第二型模糊控制之全域搜索之最佳化鋰電-超級電容混合電能管理系統

Abstract

本論文基於鋰電-超級電容混合電力系統，發展出最佳化能量控制策略來分配混和電力系統之能量分配比，以達到最佳的能耗。底層的配置是透過三台直流-直流轉換器並聯而成的系統，以提高系統輸出效率。在實務上，因電路損耗、元件匹配及外部雜訊等因素，會導致附載分配不均，且降低系統轉換效率，甚至燒毀系統電路使其無法運作。因此良好的均流控制可以確保模組間輸出的電流相同，並提高系統可靠度。本篇論文提出一滑動模式控制均流架構，其控制方式簡單，對於非線性系統有良好的控制性能，且不易受到系統參數變異的影響，有著良好的強健性。上層的部分，本論文針對鋰電-超級電容混合電力系統提出了五種最佳化能量管理策略，因應不同的負載功率需求、超級電容殘電量與鋰電池電流大小計算出最佳的功率分配比，達到提升續航力以及減少能耗之目的。本論文發展基本規則庫控制、全域搜索演算法、模糊控制、第一型模糊全域搜索以及第二型模糊全域搜索。實驗結果表明，全域搜索演算法有著最佳的能耗表現，模糊控制對於抑制電流變化有著較佳的表現，而第一型模糊全域搜索與第二型模糊全域搜索則是在最佳能耗以及抑制電流變化中取得平衡。

Based on the lithium battery-supercapacitor hybrid power system, this thesis develops an optimal energy control strategy to allocate the energy distribution ratio of the hybrid power system to achieve the best energy consumption. The configuration of the bottom layer is a system formed by connecting three DC-DC converters in parallel to improve the system output efficiency. In practice, due to factors such as circuit loss, component matching, and external noise, the load distribution will be uneven, the system conversion efficiency will be reduced, and even the system circuit will be burned to make it inoperable. Therefore, good current sharing control can ensure the same output current between modules and improve system reliability. In this thesis, a sliding mode control current sharing architecture is proposed, which has a simple control method, has good control performance for nonlinear systems, is not easily affected by system parameter variation, and has good robustness.In the upper layer, this thesis proposes five optimal energy management strategies for the lithium battery-supercapacitor hybrid power system, and calculates the best power distribution ratio according to different load power requirements, supercapacitor residual capacity, and lithium battery current. To achieve the purpose of improving battery life and reducing energy consumption. This thesis develops a basic rule-based control(RBC), global search algorithm(GSA), fuzzy logic control(FC), global search algorithm with type-1 fuzzy control(GSAT1FC), and global search algorithm with type-2 fuzzy control (GSAT2FC). The experimental results show that the GSA has the best energy consumption performance, the FC has a better performance in suppressing the current change, and the GSAT1FC and GSAT2FC strike a balance between optimal energy consumption and suppression of current variations.