Search

US-12618906-B2 - Method for estimating state of charge of battery

US12618906B2US 12618906 B2US12618906 B2US 12618906B2US-12618906-B2

Abstract

A method of estimating a state-of-charge (SOC) of a battery includes: setting an initial SOC value and an initial Kalman error covariance value; receiving an estimated G parameter value, a current current value, and a current voltage value of the battery; updating a current SOC value and a current Kalman error covariance value of the battery by inputting the estimated G parameter value, the current current value, and the current voltage value to an extended Kalman filter; and outputting the current SOC value.

Inventors

  • Christober RAYAPPAN
  • Sungwook PAEK
  • Byeonghui LIM
  • Yongjun HWANG
  • Jake Kim
  • Giheon KIM

Assignees

  • SAMSUNG SDI CO., LTD.

Dates

Publication Date
20260505
Application Date
20210614
Priority Date
20200911

Claims (20)

  1. 1 . A method of estimating a state-of-charge (SOC) of a battery, the method comprising: setting an initial SOC value and an initial Kalman error covariance value; generating, by a first filter, an estimated G parameter value based on a first current value and a first voltage value, wherein the estimated G parameter value is indicative of an estimated internal state of the battery; receiving, by an extended Kalman filter, the estimated G parameter value, the first current value, and the first voltage value of the battery; generating, by the extended Kalman filter, a current SOC value and a current Kalman error covariance value of the battery based on the estimated G parameter value, the first current value, and the first voltage value, wherein the generating includes: calculating a first-order estimated SOC value based on the first current value and the initial SOC value; calculating a first-order estimated Kalman error covariance value based on the first current value and the initial Kalman error covariance value; identifying coefficient data associated with an open circuit voltage (OCV)-SOC relationship for the battery; calculating an estimated voltage value based on the coefficient data, the first current value, and the estimated G parameter value; calculating a current Kalman gain value based on the coefficient data, the estimated G parameter value, and the first-order estimated Kalman error covariance value; computing the current SOC value based on the first-order estimated SOC value, the current Kalman gain value, the first voltage value, and the estimated voltage value; and computing the current Kalman error covariance value based on the first-order estimated Kalman error covariance value, the current Kalman gain value, the coefficient data, and the estimated G parameter value; outputting the current SOC value; and managing safety of an electrical device driven by the battery based on the current SOC value.
  2. 2 . The method of estimating a state-of-charge (SOC) of a battery of claim 1 , wherein the first-order estimated SOC value SOC est − (t) is calculated based on SOC est − (t)=SOC est (t−1)+I(t)×Ts/Q max , by using a direct previous SOC value SOC est (t−1), the first current value I (t), a sampling period Ts, and a maximum capacity Q max of the battery.
  3. 3 . The method of estimating a state-of-charge (SOC) of a battery of claim 1 , wherein the first-order estimated Kalman error covariance value P k − (t) is calculated based on P k − (t)=P k (t−1)+(Ts/Q max ) 2 ×σ w , by using a direct previous Kalman error covariance value P k (t−1), a sampling period Ts, a maximum capacity Q max of the battery, and processor noise σ w .
  4. 4 . The method of estimating a state-of-charge (SOC) of a battery of claim 1 , wherein the calculating of the estimated voltage value comprises: extracting, from the coefficient data, a first coefficient value C (t) corresponding to the estimated SOC value SOC est − (t); and calculating the estimated voltage value V est (t) based on V est (t)=C(t)×SOC est −(t)+G est (t)×I(t), by using the first coefficient value C (t), the first-order estimated SOC value SOC est − (t), the estimated G parameter value G est (t), and the first current value I (t).
  5. 5 . The method of estimating a state-of-charge (SOC) of a battery of claim 4 , wherein the first coefficient value C (t) is determined based on C(t)=OCV(N ε [SOC est − (t)])/N ε [SOC est − (t)], by using an SOC data value N ε [SOC est 31 (t)] proximate to the first-order estimated SOC value SOC est − (t) and an OCV data value OCV (N ε [SOC est − (t)]) corresponding to the SOC data value N ε [SOC est − (t)].
  6. 6 . The method of estimating a state-of-charge (SOC) of a battery of claim 1 , wherein the calculating of the estimated voltage value comprises: extracting, from the coefficient data, a second coefficient value C 1 (t) and a third coefficient value E (t) corresponding to the first-order estimated SOC value SOC est − (t); and calculating the estimated voltage value V est (t) based on V est (t)=C 1 (t)×SOC est −(t)+G est (t)×I(t)+E(t), by using the second coefficient value C 1 (t), the first-order estimated SOC value SOC est − (t), the estimated G parameter value G est (t), the first current value I (t), and the third coefficient value E (t).
  7. 7 . The method of estimating a state-of-charge (SOC) of a battery of claim 6 , wherein the second coefficient value C 1 (t) and the third coefficient value E (t) are respectively determined to be a gradient and an OCV intercept of a linear function contacting a point in a curve corresponding to an OCV-SOC relationship, the point corresponding to an SOC data value N ε [SOC est − (t)] proximate to the first-order estimated SOC value SOC est − (t).
