並聯式油電混合機車能量管理與動態規劃策略之探討

Abstract

在本論文中，試圖針對台灣的特殊交通環境，提出一個改善交通油耗與汙染的方案。 其中，台灣大量以機車為交通工具的比例，讓對於以機車做為研究系統有強烈的動機。 因此，本研究將現行極為熱門的並聯式混合動力做為動力驅動架構(Parallel Hybrid Structure)套用在機車上面，並採用固定減速比的鍊條取代原本為主流的無段變速系統(Continuous Variable Transmission, CVT)，以解決CVT傳動效率過低的問題。 而本論文的研究架構，會先建立系統的動態模型並搭配動態規劃演算法，找出此系統在行車型態ECE 40的最佳操作點與最佳省油性能。 藉由此結果，加以修正控制器的能量管理策略。 而此策略，會再搭配順向模擬分析，與其他非使用動態規劃所設計之策略進行比較。 以證明以動態規劃為基礎所設計之能量管理策略的效果是優於其他策略。 在此同時，開發了使用飛輪結構來等效車輛與駕駛者質量的實驗平台，並以數學與物理模型證明其有效性，進而提供更為方便的實驗平台且不失其擬真性，以供未來更深入的研究使用。

This research aims to study the power management via dynamic programming for a parallel hybrid scooter. The system is composed of a fixed gear ratio made with a steel chain in place of the continuous variable transmission that is commonly used in scooters. As for the research structure of this thesis, this research established the dynamic system models first. Following that, it used the dynamic programming (DP) method to find the system’s operation points for the best fuel consumption under driving cycle ECE 40. Then, the operation points were used to modify the strategy for power management set in the controller. Then, this strategy will be put into a forward simulation to make a comparison with other strategies that are not based on the result of DP. This is to make sure that the effectiveness of the one from DP is superior to others. This research simultaneously designed a flywheel structure to emulate the inertia of the mass of drivers and scooter to provide an experimental platform for practice, without losing the effectiveness of emulation. Moreover, an authentication by mathematical and physical methods was also done. These can provide a use for the more in-depth research in future.