the SOC, the system initial state needs to be set, and the battery SOC, the second error covariance and the second system noise covariance are set to SOC0, P0, Q0 respectively. The voltage U1 across the electrochemical polarization capacitor and the voltage U2 across the concentration polarization capacitor are generally set to 0.

[0099] S503, battery SOC state prediction;

[0100] Based on the state X(k-1) at the previous moment, the current state X(k) is predicted, which can be expressed as:

[0101] X(k)=A×X(k-1)+B×I(k-1); where I(k-1) is the current at the previous moment.

[0102] S504, calculating the second error covariance matrix;

[0103] Based on the second error covariance P(k-1|k-1) updated at the previous moment, the second error covariance P(k|k-1) at the current moment is calculated:

[0104] P(k|k-1)=A×P(k-1|k-1)×AT+Q(k-1); wherein Q(k-1) is the second system noise covariance matrix at the k-1th moment.

[0105] S505, calculation of the second observation value innovation;

[0106] ε(k)=Ut(k)-C×X(k)-D×I(k-1);

[0107] S506, second Kalman gain calculation:

[0108] Kg(k)=P(k|k-1)×C T / [C×P(k|k-1)×C T +R]; where R is the second observation noise covariance.

[0109] S507, calculation of the second observation value innovation rate:

[0110] e(k)=|ε(k) / Ut(k)|×100%;

[0111] S508. Introduce a second gain factor, which is positively correlated with the second observation value innovation rate. Use the second gain factor to change the weights of the state estimate and the measurement value to prevent the increase of the estimation result error due to the sudden change of the second observation value innovation, and ensure the stability of the final estimation result. Among them, an exemplary method for setting the second gain factor is: when the second observation value innovation rate is in the first preset range, the size of the second gain factor should be set to the first preset value, thereby weakening the correction of the actual measurement value to the expected result; conversely, when the second observation value innovation rate is in the second preset range, it should be set to the second preset value to strengthen the correction of the actual measurement value to the estimation result, wherein the second preset value is greater than the first preset value, and when the left side of the numerical axis is the positive direction, the first preset range is within the right range of the second preset value. Among them, the second gain factor can also add the influence of temperature on it. The specific size setting can be debugged according to the simulation test results until the final result meets the requirements.

[0112] Second gain factor: θ(k)=f(e(k), T).

[0113] S509, battery SOC state estimation correction;

[0114] X(k|k)=X(k|k-1)+θ(k)×Kg(k)×ε(k);

[0115] Among them, θ(k)=f(e(k), T), X(k|k) is the state quantity after correction at the kth moment, X(k|k-1) is the state prediction quantity at the kth moment, that is, the state quantity before correction, θ(k) is the introduced second gain factor, Kg(k) is the second Kalman gain at the kth moment, and ε(k) is the second observation value update at the kth moment.

[0116] S510, updating the second error covariance to obtain an updated second error covariance P(k|k);

[0117] P(k|k)=[E-Kg(k)×C]×P(k|k-1);

[0118] S511, updating the second system noise covariance matrix;

[0119] In order to avoid the occurrence of non-positive definiteness during the updating process of the second system noise covariance, which leads to filtering divergence, the updating calculation formula of the second system noise covariance is improved. As the operation iteration proceeds, the second error covariance matrix will gradually decrease, or even approach zero. Therefore, in the embodiment of the present invention, the subtraction operation part related to the second error covariance in the original updating calculation formula of the second system noise covariance is simplified, which can ensure the positive definiteness of the second system noise covariance on the one hand, and further simplify the calculation process on the other hand.

[0120] The second system noise covariance matrix update is expressed as:

[0121] Q(k)=[1-d(k-1)]×Q(k-1)+d(k-1)×[Kg(k)×ε(k)×ε(k) T ×Kg(k) T ]; where the weight coefficient d(k) = 1-d / [1-d(k+1)], d is the forgetting factor, and its value is generally between 0.95 and 0.99; d(k-1) is the weight coefficient at the k-1th moment, and Q(k-1) is the second system noise covariance matrix at the k-1th moment.

[0122] Through the iterative calculation of the above process, the battery SOC state and the system noise characteristics can be synchronously updated, and the SOC calculation result obtained in step S509 is used as an input condition for estimating the internal temperature of the battery.

[0123] Exemplarily, the estimation process of the internal temperature of the battery is described as follows:

[0124] S512, establishing a battery equivalent thermal circuit model;

[0125] Combination Figure 3According to the law of heat conservation, a part of the heat generated inside the battery diffuses to the battery surface, and the remaining part causes the change of the internal temperature of the battery; a part of the heat diffused to the battery surface continues to diffuse to the external environment, and the remaining part causes the change of the battery surface temperature. Therefore, the discrete state equation of the battery thermal model can be obtained as follows:

[0126] Cc×d(T in -T amb ) / dt=H-(T in -T s ) / Rc (7)

[0127] Cs×d(T s -T amb ) / dt=(T in -T s ) / Rc-(T s -T amb ) / Rs (8)

[0128] Among them, T amb is the ambient temperature, T in and T s are the internal temperature and surface temperature of the battery, H is the equivalent concentrated heat source inside the battery (heat generating power), Cc and Cs are the internal heat capacity and surface environment heat capacity, Rc and Rs are the internal thermal resistance and external thermal resistance, respectively. ia =T in -T amb , T sa =T s -T amb , then the above formula can be further simplified as:

[0129] Cc×dT ia / dt=H-(T ia -T sa ) / Rc f9)

[0130] Cs×dT sa / dt=(T ia -T sa ) / Rc-T sa / Rs (10)

[0131] The heat generation power H can be calculated using the Bernardi battery heat generation equation:

[0132] H=I×(U OCV -Ut)+I×T×(dU OCV / dT) (11)

[0133] The open circuit voltage U OCV It can be calculated through its identification relationship with SOC.

