JP2007057234A - Internal resistance estimator for secondary battery - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an internal resistance estimator for a secondary battery to accurately estimate an internal resistance value even in a state where it is difficult to perform estimating calculation by means of an adaptive digital filter as in constant current discharge or large current discharge. <P>SOLUTION: This internal resistance estimator is structured so as to choose between a first internal resistance value estimated by means of the adaptive digital filter and a second internal resistance value found from a second charging rate estimation value based on current integration, etc. in accordance with the state of current. In a state where charging rate estimation by means of the adaptive digital filter is difficult, the second charging rate estimation value based on current integration, etc. is chosen. In the other cases, a first charging rate estimation value (and the first internal resistance value) by means of the adaptive digital filter are chosen. In cases where the second internal resistance value is estimated from the second charging rate estimation value, stored values T<SB>1</SB>_<SB>ADF</SB>(k) and T<SB>2</SB>_<SB>ADF</SB>(k) of battery parameter estimation values by means of the digital filter just before choosing a second charging rate are held, and these are used for operating the second internal resistance value K_<SB>ITG</SB>(k). <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to an apparatus for estimating an internal resistance value of a secondary battery.

The charging rate estimation device for a secondary battery (also simply referred to as “battery”) described in Patent Document 1 (Japanese Patent Laid-Open No. 2004-178848) is a charge / discharge current I (also simply referred to as “current”) of the secondary battery. And internal voltage (time constant and internal resistance) of the battery using an adaptive digital filter from the terminal voltage V (also simply referred to as “voltage”), and using this, the open circuit voltage V ₀ is estimated and measured in advance. The charging rate SOC is estimated based on the open circuit voltage-charging rate characteristic data of the battery. At this time, the internal resistance value K can be estimated as a component of the parameter vector that is collectively estimated.
In Patent Document 2 (Japanese Patent Laid-Open No. 2000-323183), the detected current I and voltage V are approximated by an IV straight line (V = K · I + V ₀ ), the slope is the internal resistance value K, and the intercept is It estimates the internal resistance as the open circuit voltage V _0.
Further, in Patent Document 3 (Japanese Patent Application Laid-Open No. 2003-054035), power P _{in that} can be input from the estimated internal resistance value K, open circuit voltage V ₀ , maximum allowable terminal voltage V _MAX , and minimum allowable terminal voltage V _{MIN of the} battery. And the output power P _out is estimated.
JP 2004-178848 A JP 2000-323183 A JP 2003-054035 A

Since the battery is formulated as a linear model in the method described in Patent Document 1, an adaptive digital filter is used for slow battery parameter changes such as changes in charging rate, battery temperature, or battery deterioration. Since parameter changes can be tracked (estimated) successively with high accuracy, the charging rate and input / output power can be accurately estimated. However, in use areas where the battery exhibits strong nonlinearity, the adaptive digital filter cannot accurately estimate the battery parameters and the charging rate. Further, under the condition that the input current becomes a constant value, there is a problem that the adaptive digital filter theoretically does not update the parameter estimation value and the parameter cannot be estimated sequentially.
Note that the battery exhibits strong non-linearity, for example, corresponding to a phenomenon (see FIG. 5) in which the internal resistance value rapidly increases when energy is taken out with a large output in a lithium ion battery (current I is made to flow considerably large). The condition that the input current becomes a constant value corresponds to the case where charging / discharging at a constant current continues (several seconds or more).
Further, in the method described in Patent Document 2, although the actual battery has a dynamic characteristic indicating a transient response, the equation (T ₁ , T ₂ ) representing the transient characteristic of the battery is ignored. Since the internal resistance value is estimated, there is a problem that an error occurs in the estimation result of the internal resistance value in a transient state in which a current change occurs.
Therefore, as described in Patent Document 3, when the estimated internal resistance value is used to estimate the charging rate, the input possible power P _in , and the output possible power P _out , the internal resistance is estimated as described above. When an error occurs, there is a problem that an error also occurs in the charge rate, the input power P _in , and the output power P _out estimated using the error.
The present invention has been made to solve the above-described problems, and can accurately estimate the internal resistance value even in a state where estimation calculation by an adaptive digital filter is difficult, such as during constant current discharge or large current discharge. An object of the present invention is to provide an internal resistance estimation device for a secondary battery.

In order to achieve the above object, in the present invention, the first internal resistance value estimated by the adaptive digital filter and the second internal resistance value obtained from the second charging rate estimated value by another method such as current integration are obtained. It is configured to be selected according to the current state, and when the charge rate estimation by the adaptive digital filter is difficult, the second charge rate estimate value by other methods such as current integration is selected. When the first charging rate estimated value (and the first internal resistance value) by the adaptive digital filter is selected and the second internal resistance value is estimated from the second charging rate estimated value, the second charging rate is selected. The stored values T _{1 _ADF} (k) and T _{2 _ADF} (k) of the battery parameter estimated values by the adaptive digital filter immediately before the hold are held, and the second internal resistance value _{K_ITG} (k) is calculated using them. Configured as Yes.

In the present invention, in the case of constant current discharge or large current discharge, that is, in the state where it is difficult to estimate the charging rate by the adaptive digital filter, the second charging rate estimated value by a method such as current integration is selected. Selects the first charging rate estimated value (and the first internal resistance value) by the adaptive digital filter, so that it is possible to always perform the charging rate estimation and the internal resistance estimation with little error. Further, when the second internal resistance value is estimated from the second charging rate estimated value, the stored value T _{1 _ADF} (k), T of the battery parameter estimated value by the adaptive digital filter immediately before the second charging rate is selected. _{2 _ADF} (k) is held and the second internal resistance value _{K_ITG} (k) is calculated using the hold, up to a time constant (T ₁ and T ₂ ) representing the dynamic characteristics of the battery Therefore, the internal resistance value can be estimated with high accuracy even in a transient state where the current changes.

