基於蝙蝠演算法之二自由度PID控制器應用於直流–直流轉換器平台

Abstract

本論文以數位信號處理器 DSP TMS320F28335，實現一並聯式升、降壓直流-直流功率轉換器的智慧型控制。首先，以三台直流-直流轉換器並聯，並使用自動主僕均流控制技術，使每台轉換器在電壓有良好的穩壓性能情況下，電流能夠平均分配給三台直流轉換器。然而，轉換器因元件老化導致特性改變、負載變動、電路雜訊及外部干擾等情況下，原廠的控制器參數未能有效的匹配目前的轉換器。為了解決此一困難，本論文開發一個基於蝙蝠演算法之二自由度比例-積分-微分(Two-degree- of-freedom Proportional Integral Derivative, 2DOFPID)控制器，透過蝙蝠演算法(bat algorithm, BA)模擬蝙蝠飛行和狩獵的形式，將所有控制參數當作一個搜索空間中的微蝙蝠，蝙蝠在尋找最佳解的過程中看成是蝙蝠尋找食物的過程，以進行動態優化。從模擬和實驗結果可得知，本論文所提出的蝙蝠演算法之二自由度PID控制器，與傳統單自由度PID控制器及二自由度PID控制器相比，具有最小電壓調整率和最小均流誤差率。In this thesis, the digital signal processor DSP TMS320F28335 is used to realize the intelligent control of a parallel Buck/Boost DC-DC converter. First of all, three converters are connected in parallel, and automatic master-slave current-sharing control technology is used, so that each group of converters can be evenly distributed to three converters under the condition that the voltage has good voltage regulation performance. However, the converter's controller parameters fail to match the current converter effectively due to component changes, load changes, circuit noise, and external disturbances due to component aging. In order to solve this difficulty, this thesis develops a two-degree-of-freedom PID (Two-degree-of-freedom Proportional Integral Derivative, 2DOFPID) controller based on bat algorithm (Bat algorithm, BA), Through the bat algorithm to simulate the form of bat flying and hunting, all control parameters are treated as micro-bats in the search space. The bat is seen as the process of bats searching for food in the process of finding the best solution for dynamic optimization. From the simulation and experimental results, the bat algorithm of two-degree-of-freedom PID controller proposed in this paper has the minimum voltage regulation rate and the minimum current sharing error rate compared with the PID controller and the two-degree- of-freedom PID controller.