[0134] The battery equivalent thermal circuit system is regarded as a linear discrete system, and the difference between the battery surface temperature and the ambient temperature Tsa is taken as the observed quantity. The general expressions of the system's state equation and observation equation are listed as follows:

[0135] α(k)=E×α(k-1)+J×W(k-1)+δ(k) (12)

[0136] β(k)=L×α(k)+N×W(k)+η(k) (13)

[0137] δ(k) and η(k) are the first system noise and the first observation noise respectively. Here, the external thermal resistance parameter of the battery is regarded as a slow variable, and its discrete state equation is expressed as follows:

[0138] Rs(k)=Rs(k-1)+π(k) (14)

[0139] Where π(k) is a small disturbance of the external thermal resistance. Discretizing the state equations (9) and (10) of the equivalent thermal circuit model, we can obtain:

[0140] T ia (k) = [1-Δt / (Rc×Cc)]×T ia (k-1)+At / (Rc×Cc)×T sa (k-1)+At / Cc×H (15)

[0141] T sa (k) = Δt / (Rc×Cs)×T ia (k-1)+[1-Δt / (Rc×Cs)-Δt / (Rs×Cs)]×T sa (k-1) (16)

[0142] Where Δt is the sampling time of the discrete process.

[0143] Let the state variable of the system α = [T ia T sa Rs] T , system observation quantity β = T sa , the system input W = H, then according to the above battery thermal system discrete equations (14), (15), (16) the coefficient matrix can be obtained:

[0144]

[0145]

[0146] S513, initial state setting of battery internal temperature estimation;

[0147] The first error covariance and the first system noise covariance are set to P'0 and Q'0 respectively, and the difference between the internal temperature of the battery and the ambient temperature is T ia , the difference between the battery surface temperature and the ambient temperature Tsa is generally set to 0, and the external thermal resistance Rs is set to Rs0.

[0148] S514, battery internal temperature prediction;

[0149] Based on the state α(k-1) at the previous moment, the current state α(k) is predicted, which can be expressed as:

[0150] α(k)=E×α(k-1)+J×W(k-1); wherein W(k-1) is the heat generation power at the previous moment.

[0151] S515, calculating the first error covariance matrix;

[0152] Based on the first error covariance P′(k-1|k-1) updated at the previous moment, the first error covariance P′(k|k-1) at the current moment is calculated:

[0153] P′(k|k-1)=E×P′(k-1|k-1)×ET+Q′(k-1); wherein Q′(k-1) is the first system noise covariance matrix at the k-1th moment.

[0154] S516, calculation of the first observation value innovation;

[0155] ε'(k)=β(k)-L×α(k)-N×W(k-1);

[0156] S517, first Kalman gain calculation:

[0157] Kg′(k)=P′(k|k-1)×L T / [L×P′(k|k-1)×L T +R′]; where R′ is the first observation noise covariance.

[0158] S518, first calculate the observation quantity innovation rate:

[0159] e′(k)=|ε′(k) / βk)|×100%;

[0160] S519, introduce the first gain factor, the first gain factor is positively correlated with the first observation value innovation rate, use the first gain factor to change the weights of the state estimate and the measurement value, prevent the increase of the error of the estimation result due to the sudden change of the first observation value innovation, and ensure the stability of the final estimation result. Among them, the exemplary method of setting the first gain factor is: when the first observation value innovation rate is in the third preset range, the size of the gain factor should be set to the third preset value, so as to weaken the correction of the actual measurement value to the expected result; conversely, when the first observation value innovation rate is in the fourth preset range, it should be set to the fourth preset value to strengthen the correction of the actual measurement value to the estimation result, wherein the fourth preset value is greater than the third preset value, and when the left side of the numerical axis is the positive direction, the third preset range is within the right range of the fourth preset value. Among them, the first gain factor can also add the influence of temperature on it. The specific size setting can be debugged according to the simulation test results until the final result meets the requirements.

[0161] First gain factor: θ′(k)=f(e′(k), T);

[0162] S520, battery internal temperature state estimation correction;

[0163] α(k|k)=α(k|k-1)+θ′(k)×Kg′(k)×ε′(k);

[0164] Among them, θ′(k)=f(e′(k), T), α(k|k) is the state quantity after correction at the kth moment, α(k|k-1) is the state prediction quantity at the kth moment, that is, the state quantity before correction, θ′(k) is the first introduced gain factor, Kg′(k) is the first Kalman gain at the kth moment, and ε′(k) is the first observation value update at the kth moment.

[0165] S521, first error covariance update;

[0166] P′(k|k)=[E-Kg′(k)×L]×P′(k|k-1);

[0167] S522, updating the first system noise covariance matrix;

[0168] In order to avoid the occurrence of non-positive definiteness during the updating process of the first system noise covariance, which leads to filtering divergence, the updating calculation formula of the first system noise covariance is improved. As the operation iteration proceeds, the first error covariance matrix will gradually decrease, or even approach zero. Therefore, in the embodiment of the present invention, the subtraction operation part related to the first error covariance in the original updating calculation formula of the first system noise covariance is simplified, which can ensure the positive definiteness of the first system noise covariance on the one hand, and further simplify the calculation process on the other hand.

[0169] The first system noise covariance matrix update is expressed as:

[0170] Q′(k)=[1-d′(k-1)]×Q′(k-1)+d′(k-1)×[Kg′(k)×ε′(k)×ε′(k) T ×Kg′(k) T ]; where the weight coefficient d′(k)=1-d′ / [1-d′(k+1)], d′ is the forgetting factor, and its value is generally between 0.95-0.99. d′(k-1) is the weight coefficient at the k-1th moment, and Q′(k-1) is the first system noise covariance matrix at the k-1th moment.

[0171] Through the iterative operation of the above process, we can achieve T ia The online estimate of is then calculated through its relationship with the ambient temperature:

[0172] T in =T ia +T amb

[0173] The internal temperature of the battery T can be calculated in The internal temperature of the battery obtained in step S520 is used as an input condition in the battery SOC estimation process to update the capacitance and resistance parameters in the equivalent circuit model.

[0174] By feeding back the outputted battery internal temperature and battery SOC as the input state quantity in the battery SOC estimation process and the input state quantity in the battery internal temperature estimation process at the next moment, the respective models are updated and the estimation calculation is performed. The battery internal temperature and the battery SOC are mutually influential and interrelated. Therefore, through the iterative operation of the above process, the coupled estimation between the battery SOC and the internal temperature is realized, which further improves the accuracy of the SOC estimation result.