FIG. 1 is a functional block diagram showing an embodiment of the internal resistance estimation calculation of the present invention. FIG. 1 exemplifies a case where the present invention is applied to a device that estimates the internal resistance value and open circuit voltage of a battery, and further estimates and calculates input / output possible power using those estimated values.
The block diagram shown in FIG. 1 includes the following elements. That is, the current detection means 100 and the voltage detection means 101 for periodically detecting the current charged / discharged in the secondary battery and the terminal voltage of the battery, and the low-pass current from the measured current I (k) and voltage V (k). The filter values I ₁ (k), V ₁ (k), approximate first-order differential values I ₂ (k), V ₂ (k), and approximate second-order differential values I ₃ (k), V ₃ (k) are calculated. Pre-processing filter calculation means 102, adaptive digital filter calculation means 103 for collectively estimating parameters (K, T ₁ , T ₂ ) representing the internal state of the battery by an adaptive digital filter from the output of the pre-processing filter 102; A first open circuit voltage calculation means 104 for estimating a first open circuit voltage V _{0_ADF} (k) from the output of the preprocessing filter calculation means 102 and the battery parameter estimated by the adaptive digital filter calculation means 103; The first charging rate for estimating the first charging rate SOC _ADF (k) from the first opening voltage estimated value V _{0_ADF} (k) based on the open circuit voltage-charging rate characteristic (see FIG. 4 described later) of the battery acquired in advance. Based on the estimation means 105, the second charge rate estimation means 106 for estimating the second charge rate from the current detected by the current detection means 100 by current integration, and the pre-measured open circuit voltage-charge rate characteristic (FIG. 4). A second open-circuit voltage calculating means 107 for estimating a second open-circuit voltage from the second charge rate, and a final charge-rate estimation for selecting either the first charge rate or the second charge rate according to the current state The battery internal parameter estimated value calculated by the adaptive digital filter calculating means 103 is stored based on the value selecting means 108 and the information (final charging rate estimated value selection flag) of the final charging rate estimated value selecting means 108. Pond parameter estimated value storage means 109, second internal resistance estimation means for calculating a second internal resistance value from the stored value of the battery parameter estimated value, the second open circuit voltage estimated value, and the output of the preprocessing filter calculating means 102 110, and the final open circuit voltage estimated value that is output by switching the first open circuit voltage estimated value and the second open circuit voltage estimated value according to the information (final charge rate estimated value selection flag) of the final charge rate estimated value selection means 108 The selection unit 111, the first internal resistance value estimated by the adaptive digital filter (corresponding to the parameter K) and the second internal resistance value calculated by the second internal resistance estimation unit 110 are used to select the final charge rate estimation value. Final internal resistance estimated value selection means 112 that switches and outputs it according to information (final charge rate estimated value selection flag) of means 108, and the final internal resistance estimated value selection means And an output enable power calculating means 113 for calculating the available input and output power from the selected final internal resistance value and the final estimate open circuit voltage value at 112. However, (k) indicates a value at time point k (calculated value in the current calculation).

  In this embodiment, the second charging rate estimation unit 106 is exemplified by a configuration for estimating the second charging rate by current integration. However, the second charging rate estimation unit 106 does not use an adaptive digital filter. In other words, any method can be used as long as the charging rate can be estimated even when the estimation calculation by the adaptive digital filter is difficult, such as during constant current discharge or large current discharge. As such an estimation method, there are a charging rate estimation method using a Kalman filter, a method using an open circuit voltage (described in Patent Document 2), and the like in addition to the above-described method based on current integration.

FIG. 2 is a specific configuration diagram of an embodiment of the present invention. This embodiment has a function of estimating and calculating the internal resistance value, the charging rate, etc. of the secondary battery in a system in which a load such as a motor is driven by the secondary battery or the secondary battery is charged by regeneration of the motor. An example is shown.
In FIG. 2, 10 is a secondary battery, 20 is a load of a motor, etc., 30 is an electronic control unit (battery controller) that estimates the charging rate, full charge capacity, etc. of the secondary battery. And a microcomputer composed of a ROM and a RAM for storing calculation results and an electronic circuit. Reference numeral 40 denotes a current sensor that detects current charged / discharged from the battery, and reference numeral 50 denotes a voltage sensor that detects terminal voltage (also simply referred to as voltage) of the battery, and each is connected to the electronic control unit 30. The electronic control unit 30 includes the preprocessing filter calculation means 102, adaptive digital filter calculation means 103, first open circuit voltage calculation means 104, first charging rate estimation means 105, second charging rate estimation means 106, Two open circuit voltage calculation means 107, final charging rate estimated value selection means 108, battery parameter estimated value storage means 109, second internal resistance estimation means 110, final open circuit voltage estimated value selection means 111, final internal resistance estimated value selection means 112 and This corresponds to the input / output capable power calculation means 113. The current sensor 40 corresponds to the current detection means 100, and the voltage sensor 50 corresponds to the voltage detection means 101, respectively.

First, a battery parameter (K, T ₁ , T ₂ ) estimation method using an adaptive digital filter performed by the adaptive digital filter calculation means 103 in FIG. 1 will be described.
FIG. 3 is a circuit diagram showing a battery model used for the above estimation. In FIG. 3, the input to the model is current I (t) [A] (positive value: charge, negative value: discharge), and the output is terminal voltage V (t) [V], open circuit voltage V ₀ (t) [V]. R ₁ is a charge transfer resistance, R ₂ is a pure resistance, and C ₁ is an electric double layer capacitance. Although this model is a reduction model (primary) in which the positive electrode and the negative electrode are not particularly separated, it is possible to show the actual charge / discharge characteristics of the battery relatively accurately.
This battery model can be expressed by Equation (4) with s as a differential operator.