  8. 8 . The method of estimating a state-of-charge (SOC) of a battery of claim 1 , wherein the calculating of the current Kalman gain value comprises: extracting, from the coefficient data, a first coefficient value C (t) corresponding to the first-order estimated SOC value SOC est − (t); and calculating the current Kalman gain value L k (t) based on L k (t)=C(t)×P k − (t)/[C(t) 2 ×P k − (t)+G est (t) 2 ×σ v ], by using the first coefficient value C (t), the first-order estimated Kalman error covariance value P k − (t), the estimated G parameter value G est (t), and measured noise σ v .
  9. 9 . The method of estimating a state-of-charge (SOC) of a battery of claim 1 , wherein the current SOC value SOC est (t) is calculated based on SOC est (t)=SOC est − (t)+L k (t)×(V(t)−V est (t)), by using the first-order estimated SOC value SOC est − (t), the current Kalman gain value L k (t), the first voltage value V (t), and the estimated voltage value V est (t).
  10. 10 . The method of estimating a state-of-charge (SOC) of a battery of claim 1 , wherein the calculating of the current Kalman error covariance value comprises: extracting, from the coefficient data, a first coefficient value C (t) corresponding to the first-order estimated SOC value SOC est − (t); and calculating the current Kalman error covariance value P k (t) based on P k (t)=P k − (t)−L k (t) 2 ×[C(t) 2 ×P k − (t)+G est (t) 2 ×σ v ], by using the first-order estimated Kalman error covariance value P k − (t), the current Kalman gain value L k (t), the first coefficient value C (t), the estimated G parameter value G est (t), and measured noise σ v .
  11. 11 . The method of estimating a state-of-charge (SOC) of a battery of claim 1 , further comprising: sensing a voltage and a current of the battery for each predetermined sampling period Ts and periodically generating a voltage value and a current value of the battery; and generating the estimated G parameter value from the voltage value and the current value by using an adaptive filter, wherein the estimated G parameter value is a numerical value of a G parameter indicating a degree of sensitivity of voltage with respect to a change of current of the battery.
  12. 12 . The method of estimating a state-of-charge (SOC) of a battery of claim 11 , wherein the adaptive filter is a filter using a recursive least square (RLS) method.
  13. 13 . The method of estimating a state-of-charge (SOC) of a battery of claim 11 , further comprising generating an estimated H parameter value from the first voltage value and the first current value by using the adaptive filter, wherein the estimated H parameter value is a numerical value of an H parameter indicating a valid potential determined by local equilibrium potential distribution and resistance distribution in the battery.
  14. 14 . The method of estimating a state-of-charge (SOC) of a battery of claim 13 , further comprising setting an initial state vector value and an initial covariance matrix value of the battery, wherein the periodical generating of the voltage value and the current value of the battery comprises: generating a direct previous voltage value and a direct previous current value of the battery; and generating the first voltage value and the first current value of the battery after the sampling period Ts.
  15. 15 . The method of estimating a state-of-charge (SOC) of a battery of claim 14 , wherein the generating of the estimated G parameter value and the estimated H parameter value comprises: calculating an estimated voltage value of the battery based on the first current value and a direct previous state vector value; calculating a current gain matrix value and a current covariance matrix value based on the first current value and a direct previous covariance matrix value; calculating a voltage error based on the first voltage value and the estimated voltage value; and generating the estimated G parameter value and the estimated H parameter value by calculating a current state vector value based on the direct previous state vector value, the current gain matrix value, and the voltage error.
  16. 16 . The method of estimating a state-of-charge (SOC) of a battery of claim 15 , wherein the estimated voltage value V est (t) is calculated based on V est (t)=G est (t−1)×I(t)+H est (t−1), by using the first current value I (t), a direct previous G parameter value G est (t−1), and a direct previous H parameter value H est (t−1).
  17. 17 . The method of estimating a state-of-charge (SOC) of a battery of claim 15 , wherein the current state vector value Θ est (t) is calculated based on Θ est (t)=Θ est (t−1)+L(t)×e(t), by using the direct previous state vector value Θ est (t−1), the current gain matrix value L (t), and the voltage error e (t).
  18. 18 . The method of estimating a state-of-charge (SOC) of a battery of claim 15 , wherein the generating of the estimated G parameter value and the estimated H parameter value further comprises receiving a first forgetting factor λ 1 related to the G parameter and a second forgetting factor λ 2 related to the H parameter.
  19. 19 . The method of estimating a state-of-charge (SOC) of a battery of claim 18 , wherein the current gain matrix value is calculated by the following equation: L ⁡ ( t ) = [ L 1 ( t ) L 2 ( t ) ] = 1 1 + P 1 ( t - 1 ) ⁢ I ⁡ ( t ) 2 / λ 1 + P 2 ( t - 1 ) / λ 2 [ P 1 ( t - 1 ) ⁢ I ⁡ ( t ) / λ 1 P 2 ( t - 1 ) / λ 2 ] , and the current covariance matrix value is calculated by the following equation: P ⁡ ( t ) = [ P 1 ( t ) P 2 ( t ) ] = [ { 1 - L 1 ( t ) ⁢ I ⁡ ( t ) } ⁢ P 1 ( t - 1 ) / λ 1 { 1 - L 2 ( t ) } ⁢ P 2 ( t - 1 ) / λ 2 ] where L (t) is the current gain matrix value, P (t) is the current covariance matrix value, P (t−1) is the direct previous covariance matrix value, I (t) is the first current value, λ 1 is the first forgetting factor, and λ 2 is the second forgetting factor.