[0175] Combination Figure 4 and Figure 5 The embodiment of the present invention uses a 26Ah lithium iron phosphate battery as the experimental object. After establishing its corresponding equivalent circuit model according to the battery SOC estimation method in the embodiment of the present invention, parameter testing and identification are performed to obtain corresponding algorithm parameters. Then, the battery SOC estimation algorithm model of the adaptive extended Kalman filter algorithm in the embodiment of the present invention is established in the simulation software. The estimation results in the embodiment of the present invention are compared with the results calculated by the traditional Sage-Husa adaptive Kalman filter algorithm and the results calculated by the currently commonly used extended Kalman filter algorithm, and the actual test results of the battery are obtained to obtain the error comparison of each item under different working conditions in Table 1.

[0176] Table 1 Comparison of errors of various items under different working conditions

[0177]

[0178] It can be seen from the table that under some more complex working conditions, the results calculated by the traditional Sage-Husa adaptive Kalman filter algorithm and the results calculated by the currently commonly used extended Kalman filter algorithm will have filter divergence and non-convergence, which ultimately leads to large errors in the calculation results. The estimation method provided by the embodiment of the present invention can avoid the problem of filter divergence, can converge to the true value in a shorter time, and the maximum error is within 4%, with good accuracy and robustness.

[0179] In order to further verify the impact of internal temperature on SOC estimation, the battery SOC and internal temperature were coupled and estimated, and the battery SOC estimation result at the current moment was used as the input condition for battery internal temperature estimation; at the same time, the battery internal estimation result was used as the input condition for battery SOC estimation at the next moment to adjust the model parameters. The SOC result SOC(Tin) of the coupled estimation of battery SOC and internal temperature was compared with the SOC estimation result SOC(Ts) that only considers the battery surface temperature and the battery measured SOC result, as shown in the figure. Figure 6-8 shown.

[0180] from Figure 7 and Figure 8 It can be seen that SOC(Tin) is closer to the real SOC than SOC(Ts). From the comparison results of the absolute value of the error, the maximum error of SOC(Tin) in the whole working process is 2.25%, and the maximum error of SOC(Ts) is 2.83%. The coupled estimation result of SOC and internal temperature SOC(Tin) is more accurate, indicating that compared with the single estimation of SOC, the SOC result of the coupled estimation of battery SOC and internal temperature based on this scheme is more accurate and can better reflect the real state of the battery.

[0181] Based on the above embodiment, optionally, estimating the internal temperature of the battery at the kth moment according to the battery equivalent thermal circuit model and the first adaptive extended Kalman filter algorithm includes:

[0182] According to the discretized state equation of the battery equivalent thermal circuit model, the predicted battery internal temperature at the kth moment is obtained based on the state quantity prediction of the battery internal temperature at the k-1th moment;

[0183] Obtaining a first Kalman gain at the kth moment according to the first error covariance and the first system noise covariance;

[0184] The predicted internal battery temperature at the kth moment is corrected according to the first gain factor, the first Kalman gain at the kth moment, and the first observation value innovation rate to obtain the internal battery temperature at the kth moment, wherein the first gain factor is positively correlated with the first observation value innovation rate.

[0185] Specific, combined Figure 5 , the difference between the internal temperature of the battery and the ambient temperature in the system is estimated by using the internal temperature of the battery ia , the difference between the battery surface temperature and the ambient temperature Tsa and the external resistance Rs are state variables, the difference between the battery surface temperature and the ambient temperature Tsa is the system observable, and the heat generation power H is the external excitation. According to the discretized state equation of the battery equivalent thermal circuit model, the coefficient matrix is ​​calculated. According to the discretized state equation of the battery equivalent thermal circuit model, based on the state of the battery internal temperature at the k-1th moment (previous moment), the state of the battery internal temperature at the kth moment (current moment) is predicted, which can be expressed as: α(k) = E×α(k-1)+J×W(k-1).

[0186] The Kalman gain in the battery internal temperature estimation system, that is, the first Kalman gain can be calculated according to the first error covariance and the first system noise covariance. The first error covariance matrix is ​​calculated based on the first error covariance at the previous moment, and the first error covariance at the kth moment is calculated:

[0187] P'(k|k-1)=E×P'(k-1|k-1)×E T +Q′(k-1); where Q′(k-1) is the first system noise covariance at the k-1th moment, and P′(k-1|k-1) is the first error covariance after the update at the k-1th moment; the first Kalman gain calculation: Kg′(k)=P′(k|k-1)×L T / [L×P′(k|k-1)×L T + R′];

[0188] When correcting the state variable (internal battery temperature), a first gain factor can be introduced. The first gain factor is positively correlated with the first observation value innovation rate. The first gain factor is used to change the weights of the state estimate and the measurement value to prevent the error of the estimation result from increasing due to the sudden change of the first observation value innovation, thereby ensuring the stability of the final estimation result. Among them, an exemplary method for setting the first gain factor is: when the first observation value innovation rate is in the third preset range, the size of the gain factor should be set to the third preset value, thereby weakening the correction of the actual measurement value to the expected result; conversely, when the first observation value innovation rate is in the fourth preset range, it should be set to the fourth preset value to strengthen the correction of the actual measurement value to the estimated result, wherein the fourth preset value is greater than the third preset value, and when the left side of the numerical axis is the positive direction, the third preset range is within the right range of the fourth preset value. Among them, the first gain factor can also add the influence of temperature on it. The specific size setting can be debugged according to the simulation test results until the final result meets the requirements. According to the first gain factor, the first Kalman gain at the kth moment and the first observation value innovation, the predicted internal battery temperature at the kth moment is corrected to obtain the internal battery temperature at the kth moment:

[0189] α(k|k)=α(k|k-1)+θ′(k)×Kg(k)×ε′(k); wherein, the first gain factor: θ′(k)=f(e′(k), T).

