KR101835377B1 - Method for estimating state-of-charge of battery pack based on multivariate wavelet transform and apparatus thereof - Google Patents

Abstract

The present invention provides a method for estimating a state of charge for a battery pack including one or more cells, the method comprising: measuring a cell voltage for the cell; Performing a multivariate wavelet transform using a plurality of cell voltage measurement values as input variables; And calculating a battery pack state estimation value for the battery pack from the wavelet component values of the one or more cells.

Description

METHOD FOR ESTIMATING STATE-OF-CHARGE OF BATTERY PACK BASED ON MULTIVARIATE WAVELET TRANSFORM AND APPARATUS THEREOF BACKGROUND OF THE INVENTION 1. Field of the Invention [0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention [0002] The present invention relates to a technique for estimating a charged state of a battery pack, and more particularly, to a method and apparatus for estimating a charged state of a battery pack using multivariate wavelet transform.

In order to construct high voltage and high capacity applications such as electric vehicle (EV) and energy storage system (ESS), serial, parallel and serial / parallel combination of lithium battery cells Is required. This is called a battery pack.

Like the unit cell, a battery management system (BMS: Battery Management System) is absolutely necessary for stable operation of the battery pack. In such a battery management system, precise measurement of the state of charge (SOC) of the battery pack and balance of voltage and charge state of the cells inside the battery pack are required.

Battery Pack Charged State In terms of precise measurement, an algorithm for estimating the state of charge (SOC) of a battery pack is essential. If the state of charge of the battery pack is correctly checked by the estimation algorithm, it is possible to avoid the degradation of lifetime due to over-charge and over-discharge.

In view of the voltage and charge state balance of the cells inside the battery pack, a screening technique, a unit cell discrimination technique and a balancing technique are currently applied in which a cell having a uniform electrochemical characteristic is selected in advance and a battery pack is constructed. However, the unit cell discrimination technique is costly and time consuming, and the balancing technique has disadvantages in that the circuit design requires an additional cost. Because of these disadvantages, most battery pack manufacturers actually implement unit cell identification techniques simply and minimize balancing techniques.

In this situation, the battery pack charge state estimation algorithm is constructed by the current integration method using only the current of the battery pack, and the current integration method has problems of error accumulation and initial value error. If the charge state estimation algorithm of the battery pack is not properly constructed in a state where the voltage and the charge state imbalance of the cells inside the battery pack are deepened, the life of the battery pack due to overcharging and overdischarging is also added, and the efficient operation of the battery pack becomes more difficult. Therefore, there is a need for a new SOC estimation algorithm technology for ensuring the accuracy of battery pack state estimation.

It is an object of the present invention to provide a method and apparatus for estimating charged state of a battery pack using multivariate wavelet transform.

It is another object of the present invention to provide a battery pack charged state estimation method and apparatus for estimating an overall battery pack state estimation value by statistically analyzing cell state estimation values of cells in a battery pack.

In order to achieve the above object, in one aspect, the present invention provides a method for estimating a state of charge for a battery pack including one or more cells, comprising: measuring a cell voltage for the cell; Performing a multivariate wavelet transform using a plurality of cell voltage measurement values as input variables; And calculating a battery pack state estimation value for the battery pack from the wavelet component values of the one or more cells.

The apparatus using this method may further comprise: calculating a model parameter value for the cell from the wavelet component values for the cell, in calculating the battery pack state estimation value; Calculating a state of charge estimate for the cell from the model parameter value; And statistically analyzing the cell charge state estimation value for the cell to calculate a battery pack charge state estimation value.

The apparatus using this method may calculate an average value from the cell charge state estimation value and calculate the average value as the battery pack charge state estimation value in the statistical analysis step.

The apparatus using this method may calculate the standard deviation value from the cell state estimation value in the statistical analysis and calculate the standard deviation value as the battery pack state estimation value.

In addition, an apparatus using this method can generate a multilevel wavelet component value by the Multi-Resolution Analysis (MRA) method in the step of performing the wavelet transformation.

The apparatus using this method can calculate the model parameter value by selecting the high frequency wavelet component value in the step of calculating the model parameter value.

