CN108333528B - SOC and SOT joint state estimation method based on power-thermal coupling model of power battery - Google Patents

Abstract

The invention relates to a joint state estimation method of SOC and SOT based on the electric-thermal coupling model of a power battery, and belongs to the technical field of battery management. The method is as follows: select the power battery to be tested, establish the electric and thermal model of the power battery, and determine the parameters required for estimating the SOC and SOT of the power battery; conduct trickle charge and discharge experiments and HPPC experiments on the power battery under test at different temperatures , establish a database of equivalent circuit model parameters under charging and discharging conditions on temperature and SOC, simulate real vehicle test conditions under different road conditions, and establish a database; perform parameter identification to obtain the characteristic parameters of the electrical and thermal models, and obtain charging and discharging conditions The quantitative relationship between the parameters of the equivalent circuit model and the temperature and SOC; this model is combined with the PF algorithm and the quantitative relationship between the characteristic parameters of the equivalent circuit model under the charging and discharging conditions of the power battery on temperature and SOC to realize the power battery SOC and SOT joint state estimation.

Description

SOC and SOT joint state estimation method based on power-thermal coupling model of power battery

technical field

The invention belongs to the technical field of battery management, and relates to a combined state estimation method of SOC and SOT based on an electric-thermal coupling model of a power battery.

Background technique

As an important part of EVs, HEVs and PHEVs, it is particularly important to accurately and efficiently estimate the SOC and SOT of the power battery, because the SOC of the power battery is closely related to other states of the battery such as temperature state, power state (State ofPower , SOP) and state of health (State of Health, SOH) and other states, and the SOT of the power battery is also closely related to the safety and reliability of the battery, charge and discharge efficiency, power and capacity, life and cycle times. However, the real working conditions of electric vehicles are complex, and the measurement accuracy of current, voltage, and impedance limits the estimation accuracy of SOC and SOT.

At present, the SOC estimation methods for power batteries are mainly ampere-hour integration method, open circuit voltage method, artificial intelligence algorithm and model-based SOC estimation method. The ampere-hour integration method is a very simple SOC estimation method widely used in electric vehicle battery management systems (Battery Management System, BMS), but the estimation accuracy of this method mainly depends on the measurement accuracy of the current and the initial SOC value . The principle of the open circuit voltage method is simple, but it is difficult to obtain an accurate open circuit voltage. The algorithm of artificial intelligence method is complicated, and it needs to train a large amount of experimental data, so it cannot be used in real vehicles. The model-based SOC estimation method is currently the most widely studied. It mainly designs an observer based on the equivalent circuit model to estimate the SOC of lithium-ion batteries. The estimation accuracy of this method depends largely on the model accuracy. It is susceptible to temperature, For the influence of discharge rate and other factors, although many current methods consider temperature correction, they do not consider the characteristics of OCV variation with temperature and the real-time joint estimation of SOC and SOT online.

At present, there are mainly the following types of SOT estimation for power batteries: using a simple thermal model to estimate the average temperature of the battery. This type of method has a small amount of calculation, but the estimation accuracy cannot reflect the actual battery temperature. Using numerical solutions (such as finite element method, finite volume method, etc.) to estimate the temperature distribution of the battery, this type of method is accurate in estimation, but the calculation is complicated and difficult for practical application. Using a one-dimensional two-state thermal model combined with surface temperature measurements to estimate the temperature distribution inside the battery, this type of method has a small amount of calculation and high accuracy, but it needs to install a large number of temperature sensors, which is difficult to apply and promote. A viable alternative is to use impedance measurements combined with a suitable thermal model to estimate the battery's temperature distribution, which would eliminate the need for thermocouples on the battery cells. This method has been studied by scholars abroad, and the thermal-impedance model based on impedance measurement is used to estimate and predict the temperature distribution inside the battery cell.

At present, there have been many studies on estimating the SOC or SOT of lithium batteries alone, but the method of jointly estimating the two and applying them to the real vehicle BMS has not yet appeared.

Contents of the invention

In view of this, the object of the present invention is to provide a joint state estimation method of SOC and SOT based on electric-thermal coupling model of power battery.