上記（数４）式を変形すると下記（数５）式になる。

When the above equation (4) is modified, the following equation (5) is obtained.

開路電圧Ｖ_０(ｔ）は、電流Ｉ(ｔ）に可変な効率ｈを乗じたものを、ある初期状態から積分したものと考えれば、下記（数６）式で示すことが出来る。

The open circuit voltage V ₀ (t) can be expressed by the following equation (6), assuming that the current I (t) multiplied by the variable efficiency h is integrated from a certain initial state.

（数６）式を（数５）式に代入すれば（数７）式になり、整理すれば（数８）式になる。

Substituting equation (6) into equation (5) yields equation (7), and rearranging results in equation (8).

安定なローパスフィルタＧ_ＬＰＦ(ｓ）を（数８）式の両辺に乗じ、整理すれば（数９）式になる。なお、ローパスフィルタＧ_ＬＰＦ(ｓ）の詳細については後述する。

When a stable low-pass filter G _LPF (s) is multiplied by both sides of the equation (8) and rearranged, the equation (9) is obtained. Details of the low-pass filter G _LPF (s) will be described later.

ここで、実際に計測可能な電流Ｉ(ｔ）や端子電圧Ｖ(ｔ）を、ローパスフィルタやバンドパスフィルタ（図１の前処理フィルタ演算手段１０３で処理した値を下記（数１０）式のように定義し、これを前処理フィルタ演算手段において演算する。なお、（ｔ）は時間の関数であることを示している。

Here, the values obtained by processing the current I (t) and the terminal voltage V (t) that can be actually measured by the low-pass filter or the band-pass filter (preprocessing filter calculation means 103 in FIG. This is calculated by the preprocessing filter calculation means, where (t) indicates a function of time.

なお、本実施例においては、安定なローパスフィルタとして（数１１）式で示される特性を用いたが、その選び方はこれに限定されない。

In the present embodiment, the characteristic represented by the equation (11) is used as a stable low-pass filter, but the selection method is not limited to this.

ただし、ｐはフィルタの時定数である。
上記（数１０）式で示す変数を用いて（数９）式を書き直し、Ｖ_２(ｔ）について整理すれば、（数１２）式になる。

Here, p is a time constant of the filter.
Rewriting (Equation 9) using the variables shown in the above (Equation 10) and rearranging V ₂ (t) yields (Equation 12).

（数１２）式は、計測可能な値と未知パラメータの積和式になっているので、一般的な適応デジタルフィルタの標準形（数１３）式と一致する。

Since the equation (12) is a product-sum equation of a measurable value and an unknown parameter, it agrees with a standard form (equation 13) of a general adaptive digital filter.

ただし、
ｙ＝Ｖ_２(ｔ）、 ω^Ｔ＝［Ｖ_３(ｔ），Ｉ_３(ｔ），Ｉ_２(ｔ），Ｉ_１(ｔ）］、
θ^Ｔ＝［−Ｔ_１，Ｋ・Ｔ_２，Ｋ，ｈ］
である。
従って、電流Ｉと端子電圧Ｖにフィルタ処理を施した信号を、適応デジタルフィルタ演算に用いることで、未知パラメータベクトルθを推定することが出来る。

However,
y = V ₂ (t), ω ^T = [V ₃ (t), I ₃ (t), I ₂ (t), I ₁ (t)],
θ ^T = [− T ₁ , K · T ₂ , K, h]
It is.
Therefore, the unknown parameter vector θ can be estimated by using a signal obtained by filtering the current I and the terminal voltage V in the adaptive digital filter calculation.

The present embodiment has improved the logical disadvantage of a simple “adaptive digital filter based on the least squares method”, that is, once the estimated value has converged, accurate estimation cannot be performed again even if the parameter changes thereafter. "Limit trace gain method" is used.
A sequential estimation algorithm for estimating the unknown parameter vector θ based on the above equation (13) is as shown in equation (14). However, the parameter estimation value at the time point k is θ ^ (k). In addition, although ^ attached to the right shoulder of θ represents an estimated value, it is attached immediately above θ in the mathematical formula (hereinafter the same).

ここで、trace｛Ｑ(ｋ)｝は行列Ｑ(ｋ)のトレースを意味する。また、λ_１、λ_３、γ_Ｕ、γ_Ｌは設計パラメータ（初期設定値）で、０＜λ_１＜１、０＜λ_３＜∞とする。λ_３は適応デジタルフィルタの推定演算の応答速度を設定する定数（調整ゲイン）であり、値を大きくすることにより応答速度は速くなるが、その反面ノイズの影響を受けやすくなる。γ_Ｕおよびγ_Ｌはそれぞれ行列Ｑ(ｋ)のトレースの上下限を規定するパラメータであり、０＜γ_Ｌ＜γ_Ｕとなるように設定する。また、Ｐ(０）は十分大きな値、θ(０）は非ゼロで十分小さな値を初期値とする。
以上が適応デジタルフィルタ演算部において行われる適応デジタルフィルタを用いた電池パラメータ推定方法である。

Here, trace {Q (k)} means tracing of the matrix Q (k). Also, λ ₁ , λ ₃ , γ _U , and γ _L are design parameters (initial setting values), and 0 <λ ₁ <1, 0 <λ ₃ <∞. lambda ₃ is a constant (adjustment gain) to set the response speed of the estimation calculation of the adaptive digital filter, the response speed is faster by increasing the value becomes susceptible to the contrary noise. γ _U and γ _L are parameters that define the upper and lower limits of the trace of the matrix Q (k), and are set such that 0 <γ _L <γ _U. Further, P (0) is a sufficiently large value, and θ (0) is a non-zero and sufficiently small value as an initial value.
The above is the battery parameter estimation method using the adaptive digital filter performed in the adaptive digital filter calculation unit.