In the apparatus using this method, in the step of calculating the model parameter value, the low-frequency wavelet component value may be selected to calculate the model parameter value.

In an apparatus using this method, in measuring the cell voltage, the cell voltage can be measured while charging / discharging the cell with a predetermined current profile.

In an apparatus using this method, in the step of calculating the estimated state of charge of the battery pack, the estimated state of charge of the battery pack may be calculated based on Adaptive Control.

Further, in the apparatus using this method, in the current profile, the charge / discharge current value can be changed into a plurality of levels.

Also, in the apparatus using this method, in the current profile, the charge-discharge time may be shorter than the time required to fully discharge the full-cell.

In another aspect, the present invention provides an apparatus for estimating a state of charge for a battery pack including one or more cells, the apparatus comprising: a cell voltage measurement unit for measuring a cell voltage for the cell; A wavelet transform unit for performing a multivariate wavelet transform using a plurality of cell voltage measurement values as input variables; And calculating a battery pack state estimation value for the battery pack from the wavelet component values of the one or more cells.

In this apparatus, the charge state estimation unit may include a model parameter calculation unit for calculating a model parameter value for the cell from the wavelet component values for the cell; A cell charge state calculation unit for calculating a charge state estimation value for the cell from the model parameter value; And a statistical analysis unit for statistically analyzing the cell state estimation value for the cell to calculate the battery pack state estimation value.

The present invention provides a battery pack charge state estimation value at a high speed using multivariate wavelet transform that simultaneously receives a plurality of cell voltages.

In addition, the present invention has an effect of providing a reliable battery pack state estimation value by statistically analyzing state of charge state estimation values of cells.

1 is a configuration diagram of a battery pack charge state estimation system according to an embodiment.
2 is an exemplary waveform diagram of a current profile with respect to charge / discharge current.
FIG. 3 is a diagram illustrating a wavelet transform process according to an exemplary embodiment. Referring to FIG.
4 is a diagram showing the frequency band of the wavelet component values.
5 is a diagram showing an equivalent circuit model of a unit cell of the battery pack.
FIG. 6 is an internal flowchart of a battery pack charge state estimating apparatus according to an embodiment.
FIG. 7 is a flowchart of a method of estimating charged state of a battery pack according to an embodiment.

Hereinafter, some embodiments of the present invention will be described in detail with reference to exemplary drawings. It should be noted that, in adding reference numerals to the constituent elements of the drawings, the same constituent elements are denoted by the same reference symbols as possible even if they are shown in different drawings. In the following description of the present invention, a detailed description of known functions and configurations incorporated herein will be omitted when it may make the subject matter of the present invention rather unclear.

In describing the components of the present invention, terms such as first, second, A, B, (a), and (b) may be used. These terms are intended to distinguish the constituent elements from other constituent elements, and the terms do not limit the nature, order or order of the constituent elements. When a component is described as being "connected", "coupled", or "connected" to another component, the component may be directly connected to or connected to the other component, It should be understood that an element may be "connected," "coupled," or "connected."

The present invention secures charging / discharging voltage data due to voltage sensing of a unit cell in a battery pack, decomposes the data into low frequency and high frequency voltage components by using wavelet transform, calculates parameter values of the electrical equivalent circuit model from the decomposed voltage components Estimates the state of charge of the unit cell by applying the adaptive control method based on the parameter value of the model, and statistically processes the state of charge of the unit cell to estimate the state of charge of the entire battery pack. Here, the battery pack is composed of cells in a serial or parallel combination, and the degree of imbalance of each cell is not considered.

As a premise for calculating the state-of-charge (SOC) estimation value of the battery pack, the estimated state of charge of the unit cell must be calculated. The present invention uses Multivariate Wavelet Transform (MWT) to calculate a unit cell charge state estimation value.

The wavelet transform is a method of decomposing an original signal, if any, into a low frequency (approximation, Am) and a high frequency (detail, Dm) component.

Assuming that the unit cell voltage is the original signal, it can be decomposed into a low-frequency voltage component and a high-frequency voltage component using a multi-resolution analysis (MRA) method of wavelet transform. Then, when the model parameter value of each unit cell is obtained from the decomposed components, the unit cell charge state estimation value can be calculated.