To achieve the above object, the present invention provides the following technical solutions:

A state of charge (State of Charge, SOC) and temperature state (State of Temperature, SOT) joint state estimation method based on the electric-thermal coupling model of the power battery, the method includes the following steps:

S1: Select the power battery to be tested, collect and organize the technical parameters of the power battery, establish the electric and thermal model of the power battery, and determine the model parameters required for joint estimation of the power battery SOC and SOT;

S2: Conduct trickle charge and discharge experiments and hybrid pulse power characteristic (Hybrid Pulse Power Characteristic, HPPC) experiments on the power battery under test at different temperatures, and establish an experimental database of equivalent circuit model parameters under charge and discharge conditions with respect to temperature and SOC , to simulate pure electric vehicles (Electric Vehicles, EVs), hybrid electric vehicles (Hybrid Electric Vehicles, HEVs) and plug-in hybrid electric vehicles (Plug-Hybrid Electric Vehicles, PHEVs) under different road conditions in cities, suburbs, villages and high speeds Real vehicle test conditions, establish a real vehicle test experimental database, including current, voltage, temperature and impedance data;

S3: Perform parameter identification to obtain the characteristic parameters of the electrical and thermal models, and obtain the quantitative relationship between the parameters of the equivalent circuit model under charge and discharge conditions, temperature and SOC through data fitting;

S4: Combining the electric-thermal coupling model of the power battery with the particle filter (Particle Filter, PF) algorithm and the quantitative relationship between the characteristic parameters of the equivalent circuit model under the charging and discharging conditions of the power battery with respect to temperature and SOC to realize the SOC and SOT of the power battery Joint state estimation.

Further, in step S1, the thermal model of the power battery is a one-dimensional (One-Dimension, 1-D) non-steady-state heat generation and heat transfer model or a one-dimensional concentrated heat generation model, and the electric power of the power battery The model is one or a combination of impedance models or equivalent circuit models.

Further, the step S2 is specifically:

S21: Leave the power battery to be tested in a constant temperature environment of 25°C for 2 hours;

S22: Charge and discharge the power battery at a charge-discharge rate of C/20, measure the relationship curve between the open circuit voltage (Open Circuit Voltage, OCV) and SOC of the power battery, and determine the available capacity of the power battery at the current stage;

S23: Perform HPPC test to obtain the current and voltage data of the power battery at the current temperature;

S24: Repeat steps S21-S23 every 10°C within the full temperature range of the power battery, and record current and voltage data at different temperatures;

S25: Simulate the real vehicle test conditions of EVs, HEVs and PHEVs under different road conditions in cities, suburbs, villages and high speeds to obtain experimental data such as current, voltage, temperature and impedance of the power battery;

S26: Summarize and process the acquired experimental data to form a usable experimental database.

Further, the step S3 is specifically:

S31: Using the experimental data obtained in step S2, using a parameter identification method to identify the characteristic parameters of the electrical and thermal models;

S32: Using the experimental data obtained in step S2, obtain the quantitative relationship between the equivalent circuit model parameters and the temperature and SOC under the charging and discharging condition through data fitting.

Further, in step S4, in step S4, the PF algorithm can be replaced by extended Kalman filter, unscented Kalman filter or H infinite filter optimal estimation algorithm.

Further, in step S31, the parameter identification method is the least square method, but it is not limited to this algorithm.

The beneficial effect of the present invention is that: the present invention provides the average temperature state obtained by the online estimation of the thermal model to the electric model to correct the characteristic parameters in the electric model, thereby realizing higher-precision SOC estimation, and then using the high-precision SOC value to calculate the current The open circuit voltage can then be used to calculate the heat generation power of the battery, which is fed back to the thermal model to correct the SOT estimate. Advantage of the present invention has:

(1) Establish an electric-thermal coupling model based on temperature and current correction for vehicle power batteries, which can accurately obtain the electric and thermal characteristics of power batteries in the full temperature range;

(2) Considering the relationship between the equivalent circuit model parameters of the power battery under charging and discharging conditions, temperature and SOC, accurate estimation of SOC under real vehicle conditions can be realized;

(3) The calculation complexity of the electric-thermal coupling model is moderate, and the joint state estimation accuracy of SOC and SOT is enough to be applied to the BMS of the real vehicle;

(4) An electric-thermal coupling model based on the power battery is proposed, combined with a nonlinear filtering method, to realize a real-time joint estimation method of the power battery SOC and SOT dual-state online.