[0190] α(k|k) is the state quantity after correction at the kth moment, that is, the internal temperature of the battery at the kth moment, α(k|k-1) is the state estimation quantity at the kth moment, that is, the state quantity before correction, θ′(k) is the introduced gain factor, Kg′(k) is the first Kalman gain at the kth moment, and ε′(k) is the first observation quantity innovation at the kth moment. The first observation quantity innovation rate at the kth moment is e′(k)=|ε′(k) / βk)|×100%.

[0191] Based on the above embodiment, optionally, the internal temperature of the battery at the kth moment is used as the input state quantity at the k+1th moment in the battery state of charge estimation process, and the battery state of charge at the k+1th moment is estimated based on the second adaptive extended Kalman filter algorithm, including:

[0192] According to the battery internal temperature and capacitance and resistance parameter identification conditions at the kth moment, the capacitance and resistance parameters in the battery equivalent circuit model are updated;

[0193] The battery state of charge at the k+1th moment is estimated according to the updated battery equivalent circuit model and the second adaptive extended Kalman filter algorithm.

[0194] Specifically, the internal temperature of the battery at the kth moment is used as the input state quantity at the k+1th moment in the battery SOC estimation process, and the capacitance and resistance parameters of the battery equivalent circuit model at the internal temperature of the battery are identified. According to the identification relationship, the corresponding identification parameters that need to be updated are obtained, for example, the corresponding parameters of the variables such as the electrochemical polarization resistance R1 and the concentration polarization resistance R2 of the battery, the actual capacity Ca of the battery, the electrochemical polarization capacitance C1 and the concentration polarization capacitance C2 of the battery. Thus, the parameters in the battery equivalent circuit model are updated. In the SOC estimation process, the SOC, the voltage U1 across the electrochemical polarization capacitor, and the voltage U2 across the concentration polarization capacitor are set as state variables, the output terminal voltage Ut of the battery is set as an observed quantity, and the current I is used as an external excitation. The estimated value of the battery SOC is updated using the second adaptive extended Kalman filter algorithm, and the battery SOC at the k+1th moment is estimated, so as to realize the coupled estimation of the battery SOC and the internal temperature.

[0195] Based on the above embodiment, optionally, estimating the battery state of charge at the k+1th moment according to the updated battery equivalent circuit model and the second adaptive extended Kalman filter algorithm includes:

[0196] According to the updated state equation of the battery equivalent circuit model, the predicted battery state of charge at the k+1th moment is obtained based on the state quantity prediction of the battery state of charge at the kth moment;

[0197] Updating the second error covariance and the second system noise covariance, and obtaining the second Kalman gain at the k+1th moment according to the updated second error covariance and the second system noise covariance; wherein the subtraction operation of the second error covariance is not included in the updating calculation of the second system noise covariance;

[0198] The predicted battery state of charge at the k+1th moment is corrected according to the second gain factor, the second Kalman gain at the k+1th moment, and the second observation value innovation rate to obtain the battery state of charge at the k+1th moment, wherein the second gain factor is positively correlated with the second observation value innovation rate.

[0199] Specific, combined Figure 5, after adjusting the model parameters, the battery SOC state at the k+1th moment is predicted based on the battery SOC state at the kth moment, and the predicted battery SOC at the k+1th moment is obtained. The second error covariance and the second system noise covariance matrix in the second adaptive extended Kalman filter algorithm are updated. In order to avoid the non-positive definiteness that causes the filter divergence during the update of the second system noise covariance, the update calculation formula of the second system noise covariance is improved. As the operation iteration proceeds, the second error covariance matrix will gradually decrease, or even approach zero. Therefore, in the embodiment of the present invention, the subtraction operation part related to the second error covariance in the original second system noise covariance update calculation formula is simplified. On the one hand, it can ensure the positive definiteness of the matrix system noise covariance, and on the other hand, it simplifies the operation process.

[0200] When correcting the state variable (battery SOC), a second gain factor can be introduced. The second gain factor is positively correlated with the second observation value innovation rate. The second gain factor is used to change the weights of the state estimate and the measurement value to prevent the error of the estimation result from increasing due to the sudden change of the second observation value innovation, thereby ensuring the stability of the final estimation result. Among them, an exemplary second gain factor setting method is: when the second observation value innovation rate is in the first preset range, the size of the second gain factor should be set to the first preset value, thereby weakening the correction of the actual measurement value to the expected result; conversely, when the second observation value innovation rate is in the second preset range, it should be set to the second preset value to strengthen the correction of the actual measurement value to the estimation result, wherein the second preset value is greater than the first preset value, and when the left side of the numerical axis is in the positive direction, the first preset range is within the right range of the second preset value. Among them, the gain factor can also add the influence of temperature on it. The specific size setting can be debugged according to the simulation test results until the final result meets the requirements. According to the second gain factor, the second Kalman gain at the k+1th moment and the second observation value innovation, the predicted battery SOC at the k+1th moment is corrected to obtain the battery SOC at the k+1th moment:

[0201] X(k+1|k)=X(k+1|k-1)+θ(k+1)×Kg(k+1)×ε(k+1);

[0202] Among them, θ(k+1)=f(e(k+1), T); e(k+1)=|ε(k+1) / Ut(k+1)|*100%;

[0203] X(k+1|k) is the state quantity after correction at the k+1th moment, that is, the battery SOC at the k+1th moment, X(k+1|k-1) is the state prediction quantity at the k+1th moment, that is, the state quantity before correction, θ(k+1) is the introduced gain factor, Kg(k+1) is the second Kalman gain at the k+1th moment, ε(k+1) is the second observation quantity innovation at the k+1th moment, and e(k+1) is the second observation quantity innovation rate at the k+1th moment.

[0204] Based on the above embodiment, optionally, estimating the battery state of charge at the kth moment according to the battery equivalent circuit model and the second adaptive extended Kalman filter algorithm includes:

[0205] According to the state equation of the battery equivalent circuit model, the predicted battery state of charge at the kth moment is obtained based on the state quantity of the battery state of charge at the k-1th moment;

[0206] Obtaining a second Kalman gain at the kth moment according to the second error covariance and the second system noise covariance;

[0207] The predicted battery state of charge at the kth moment is corrected according to the second gain factor, the second Kalman gain at the kth moment and the second observation value innovation rate to obtain the battery state of charge at the kth moment, wherein the second gain factor is positively correlated with the second observation value innovation rate.