Next, processing in the first open circuit voltage calculation means 104 in FIG. 1 will be described.
First, when formula (5) is rearranged with respect to the open circuit voltage V ₀ , formula (15) is obtained.

開路電圧Ｖ_０の変化は穏やかであると考え、この両辺に安定なローパスフィルタＧ_ＬＰＦ(ｓ）を乗じた値を開路電圧推定値として（数１６）式で推定する。

The change in the open circuit voltage V ₀ is considered to be gentle, and a value obtained by multiplying both sides by a stable low-pass filter G _LPF (s) is estimated by the equation (16) as an open circuit voltage estimated value.

（数１６）式に（数１０）式を代入すると、（数１７）式となる。

Substituting (Equation 10) into (Equation 16) yields (Equation 17).

従って、（数１７）式に前記適応デジタルフィルタを用いて推定した電池パラメータ推定値（Ｔ^_１，Ｔ^_２，Ｋ^）と前処理フィルタの出力（Ｉ_１(ｋ）、Ｉ_２(ｋ）、Ｖ_１(ｋ）、Ｖ_２(ｋ））を代入することで第一開路電圧推定値Ｖ_０__ＡＤＦが算出できる。
さらに図１中の第一充電率推定手段１０５において、予め取得した当該電池の開路電匠−充電率特性（図４参照）に基づいて、前記第一開路電圧推定値から第一充電率を算出することができる。

Therefore, (Equation 17) the adaptive digital filter a battery parameter value estimated by using the formula _{(T ^ 1, T ^ 2} , K ^) and the output of the preprocessing filter _{(I 1 (k), I} 2 (k ), V ₁ (k), V ₂ (k)) can be substituted to calculate the first open circuit voltage estimated value V _{0 _ADF} .
Further, in the first charging rate estimation means 105 in FIG. 1, the first charging rate is calculated from the first open circuit voltage estimated value based on the previously obtained open circuit design-charging rate characteristic of the battery (see FIG. 4). can do.

Next, processing performed by the electronic control unit 30 in FIG. 2 will be described with reference to the flowchart shown in FIG. Note that the processing of FIG. 6 is performed at regular intervals T ₀ (T ₀ = 50 msec in this embodiment). In the following description, I (k) is a current value (current measurement value) of the current execution cycle, I (k−1) is a current value (previous measurement value) of the previous execution cycle, The same applies to values other than current.

First, in step 1, the charge / discharge current I (k) is detected based on the signal from the current sensor 40, and the terminal voltage V (k) of the secondary battery is detected based on the signal from the voltage sensor 50.
In step 2, a selection flag f_SOCSW for selecting one of the charging rates estimated by the two methods described later as the final estimated value from the current value I (k) and the current change rate ΔI (k) is as follows. Generate.
The current change rate ΔI (k) is, for example, the absolute value of the difference between the previous value and the current value, as shown by the equation (18), or the change within a predetermined time as shown in the equation (19). It can be the width (difference between the maximum and minimum values).

ただし、（数１９）式においてＭＡＸおよびＭＩＮはそれぞれ、最大値および最小値を示す関数であり、ｎは設定する所定時間を表す。

In Equation (19), MAX and MIN are functions indicating the maximum value and the minimum value, respectively, and n represents a predetermined time to be set.

(Conditions for selection flag f_SOCSW)
(1) In the case of constant current discharge where the current change rate ΔI (k) is equal to or less than the predetermined value β continuously for the predetermined time ts, the final charge rate estimated value selection flag f_SOCSW is set. In this case, the second charging rate estimated value (current integrated SOC estimated value) is selected.
(2) When the current becomes a large current exceeding the predetermined value γ, the selection flag f_SOCSW is set. Also in this case, the second charging rate estimated value (current integrated SOC estimated value) is selected.
FIG. 5 is a characteristic diagram showing the relationship between the discharge current and the internal resistance value in the lithium ion battery. As shown in FIG. 5, when energy is extracted at a high output in a lithium ion battery (current I is made to flow considerably large), a phenomenon occurs in which the internal resistance value increases rapidly. In such a case, an adaptive digital filter ( SOC estimation by ADF) becomes difficult.
(3) In cases other than the above, the selection flag f_SOCSW is cleared. In this case, the first charging rate estimated value (SOC estimated value by adaptive digital filter ADF) is selected.
In step 3, the current I (k) and voltage V (k) detected in step 1 are subjected to low-pass filter processing and approximate differential filter processing shown in the following equation (20), and I ₁ (k), I ₂ (k), I ₃ (k), V ₁ (k), V ₂ (k), and V ₃ (k) are calculated.

これらは、実際には（数２０）式をタスティン近似などで離散化して得られた漸化式を用いて演算する。
ｓｔｅｐ４では、電流積算によるＳＯＣ推定を行うか否かを選択する。選択フラグｆ_ＳＯＣＳＷがクリアされている場合には、ｓｔｅｐ５に進み、電流積算によるＳＯＣ（第二充電率）推定の初期化処理のみを行い、セットされている場合にはｓｔｅｐ６に進み、電流積算によるＳＯＣ（第二充電率）推定を行う。
ｓｔｅｐ５では、（数２１）式に基づき前回のＡＤＦによる第一充電率推定値ＳＯＣ_ＡＤＦ(ｋ−１）を用いて電流積算による第二充電率推定値ＳＯＣ_ＩＴＧ’(ｋ）を初期化することで、電流積算用積分器の初期値とする。

These are actually calculated using a recurrence formula obtained by discretizing Formula (20) by Tustin approximation or the like.
In step 4, it is selected whether or not to perform SOC estimation by current integration. If the selection flag f_SOCSW is cleared, the process proceeds to step 5, and only the initialization process of SOC (second charge rate) estimation by current integration is performed. If it is set, the process proceeds to step 6, and SOC by current integration is performed. (Second charging rate) is estimated.
In step 5, the second charge rate estimated value SOC _ITG '(k) by current integration is initialized using the first charge rate estimated value SOC _ADF (k-1) by the previous ADF based on the equation (21). Thus, the initial value of the integrator for current integration is used.

