Please use this identifier to cite or link to this item:
Title: 電動車動力測試系統之建模與性能改善分析
Systems modeling and performance improvement of test platform for electric vehicles
Authors: 洪翊軒
Yi-Hsuan Hung
Yeou-Feng Lue
Keywords: 硬體嵌入式模擬
power systems
motor control
driving tests
electric Vehicles
Issue Date: 2014
Abstract: 本論文針對電動車動力測試系統之建模與性能改善分析進行研究。為提升高準確度與快速反應之虛擬電動車動力系統測試技術,本研究首先透過Matlab/Simulink與參數分析建立離線之軟體模擬系統(Model In The Loop,MIL),比對基礎PI值之離線模擬(off-line)與即時模擬(on-line)平台精準度,確認離線模型(off-line)代表實際動力計HIL(Hardware-In-the-Loop,HIL)平台,接著透過全域搜尋法則(Global Search Algorithm)進行一For迴圈最佳化參數搜尋,並獲得最佳化PI控制器參數值,之後便可將此組PI值注入on-line平台,提升實際車速跟隨精準度。通訊介面取樣時間分析部分,透過準確的離線模型建立,可針對動力計測試平台之提升效益進行敏感度分析評估,並藉由馬達控制參數調校與硬體的改善,進而達成提升虛擬電動車動力計測試平台之效益。 結果顯示本研究可成功透過離線模擬執行動力計測試平台之車速跟隨精準度與性能提升效益分析;調校控制參數部分:基礎off-line與on-line平台在相同條件測試下,實際車速模擬誤差平均0.1 km/h;透過off-line之全域搜尋法則讓PI控制器於最佳化參數下,預測可使on-line平台平均車速誤差跟隨改善33 %以上;將最佳化之PI值注入on-line平台驗證,平均車速誤差跟隨改善38 %,透過本研究方法與搜尋法則有助於未來研究人員先期評估平台設備的性能提升與改善。
This thesis aims at modeling and improving the performance of an electric vehicle power testing system. To reach the high accuracy and fast response of the system, this study firstly established a off-line software system (Model In The Loop,MIL) on the Matlab/Simulink platform with real-plant parameter input. By comparing the off-line/on-line performance under the baseline PI control case, we ensured that the off-line model can represent the HIL(Hardware-In-the-Loop,HIL) platform. Next, a global search algorithm derived the optimal PI control parameters by for-loop search. The set of PI values them was employed to the on-line platform to enhance the vehicle speed tracking accuracy. For the bandwidth and delay analysis of the data communication, through the precise offline model, the benefit can be analyzed. Results show that this research successfully evaluates the vehicle speed tracking accuracy and performance improvement by the off-line simulation of the dynamometer test platform. For the control parameter tuning, the average vehicle speed error between the off-line and on-line platform was 0.1 km/h. By using the global search algorithm on the off-line simulation, the vehicle speed tracking for the platform can be predicted to be improved by 33+%. The same PI can improve 38 % of the vehicle tracking error for the on-line platform. This research method and the parameter search algorithm truly help for the preliminary study of the performance enhancement and improvement of the testing platform.
Other Identifiers: GN060170026H
Appears in Collections:學位論文

Files in This Item:
File SizeFormat 
n060170026h01.pdf3.7 MBAdobe PDFView/Open

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.