[0208] Specific, combined Figure 5 , the battery SOC, the voltage U1 across the electrochemical polarization capacitor, and the voltage U2 across the concentration polarization capacitor in the battery SOC estimation system are used as state variables, the terminal voltage Ut(k) of the battery is used as the system observable, and the current I(k) is used as the external excitation. According to the discretized state equation of the second-order RC equivalent circuit model, the coefficient matrix is ​​calculated. According to the discretized state equation of the second-order RC equivalent circuit model, based on the state of the battery SOC at the k-1th moment (previous moment), the state of the battery SOC at the kth moment (current moment) is predicted, which can be expressed as: X(k) = A×X(k-1) + B×I(k-1);

[0209] The Kalman gain in the battery SOC estimation system, that is, the second Kalman gain, can be calculated based on the second error covariance and the second system noise covariance. The second error covariance matrix is ​​calculated based on the second error covariance at the previous moment, and the second error covariance at the kth moment is calculated:

[0210] P(k|k-1)=A×P(k-1|k-1)×A T+Q(k-1); where Q(k-1) is the second system noise covariance at the k-1th moment, and P(k-1|k-1) is the second error covariance after the update at the k-1th moment; the second Kalman gain calculation: Kg(k)=P(k|k-1)×C T / [C×P(k|k-1)×C T +R];

[0211] When correcting the state variable (battery SOC), a second gain factor can be introduced. The second gain factor is positively correlated with the second observation value innovation rate. The second gain factor is used to change the weights of the state estimate and the measurement value to prevent the increase of the estimation result error due to the sudden change of the observation value innovation, thereby ensuring the stability of the final estimation result. Among them, an exemplary second gain factor setting method is: when the second observation value innovation rate is in the first preset range, the size of the second gain factor should be set to the first preset value, thereby weakening the correction of the actual measurement value to the expected result; conversely, when the second measurement innovation rate is in the second preset range, it should be set to the second preset value to strengthen the correction of the actual measurement value to the estimation result, wherein the second preset value is greater than the first preset value, and when the left side of the numerical axis is the positive direction, the first preset range is within the right range of the second preset value. Among them, the second gain factor can also add the influence of temperature on it. The specific size setting can be debugged according to the simulation test results until the final result meets the requirements. According to the second gain factor, the second Kalman gain at the kth moment and the second observation value innovation, the predicted battery internal temperature at the kth moment is corrected to obtain the battery SOC at the kth moment:

[0212] X(k|k)=X(k|k-1)+θ(k)×Kg(k)×ε(k); wherein, the second gain factor: θ(k)=f(e(k), T).

[0213] X(k|k) is the state quantity after correction at the kth moment, that is, the battery SOC at the kth moment, X(k|k-1) is the state prediction quantity at the kth moment, that is, the state quantity before correction, θ(k) is the introduced second gain factor, Kg(k) is the second Kalman gain at the kth moment, and ε(k) is the second observation quantity innovation at the kth moment. The second observation quantity innovation rate at the kth moment is e(k) = |ε(k) / β(k)|×100%.

[0214] Based on the above embodiment, optionally, the battery state of charge at the kth moment is used as an input state quantity in the battery internal temperature estimation process, and the battery internal temperature at the k+1th moment is estimated based on the adaptive extended Kalman filter algorithm, including:

[0215] The heat generation power in the battery equivalent thermal circuit model is updated according to the battery state of charge at the kth moment;

[0216] The internal temperature of the battery at the k+1th moment is estimated according to the updated battery equivalent thermal circuit model and the first adaptive extended Kalman filter algorithm.

[0217] Specifically, the battery SOC at the kth moment is used as the input state quantity at the k+1th moment in the process of estimating the internal temperature of the battery, the parameters of the battery equivalent thermal circuit model under the battery SOC are identified, the corresponding identification parameters that need to be updated are obtained according to the identification relationship, and the parameters in the battery equivalent thermal circuit model are updated. In the process of estimating the internal temperature of the battery, the difference between the internal temperature of the battery and the ambient temperature, and the external thermal resistance Rs of the battery are used as state variables of the system, the difference between the surface temperature of the battery and the ambient temperature is used as the observation variable of the system, and the heat generation power is used as an external excitation. The estimated values ​​of the internal temperature of the battery and the external thermal resistance Rs of the battery are updated using the first adaptive extended Kalman filter algorithm, and the internal temperature of the battery at the k+1th moment is estimated.

[0218] Based on the above embodiment, optionally, estimating the internal temperature of the battery at the k+1th moment according to the updated battery equivalent thermal circuit model and the first adaptive extended Kalman filter algorithm includes:

[0219] According to the discretized state equation of the battery equivalent thermal circuit model, the predicted battery internal temperature at the k+1th moment is obtained based on the state quantity of the battery internal temperature at the kth moment;

[0220] Updating the first error covariance and the first system noise covariance, and obtaining the first Kalman gain at the k+1th moment according to the updated first error covariance and the second system noise covariance; wherein the update calculation of the first system noise covariance does not include a subtraction operation of the first error covariance;

[0221] The predicted internal battery temperature at the k+1th moment is corrected according to the first gain factor, the first Kalman gain at the k+1th moment, and the first observation value innovation rate to obtain the internal battery temperature at the k+1th moment, wherein the first gain factor is positively correlated with the first observation value innovation rate.

[0222] Specific, combined Figure 5, after adjusting the model parameters, the internal temperature state of the battery at the k+1th moment is predicted based on the internal temperature state of the battery at the kth moment, and the predicted internal temperature of the battery at the k+1th moment is obtained. The first error covariance and the first system noise covariance matrix in the first adaptive extended Kalman filter algorithm are updated. In order to avoid the non-positive definiteness that causes the filter divergence during the update of the first system noise covariance, the update calculation formula of the first system noise covariance is improved. As the operation iteration proceeds, the first error covariance matrix will gradually decrease, or even approach zero. Therefore, in the embodiment of the present invention, the subtraction operation part related to the first error covariance in the update calculation formula of the original first system noise covariance is simplified. On the one hand, it can ensure the positive definiteness of the matrix system noise covariance, and on the other hand, it simplifies the operation process.

