METHOD FOR ESTIMATING OR PREDICTING AN INTERNAL BATTERY STATE OF A BATTERY UNIT

Abstract

A method for estimating or predicting an internal battery state of at least one battery unit within an electric energy storage system, such as in a vehicle. The method includes obtaining operational data of the electric energy storage system relating to operating conditions of the electric energy storage system, feeding the obtained operational data to a non-linear state observer adapted to estimate and/or predict the internal battery state of the at least one battery unit in a series of time steps, such that an observer error of the non-linear state observer converges towards zero, or towards a value close to zero, based on at least the obtained operational data, estimating or predicting the internal battery state using the non-linear state observer.

Claims

1. A method for estimating or predicting an internal battery state of at least one battery unit within an electric energy storage system, the method comprising: obtaining operational data of the electric energy storage system relating to operating conditions of the electric energy storage system, feeding the obtained operational data to a non-linear state observer adapted to estimate and/or predict the internal battery state of the at least one battery unit in a series of time steps, such that an observer error of the non-linear state observer converges towards zero, or towards a value close to zero, based on at least the obtained operational data, estimating or predicting the internal battery state using the non-linear state observer, wherein the non-linear state observer is a switched multi-gain observer switching between at least two different static observer gains, wherein the observer gain to be used is selected based on a predicted or estimated value of the internal battery state as determined by the nonlinear state observer.

2. The method according to claim 1, further comprising: pre-calculating the at least two static observer gains offline, wherein the internal battery state is estimated or predicted online using the non-linear state observer.

3. The method according to claim 1, wherein the at least two static observer gains are pre-calculated offline based on a description of an open circuit voltage of the at least one battery unit as a function of the internal battery state.

4. The method according to claim 1, wherein the observer gain to be used is set to a first observer gain if the estimated or predicted value of the battery state is below a predetermined threshold, and to a second observer gain if the estimated or predicted value of the battery state is above the predetermined threshold.

5. The method according to claim 1, further comprising: based on the obtained operational data relating to a terminal voltage of the at least one battery unit, determining an initial estimate of the internal battery state, wherein the initial estimate is used as an initial value in the estimation or prediction of the internal battery state using the non-linear state observer.

6. The method according to claim 1, wherein, in each time step, the non-linear observer predicts a value of the battery state, wherein the selected observer gain is used for correcting the predicted value.

7. The method according to claim 1, wherein the multi-gain observer is a bimodal observer configured to switch between two static observer gains.

8. The method according to claim 1, wherein the internal battery state is a state-of-charge, SoC, of the at least one battery unit.

9. The method according to claim 1, wherein the non-linear state observer is configured to use a battery model for determining the internal battery state, in which battery model an open circuit voltage of the at least one battery unit is a non-linear function of the state-of-charge of the at least one battery unit.

10. A computer program comprising program code means for performing the method according to claim 1 when said computer program is run on a computer.

11. A computer readable medium carrying a computer program comprising program means for performing the method according to claim 1 when said program means is run on a computer.

12. A control unit configured to perform the method according to claim 1.

13. A battery management system for an electric energy storage system comprising the control unit according to claim 12.

14. A vehicle, such as a hybrid vehicle or a fully electrified vehicle, comprising an electric energy storage system and a control unit according to claim 12.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0033] With reference to the appended drawings, below follows a more detailed description of embodiments of the invention cited as examples.

[0034] In the drawings:

[0035] FIG. 1 shows a vehicle in which a method according to the invention may be implemented,

[0036] FIG. 2 schematically illustrates parts of a battery model describing a battery unit;

[0037] FIG. 3 is a flow-chart illustrating a method according to an embodiment of the invention, and

[0038] FIG. 4 is a diagram showing an OCV curve of a battery unit.

[0039] The drawings are schematic and not necessarily drawn to scale.

DETAILED DESCRIPTION OF EXAMPLE EMBODIMENTS OF THE INVENTION

[0040] In the present detailed description, embodiments of the method according to the present invention are mainly described with reference to an all-electric bus, comprising a propulsion system in the form of battery powered electric motors. However, it should be noted that various embodiments of the described invention are equally applicable for a wide range of hybrid and electric vehicles as well as for stationary applications.

