以擴散模型實現之獨立電池仿真儀

Abstract

本論文提出以擴散模型實現之獨立電池仿真儀，能有效模仿鋰電池在動態負載下的放電行為，提供電路設計者在設計電路時，重複試探不同的放電演算法，以達最大放電電量。仿真技術的建構，係利用微控器計算輸出電壓，透過12Bits DAC調整穩壓IC PQ30RV21產生仿真的電池電壓，微控器中以鋰電池等效電路作為電池模型，透過定電流放電實驗量測電池在不同的電量狀態(SoC)下模型內部等效參數：電池電動勢(EMF)、過電壓(Vop)及過電壓內電阻(Rop)，並且將取得的參數關係建立查找表。放電過程中，對設定之負載電流積分，引入擴散法分析模型估測電池的初始電容量與運行時間，取得電池在此負載電流下的電量狀態(SoC)，藉由找查表轉換出與之相對應的EMF及Vop，並且計算出鋰電池在特定電量狀態下的電池端電壓。根據多組多組動態負載進行驗證，本文所提出之電池仿真儀(額定輸出電壓4V，最高輸出電流2A，電流取樣頻率1kHz)與實際電池相比，輸出電壓Normalized root mean square deviation均低於3%，放電時間的誤差皆低於1%，混合切換式模型可減少11%的程式記憶體。

A new battery emulator based on electrochemical diffusion model is proposed in this thesis. Researchers are able to vary the discharge patterns to maximize effective lifetime with reproducible result. The emulator is an intelligent power supply that mimics a battery’s behavior by running a battery simulation program in Arduino. It senses the current load and responds by controlling the output voltage as an actual battery would. The battery simulation is based on electrochemical diffusion model and switching overpotential hybrid model. The propose method emulates battery’s behavior with less computational efforts. For model validation, experiments are conducted with six different profiles, including constant loads, an interrupted load, an increasing load, a decreasing load and a varying load. The experiment results show the usefulness and accuracy of the proposed emulator, with 4 Volts of the rated output voltage, 2 Amps of the maximum output current, and 1 kHz of current sampling rate. The normalized root mean square deviation of output voltage are all below 3%. The runtime errors are below 1%. The switching overpotential hybrid model saves 11% of program memory compares to convention model.