The multivariate wavelet transform is a parallel combination of discrete wavelet transform (DWT). The discrete wavelet transform uses one input. In other words, the voltage of one unit cell is regarded as the original signal, and it is decomposed into low frequency and high frequency voltage components through resolution analysis. However, if one input is applied and the next input is repeated, there is a disadvantage that the time required to estimate the state of charge of the unit cell to calculate the estimated state of charge of the battery pack is excessive. Therefore, in order to calculate the estimation value of the state of charge of the battery pack at a high speed, it is necessary to apply a plurality of inputs to the discrete wavelet transform, that is, a technique of simultaneously applying the voltages of several unit cells and decomposing them into low frequency and high frequency voltage components.

Accordingly, the charged state of the battery pack can be quickly estimated by applying the multivariate wavelet transform as in the following embodiment. Hereinafter, a method and an apparatus for estimating a charged state of a battery pack based on a multivariate wavelet will be described.

1 is a configuration diagram of a battery pack charge state estimation system according to an embodiment.

Referring to FIG. 1, the battery pack charging state estimating system may include a battery pack 10 and a battery pack charging state estimating apparatus 100. The battery pack 10 is configured in series and in parallel with a plurality of cells 11a, 11b, ..., 11n.

The battery pack charge state estimation apparatus 100 may internally include a charge collector unit 110, a cell voltage measurement unit 120, a wavelet transform unit 130, a charge state estimation unit 140, and the like.

The charging and discharging unit 110 can supply a charging current or a discharging current to each of the plurality of cells 11a, 11b, ..., 11n. At this time, the currents Ia, Ib, ..., In flowing into the respective cells 11a, 11b, ..., 11n can be determined according to a preset current profile. The current profile can be characterized by a charge / discharge scheme, a rated discharge current (C-rate), and a pulse application time.

2 is an exemplary waveform diagram of a current profile with respect to charge / discharge current.

The charging and discharging unit 110 can charge and discharge each cell 11a, 11b, ..., 11n using a current profile in which the charging / discharging current value changes at a plurality of levels as shown in Fig.

In the current profile, the charge-discharge time may be shorter than the time required to fully discharge the full-charge cell. The discharge capacity of the battery can be expressed as the time required to reach the rated discharge current. For example, when the battery is discharged for 1 hour at the rated discharge current of 1A, the discharge capacity of the battery that is in the full state can be indicated as 1C. As another example, when the battery is discharged for 2 hours at a rated discharge current of 0.5 A, the discharge capacity of the battery that is in the full state can be indicated by 2C.

In the current profile, the charge-discharge time may be shorter than the time required to discharge the full-charged cell to such a rated discharge current to free up the cell.

Normally, to check the discharge capacity of a battery, it is necessary to measure the time while the battery is fully charged and discharged again. This method is not suitable for a mass production system because it consumes a lot of time (for example, 1 hour, 2 hours) to check the discharge capacity of the battery cell.

The battery pack charge state estimation apparatus 100 according to an embodiment estimates the charge state of each cell 11a, 11b, ..., 11n using a current profile having a charge / discharge time shorter than the time required for checking the conventional discharge capacity, And the charge state of the entire battery pack can be estimated.

1, the cell voltage measuring unit 120 includes a plurality of cells 11a, 11b, ..., 11n while the charging and discharging unit 110 charges and discharges the cells 11a, 11b, ..., 11n. Vn can be measured and the measured cell voltages Va, Vb, ..., Vn can be converted into digital data.

The wavelet transform unit 130 performs a multivariate wavelet transform using the measured values (digital data) of the cell voltages Va, Vb, ..., Vn as input variables, Component values can be generated. Since the plurality of cells 11a, 11b, ..., 11n are charged and discharged at the same time and the measured values are simultaneously subjected to wavelet transform by the multivariate wavelet transform, the calculation time of the battery pack charge state estimation value is shortened by 1 / N .

FIG. 3 is a diagram illustrating a wavelet transform process according to an exemplary embodiment. Referring to FIG.

