State-of-Charge Estimation Based on Kalman Filter for Lithium Ion Batteries

• Main Authors: Ruei-Ji Chen, 陳瑞基
• Other Authors: Yao-Ching Hsieh
• Format: Others
• Language: zh-TW
• Published: 2011
• Online Access: http://ndltd.ncl.edu.tw/handle/93493958931307903146

碩士 === 國立東華大學 === 電機工程學系 === 99 === Battery researchers often spend a lot of time in the processes of charging and discharging cycles. Evidently, the reason is that to study upon the batteries, charging and discharging on the batteries for several times are inevitable. However, if there is an appropriate and accurate battery model, simulations are possible and so that battery experiments are no longer so indispensable. On the other hand, the induced over-voltage while the battery is charged or discharged renders the directly measured loaded battery voltage inadequate to be an accurate index for the battery’s state-of-charge. Suppose that there is an appropriate battery model, there will be adequate estimations on the over-voltages; and therefore, better real-time estimates upon state-of-charge directly from the loaded voltages can be derived. In this thesis, the electrochemical properties of lithium-ion batteries are studied. Primary factors, such as state-of-charge, battery current, aging effects, are separately experimented. Kalman filter method is applied to analyze the long-time recorded data, so as to identify the over-potential features during each stage. According to the features, lumped circuit elements can be adopted to mimic the battery’s dynamic behavior. The battery model will then be used to form a real-time state-of-charge estimation algorithm based on Kalman filter. The experimental tests show that satisfactory accuracy of the proposed method can attained. Without considering the aging of the battery, the state-of-charge estimation error by this method can be restrained within 5%.