[0223] When correcting the state variable (internal battery temperature), a first gain factor can be introduced. The first gain factor is positively correlated with the first observation value innovation rate. The first gain factor is used to change the weights of the state estimate and the measurement value to prevent the error of the estimation result from increasing due to the sudden change of the first observation value innovation, thereby ensuring the stability of the final estimation result. Among them, an exemplary method for setting the first gain factor is: when the first observation value innovation rate is in the third preset range, the size of the gain factor should be set to the third preset value, thereby weakening the correction of the actual measurement value to the expected result; conversely, when the first observation value innovation rate is in the fourth preset range, it should be set to the fourth preset value to strengthen the correction of the actual measurement value to the estimated result, wherein the fourth preset value is greater than the third preset value, and when the left side of the numerical axis is the positive direction, the third preset range is within the right range of the fourth preset value. Among them, the first gain factor can also add the influence of temperature on it. The specific size setting can be debugged according to the simulation test results until the final result meets the requirements. According to the first gain factor, the first Kalman gain at the k+1th moment and the first observation value innovation, the predicted internal battery temperature at the k+1th moment is corrected to obtain the internal battery temperature at the k+1th moment:

[0224] α(k+1|k)=α(k+1|k-1)+θ′(k+1)×Kg(k+1)×ε′(k+1);

[0225] Among them, θ′(k+1)=f(e′(k+1), T).

[0226] α(k+1|k) is the state quantity after correction at the k+1th moment, that is, the internal temperature of the battery at the k+1th moment, α(k+1|k-1) is the state estimation quantity at the k+1th moment, that is, the state quantity before correction, θ′(k+1) is the introduced gain factor, Kg′(k+1) is the first Kalman gain at the k+1th moment, and ε′(k+1) is the first observation quantity innovation at the k+1th moment. The first observation quantity innovation rate at the k+1th moment is e′(k+1)=|ε′(k+1) / (β+1)|*100%.

[0227] Fig. 9 A schematic diagram of a battery SOC estimation device provided by an embodiment of the present invention is shown in FIG. Fig. 9 ,include:

[0228] A model unit 910 is used to establish an equivalent model; wherein the equivalent model includes a battery equivalent circuit model and a battery equivalent thermal circuit model;

[0229] The first estimation unit 920 is used to estimate the internal temperature of the battery at the kth moment according to the battery equivalent thermal circuit model and the first adaptive extended Kalman filter algorithm; and to estimate the battery state of charge at the kth moment according to the battery equivalent circuit model and the second adaptive extended Kalman filter algorithm;

[0230] The second estimation unit 930 is used to use the battery internal temperature at the kth moment as the input state quantity at the k+1th moment in the battery state of charge estimation process, and estimate the battery state of charge at the k+1th moment based on the second adaptive extended Kalman filter algorithm; use the battery state of charge at the kth moment as the input state quantity in the battery internal temperature estimation process, and estimate the battery internal temperature at the k+1th moment based on the first adaptive extended Kalman filter algorithm; wherein k is a positive integer greater than 0;

[0231] The circulation unit 940 is used to feed back the estimated results of the battery internal temperature at the k+1th moment and the battery state of charge at the k+1th moment as the input state quantity in the subsequent corresponding battery state of charge estimation process and the input state quantity in the battery internal temperature estimation process, and calculate based on the second adaptive extended Kalman filter algorithm, so as to estimate the battery state of charge at each moment.

[0232] Specific, combined Figure 2 and Figure 3The battery equivalent circuit model adopts the second-order RC equivalent circuit model. In the battery equivalent circuit model, R0 represents the ohmic internal resistance of the battery, R1 and R2 represent the electrochemical polarization resistance and concentration polarization resistance of the battery, C1 and C2 represent the electrochemical polarization capacitance and concentration polarization capacitance of the battery, respectively. U1 and U2 represent the voltage across the electrochemical polarization capacitance and the voltage across the concentration polarization capacitance, respectively. I is the current, Ut is the battery output voltage, and U OCV It represents the open circuit voltage of the battery. In the battery equivalent thermal circuit model, H is the equivalent concentrated heat source inside the battery (heat generating power), Cc and Cs are the internal heat capacity and surface environmental heat capacity, Rc and Rs are the internal thermal resistance and external thermal resistance, Tin and Ts are the internal temperature and surface temperature of the battery, respectively.

[0233] The parameters of the internal heat capacity Cc, the surface environment heat capacity Cs, the internal thermal resistance Rc and the external thermal resistance Rs in the battery equivalent thermal circuit model change according to the ambient temperature or the internal temperature. The parameters corresponding to the internal heat capacity Cc, the surface environment heat capacity Cs, the internal thermal resistance Rc and the external thermal resistance Rs can be obtained according to known identification means. Since the parameters of the internal heat capacity Cc, the surface environment heat capacity Cs, the internal thermal resistance Rc and the external thermal resistance Rs are mainly related to the material properties of the battery itself, and the external thermal resistance Rs is also related to the external environment in which the battery is located, the identification results of the internal heat capacity Cc, the surface environment heat capacity Cs and the internal thermal resistance Rc can be used as fixed values, and the identification results of the external thermal resistance Rs can be used only as the initial value of the external thermal resistance Rs. The parameters of the external thermal resistance Rs can be used as a state variable and estimated simultaneously with the internal temperature of the battery in the subsequent process to improve the adaptability of the model. The internal heat capacity Cc, the surface environment heat capacity Cs, the internal thermal resistance Rc and the external thermal resistance Rs are used together with other variables as the input conditions of the adaptive extended Kalman filter algorithm, and the battery equivalent thermal circuit model is discretized to obtain the discretized state equation. According to Kirchhoff's law, the state equation of the equivalent circuit model is listed, and the discretized state equation of the equivalent circuit model is further obtained according to the state equation.