ｓｔｅｐ６では、後述する総容量推定値Ｑ_ＭＡＸを用いて、（数２２）式に基づき電流を積算（積分）することにより、電流積算による第二充電率推定値ＳＯＣ_ＩＴＧ'(ｋ）を演算する。

In step 6, the second charge rate estimated value SOC _ITG ′ (k) based on the current integration is calculated by integrating (integrating) the current based on the equation (22) using a total capacity estimation value Q _MAX described later. .

ｓｔｅｐ７では、ｓｔｅｐ６で算出した第二充電率推定値ＳＯＣ_ＩＴＧ'(ｋ）に対して、（数２３）式に示すｓｔｅｐ３で述べた前処理フィルタを用いて、それと同等の遅れＧ_ＬＰＦ(ｓ）を持つローパスフィルタ処理を施し、（数２４）式に示すようにフィルタ処理後の第二充電率推定値ＳＯＣ_ＩＴＧ(ｋ）を算出する。

In step 7, the second charging rate estimated value SOC _ITG ′ (k) calculated in step 6 is used for the delay G _LPF (s) equivalent to that by using the preprocessing filter described in step 3 shown in equation (23). Is applied to calculate the second charge rate estimated value SOC _ITG (k) after the filter process as shown in the equation (24).

ただし、ｐはフィルタの時定数である。
実際には（数２４）式をタスティン近似などで離散化して得られた漸化式を用いて演算する。
ｓｔｅｐ８では、予め計測した開路電圧−充電率特性（図４）に基づいて、ｓｔｅｐ７で電流積算によって求めた第二充電率推定値ＳＯＣ_ＩＴＧ(ｋ）から第二開路電圧推定値Ｖ_０__ＩＴＧ(ｋ）を求める。

Here, p is a time constant of the filter.
Actually, the calculation is performed using a recurrence formula obtained by discretizing Formula (24) by Tustin approximation or the like.
In step8, previously the measured open circuit voltage - based on the charging rate characteristic (FIG. 4), the second charging rate estimate _{SOC ITG} (k) determined by the current integration in step7 second open-circuit voltage estimated value _{V 0} _ _ITG ( k).

In step 9, the second internal resistance value of the battery is estimated from the second open circuit voltage estimated value V _{0 —ITG} (k) calculated in step 8 and the parameter estimated by ADF. The detailed method will be described later.
In step 10, using I ₁ (k), I ₂ (k), I ₃ (k), V ₂ (k), and V ₃ (k) calculated in step ₃ , the adaptation shown in the equation (14) is used. The battery parameter estimated value θ ^ (k) by the digital filter is calculated. However, in Equation (14),
y (k) = V ₂ (k)
ω ^T (k) = [V ₃ (k) I ₃ (k) I ₂ (k) I ₁ (k)]
θ ^T (k) = [− T ₁ (k) K (k) · T ₂ (k) K (k) h (k)]
It is.
Then, the internal resistance value K (k) that is a component of the estimated battery parameter estimated value θ ^ (k) is set as the first internal resistance estimated value.

In step 11, T ₁ (k), K (k) · T ₂ (k), K (k), and I ₁ (k) calculated in step 9 among the battery parameter estimated values θ ^ (k) calculated in step 10 are used. ), _{I 2} (k) and _{V 1} (k), by substituting _{V 2} a (k) (the number 17), calculates a first open-circuit voltage estimated value _{V 0} _ _ADF based on the adaptive digital filter.
Equation (17) is an equation obtained by modifying the battery model (Equation 15) and multiplying both sides by a low-pass filter, and is actually estimated as a value obtained by applying a low-pass filter to the open circuit voltage. However, since the open circuit voltage changes moderately, V _{0 — ADF} (k) can be substituted with G _LPF (s) · V _{0 — ADF} (k).

In step 12, the first charge rate estimated value SOC _ADF by _ADF is calculated from the first open circuit voltage estimated value V ₀ _ADF (k) calculated in step 11 using the correlation characteristic of the open circuit voltage and the charge rate shown in FIG. 4 measured in advance. (k) is calculated. In FIG. 4, V _L is an open circuit voltage corresponding to SOC = 0%, and V _H is an open circuit voltage corresponding to SOC = 100%.

In step 13, the final charging rate estimated value selection flag calculated by the current I and (k) step2, and a second charging rate estimated value SOC_ _ITG (k) calculated in step6, first charging rate estimated value SOC calculated in step12 Based on the selection flag f_SOCSW calculated at step 2 from _ADF (k), the final charge rate estimated value SOC _est (k) is determined.
In other words, as described in the above (selection flag f_SOCSW condition), when the selection flag f_SOCSW is set, the second charging rate estimated value (current integrated SOC estimated value) is selected, and the selection flag f_SOCSW is cleared. If so, the first charging rate estimated value (SOC estimated value by adaptive digital filter (ADF)) is selected.

In step 14, the final open circuit voltage estimated value V _{0 _est} (k) is calculated from the final charging rate estimated value SOC _est (k) selected in step 13 based on FIG. Incidentally, in response to said selection flag F_SOCSW, select either the computed second open-circuit voltage estimated value SOC_ _ITG (k) and the first open-circuit voltage estimated value _{V 0} _ _ADF calculated in step 11 (k) at step7 Then, it may be configured to be the final open circuit voltage estimated value V _{0 — est} (k).
That is, when the selection flag f_SOCSW is set, select the second open-circuit voltage estimated value SOC_ _ITG (k), when the selection flag f_SOCSW is cleared, the first open-circuit voltage estimated value _{V 0} _ _ADF ( k) is selected. FIG. 1 shows this method.