Description of drawings

In order to make the purpose, technical scheme and beneficial effect of the present invention clearer, the present invention provides the following drawings for illustration:

Fig. 1 is the whole method flowchart of the present invention;

FIG. 2 is a detailed flow chart of step S1 in an embodiment of the present invention;

3 is an equivalent circuit model diagram of a power battery in an embodiment of the present invention;

Fig. 4 is the establishment process figure of impedance model of the embodiment of the present invention;

Fig. 5 is a schematic diagram of the thermal model of the power battery in the embodiment of the present invention;

Fig. 6 is the flowchart of experimental data acquisition in step S2 of the embodiment of the present invention;

Fig. 7 is a detailed flowchart of step S3 in the embodiment of the present invention;

FIG. 8 is a flow chart of the particle filter algorithm in step S4 of the embodiment of the present invention.

Detailed ways

The preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

Please refer to Figure 1, the joint state estimation method of SOC and SOT based on the electric-thermal coupling model of the power battery is divided into the following steps:

S1: Select the power battery to be tested, collect and organize the technical parameters of the power battery, establish the electric and thermal model of the power battery, and determine the model parameters required for joint estimation of the SOC and SOT of the power battery;

S2: Conduct trickle (for example, C/20A) charge and discharge experiments and HPPC experiments on the power battery under test at different temperatures, establish an experimental database of equivalent circuit model parameters under charge and discharge conditions on temperature and SOC, and simulate the city's intense Driving conditions (Urban Assault Cycle, UAC) or Artemis hybrid electric vehicle conditions (Artemis HEV), collecting data such as current, voltage, temperature and impedance of the power battery;

S3: Perform parameter identification to obtain the characteristic parameters of the electrical and thermal models, and obtain the quantitative relationship between the parameters of the equivalent circuit model under charge and discharge conditions, temperature and SOC through data fitting;

S4: Combining the electrical-thermal coupling model of the power battery with the PF algorithm and the quantitative relationship between the characteristic parameters of the equivalent circuit model under the charging and discharging conditions of the power battery with respect to temperature and SOC to achieve accurate joint state estimation of the SOC and SOT of the power battery.

Referring to FIG. 2, step S1 specifically includes steps S11-S13.

S11: Select the traction battery to be tested, establish a continuous electrical and thermal model of the traction battery in the time domain, and determine the model parameters required to jointly estimate the SOC and SOT of the traction battery. specifically,

The SOC of the power battery is calculated by the following formula:

Among them, SoC(t) and I(t) refer to the time-varying state of charge and current of the power battery, respectively, η is the Coulombic efficiency, and Q _n is the capacity of the power battery, which will change with the number of battery cycles, temperature and other conditions.

The electrical model of the power battery includes an equivalent circuit model and an impedance model.

Please refer to Figure 3 for the equivalent circuit model. An ohmic internal resistance _Re , two polarized RC pairs including resistors R _s , R _l , C _s , and C _l are connected in series, and the open circuit voltage OCV. The mathematical model can be expressed as :

V _T (t)＝U _OCV (SoC,t)-V _s (t)-V _l (t)-R _e I(t)

Among them, I(t) is the measured battery current, V _s (t), V _l (t) is the polarization voltage of the battery, τ _s = R _s C _s is the short time constant of the battery, τ _l = R _l C _l is the long-time constant, R _s , R _l , C _s and C _l are the polarization resistance and polarization capacitance of the battery, U _OCV (SoC,t) represents the open circuit voltage OCV of the battery, which is a function of the state of charge SOC and time , the expression of V _T (t) can be obtained through Thevenin's theorem of the equivalent circuit, which is a nonlinear relational expression.

For the establishment of the impedance model, please refer to Figure 4. It is approximately assumed that the distribution of the admittance of the power battery along the radial direction is:

where a ₁ , a ₂ and a ₃ are the first, second and third coefficients of the polynomial fit on the admittance and the average battery temperature T(r) is the temperature distribution of the battery along the radial direction.