[0041] FIG. 1 shows a simplified perspective view of an all-electric vehicle in the form of a bus 200, which according to an embodiment is equipped with an electric propulsion unit 204 for propulsion of the bus. Of course, other loads may be provided in addition to or instead of the electric propulsion unit 204, for example auxiliary systems requiring electric power, and/or an on-board charger, and/or a power take-off.

[0042] The bus 200 carries an electric energy storage system (ESS) 201 comprising a battery pack with a plurality of battery units 203. The ESS 201 is herein illustrated with a single battery pack, but it is to be noted that it may comprise two or more battery packs, each battery pack comprising a plurality of battery cells. The battery units 203 may e.g. be in the form of battery cells. Some or all of the battery cells 203 within the battery pack may be connected in series to provide an output DC voltage having a desired voltage level. It is possible to arrange the battery cells as a plurality of cell strings, each cell string forming a battery module. The battery modules of the battery pack may in turn be parallel connected. Suitably, the battery cells 203 are of lithium-ion type, but other types may also be used. The number of battery cells 203 per battery pack may be in the range of 50 to 500 cells, or up to many thousands of cells in the case of small format cells. The battery packs of the ESS 201 may be connected in parallel or in series.

[0043] Sensor units (not shown) may be arranged for collecting measurement data relating to operating conditions of the ESS 201, for example by measuring temperature, voltage and current level of the battery cells 203. Measurement data from each sensor unit is transmitted to an associated ESS control unit 208, which is configured for managing the ESS 201 during operation of the bus 200. A single control unit 208 is shown, which may be e.g., a so-called Domain Control Unit, DCU, configured to implement complete control functionality on all levels of the ESS 201. In particular when the ESS 201 comprises more than one battery packs, separate battery management units may however be provided for individual battery pack management. As used herein, the term ESS control unit may also be understood as encompassing such individual battery management units. Thus, the method as described herein may be performed either in a battery management unit of an individual battery pack, or in an ESS control unit handling data from, and operation of, several battery packs.

[0044] The ESS control unit 208 may include a microprocessor, a microcontroller, a programmable digital signal processor or another programmable device. Thus, the ESS control unit 208 comprises electronic circuits and connections (not shown) as well as processing circuitry (not shown) such that the ESS control unit 208 can communicate with different parts of the bus 200 or with different control units of the bus 200. The ESS control unit 208 may comprise modules in either hardware or software, or partially in hardware or software, and communicate using known transmission buses such a CAN-bus and/or wireless communication capabilities. The processing circuitry may be a general-purpose processor or a specific processor. The ESS control unit 208 comprises a non-transitory memory for storing computer program code and data. Thus, the skilled person realizes that the ESS control unit 208 may be embodied by many different constructions. This is also applicable to other control units of the ESS 201.

[0045] Turning now to FIG. 2, there is depicted a battery model comprising an equivalent circuit of the battery unit 203, also known as a Thevenin battery model. The exemplary equivalent circuit model comprises two RC circuits to model the battery unit 203, although a different number of RC circuits may be used in the model, such as one RC circuit or three RC circuits, depending on battery dynamics and application. The exemplary equivalent circuit model may be used in estimation of internal battery states of the battery unit 203 and is typically implemented by the above mentioned ESS control unit 208. The exemplified equivalent circuit model illustrated in FIG. 2 may be used for estimating the open circuit voltage of the battery unit 203 based on direct battery measurements. The battery unit open circuit voltage estimation may for example be based on measured battery current inputs I.sub.b and a battery terminal voltage V.sub.b.

[0046] The equivalent circuit model described in relation to FIG. 2 consists of an active electrolyte resistance and conductive resistance of electrodes (or internal ohmic resistance) R.sub.0, connected in series with two RC branches. A first RC branch and a second RC branch comprise, respectively, capacitances c.sub.1, c.sub.2 and active charge transfer resistances R.sub.1, R.sub.2 connected in parallel. V.sub.b refers to terminal voltage output, l.sub.b refers to the current in the circuit and V.sub.OC refers to the battery open circuit voltage. For given values of the terms V.sub.OC, R.sub.0, R.sub.1, R.sub.2, c.sub.1 and c.sub.2, the terminal voltage V.sub.b can be expressed as a function of the current I.sub.b. Voltages across the internal ohmic resistance R.sub.0 and the first and second RC branches, respectively, are expressed as v.sub.0, v.sub.1 and v.sub.2, wherein v.sub.1 and v.sub.2 are also referred to as voltage polarizations. Normally R.sub.0, R.sub.1 and R.sub.2 increase with age, while battery cell capacity Q (not illustrated in the figure) decreases with age.

