以觸媒不完全氧化之燃料處理系統的負載追蹤最佳化設計

Optimal Load Tracking Design for a Catalytic-Partial-Oxidation-Based Fuel Processing System

本論文運用文中推導的通用型線性二次高斯及迴路轉移函數回歸法的設計法則，進行一個以天然氣為燃料、觸媒不完全氧化反應之燃料處理系統的負載追蹤最佳化控制系統設計，最佳化控制法則係以燃料電池輸出的電流大小為負載追蹤目標。控制的目標集中在調整燃料處理系統輸出向量的反應。首先要設計一個最佳控制器滿足通用型二次線性性能指標，然後設計卡門濾波器來提供一個最佳化狀態變數的估測。本文所提出的方法可以適用於極小相和非極小相的系統，同時提供一個額外的設計自由度及特定的穩定度。最後，本文以觸媒不完全氧化的天然氣燃料處理系統為例，經由數值模擬分析，顯示本方法在時域和頻域響應有較好的性能和穩定性特性。 This report presents an optimal control system design for a fuel processing system (FPS) using a generalized linear quadratic Gaussian and loop transfer recovery (LQG/LTR) method. The FPS uses natural gas as fuel and works with a catalytic partial oxidation (CPO) reaction. The control objective focuses on the regulation performances of an output vector in response to a stack current command. First, an optimal controller was designed subject to a generalized linear quadratic performance index to shape the target feedback loop function. Then a Kalman filter was designed to provide an optimal estimation of state variables for minimum and non-minimum phase plants in an LTR process. The proposed method provides another degree-of-freedom in optimal controller design and enables the compensated system to maintain a prescribed degree of stability. Finally, numerical simulations reveal that the proposed method achieves better performance and robustness properties in time- and frequency-domain responses.