Apparatus and method for estimating state of charge of battery using neural network

• US 7,583,059 B2
• Filed: 12/17/2004
• Issued: 09/01/2009
• Est. Priority Date: 12/18/2003
• Status: Active Grant

• Current Assignee: LG Energy Solution Limited (LG Corporation)
• Original Assignee: LG Chem Limited (LG Corporation)
• Inventors: Cho, Il
• Primary Examiner(s): Ullah; Akm E
• Assistant Examiner(s): Zhang; Jue
• Application Number:
  US11/016,028
• Publication Number:
  US 20050194936A1
• Time in Patent Office: 1,719 Days
• Field of Search: 320/132, 324426-450, 702/60, 702/181, 702/182, 702/193
• US Class Current: 320/132
• CPC Class Codes: B60L 2260/44 by parameter estimation
• CPC Class Codes: B60L 2260/46 by self learning
• CPC Class Codes: B60L 2260/48 by fuzzy logic
• CPC Class Codes: B60L 58/12 responding to state of charge [SoC]
• CPC Class Codes: G01R 31/367 Software therefor, e.g. for battery testing using modelling or look-up tables
• CPC Class Codes: H01M 10/48 Accumulators combined with arrangements for measuring, testing or indicating the condition of cells, e.g. the level or density of the electrolyte constructional details of current conducting connections for detecting conditions inside cells or batteries, e.g. details of voltage sensing terminals, H01M50/569
• CPC Class Codes: H01M 2200/10 Temperature sensitive devices
• CPC Class Codes: Y02E 60/10 Energy storage using batteries
• CPC Class Codes: Y02T 10/70 Energy storage systems for electromobility, e.g. batteries

Abstract

Disclosed are an apparatus and a method for estimating a state of charge of a battery representing a non-linear characteristic by using a neural network. The apparatus includes a sensing section for detecting current, voltage and a temperature from a battery cell, a neural network performing a neural network algorithm and a learning algorithm based on data of the current, voltage and temperature transmitted thereto from the sensing section and present time data, thereby outputting the SOC of the battery estimated through a final learning algorithm, and a comparator for comparing an output value of the neural network with a predetermined target value and making the neural network iteratively perform the learning algorithm if a difference between the output value of the neural network and the predetermined target value is out of a predetermined limit, thereby update the learning algorithm to generate the final learning algorithm. The state of charge of the battery is precisely estimated through the neural network algorithm.

Claims

1. An apparatus for estimating a state of charge (SOC) of a battery, the apparatus comprising:

a sensing section for detecting current, voltage and a temperature from a battery cell;

a neural network performing a neural network algorithm for calculating SOC of the battery by processing data of the current, voltage and temperature transmitted thereto from the sensing section and time data at which the current, voltage and temperature are measured, according to a function having a predetermined weight, the predetermined weight being updated by a learning algorithm that directs the calculated SOC of the battery to follow a predetermined target value; and

a comparator in communication with the neural network to monitor the output of the neural network during calculation of the SOC of the battery, which compares the SOC of the battery calculated through the neural network with the predetermined target value varying with charging or discharging of the battery and provides a control signal to the neural network to perform the learning algorithm if a difference between the calculated SOC of the battery and the predetermined target value is out of a predetermined limit.

2. The apparatus as claimed in claim 1, wherein the predetermined target value is a reference value obtained through performing an experiment under specific conditions.

3. The apparatus as claimed in claim 2, wherein the reference value is obtained based on a value of subtracting ampere-hour data of a charger/discharger from rated capacity of the battery, or an open circuit voltage value of the battery, respectively or mathematically compensating each other.

4. The apparatus as claimed in claim 1, wherein the neural network includes a dynamic wavelet neural network algorithm or a static wavelet neural network algorithm.

5. The apparatus as claimed in claim 4, wherein the neural network includes a dynamic multi-dimensional wavelet neural network algorithm.

6. The apparatus as claimed in claim 1, wherein the learning algorithm includes a back-propagation learning algorithm, a Kalman filter scheme, GA, or a fuzzy learning algorithm.

7. A method for estimating a state of charge (SOC) of a battery by using a neural network in communication with a comparator, the method comprising the steps of:

(a) sensing current, voltage and a temperature from the battery;

(b) calculating a SOC of the battery according to a neural network algorithm, defined by a function with a predetermined weight, by processing data of the sensed current, voltage and temperature and time data at which the current, voltage and temperature are measured, and outputting the SOC of the battery to a battery management means;

(c) calculating the difference value between the calculated SOC of the battery and a predetermined target value varying with charging or discharging of the battery and (d) monitoring the difference value during calculation of the SOC of the battery; and

(e) performing a learning algorithm so as to update the weight of the neural network such that the calculated SOC of the battery can follow the predetermined target value if the difference value is out of a predetermined limit.

8. The method as claimed in claim 7, wherein the predetermined target value is a reference value obtained through performing an experiment under specific conditions.

9. The method as claimed in claim 8, wherein the reference value is obtained based on a value of subtracting ampere-hour data of a charger/discharger from rated capacity of the battery, or an open circuit voltage value of the battery, respectively or mathematically compensating each other.

10. The method as claimed in claim 7, wherein the neural network includes a dynamic wavelet neural network algorithm or a static wavelet neural network algorithm.

11. The method as claimed in claim 10, wherein the wavelet neural network includes a dynamic multi-dimensional wavelet neural network algorithm.

12. The method as claimed in claim 7, wherein the learning algorithm includes a back-propagation learning algorithm, a Kalman filter scheme, GA, or a fruzzy learning algorithm.