[0234] The first estimation unit 920 obtains a discretized state equation based on the discretization of the battery equivalent thermal circuit model, takes the difference between the internal temperature of the battery and the ambient temperature, and the external thermal resistance Rs of the battery as the state variables of the system, takes the difference between the surface temperature of the battery and the ambient temperature as the observed variable of the system, and takes the heat generation power as the external excitation, and updates the estimated values ​​of the internal temperature of the battery and the external thermal resistance Rs of the battery using the first adaptive extended Kalman filter algorithm to estimate the internal temperature of the battery at the kth moment. The first estimation unit 920 also uses the battery SOC, the voltage U1 across the electrochemical polarization capacitor, and the voltage U2 across the concentration polarization capacitor in the battery SOC estimation system as the state variables, the terminal voltage Ut(k) of the battery as the system observed variable, and the current I(k) as the external excitation, and uses the second adaptive extended Kalman filter algorithm to estimate the battery SOC at the kth moment.

[0235] In some embodiments, when correcting the state variable (internal battery temperature), a first gain factor can also be introduced. The first gain factor is positively correlated with the first observation value innovation rate. The first gain factor is used to change the weights of the state estimate and the measurement value to prevent the increase in the error of the estimation result due to the sudden change of the first observation value innovation, thereby ensuring the stability of the final estimation result. Among them, an exemplary method for setting the first gain factor is: when the first observation value innovation rate is in the third preset range, the size of the gain factor should be set to the third preset value, thereby weakening the correction of the actual measurement value to the expected result; conversely, when the first observation value innovation rate is in the fourth preset range, it should be set to the fourth preset value to strengthen the correction of the actual measurement value to the estimation result, wherein the fourth preset value is greater than the third preset value, and when the left side of the numerical axis is the positive direction, the third preset range is within the right range of the fourth preset value. Among them, the first gain factor can also add the influence of temperature on it. The specific size setting can be debugged according to the simulation test results until the final result meets the requirements.

[0236] When correcting the state variable (battery SOC), a second gain factor can also be introduced. The second gain factor is positively correlated with the second observation value innovation rate. The second gain factor is used to change the weights of the state estimate and the measurement value to prevent the increase in the error of the estimation result due to the sudden change of the observation value innovation, thereby ensuring the stability of the final estimation result. Among them, an exemplary method for setting the second gain factor is: when the second observation value innovation rate is in the first preset range, the size of the second gain factor should be set to the first preset value, thereby weakening the correction of the actual measurement value to the expected result; conversely, when the second observation value innovation rate is in the second preset range, it should be set to the second preset value to strengthen the correction of the actual measurement value to the estimation result, wherein the second preset value is greater than the first preset value, and when the left side of the numerical axis is in the positive direction, the first preset range is within the right range of the second preset value. Among them, the second gain factor can also add the influence of temperature on it.

[0237] The second estimation unit 930 uses the internal temperature of the battery at the kth moment as the input state quantity at the k+1th moment in the battery SOC estimation process, identifies the parameters of the battery equivalent circuit model at the internal temperature of the battery, and obtains the corresponding identification parameters that need to be updated according to the identification relationship, for example, the corresponding parameters of the variables such as the electrochemical polarization resistance R1 and the concentration polarization resistance R2 of the battery, the actual capacity Ca of the battery, the electrochemical polarization capacitance C1 and the concentration polarization capacitance C2 of the battery. Thus, the parameters in the battery equivalent circuit model are updated. In the SOC estimation process, the second estimation unit 930 uses the SOC, the voltage U1 across the electrochemical polarization capacitor, and the voltage U2 across the concentration polarization capacitor as state variables, the output terminal voltage Ut of the battery as an observed quantity, and uses the current I as an external excitation. The estimated value of the battery SOC is updated using the second adaptive extended Kalman filter algorithm, and the battery SOC at the k+1th moment is estimated to achieve coupled estimation of the battery SOC and the internal temperature. The battery SOC calculation result output at the kth moment is used to update the heat generation power of the estimation process of the battery internal temperature at the k+1th moment, and the battery internal temperature at the k+1th moment is estimated based on the first adaptive extended Kalman filter algorithm.

[0238] The circulation unit 940 feeds back the outputted battery internal temperature and battery SOC as the input state quantity in the battery SOC estimation process and the input state quantity in the battery internal temperature estimation process at the next moment, respectively, and performs estimation calculation after updating the respective models. The battery internal temperature and the battery SOC are mutually influential and interrelated. Therefore, through the iterative operation of the above process, the coupled estimation between the battery SOC and the internal temperature is realized, and the accuracy of the SOC estimation result is further improved.

[0239] Fig.10A schematic diagram of the structure of an electronic device 10 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0240] like Fig.10 As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0241] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0242] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as a method for estimating the state of charge of a battery.

[0243] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for estimating a battery state of charge, characterized in that: include: Establishing an equivalent model; wherein the equivalent model includes a battery equivalent circuit model and a battery equivalent thermal circuit model; The internal temperature of the battery at the kth moment is estimated according to the battery equivalent thermal circuit model and the first adaptive extended Kalman filter algorithm; the state of charge of the battery at the kth moment is estimated according to the battery equivalent circuit model and the second adaptive extended Kalman filter algorithm; The battery internal temperature at the kth moment is used as the input state quantity at the k+1th moment in the battery state of charge estimation process, and the battery state of charge at the k+1th moment is estimated based on the second adaptive extended Kalman filter algorithm; the battery state of charge at the kth moment is used as the input state quantity in the battery internal temperature estimation process, and the battery internal temperature at the k+1th moment is estimated based on the first adaptive extended Kalman filter algorithm; wherein k is a positive integer greater than 0; The results of the battery internal temperature at the k+1th moment and the battery state of charge at the k+1th moment, which are estimated respectively, are fed back as the input state quantity in the subsequent corresponding battery state of charge estimation process and the input state quantity in the battery internal temperature estimation process, and are calculated based on the second adaptive extended Kalman filter algorithm, so as to estimate the battery state of charge at each moment; the battery internal temperature at the kth moment is estimated according to the battery equivalent thermal circuit model and the first adaptive extended Kalman filter algorithm, including: According to the discretized state equation of the battery equivalent thermal circuit model, the predicted battery internal temperature at the kth moment is obtained based on the state quantity prediction of the battery internal temperature at the k-1th moment; Obtaining a first Kalman gain at the kth moment according to the first error covariance and the first system noise covariance; The predicted internal battery temperature at the kth moment is corrected according to the first gain factor, the first Kalman gain at the kth moment, and the first observation quantity innovation information to obtain the internal battery temperature at the kth moment, wherein the first observation quantity innovation rate is obtained according to the first observation quantity innovation information, and the first gain factor is positively correlated with the first observation quantity innovation rate.