In step 15, as the internal resistance estimated value for input and output can power operation later in step16, in the same manner as in step 13, a first inner calculated by the second internal resistance estimated value K_ _ITG and (k) step 10 calculated in step9 Either one of the estimated resistance value K _ADF (k) is selected and determined as the final estimated internal resistance value K_ _est (k).
That is, when the selection flag f_SOCSW has been set, the second internal resistance estimated value K_ _ITG (k) of selecting, when the selection flag f_SOCSW is cleared, the first internal resistance estimated value _{K ADF} (k) Select.

In step 16, using the final open-circuit voltage estimated value V _{0 — est} (k) calculated in step 14 and the final internal resistance estimated value K_ _est (k) calculated in step 15, for example, input that is defined by equation (25) is possible. The power P _in (k) and the output power P _out (k) are calculated.

ただし、Ｖ_ＭＡＸおよびＶ_ＭＩＮは、それぞれ電池の最大許容端子電圧および最小許容端子電圧であり、電池に応じて設定された値である。
ｓｔｅｐ１７では、ｓｔｅｐ１３で算出した最終充電率推定値ＳＯＣ_ｅｓｔ(ｋ）と電流Ｉ(ｋ）から総容量推定値Ｑ_ＭＡＸ(ｋ）（満充電容量ともいう）を演算する。この値は、例えば（数２６）式に示すように、電流Ｉ(ｋ）を充電率推定値の微分値で除算して求める。

However, V _MAX and V _MIN are the maximum allowable terminal voltage and the minimum allowable terminal voltage of the battery, respectively, and are values set according to the battery.
In step 17, a total capacity estimated value Q _MAX (k) (also referred to as a full charge capacity) is calculated from the final charging rate estimated value SOC _est (k) calculated in step 13 and the current I (k). This value is obtained, for example, by dividing the current I (k) by the differential value of the estimated charge rate as shown in the equation (26).

ｓｔｅｐ１８ではｓｔｅｐ２で求めた選択フラグｆ_ＳＯＣＳＷに基づいて、次回の演算に必要な数値を保存して、今回の演算を終了する。

In step 18, based on the selection flag f_SOCSW obtained in step 2, a numerical value necessary for the next calculation is stored, and the current calculation is terminated.

Next, a method for calculating the second internal resistance estimated value _{K_ITG} (k) from the second charging rate estimated value (current integrated SOC estimated value) in step 8 will be described in detail.
There are several possible methods, but three examples will be described here. In any of the embodiments, the above formula (12) is rearranged with respect to the internal resistance value K to use the formula (27), and the open circuit voltage estimated value V ₀ (k) in the formula (27) is used. Is used as the final open circuit voltage estimated value V _{0 — est} (k) calculated in step 14. The output of the preprocessing filter calculated in step 3 is used for I ₁ (k), I ₂ (k), V ₁ (k), and V ₂ (k).

まず、電流積算ＳＯＣから第二内部抵抗値を推定する第一の方法としては、前記電流積算が選択される直前の適応デジタルフィルタＡＤＦによる電池パラメータ推定値の記憶値Ｔ_１__ＡＤＦ(ｋ）、Ｔ_２__ＡＤＦ(ｋ）を第二充電率（電流積算ＳＯＣ）演算中ホールドし、（数２７）式の右辺に代入することで得られる（数２８）式で第二内部抵抗値Ｋ__ＩＴＧ(ｋ）を演算する。

First, as a first method for estimating the second internal resistance value from the current integration SOC, a storage value T _{1 _ADF} (k) of a battery parameter estimation value by the adaptive digital filter ADF immediately before the current integration is selected, T _{2 _ADF} (k) is held during the calculation of the second charging rate (current integrated SOC), and is substituted into the right side of Equation (27), and the second internal resistance value _{K_ITG} ( k) is calculated.

電流積算ＳＯＣから第二内部抵抗値を推定する第二の方法としては、１回前の演算実行時に推定された第二内部抵抗推定値Ｋ__ＩＴＧ(ｋ−１）と適応デジタルフィルタＡＤＦで推定し記憶・ホールドした第一内部抵抗推定値Ｋ__ＡＤＦ(ｋ）との比に基づいて、電池パラメータ推定値の記憶値の時定数Ｔ_１__ＡＤＦ、Ｔ_２__ＡＤＦを（数２９）式と（数３０）式により補正する。

As a second method for estimating the second internal resistance value from the current integration SOC, the second internal resistance estimated value _{K_ITG} (k-1) estimated at the time of the previous calculation and the adaptive digital filter ADF is used. on the basis of the ratio of the storage hold the first internal resistance estimated value K_ _ADF (k), a constant _{T 1} _ _ADF when stored value of the battery parameter estimates, the _{T 2} _ _ADF (number 29) and (expression 30) Correct according to the equation.

そして、この補正された時定数Ｔ_１__ＡＤＦ’、Ｔ_２__ＡＤＦ’を前記（数２７）式の右辺に代入することで得られる（数３１）式を用いて第二内部抵抗値Ｋ__ＩＴＧ(ｋ）を演算する。

Then, the corrected constant _{T 1} _ _ADF when said a ', _{T 2} _ _ADF' (number 27) the second internal resistance value by using the equation obtained by substituting the right-hand side (number 31) below K_ _ITG (k) is calculated.

