VRLA Battery SOC And SOH On-line Estimation In Backup Power System

ABSTRACT:VRLA batteries play a very important role in back power supply system as the last line of defense against AC power outage for crucial equipment. Due to long time float charge and lack of maintenance, the performance of some batteries would gradually deteriorate. Being used in series in backup power systems, the defective batteries would lead an unbalanced cell voltage distribution in battery strings, which makes charge and discharge processes nonstandard for individuals, accelerates cells aging and leads the whole battery group unreliable. So understanding the State-of-Health (SOH) of a battery and the State-of-Charge (SOC) in discharge processes is important. However, on-line SOC and SOH monitoring with high accuracy has long been a hard task because of the highly nonlinear electrochemical and dynamic characteristic of VRLA batteries.Currently the only reliable SOH examine method is complete load test, which is off-line, time-consuming and laborious. Among the existing on-line SOC and SOH examine methods, the accuracy of voltage curve fit method is low; Quantitative indicators cannot be given by AC conductance (resistance) method; Multi-frequency electrochemical impedance spectroscopy (EIS) method, which only exists in portable devices and is too costly, is not suitable for backup power system application.On the basis of two existing discharging voltage curve models which were insensitive to current rate, ambient temperature and cell aging etc. condition variation, this thesis presented a new mode that describes the discharge voltage as a function of discharge current, depth-of-discharge (DOD) and resistance. The parameters in the model can be estimated by recursive least squares (RLS) method in the real applications, which gave the model the ability of adaptive adjustment to condition variation. Combining the characteristics of the two existing models and the new model proposed, this thesis presented a new on-line SOC and SOH estimation method which was suitable for backup power system application. Experiment results proved that it has high estimation accuracy.