2. The method for estimating the state of charge of a battery according to claim 1, characterized in that: The method uses the internal temperature of the battery at the kth moment as the input state quantity at the k+1th moment in the battery state of charge estimation process, and estimates the battery state of charge at the k+1th moment based on the second adaptive extended Kalman filter algorithm, including: updating the capacitance and resistance parameters in the battery equivalent circuit model according to the battery internal temperature and capacitance and resistance parameter identification conditions at the kth moment; The battery state of charge at the k+1th moment is estimated according to the updated battery equivalent circuit model and the second adaptive extended Kalman filter algorithm.

3. The method for estimating the state of charge of a battery according to claim 2, characterized in that: Estimating the battery state of charge at the k+1th moment according to the updated battery equivalent circuit model and the second adaptive extended Kalman filter algorithm, comprising: According to the updated state equation of the battery equivalent circuit model, a predicted battery state of charge at the k+1th moment is obtained based on the state quantity prediction of the battery state of charge at the kth moment; Updating the second error covariance and the second system noise covariance, and obtaining the second Kalman gain at the k+1th moment according to the updated second error covariance and the second system noise covariance; wherein the update calculation of the second system noise covariance does not include a subtraction operation of the second error covariance; The predicted battery state of charge at the k+1th moment is corrected according to the second gain factor, the second Kalman gain at the k+1th moment, and the second observation quantity innovation information to obtain the battery state of charge at the k+1th moment, wherein the second observation quantity innovation rate is obtained according to the second observation quantity innovation information, and the second gain factor is positively correlated with the second observation quantity innovation rate.

4. The method for estimating the state of charge of a battery according to claim 3, characterized in that: The battery state of charge at the kth moment is estimated according to the battery equivalent circuit model and the second adaptive extended Kalman filter algorithm, including: According to the state equation of the battery equivalent circuit model, based on the state quantity of the battery state of charge at the k-1th moment, a predicted battery state of charge at the kth moment is obtained; Obtaining a second Kalman gain at the kth moment according to the second error covariance and the second system noise covariance; The predicted battery state of charge at the kth moment is corrected according to the second gain factor, the second Kalman gain at the kth moment and the second observation value innovation rate to obtain the battery state of charge at the kth moment, wherein the second gain factor is positively correlated with the second observation value innovation rate.

5. The method for estimating the state of charge of a battery according to claim 4, characterized in that: The battery state of charge at the kth moment is used as an input state quantity in the battery internal temperature estimation process, and the battery internal temperature at the k+1th moment is estimated based on the first adaptive extended Kalman filter algorithm, including: Updating the heat generation power in the battery equivalent thermal circuit model according to the battery state of charge at the kth moment; The internal temperature of the battery at the k+1th moment is estimated according to the updated battery equivalent thermal circuit model and the first adaptive extended Kalman filter algorithm.

6. The method for estimating the state of charge of a battery according to claim 5, characterized in that: Estimating the internal temperature of the battery at the k+1th moment according to the updated battery equivalent thermal circuit model and the first adaptive extended Kalman filter algorithm, including: According to the discretized state equation of the battery equivalent thermal circuit model, the predicted battery internal temperature at the k+1th moment is obtained based on the state quantity of the battery internal temperature at the kth moment; Updating a first error covariance and a first system noise covariance, and obtaining a first Kalman gain at the k+1th moment according to the updated first error covariance and the second system noise covariance; wherein the update calculation of the first system noise covariance does not include a subtraction operation of the first error covariance; The predicted internal temperature of the battery at the k+1th moment is corrected according to the first gain factor, the first Kalman gain at the k+1th moment and the first observation value innovation rate to obtain the internal temperature of the battery at the k+1th moment, wherein the first gain factor is positively correlated with the first observation value innovation rate.

7. A battery state of charge estimation device, characterized in that: include: A model unit, used to establish an equivalent model; wherein the equivalent model includes a battery equivalent circuit model and a battery equivalent thermal circuit model; a first estimation unit, configured to estimate the internal temperature of the battery at the kth moment according to the battery equivalent thermal circuit model and the first adaptive extended Kalman filter algorithm; and to estimate the state of charge of the battery at the kth moment according to the battery equivalent circuit model and the second adaptive extended Kalman filter algorithm; a second estimation unit, configured to use the battery internal temperature at the kth moment as an input state quantity at the k+1th moment in the battery state of charge estimation process, and estimate the battery state of charge at the k+1th moment based on the second adaptive extended Kalman filter algorithm; use the battery state of charge at the kth moment as an input state quantity in the battery internal temperature estimation process, and estimate the battery internal temperature at the k+1th moment based on the first adaptive extended Kalman filter algorithm; wherein k is a positive integer greater than 0; a circulation unit, configured to feed back the results of the battery internal temperature at the k+1th moment and the battery state of charge at the k+1th moment as input state quantities in subsequent corresponding battery state of charge estimation processes and input state quantities in battery internal temperature estimation processes, and calculate based on the second adaptive extended Kalman filter algorithm, thereby estimating the battery state of charge at each moment; The first estimation unit is further used to obtain a predicted battery internal temperature at the kth moment based on the state quantity prediction of the battery internal temperature at the k-1th moment according to the discretized state equation of the battery equivalent thermal circuit model; Obtaining a first Kalman gain at the kth moment according to the first error covariance and the first system noise covariance; The predicted internal battery temperature at the kth moment is corrected according to the first gain factor, the first Kalman gain at the kth moment, and the first observation quantity innovation information to obtain the internal battery temperature at the kth moment, wherein the first observation quantity innovation rate is obtained according to the first observation quantity innovation information, and the first gain factor is positively correlated with the first observation quantity innovation rate.

8. An electronic device, characterized in that: The electronic device comprises: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can perform the method for estimating the state of charge of a battery according to any one of claims 1 to 6.

Images