Please refer to Figure 5 for the establishment of the thermal model. After making reasonable assumptions about the power battery,

The governing equation for establishing the battery is:

Its boundary conditions are:

Where t represents the time, ρ, c _p , k _t represent the volume average density, specific heat capacity and thermal conductivity respectively, V _b represents the volume of the battery, r _o represents the maximum radius of the battery, Q is the heating rate of the battery, h is the convective conversion Thermal coefficient, T _∞ is the temperature of the heat transfer medium.

S12: discretize the calculation formula and equivalent circuit model of the power battery SOC in step S11, and perform order reduction processing on the thermal model to convert it into a control-oriented state space expression.

The calculation formula of power battery SOC and the equivalent circuit model are discretized to obtain the following state space expression:

Equation of state:

Output equation: V _T (k)＝U _OCV (SoC(k))-V _s (k)-V _l (k)-R _e I(k)+v _k

Among them, Δt represents the sampling interval, k represents the sampling time, w _k and v _k are process noise and measurement noise respectively.

After reducing the order of the thermal model of the power battery, the following state space expression is obtained:

y=Cx+Du

in u=[QT _∞ ] ^T , y=[T _c T _s ] ^T are the state, input and output of the control system respectively. The system matrices A, B, C, D are defined as follows:

Wherein, α=k _t /ρc _p is the thermal diffusivity of the battery.

Step S13: Based on the impedance characteristics of the battery, make reasonable assumptions about the battery, and obtain the nonlinear functional relationship of the impedance with respect to the average temperature of the battery, the temperature gradient, and the ambient temperature.

The real part of the impedance represents:

Please refer to FIG. 6 , step S2 specifically includes steps S21-S26.

S21: Leave the power battery to be tested in a constant temperature environment of 25°C for 2 hours;

S22: Charge and discharge the power battery at a charge-discharge rate of C/20, measure the relationship curve between OCV and SOC of the power battery, and determine the available capacity of the power battery at the current stage;

S23: Perform HPPC test to obtain the current and voltage data of the power battery at the current temperature;

S24: Repeat steps S21-S23 every 10°C within the full temperature range of the power battery, and record current and voltage data at different temperatures;

S25: Simulate the UAC or Artemis HEV real vehicle test conditions, and collect experimental data such as current, voltage, temperature and measured impedance of the power battery;

S26: Summarize and process the experimental data obtained before this step to form a usable experimental database.

Please refer to FIG. 7 , step S3 specifically includes steps S31-S33.

S31: Using the experimental data obtained in step S2, determine the quantitative relationship between OCV under charge and discharge conditions with respect to temperature and SOC;

S32: Using the experimental data obtained in step S2 and the quantitative relationship obtained in S31, use the parameter identification method to identify the characteristic parameters of the electrical and thermal models. Step S32 includes S321-S322, specifically,

S321: The identification process of the characteristic parameters R _e , R _s , R _l , C _s , and C _l in the equivalent circuit model is:

The terminal voltage of the power battery can be described in the complex frequency domain as:

where s is the complex frequency domain symbol.

According to the principle of the least square method, the following equation can be constructed by using the equation difference:

Among them, z(k)=U _OCV (k,T _k )-V _T (k)

θ＝[k ₁ k ₂ k ₃ k ₄ k ₅ ]

In the formula, z(k) is the output matrix, θ is the intermediate variable matrix, which is the vector to be identified, is the input matrix. Then use the recursive least squares (Recursive Least Squares, RLS) algorithm to get the characteristic parameters in the battery electric model.

S322: The identification process of the characteristic parameters h, k _t and c _p in the thermal model is:

The objective function used for optimization can be expressed as follows:

Among them, N _f is the number of measurements in this experiment, and θ ^* is the battery parameter value corresponding to the minimum Euclidean distance. e(k,θ) can be expressed in the form of vector difference as follows:

e(k,θ)=[T _c,e (k,θ)T _s,e (k,θ)] ^T -[T _c,exp (k)T _s,exp (k)] ^T

where θ=[k _t c _p h] ^T , T _c,e (k,θ) and T _s,e (k,θ) denote the model estimates of core temperature and surface temperature, respectively, T _c,exp (k) and T _s,exp (k) denote the measured values of core temperature and surface temperature, respectively. The optimization toolbox function fmincon in MATLAB can be used to minimize the Euclidean distance in the vector space, thereby identifying the characteristic parameters of the thermal model.