電流積算ＳＯＣから第二内部抵抗値を推定する第三の方法としては、電流と電圧の変化率に応じて前記第一の方法と第二の方法とを切り替えて用いる方法である。つまり、電流と電圧の変化率が共に所定値以下の場合には、前記第二の方法により第二内部抵抗値を演算するが、それ以外の場合には前記第一の方法により第二内部抵抗値を演算する。ここで、電流の変化率は、ｓｔｅｐ２において電流の変化率を求めたときと同じ方法により求める。電圧に関しても電流と同様に変化率を求めることができる。

A third method of estimating the second internal resistance value from the current integration SOC is a method of switching between the first method and the second method according to the rate of change of current and voltage. That is, when both the current and voltage change rates are less than or equal to a predetermined value, the second internal resistance value is calculated by the second method. In other cases, the second internal resistance value is calculated by the first method. Calculate the value. Here, the rate of change of current is obtained by the same method as that for obtaining the rate of change of current in step 2. Regarding the voltage, the rate of change can be obtained in the same manner as the current.

7 to 10 show the results of verifying the effect of the present invention by simulation using a battery model. In each figure, the input current I from the top, the voltage V, when the true value and the estimated value of the constant T _1, when the true value and the estimated value of the constant T _2, the true value and the estimated value of the internal resistance value K, the charging rate SOC , True value and estimated value of input power P _in , and true value and estimated value of output power P _out are shown. The true value is indicated by a broken line and the estimated value is indicated by a solid line, except for current and voltage waveforms. The parameters in the battery model were set so as to change suddenly with a delay of 5 seconds after the discharge current exceeded −100 [A] in accordance with the characteristics of the actual battery.

FIG. 7 is a simulation result in the prior art, and is a result when the internal resistance value is calculated as K in (V = K · I + V ₀ ) ignoring the transient characteristics of the battery. A large estimation error occurs in the internal resistance value at the timing when the current changes (near 100 seconds and 120 seconds), and this appears as an estimation error of input / output power.

  FIG. 8 shows a result when the internal resistance value is calculated by the first method of estimating the second internal resistance value from the current integration SOC. As compared with FIG. 7, it is clear that the estimation accuracy of the internal resistance value is improved at the timing when the current changes (around 100 seconds and 120 seconds).

  FIG. 9 shows the result of calculating the internal resistance value when the time constant also changes according to the internal resistance value in the first method of estimating the second internal resistance value from the current integration SOC. Actually, the time constant increases with the internal resistance value from around 105 seconds, but since the internal resistance value is estimated by continuously storing the value estimated by the adaptive digital filter ADF immediately before, it is large at around 125 seconds. An error is generated. This appears as an estimation error of power that can be input and output.

  FIG. 10 shows a result when the internal resistance value is calculated by the second method of estimating the second internal resistance value from the current integration SOC under the same conditions as in FIG. The estimation error of the large internal resistance value near 125 seconds seen in FIG. 9 can be suppressed, and the changing internal resistance value can be estimated with high accuracy. As a result, it can be seen that the input / output power can be estimated with high accuracy.

As described above, in the present invention, the first internal resistance value estimated by the adaptive digital filter and the second internal resistance value obtained from the second charging rate estimated value by current integration or the like are determined according to the current state. When the current change rate ΔI (k) is constant current discharge in which the current change rate ΔI (k) is continuously equal to or less than the predetermined value β for a predetermined time ts or when the current becomes a large current exceeding the predetermined value γ In other words, when the SOC estimation by the adaptive digital filter (ADF) is difficult, the SOC estimation value (second charge rate estimation value) by another method such as current integration is selected. Otherwise, by the adaptive digital filter Since the first charging rate estimated value (and the first internal resistance value) is selected, it is possible to always perform the SOC estimation and the internal resistance value estimation with little error. Then, when estimating the second internal resistance value from the second charging rate estimated value, the stored value T _{1 — ADF} (k) of the battery parameter estimated value by the adaptive digital filter ADF immediately before the second charging rate is selected, When T _{2 _ADF} (k) is held and the second internal resistance value _{K_ITG} (k) is calculated using the T _{2 _ADF} (k), time constants (T ₁ and T ₂₎ representing the dynamic characteristics of the battery are used. Therefore, the internal resistance value can be estimated with high accuracy even in a transient state where the current changes. Therefore, for example, the estimation accuracy of the input / output possible power estimation value calculated using the internal resistance estimation value is improved.

Further, when the second internal resistance value is estimated from the second charging rate estimated value (SOC estimated value by current integration), the second internal resistance estimated value _{K_ITG} (k−1) estimated at the previous execution. based on the ratio between the estimated and stored and hold the first internal resistance estimated value K_ _ADF (k) by the adaptive digital filter ADF and the constant _{T 1} _ _ADF when stored value of the battery parameter estimates, the _{T 2} _ _ADF When the correction is made by the equations (29) and (30), even if the time constant changes due to the change in the internal resistance value of the battery, the amount of change of the time constant is corrected. Therefore, the time constant estimation accuracy is improved. Therefore, since the internal resistance value is calculated using a time constant closer to the true value, the estimation accuracy of the internal resistance value is further improved. The relational expressions of the battery parameters are K = R ₁ + R ₂ , T ₁ = C ₁ · R ₁ , T ₂ = C ₁ · R ₁ · R ₂ / (R ₁ + R ₂ ). Therefore, assuming that the electric double layer capacitance C ₁ is constant and the ratio of the elements R ₁ and R ₂ of the internal resistance value K is constant regardless of the fluctuation of the overall internal resistance value K, the time constants T ₁ , T ₂ Changes in proportion to the change in the internal resistance value K.

