本論文目的在於針對運用在電動車上的密閉式鉛酸電池，設計快速充電器，充電策略採用正負脈衝充電法（Reflex TM）充電，透過短時間大電流放電以提升充電速率，有鑒於傳統正負脈衝充電法使用固定電阻放電，本論文透過全橋式轉換器與反馳式轉換器結合，提出一新型具有高功率快速充電與能量回送之快速充電器，提高整體充電效率，系統能量轉換效率為85%，且其具有功率因數校正能力，使功率因數達到0.99，以落實環保、減少能源的消耗。此外，設計新型模糊電量估測器，以自行設計之標準化電量參數與電壓為輸入值，準確估測電池電量，使快速充電器能快速且有效率的進行充電，電量從0%充電至70%約10分鐘左右。在大電流充電下，使電池不至於充電不足或是過度充電，避免電池壽命減損。系統以模糊理論為基礎，透過數位訊號控制器（dsPIC30F4012）來實現充電系統的控制與電量估測。在快速充電的同時，完成電量估測，達成充電及時估測的目的，並保持估測誤差在可接受的範圍之內（±5%）。本系統亦利用等效內阻估測法來估測電池內阻，再根據內阻的大小自行調節負脈衝的時間，有效地降低充電時電池的溫升，使其控制在8°C以內，以避免損害電池壽命。最後本文利用實驗結果，以驗證系統的性能與估測器的可行性。 This thesis proposes a new rapid charger based on a fuzzy state-of-charge (SOC) estimation in Reflex TM charging method for the sealed lead-acid battery of the electric vehicle (EV). It includes an energy recovery stage which can save energy instead of conventional one which uses static resistance to discharge the batteries. The efficiency of conversion is up to 85% and power factor is up to 0.99. Charging time from SOC 0% to 70% is about 10 minutes. Moreover, this thesis proposes a novel fuzzy SOC estimator which uses normalized parameter and open circuit voltage to estimate SOC of batteries accurately. In the high current charging, the proposed SOC estimation can avoid a situation of undercharge or overcharge; note that the overcharge will shorten the battery lifetime. The charging control and the SOC estimation are based on fuzzy theorem, implemented by a digital signal controller (dsPIC30F6010). The SOC of batteries is estimated while the system is charging the battery and the estimated error is maintained within an acceptable range (±5%). The thermal rise can also limited under 8°C to avoid damaging the batteries. Moreover, this system uses an equivalent internal resistance to calculate the internal resistance of the batteries. During the Reflex TM charging, the controller can self-tune the discharging time according to the internal resistances, which reduces the thermal rises and the lifetimes for the batteries. Finally, experimental results are provided to verify the performance of the system and the feasibility of the proposed SOC estimation in this thesis.