S33: Based on the experimental data obtained in step S2 and the parameter identification method in step S321, the quantitative relationship between the electrical model parameters and the temperature and SOC under the charging and discharging conditions is obtained through data fitting.

S4, please refer to Figure 8, the relationship data between the open circuit voltage and temperature and SOC under the charge and discharge conditions is used as the open circuit voltage database for the particle filter observer to query during operation, and iteratively based on the electric-thermal coupling model, after The importance sampling stage and the resampling stage obtain the estimated values of the particle filter observers. In the particle filter algorithm, p(X _k |X _k-1 ) is usually used to represent the state transition model, and p(Z _k |X _k ) is used to represent the state observation model, which essentially corresponds to the state equation and the observation equation. Specifically, step S4 includes steps S41-S46.

S41: Initialize relevant parameters, such as number of sampling points, sampling period, variance of process noise and variance of measurement noise, number of particles, etc.

S42: Load the measurement data of the sensor, such as battery surface temperature, core temperature and measured impedance data. During the real-time online observation process, this step can be omitted, and the sensor collected data can directly enter step S44 after processing.

S43: Initialize the particle filter observer, such as initializing the particle set, particle weight array, etc.

S44: Particle set importance sampling stage, this step includes steps S441-S444, specifically,

S441: Particle Importance Sampling

Indicates that the particle set obeys the probability distribution of the reference condition at time k, where Z _1:k = {Z ₁ , Z ₂ ,···,Z _k }.

S442: Calculating particle weights

The weight calculation formula comes from sequence importance sampling, where is the probability distribution of the observed value of the i-th particle in the particle set at time k, is the prior probability distribution of the i-th particle at time k calculated from time k-1, is the probability density function of the reference distribution, and assuming that the state estimation of each step is the optimal estimation, this function only depends on X _k-1 and Z _k .

S443: At the current sampling moment, iterate steps S441-S442.

S444: Particle weight normalization processing

in is the weight of the i-th particle in the particle set at each sampling moment, is the normalized weight of the i-th particle in the particle set at each sampling moment.

S45: Resampling stage, this step includes steps S451-S452, specifically,

S451: According to approximate distribution Generate N random sample sets Calculate the weights according to the selected sampling strategy and normalize the weights Then for the collection of particles Eliminate and replicate.

S452: At the current sampling moment, reset the weight of each particle

in is the normalized weight of the i-th particle at time k, is the weight of the particle at the current moment, and N is the number of random samples generated.

S46: Calculate the particle filter output, the formula is as follows

Among them, p(X _0:k |Z _1:k ) is the posterior probability density function, and δ(dX _0:k ) is the Dirac-delta function.

S47: Steps S44-S46 are iterated, the importance sampling and re-sampling process of the particle set is repeated at each sampling moment, and the estimated value of the particle filter is calculated.

The algorithm of the observer is superior to the extended Kalman filter, and the particle filter is based on probability and statistics, and there is no limit to the process noise and measurement noise of the system. And it should be noted that the algorithm flow described in step S4 is a basic particle filter algorithm. For the actual battery management system, the algorithm can be extended to Extended Kalman Particle Filter (EPF) according to different observation accuracy requirements. , Unscented Kalman Particle Filter (UPF) or Adaptive Particle Filter (Adaptive Particle Filter, APF).

Function and effect of embodiment

According to the SOC and SOT joint state estimation method based on the electric-thermal coupling model of the power battery involved in the present invention, the average temperature state obtained by the online estimation of the thermal model is provided to the electric model to correct the characteristic parameters in the electric model, thereby achieving higher accuracy Then the current open circuit voltage can be calculated by using the high-precision SOC value, and then the heat production power of the battery can be calculated, which is fed back to the thermal model to correct the SOT estimate.