[0047] The zero-order hold discrete time state space model of the equivalent circuit in FIG. 2 is:

[00001]$\begin{array}{cc}\left[\begin{array}{c}{\upsilon }_1\left(k+1\right)\\ {\upsilon }_2\left(k+1\right)\\ z\left(k+1\right)\end{array}\right]=\left[\begin{array}{ccc}{e}^{\frac{-\Delta t}{}}\text{?}& 0& 0\\ 0& e^{-\frac{\Delta t}{}\text{?}}& 0\\ 0& 0& 1\end{array}\right]\left[\begin{array}{c}{\upsilon }_1\left(k\right)\\ {\upsilon }_2\left(k\right)\\ z\left(k\right)\end{array}\right]+\left[\begin{array}{c}{R}_1\left(1-e^{-\frac{\Delta t}{}}\text{?}\right)\\ R_2\left(1-e^{-\frac{\Delta t}{}}\text{?}\right)\\ \frac{{\eta }_i\Delta t}{Q}\end{array}\right]u\left(k\right);& \left(1a\right)\end{array}$$\text{?}\text{indicates text missing or illegible when filed}$$\begin{array}{cc}y\left(k\right)={\upsilon }_{\mathrm{ocv}}\left(z\left(k\right)\right)+{\upsilon }_1\left(k\right)+{\upsilon }_2\left(k\right)+R_{0}u\left(k\right),& \left(1b\right)\end{array}$

where z is the state-of-charge (SoC), τ.sub.1=R.sub.1c.sub.1 and τ.sub.2=R.sub.2c.sub.2 are time constants of the circuit, Δt is the sampling time, η.sub.i is the Coulombic efficiency, and v.sub.ocv is the OCV-curve. The applied current u(k) is defined to be positive for charging and negative for discharging.

[0048] For notational simplicity, the following notation for the state space model is introduced:

x(k+1)=Ax(k)+Bu(k)  (2a)

y(k)=h(x(k))+Du(k)  (2b)

where x=[v.sub.1, v.sub.2, z].sup.T. A∈.sup.3×3, B∈.sup.3×3, h: .sup.3.fwdarw., and D∈ correspond to their counterparts in the state space model defined in equations (1a, 1b) above.

[0049] FIG. 3 is a flow-chart illustrating a method for determining an internal battery state of at least one battery unit 203 according to an embodiment of the invention. In short, the method comprises: [0050] the step 101 of obtaining operational data of the ESS 201 relating to operating conditions of the ESS 201; [0051] the step 102 of feeding the obtained operational data to a non-linear state observer adapted to estimate and/or predict the internal battery state of the at least one battery unit in a series of time steps, such that an observer error of the non-linear state observer converges towards zero, or towards a value close to zero; and [0052] the step 103 of, based on at least the obtained operational data, estimating or predicting the internal battery state using the non-linear state observer.

[0053] The steps 101-103 will be described in further detail in the following. For illustrational purposes, the internal battery state will be exemplified by the state-of-charge, SoC, of a single battery unit 203, although the invention is not limited to the determination of SoC but is also applicable to other internal battery states. Furthermore, the invention is not limited to the determination of the internal battery state of a single battery unit 203 but may be used for simultaneous determination of the internal battery state of a plurality of battery units 203. Furthermore, for illustrational purposes, the battery cell 203 may be modelled using an equivalent circuit model with two resistor-capacitor circuits such as the one shown in FIG. 2.

[0054] Step 101: Operational data of the ESS 201 relating to operating conditions of the ESS 201 are obtained by means of the sensor units as discussed above. The operational data may include measurement data relating to current, voltage and temperature of each individual battery unit 203.

[0055] Step 102: The obtained operational data are fed to a non-linear state observer adapted to estimate and/or predict the SoC of the battery unit 203 in a series of time steps, such that an observer error of the non-linear state observer converges towards zero, or towards a value close to zero. The non-linear state observer used herein is a switched multi-gain observer switching between at least two different static observer gains K.sub.1, K.sub.2. Which observer gain to be used is selected based on a predicted or estimated value of the SoC of the battery unit 203 as determined by the nonlinear state observer.

