| US 7,554,296 B2 | ||
| Method and apparatus for detecting charged state of secondary battery based on neural network calculation | ||
| Satoru Mizuno, Nishio (Japan); Atsushi Hashikawa, Okazaki (Japan); Shoji Sakai, Toyota (Japan); Atsushi Ichikawa, Toyoake (Japan); Takaharu Kozawa, Kounan (Japan); Naoki Mizuno, Nagoya (Japan); and Yoshifumi Morita, Gifu (Japan) | ||
| Assigned to Denso Corporation, Kariya (Japan); Nippon Soken, Inc., Nishio (Japan); and National University Corporation Nagoya Institute of Technology, Nagoya (Japan) | ||
| Filed on Feb. 14, 2006, as Appl. No. 11/353,220. | ||
| Claims priority of application No. 2005-036437 (JP), filed on Feb. 14, 2005; application No. 2005-036442 (JP), filed on Feb. 14, 2005; application No. 2005-039614 (JP), filed on Feb. 16, 2005; application No. 2005-122004 (JP), filed on Apr. 20, 2005; application No. 2005-122011 (JP), filed on Apr. 20, 2005; and application No. 2005-151059 (JP), filed on May 24, 2005. | ||
| Prior Publication US 2006/0181245 A1, Aug. 17, 2006 | ||
| Int. Cl. H01M 10/44 (2006.01); H02J 7/00 (2006.01); G01N 27/416 (2006.01) | ||
| U.S. Cl. 320—132 [320/106; 320/128; 320/136; 702/63; 324/426] | 9 Claims |
| 1. A neural network type of apparatus for detecting an internal state of a secondary battery implemented in a battery system,
the apparatus comprising:
detecting means for detecting, in real time, electric signals indicating an operating state of the battery during a predetermined
period of time, the electric signals being voltage and current of the battery;
producing means for producing, using the electric signals, an input parameter required for estimating the internal state of
the battery, the input parameter reflecting calibration of a present charged state of the battery and consisting of a first
input parameter indicating the operating state of the battery and a second input parameter indicating a degraded state of
the battery,
wherein the producing means includes:
means for calculating the first input parameter on the basis of data of the voltage and current of the battery, and
means for calculating the second input parameter in response to a state of predetermined charge of the battery; and
calculating means for calculating an output parameter indicating the charged state of the battery by processing both the first
and second input parameters based on a neural network calculation technique,
wherein
the second input parameter calculating means includes
means for calculating data of history of both the voltage and the current into an approximate expression on a least-squares
methods,
means for calculating a present value of an open-circuit voltage of the battery on the approximate expression, the present
value being included in the first input parameters, and
the output parameter calculating means is configured to calculate the output parameter by using both the first parameter consisting
of the data of the voltage history, the data of the current history, and the present value of the open-circuit voltage and
the second input parameter;
the second input parameter indicating the degraded state of the battery is both an open-circuit voltage and an internal resistance
of the battery detected in response to a state of predetermined charge of the battery,
the output parameter indicating the present charged state is one of and SOC (state of charge) of the battery, an SOH (state
of health) of the battery, and a function whose variables including information indicative of the SOC and SOH, and
the function is a degree of degradation of the battery which is defined by an expression of:
the degree of desegregation=SOH/(initial full charge capacity×SOC).
|