Referring to FIG. 3, the wavelet transform unit 130 down-samples the cell voltage measurement value to generate an input variable X. At this time, the wavelet transform unit 130 generates input variables Xa, Xb, ..., Xn down-sampled for the plurality of cells 11a, 11b, ..., 11n to perform multivariate wavelet transform .

The wavelet transform unit 130 generates a multilevel wavelet component value using a Multi-Resolution Analysis (MRA) technique. In FIG. 3, it is exemplarily shown that the wavelet transformation proceeds to five steps.

Referring to FIG. 3, the wavelet transform unit 130 generates a low-frequency wavelet component value A1 and a high-frequency wavelet component value D1 through wavelet transformation of the input variable X in a first step. Then, in the second step, the wavelet transform unit 130 generates a low-frequency wavelet component value A2 and a high-frequency wavelet component value D2 through wavelet transformation of A1. In this manner, the wavelet transform unit 130 sequentially performs the third, fourth and fifth wavelet transforms to generate low-frequency wavelet component values A1 to A5 and high-frequency wavelet component values D1 to D5. In this case, since the wavelet transform is performed using the low-frequency wavelet component values of the previous stage in each of the two or more stages, the final low-frequency wavelet component value (for example, A5) and the high- For example, D1 to D5.

4 is a diagram showing the frequency band of the wavelet component values.

Referring to FIG. 4, the wavelet transform unit 130 generates a low frequency wavelet component value Am and high frequency wavelet component values D1 to Dm through M (M is a natural number of 2 or more) wavelet transform steps. At this time, the low-frequency wavelet component value Am generated in the M-th stage is located in the lowest frequency band, and the high-frequency wavelet component value Dm generated in the M-th stage is positioned in the next lowest frequency band.

The cell can be modeled as a capacitor and a resistor. Since the capacitor corresponds to the low frequency band in the frequency band, the characteristic of the low frequency wavelet component value is well represented among the wavelet component values. The model parameter calculator 141 calculates a low-frequency wavelet component value, particularly, a low-frequency wavelet component value generated in the last stage (M-th stage) to select a discharge capacity of the battery, that is, The wavelet component value - can be selected.

The resistance is better in the frequency band higher than the discharge capacity. Accordingly, the model parameter calculator 141 can statistically analyze the high frequency wavelet component value, particularly, the high frequency wavelet component value generated in the last stage (Mth step), out of the wavelet component values.

1, the wavelet component values (for example, A5 and D1 to D5) generated by the wavelet transform unit 130 are transmitted to the charge state estimation unit 140, and the charge state estimation unit 140 selects a wavelet component value for each of the cells 11a, 11b, ..., 11n to calculate a battery pack charged state estimation value.

The charge state estimation unit 140 estimates the state of charge of each cell by applying an adaptive control method or a current integration method based on the wavelet component value of each cell, Estimate the state of charge of the pack.

The adaptive control means that the processing of estimating the internal state with the input is performed first, and the processed value is corrected by the correction algorithm. Extended Kalman Filter (EKF) is a typical example. Sliding Mode Observer (SMO) can also be used to estimate the state of charge of the cell.

A method for estimating the state of charge of each cell using the extended Kalman filter will be described below.

The charge state estimating unit 140 includes a model parameter calculating unit 141, a cell charge state calculating unit 143, and a statistical analyzing unit 145.

The model parameter calculating unit 141 receives the wavelet component values of the cell from the wavelet transform unit 130 and calculates a parameter value for the equivalent circuit model of the cell.

5 is a diagram showing an equivalent circuit model of a unit cell of the battery pack.

5, an equivalent circuit model of a cell includes an internal resistance 510, a charge transfer 520, an electric double layer 530, a diffusion 540, and a parasitic inductance 550 , Parasitic Inductance) as parameters. The internal resistance 510 includes both ends of the battery and is an obstacle to the movement of charges inside the battery. The internal resistance 510 is a value that includes both the resistance at the electrode and the resistance to the movement of ions in the electrolyte. The charge transfer 520 means a chemical energy loss resulting from a change in state between the electrode and the electrolyte occurring in the battery. The electric double layer 530 corresponds to a capacitor-like effect in which the charges between the interfaces are arranged in a constant arrangement. The diffusion 540 means a phenomenon in which a substance is moved to form a concentration gradient inside the electrolyte due to a chemical reaction in the vicinity of the electrode or to solve the concentration gradient. Parasitic inductance (550, parasitic inductance) is the value that appears in the high frequency component and is due to the wire connecting the battery terminal rather than the inherent component of the battery.