  20. 20 . The method of estimating a state-of-charge (SOC) of a battery of claim 1 , further comprising: sensing a voltage and a current of the battery for each predetermined sampling period Ts and generating a sensed voltage value and a sensed current value of the battery; periodically generating a voltage value and a current value of the battery by inputting each of the sensed voltage value and the sensed current value to a noise filter; and generating the estimated G parameter value and an estimated H parameter value from the voltage value and the current value by using an adaptive filter, wherein the estimated G parameter value is a numerical value of a G parameter indicating a degree of sensitivity of voltage with respect to a change of current of the battery, and the estimated H parameter value is a numerical value of an H parameter indicating a valid potential determined by local equilibrium potential distribution and resistance distribution in the battery.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS This application is a U.S. National Phase Patent Application of International Application Number PCT/KR2021/007413, filed on Jun. 14, 2021, which claims priority of Korean Patent Application Number 10-2020-0117047, filed on Sep. 11, 2020, the entire content of each of which is incorporated herein by reference. TECHNICAL FIELD The disclosure relates to a method of estimating a state-of-charge (SOC) of a battery in use. BACKGROUND ART Batteries are easily applied to electrical devices and have relatively higher energy, power density, etc., compared with other energy storage devices, and thus, have been widely applied not only to portable electronic devices, but also to electric vehicles (EVs), hybrid electric vehicles (HEVs), or the like, driven by an electrical driving source. Particularly, when an intense output is required, a battery pack in which a plurality of battery cells are serially or parallelly connected with each other may be used. In order to energy-efficiently and safely use an electrical device driven by a battery or a battery pack, battery management is important, and to this end, accurate measurement and diagnosis of a state of the battery are essential. Currently widely used estimation values include a state-of-charge (SOC) value, a state-of-health (SOH) value, a power limit estimation (PLE) value, etc. An SOC value according to the related art is represented by the amount of charge of a battery, a remaining capacity of a battery, etc., and is defined by a percentage of a current capacity of a battery to a fully charged capacity of the battery. Methods of estimating an SOC include a method of measuring the amount of charge which is discharged or flows into by using a current sensor and then integrating the amount of charge, a method of using a relationship of an open-circuit voltage (OCV) and an SOC, a method of estimating an SOC by using a battery model, etc. According to the method of estimating an SOC by measuring the amount of charge which is discharged or flows into by using a current sensor, a current SOC is estimated by adding, to an initial SOC, a value obtained by dividing an integrated current value by a fully charged capacity. This method is also referred to as an Ah counting method or a Coulomb counting method and is widely used because of its simplicity, but this method is affected by the accuracy of the current sensor. The method of estimating an SOC by measuring an OCV uses an OCV-SOC relationship, which is unique for each battery. The OCV-SOC relationship is known to seldom change even if a battery deteriorates, and thus, the method using this is highly reliable. However, to measure the OCV, a battery has to be left in a zero current state for long, and thus, while the battery is used, the OCV is not able to be measured, and it is difficult to accurately predict the SOC. When a battery model is used, the effects due to an error (noise) of a current sensor may be minimized, and without a long idle state of a battery, the SOC may be estimated in real time. The battery model includes an equivalent circuit model (ECM), a physics-based model, etc. The ECM is not able to provide an insight into what happens in a battery cell, and a parameter used in the ECM does not actually have a physical indication. The physics-based model is more accurate than the ECM, but has a problem of complexity and convergence. As described above, previous methods of estimating an SOC have limitations, such as the complexity of a battery model or the suspension of the use of a battery. For preventing over charge and over discharge of a battery and performing cell balancing, it is essential to accurately estimate an SOC. However, the Ah counting method may not guarantee accuracy due to a current measurement error. For example, just based on the assumption that a current sensor has merely an error of 0.1 A, when an electric vehicle is used for 8 hours, an SOC estimation error of 0.8 Ah, that is, the SOC estimation error per week that is equal to or greater than 5 Ah, occurs. When a battery capacity is 100 Ah, the error reaches 5%. With respect to the SOC estimation, not only accuracy, but also a low calculation load and a high calculation speed are important. When it is possible to accurately estimate and control an internal state of a battery, safety and performance of a battery pack based on price and weight may be improved, so that the battery may be applied not only to vehicles, but also to means of transportation, such as aviation, and other various fields. DESCRIPTION OF EMBODIMENTS Technical Problem The disclosure provides a method of accurately estimating a state-of-charge (SOC) of a battery in real time by using a voltage value and a current value of the battery. The disclosure provides a method of accurately estimating an SOC of a battery in real time by using a G parameter indicating an internal state of the battery. The method of estimating an SOC, accordi