Further, when the current and voltage change rates are equal to or lower than predetermined values, the time constant T _{1 _ADF} and T _{2 _ADF} are corrected, and the estimated value of the time constant is improved. There is. That is, as can be seen from the equation (28), when the rate of change of current and voltage is small (differential values: I ₂ ≈0, V ₂ ≈0), the internal resistance estimation value is the current (I ₁ ) Since it is determined only by the voltages (V ₁ , V _{0 _ITG} ), even if a slight error occurs in the time constants (T _{1 _ADF} , T _{2 _ADF} ) of the battery parameter stored values, the influence on the internal resistance estimation value The internal resistance value can be estimated with high accuracy. Therefore, the estimated value of the time constant is improved by correcting the time constant of the battery parameter stored value using the estimated internal resistance value estimated under such a condition that the rate of change of current and voltage is small. As a result, the estimation accuracy of the internal resistance value is improved even when the rate of change in current and voltage subsequently increases. Therefore, for example, the estimation accuracy of the input / output possible power estimation value calculated using the internal resistance estimation value is improved.

The functional block diagram which shows one Example of the internal resistance estimation calculation of this invention. The specific block diagram of one Example of this invention. The circuit diagram which shows the battery model used for the calculation of an adaptive digital filter. The figure which shows the characteristic of the charging rate and open circuit voltage in a lithium ion battery. The figure which shows the characteristic of the electric current in a lithium ion battery, and internal resistance. The flowchart which shows the arithmetic processing performed with an electronic control unit. The figure which shows the battery parameter in a prior art, a charging rate, and the input / output possible power estimation result by simulation. The figure which shows the result at the time of calculating an internal resistance value with the 1st method of estimating a 2nd internal resistance value from electric current integration SOC by simulation. The figure which shows the result at the time of calculating an internal resistance value in case the time constant also changes according to an internal resistance value in the 1st method of estimating a 2nd internal resistance value from electric current integration SOC. The figure which shows the result at the time of calculating an internal resistance value by the 2nd method of estimating 2nd internal resistance value from electric current integration SOC on the same conditions as FIG.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 ... Current detection means 101 ... Voltage detection means 102 ... Pre-processing filter calculation means 103 ... Adaptive digital filter calculation means 104 ... First open circuit voltage calculation means 105 ... First charge rate estimation means 106 ... Second charge rate estimation means 107 ... Second open circuit voltage calculation means 108 ... final charge rate estimated value selection means 109 ... battery parameter estimated value storage means 110 ... second internal resistance estimation means 111 ... final open circuit voltage estimated value selection means 112 ... final internal resistance estimated value selection means 113 ... Input / output possible power calculation means 10 ... secondary battery 20 ... load 30 ... electronic control unit 40 ... current sensor 50 ... voltage sensor

Claims (4)

Current detecting means for detecting current during charging and discharging of the secondary battery;
Voltage detection means for detecting the terminal voltage of the secondary battery;
Preprocessing filter means for calculating a low-pass filter value and a differential value of the current and voltage;
An adaptive digital filter that defines a battery model of a secondary battery, and uses the battery model to input a low-pass filter value and a differential value of the current and voltage, and collectively estimate and calculate parameters including the first internal resistance value of the battery model Computing means;
A first open-circuit voltage calculating means for estimating and calculating a first open-circuit voltage with the battery parameter and the pre-processed filter current and voltage as inputs;
First charge rate estimating means for estimating and calculating the first charge rate from the first open circuit voltage;
A second charging rate estimating means for estimating and calculating a second charging rate by a method capable of estimating the charging rate even in a current state where it is difficult to estimate the charging rate by the adaptive digital filter;
Second open circuit voltage calculation means for calculating a second open circuit voltage estimated value from the second charging rate estimated value;
A final charge rate estimated value selection means for selecting one of the first charge rate and the second charge rate as a final charge rate estimated value;
Battery parameter estimated value storage means for storing a battery parameter estimated value by the adaptive digital filter calculation means immediately before the second charging rate is selected;
Final open circuit voltage estimated value selection means for selecting either the first open circuit voltage or the second open circuit voltage based on the information of the selection flag of the final charge rate estimated value determined by the state of the current, and setting it as the final open circuit voltage;
Second internal resistance estimating means for estimating and calculating a second internal resistance value from the second charging rate estimated value;
Based on the information of the selection flag of the final charge rate estimated value determined by the current state, the first internal resistance value estimated by the adaptive digital filter calculation means and the second internal resistance value estimated by the second internal resistance estimation means A final internal resistance estimated value selection means that selects any one of them and sets it as a final internal resistance estimated value;
The second internal resistance estimation means includes
When the second charge rate estimated value is selected, the second internal resistance estimated value _{K_ITG} is used as the time constant (T _{1 _ADF} , T _{2 _ADF} ) of the battery parameter stored value in the battery parameter estimated value storage means. , The second open circuit voltage estimated value V _{0 _ITG} , and the low-pass filter values (I ₁ , V ₁ ) and the differential values (I ₂ , V ₂ ) of the current and voltage that are the outputs of the preprocessing filter An internal resistance estimation device for a secondary battery, wherein an estimation calculation is performed using the equation (1).

In the internal resistance estimation apparatus of the secondary battery according to claim 1,
The second charging rate estimation means estimates the second charging rate by integrating the current detected by the current detection means, and the internal resistance estimation device for a secondary battery. In the internal resistance estimation apparatus of the secondary battery according to claim 1,
A first internal resistance value K_ _ADF is a battery parameter stored value in said battery parameter estimation value storage means, the second internal resistance estimated value K_ _ITG in the previous operation (k-1), (Formula 2) and (number 3) An internal resistance estimation device for a secondary battery, wherein the time constants (T _{1 —ADF} , T _{2 —ADF} ) of the battery parameter stored value are corrected based on the equation.

_{However, T 1 _ ADF ', T} 2 _ ADF' when after the correction constant In the internal resistance estimation apparatus of the secondary battery according to claim 3,
An internal resistance estimation device for a secondary battery, wherein the time constants (T _{1 _ADF} , T _{2 _ADF} ) are corrected when the rate of change in current and voltage is equal to or less than a predetermined value.

Images