The model parameter calculator 141 calculates the values of the internal equivalent resistance 510, the charge transfer 520, the electric double layer 530 and the diffusion 540, which are cell equivalent circuit model parameters, from the wavelet component values. In detail, the model parameter calculating unit 141 can calculate the model parameter values using the EIS (Electrochemical Impedance Spectroscopy) equipment or the response to the step current.

For example, the calculation of model parameter values using the EIS is as follows. An EIS is a device that measures impedance in response to an AC input at a particular frequency. The cell model parameters (internal resistance, charge transfer, electrical double layer, diffusion, etc.) usually appear in different frequency bands. For this reason, when measuring the impedance while changing the frequency, each of the parameters can be easily obtained.

In particular, the model parameter calculator 141 can appropriately select the high-frequency wavelet component value or the low-frequency wavelet component value from the wavelet component values according to the model parameter to be calculated. Since the parasitic inductance can be seen at a large high frequency, the model parameter calculation unit 141 uses the high frequency wavelet component value when the parasitic inductance value is obtained. It is desirable that the internal resistance is measured together with the parasitic inductance in the high-frequency range because the internal resistance appears in the entire frequency range but the other components appear in the low-frequency range. Therefore, when obtaining the internal resistance value, the model parameter calculating section 141 uses the high frequency wavelet component value. On the other hand, since the diffusion is well seen in the low frequency region, the model parameter calculation unit 141 uses the low frequency wavelet component value when the diffusion value is obtained.

The cell charge state calculating unit 143 receives the parameter values for the equivalent model of the cell from the model parameter calculating unit 141 and estimates and calculates the charge state value for the cell.

The cell charging state calculating unit 143 may calculate the cell charging state estimation value based on the adaptive control method. The cell charge state calculator 143 estimates and calculates the charge state value of the unit cell by applying the model parameter values to the Kalman filter. Among the above-described adaptive control methods, an extended Kalman filter (EKF), i.e., a Kalman filter, can be applied. Kalman filter is a model-based internal state estimation method. Thus, the accuracy of the model determines the accuracy of the estimated value. In addition, the Kalman filter is a method that requires a long calculation time. In order to shorten the calculation time, a simplified model is required. In this case, the model parameter calculator 141 calculates the parameter values for the simplified model, and the cell charge state calculator 143 can calculate the cell charge state estimate using the simplified parameter values.

The statistical analysis unit 145 statistically analyzes the estimated state of the cell charge state for the estimated cell to calculate the estimated state of charge of the entire battery pack.

In detail, the statistical analysis unit 145 may finally calculate the representative state of charge of the battery pack from the probability distributions of the cell charge state estimation values of the unit cells. The statistical analysis unit 145 may calculate an average value from the cell charge state estimation values of each unit cell and calculate the average value as the estimated state of charge of the entire battery pack. Also, the statistical analyzer 145 may calculate a standard deviation value as the cell state estimation values of each unit cell, and calculate the standard deviation value as the estimated state of the entire battery pack state.

FIG. 6 is an internal flow chart of a battery pack charge state estimating apparatus according to an embodiment, and FIG. 7 is a flowchart of a battery pack charge state estimating method according to an embodiment.

Referring to FIGS. 6 and 7, when the charging unit 110 charges and discharges a plurality of cells with a predetermined current profile, the cell voltage measuring unit 120 measures the cell voltage for each cell (S702). The wavelet transform unit 130 may perform a multivariate wavelet transform using a plurality of cell voltage measurement values as input variables (S704). The wavelet transform unit 130 is a combined form of a plurality of discrete wavelet transform units, and can perform discrete wavelet transform at the same time by receiving various inputs.

Then, the charge state estimation unit 140 calculates model parameter values from the wavelet component values for each cell (S706), and calculates a cell charge state estimation value based on the calculated model parameter values (S708). Finally, the battery pack state estimation unit 100 statistically analyzes each cell state estimation value to calculate a final state of charge estimation of the battery pack state (S710).

