使用單電池等校電路模型和模糊回歸模型於鋰電池組的電量估測

THE STATE OF CHARGE ESTIMATION FOR LI-ION BATTERY PACK BY THE SINGLE BATTERY EQUIVALENT CIRCUIT-BASED MODEL AND THE FUZZY C-REGRESSION MODEL

廖緯祥(Wei-Siang Liao)

指導教授 ： 龔宗鈞

大同大學/電機資訊學院/電機工程學系所/碩士(2012年)

摘要

在本論文中，我們使用模糊 C-回歸模型、單一鋰電池等效電路模型和串聯鋰電池組等效電路模型去估測電池的殘電量。單一鋰電池等效電路模型由各種電子元件組成，並可用數學分析來描寫其電路的行為，其中也考慮了一階RC電路來模擬電池內部的化學變化，也就是極化效應。
但等效電路模型中的各項參數都是未知的，而我們為了要得知各項參數的特性，我們量測電池的額定電量，以便選用特性相近的電池來做開路電壓和直流內電阻的測試，並藉由量測結果和電池端電壓去計算這些參數。根樹實驗結果，我們可以知道等效電路模型中的參數都是跟殘電量有關，而因為量測到的參數其資料曲線是非線性的，所以我們採用模糊 C-回歸模型去制訂這些參數和殘電量的關係。
其中所有的電池資料都是由電池測試系統Gwinstek GBT-2211和Maccor 4300在室溫下所量測，我們選用4顆特性相近的鋰電池Molicel ICR 18650H來做單顆鋰電池的實驗和4顆串聯鋰電池組的實驗，藉由前面方法得到的參數可以代入等效電路模型中去估測單顆鋰電池的殘電量和4顆串聯鋰電池組的殘電量以來證明其等效電路模型和殘電量的估測是有效的。

關鍵字

模糊回歸模型 ； 殘電量估測

並列摘要

In this thesis, the fuzzy c-regression model of the equivalent circuit-based model (ECM) shown in Fig 2.2 and series battery pack model shown in Fig. 2.3 are adopted to estimate the state of charge (SOC) of the battery. The ECM has been built by the electric circuit parameters, but it can be possible to express as mathematical formulas. Then, the one set of RC circuit can model the polarization effect that will provide us to simulate about the internal chemical variation of the battery to estimate the SOC in detail.
In order to obtaining the parameters such as V_OC, R_t, R_p, and C_p are not known in the ECM, we set the rated capacity, the open circuit voltage test, and the direct current internal resistance test to calculate the parameters with terminal voltage. Fallowing that, the parameters of ECM are known as with the function of SOC. Since the data line of those parameters varying with the SOC is nonlinear and piecewise continue, the fuzzy c-regression model (FCRM) will be used to formulate the test data with the function of SOC.
The all test data is measured by the battery testing system, Gwinstek GBT-2211 and Maccor 4300, and test in room temperature with 4 batteries, Molicel ICR 18650H, with similar characteristics. We will do the test with each 4 batteries and also with 4 series battery pack. Then, the all parameters will be obtained by mention method, and we can use it to estimate the SOC of the battery. Finally, the proposed method will be proved it is work by simulation result to estimate the SOC of the ECM of single battery and battery pack.

並列關鍵字

fuzzy c-regression model ； state of charge

參考文獻

pack simulation for Li-ion batteries,” J. Power Sources, vol.186,

open-circuit voltage of lithium-ion batteries in electric vehicles,”

J. Power Sources, vol. 196, pp. 3921-3932, 2011.

pack based on a screening process for improved voltage/SOC balancing,”

J. Power Sources, vol. 27, pp. 411-424, 2012