[0056] The multi-gain observer may be a bimodal observer configured to switch between two static observer gains K.sub.1 and K.sub.2, i.e. a so-called bimodal Luenberger observer:

[00002]$\begin{array}{cc}\hat{x}\left(k+1\right)=\{\begin{array}{c}A\hat{x}\left(k\right)+Bu\left(k\right)+K_1\left(y\left(k\right)-\hat{y}\left(k\right)\right),\hat{z}\left(k\right)\le \stackrel{\_}{z}\\ A\hat{x}\left(k\right)+Bu\left(k\right)+K_2\left(y\left(k\right)-\hat{y}\left(k\right)\right),\hat{z}\left(k\right)>\stackrel{\_}{z}\end{array}& \left(3a\right)\end{array}$$\begin{array}{cc}\hat{y}\left(k\right)=h\left(\hat{x}\left(k\right)\right)+Du\left(k\right),& \left(3b\right)\end{array}$

[0057] The objective could for example be to determine K.sub.1∈.sup.3, K.sub.2∈.sup.3, and z∈ such that the estimation error e(k)={circumflex over (x)}(k)−x(k) converges exponentially and asymptotically to zero. It is noted that the SoC to be estimated or predicted, herein expressed as z, forms part of the vector x=[v.sub.1, v.sub.2, z].sup.T.

[0058] The observer gain may be set to a first observer gain K.sub.1 if the estimated or predicted value of the SoC is below or equal to a predetermined threshold, i.e. {circumflex over (z)}(k)≤z, and to a second observer gain K.sub.2 if the estimated or predicted value of the battery state is above the predetermined threshold, i.e. {circumflex over (z)}(k)>z.

[0059] Step 103: The SoC of the battery unit 203 is predicted or estimated using the non-linear state observer based on, at least, the obtained operational data of the battery unit 203. The step 103 of predicting or estimating the SoC is determined online during operation of the ESS 201. For this purpose, the only equations that need to be calculated online are the equations corresponding to (3a)-(3b) listed above. The number of computations needed to be performed online are thereby significantly reduced in comparison to e.g. methods using an Extended Kalman Filter, EKF, for SoC estimation.

[0060] The static observer gains K.sub.1, K.sub.2 may be pre-calculated offline in a step 104. This may be carried out in different ways, but it may typically be performed based on the OCV curve of the battery unit 203, i.e. the curve describing the open circuit voltage V.sub.OC as a function of SoC. This curve may be experimentally defined beforehand. An example of an OCV curve for a battery cell is shown in FIG. 4, wherein the circles represent experimentally determined values. As can be seen from FIG. 4, the OCV curve can relatively accurately be described as a piecewise linear function, comprising a first section for SoC≤0.12 that can be approximated as a first line L.sub.1 having a relatively large slope of 4.7, and a second section for SoC>0.12 that can be approximated as a second line L.sub.2 having a relatively small slope 0.7. The static observer gains K.sub.1, K.sub.2 are in the example calculated using the slopes of the lines L.sub.1 and L.sub.2. This may for example be carried out offline by trial and error, during which a number of different observer gains are tested to find the observer gains K.sub.1, K.sub.2 that result in a convergence of the estimated SoC value to the experimentally determined “true” SoC value. The thus determined observer gains K.sub.1, K.sub.2 are subsequently used in the online estimations carried out in step 103. In each time step, the non-linear observer may predict a value of the SoC, and the observer gain K.sub.1 or K.sub.2 is selected based on the predicted SoC value and used for correcting the predicted SoC value. In the present example, the gain K.sub.1 is selected if the predicted SoC value is ≤0.12, and the gain K.sub.2 is selected if SoC>0.12. Alternatively, the observer gain to be used may be selected based on the SoC estimate as estimated by the non-linear state observer in a preceding time step.

[0061] As an initial value for the estimation of the battery unit SoC, an initial estimate of the battery unit SoC may be determined in a step 105. The initial SoC estimate is used as an initial value in the estimation or prediction of SoC using the non-linear state observer. To provide a suitable initial value for the estimation, the measured terminal voltage V.sub.b can be considered as an estimate of the open circuit voltage (OCV). In turn, this gives an initial SoC estimate using the OCV curve, which initial estimate can be input as an initial value to the non-linear state observer.

[0062] It is to be understood that the present invention is not limited to the embodiments described above and illustrated in the drawings; rather, the skilled person will recognize that many changes and modifications may be made within the scope of the appended claims.