It is to be understood that the terms "comprises", "comprising", or "having" as used in the foregoing description mean that the constituent element can be implanted unless specifically stated to the contrary, But should be construed as further including other elements. All terms, including technical and scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs, unless otherwise defined. Commonly used terms, such as predefined terms, should be interpreted to be consistent with the contextual meanings of the related art, and are not to be construed as ideal or overly formal, unless expressly defined to the contrary.

The foregoing description is merely illustrative of the technical idea of the present invention, and various changes and modifications may be made by those skilled in the art without departing from the essential characteristics of the present invention. Therefore, the embodiments disclosed in the present invention are intended to illustrate rather than limit the scope of the present invention, and the scope of the technical idea of the present invention is not limited by these embodiments. The scope of protection of the present invention should be construed according to the following claims, and all technical ideas within the scope of equivalents should be construed as falling within the scope of the present invention.

Claims (13)

A method for estimating a state of charge for a battery pack comprising one or more cells,
Measuring a cell voltage for the cell;
Performing a multivariate wavelet transform using a plurality of cell voltage measurement values as input variables;
Calculating a battery pack state estimation value for the battery pack from the wavelet component values of the one or more cells,
Wherein the step of calculating the battery pack state estimation value comprises:
Calculating a model parameter value for the cell from the wavelet component values for the cell,
Calculating a state of charge estimate for the cell from the model parameter value, and
Statistically analyzing a cell charge state estimation value for the cell to calculate a battery pack charge state estimation value,
In the statistical analysis,
Calculating a standard deviation value from the cell charge state estimation value and calculating the standard deviation value as the battery pack charge state estimation value. delete delete delete The method according to claim 1,
In performing the multivariate wavelet transform,
A method for estimating a charged state of a battery pack that generates a multilevel wavelet component value by a Multi-Resolution Analysis (MRA) technique. The method according to claim 1,
In the step of calculating the model parameter value,
And calculating a model parameter value by selecting a high frequency wavelet component value. The method according to claim 1,
In the step of calculating the model parameter value,
And selecting a low-frequency wavelet component value to calculate a model parameter value. delete delete A method for estimating a state of charge for a battery pack comprising one or more cells,
Measuring a cell voltage for the cell;
Performing a multivariate wavelet transform using a plurality of cell voltage measurement values as input variables;
Calculating a battery pack state estimation value for the battery pack from the wavelet component values of the one or more cells,
In the step of measuring the cell voltage,
The cell voltage is measured while charging / discharging the cell with a preset current profile,
In the current profile,
Wherein the charge / discharge current value is changed into a plurality of levels. A method for estimating a state of charge for a battery pack comprising one or more cells,
Measuring a cell voltage for the cell;
Performing a multivariate wavelet transform using a plurality of cell voltage measurement values as input variables;
Calculating a battery pack state estimation value for the battery pack from the wavelet component values of the one or more cells,
In the step of measuring the cell voltage,
The cell voltage is measured while charging / discharging the cell with a preset current profile,
In the current profile,
Wherein the charging and discharging time is shorter than the time required for full discharge of the fully charged cell. An apparatus for estimating a state of charge for a battery pack including one or more cells,
A cell voltage measuring unit for measuring a cell voltage for the cell;
A wavelet transform unit for performing a multivariate wavelet transform using a plurality of cell voltage measurement values as input variables;
And a charge state estimation unit for calculating a battery pack charge state estimation value for the battery pack from the wavelet component values of the one or more cells,
The charge state estimating unit may include:
A model parameter calculator for calculating a model parameter value for the cell from the wavelet component values for the cell,
A cell charge state calculation unit for calculating a charge state estimation value for the cell from the model parameter value,
And a statistical analysis unit for statistically analyzing a state of charge of the cell to estimate a state of charge of the cell to calculate a state of charge of the battery pack,
Wherein the statistical analysis unit comprises:
A standard deviation value is obtained from the cell state estimation value, and the standard deviation value is calculated as the battery pack state estimation value. delete

Images