The advantages of the invention of the SOC and SOT joint state estimation method based on the electric-thermal coupling model of the power battery are:

1) Establish an electrical-thermal coupling model based on temperature and current correction for vehicle power batteries, which can accurately obtain the electrical and thermal characteristics of the power battery in the full temperature range;

2) Considering the relationship between the parameters of the equivalent circuit model and the temperature and SOC under the charging and discharging conditions of the power battery, the accurate estimation of the SOC under the actual vehicle working conditions can be realized;

3) The calculation complexity of the electric-thermal coupling model is moderate, and the joint state estimation accuracy of SOC and SOT is enough to be applied to the BMS of real vehicles;

4) An electric-thermal coupling model based on the power battery is proposed, combined with a nonlinear filtering method, to realize a real-time joint estimation method of the power battery SOC and SOT dual-state online.

Finally, it should be noted that the above preferred embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail through the above preferred embodiments, those skilled in the art should understand that it can be described in terms of form and Various changes may be made in the details without departing from the scope of the invention defined by the claims.

Claims (5)

1. state-of-charge SOC and state of temperature SOT united state estimation method based on power battery electric-thermal coupling model, It is characterized in that: method includes the following steps:

S1: selecting power battery to be measured, collects the technical parameter for arranging the power battery, establishes electricity, the hot-die of the power battery Type, and model parameter needed for determining Combined estimator power battery SOC and SOT；

S2: trickle charge-discharge test is carried out to tested power battery at different temperatures and mixed pulses power characteristic HPPC is real Test, establish experimental data base of the equivalent circuit model parameter under the conditions of charge and discharge about temperature and SOC, simcity, suburb, Pure electric automobile EVs, hybrid vehicle HEVs and plug-in hybrid-power automobile under rural road conditions different with high speed PHEVs real steering vectors operating condition establishes real vehicle working condition measurement experimental data base, including electric current, voltage, temperature and impedance data；

S3: carrying out parameter identification and obtain the characterisitic parameter of electricity, thermal model, is fitted by data equivalent under the conditions of obtaining charge and discharge Quantitative relationship between circuit model parameters and temperature and SOC；

S4: under the conditions of the electric-thermal coupling model combination particle filter PF algorithm of power battery and power battery charge and discharge Equivalent-circuit model characterisitic parameter realizes that power battery SOC and SOT united state estimate about the quantitative relation formula of temperature and SOC Meter；

The step S2 specifically:

S21: power battery to be measured is stood into 2h in 25 DEG C of isoperibol；

S22: charge and discharge are carried out to power battery with C/20 charge-discharge magnification, measure the open-circuit voltage OCV and SOC of the power battery Relation curve and determine the current generation power battery active volume；

S23: electric current, voltage data that HPPC test obtains power battery under Current Temperatures are carried out；

S24: every 10 DEG C of repetition step S21-S23 within the scope of the total temperature of the power battery, the electricity under different temperatures is recorded Stream, voltage data；

S25: simulate EVs, HEVs and PHEVs real steering vectors operating condition under different road conditions obtain the power battery electric current, The experimental datas such as voltage, temperature, impedance；

S26: the experimental data that will acquire summarizes and handles, and forms available experimental data base.

2. SOC the and SOT united state estimation method according to claim 1 based on power battery electric-thermal coupling model, It is characterized by: in step sl, the thermal model of the power battery is the unstable state heat heat transfer model or one-dimensional of one-dimensional 1-D Concentration heat model, the electric model of the power battery is the group of one or more of impedance model or equivalent-circuit model It closes.

3. SOC the and SOT united state estimation method according to claim 1 based on power battery electric-thermal coupling model, It is characterized by:

The step S3 specifically:

S31: it using the experimental data obtained in step S2, recognizes to obtain the characteristic ginseng of electricity, thermal model using parameter identification method Number；

S32: using the experimental data obtained in step S2, it is fitted equivalent-circuit model under the conditions of obtaining charge and discharge by data and joins The several and quantitative relationship between temperature and SOC.

4. SOC the and SOT united state estimation method according to claim 1 based on power battery electric-thermal coupling model, It is characterized by: in step s 4, the PF algorithm can replace with Extended Kalman filter, Unscented kalman filtering or H infinity Filter optimal estimation algorithm.

5. SOC the and SOT united state estimation method according to claim 1 based on power battery electric-thermal coupling model, It is characterized by: the parameter identification method is least square method, but is not limited to the algorithm in step S31.