應用模糊邏輯與基因演算法在混成動力機車控制策略的最佳化

Optimization of Control Strategy for a Hybrid Electric Motorcycle Using Fuzzy Logic and Algorithm Genetic

趙嘉斌(Chia-Bin Chau)

國立虎尾科技大學/工程學院/動力機械工程研究所/碩士(2006年)

摘要

摘 要
本研究主要是構想出一套模糊控制的混成動力機車系統，並根據不同行車型態來做性能模擬分析，進而利用基因演算法來尋求最佳的模糊控制策略。此動力系統屬於一種並聯式雙軸配置，也可另外歸類為高量級混成動力系統。基本上低速純由電動馬達帶動，高速時以引擎動力為主，馬達動力為輔，必要時亦可發電回充電瓶。系統的主控制器控制法則主要是以引擎轉速和電瓶的放電深度為主要輸入參數建立一模糊控制架構來決定控制指令，主要的輸出控制參數為引擎瞬時操作點。另外車速切換點和發電機的瞬間最大發電量也隨電瓶的放電深度來調整。基因演算法是利用來尋找模糊規則庫裡的最佳解。
由模擬的結果顯示，整體系統最佳操作點的解有多組解而非單一解，因此導致基因演算過程中收斂狀況不理想；而且其與引擎最佳操作點差異大，所以傳統以引擎最佳操作點為基準來訂定最佳控制策略的原則並不適合；另外得知整體系統最佳操作點與所模擬的行車型態有很大的關連性，因此最佳的控制策略必須配合每部車輛的行車狀況來做即時最佳化的個別調整。
關鍵字：模糊控制、混成動力機車系統、基因演算法、
並聯式雙軸配置、高量級混成動力系統、引擎最佳操作點、
整體系統最佳操作點

並列摘要

Abstract
This research proposes a hybrid electric system for motorcycle with fuzzy control with corresponding driving performance is simulated based on various driving cycles. Then, the genetic algorithm is further applied to search for an optimal fuzzy control strategy. The system is of a parallel type with double-shaft layout and can also be classified as full hybrid type. Basically, the vehicle is driven purely by electric motor at low speed, while, at high speed, by a gasoline engine assisted by electric motor which can be switched to generator mode to recharge battery if required. The main control strategy uses engine speed and depth of discharge for battery as major input parameters. The corresponding control commands are determined through fuzzy logic reasoning. The sole output control parameter is instant engine operating point. Furthermore, the switching point of velocity and instant maximum load of generator is adjusted according to the instant depth of discharge for battery. The genetic algorithm is adopted to optimize the data in the fuzzy rule table of the fuzzy control.
The results from simulation show that there are multiple solutions, rather than a single solution, for the optimal global operating points. This may cause an undesirable condition of convergence in genetic searching process. Furthermore, since the optimal global operating point is largely deviated from the optimal engine operating points, the conventional way to construct the optimal control strategy based on the engine optimal operating points is impropriate. In addition, the resulting solutions of the optimal global operating points are closely related to the driving cycle used for simulation. Hence, the optimal control strategy need to be obtained by a real-time process of optimization depending on various specific driving conditions for every individual vehicle.
Keyword: fuzzy control, hybrid electric system for motorcycle, parallel type with double-shaft layout, full hybrid type, the optimal global operating points, the optimal engine operating points

參考文獻

【6】 許騌譁，“混成動力機車控制系統設計與製作”，國立虎尾科技大學碩士論文，民國94年。

【7】 楊書楷，“混成動力機車之性能模擬與實驗分析”，國立虎尾科技大學碩士論文，民國94年。

【12】 Bernd Baumann, Giorgio Rizzoni and Gregory Washington, “Intelligent Control of Hybrid Vehicles Using Neural Networks and Fuzzy Logic”, SAE Paper, No.981061, 1998.

【13】 Bradley Glenn, Gregory Washington and Giorgio Rizzoni, “Operation and Control Strategies for Hybrid Electric Automobiles”, SAE Paper, No.011537, 2001.

【14】 Valerie H. Johnson, Keith B. Wipke and David J. Rausen,“HEV Control Strategy for Real-Time Optimization of Fuel Economy and Emissions”, SAE Paper, No.011543, 2000.