智慧型控制器非常適合應用於未知變數的非線性系統,與不易由傳統控制方法實現之控制器。 本文的目的是研究通用的直流轉直流電源轉換器之適應性小波類神經網路控制,所提的方法較傳統類神經網路更具有良好的穩定性與收斂性。首先針對直流轉直流電源轉換器與所提出之演算法逐一說明,最後從前饋式直流轉直流電源轉換器之實作上,列表如安定時間及動態超越量與傳統控制器做一比較,證明利用適應性小波類神經網路控制器來實現直流轉直流電源轉換器之可行性。
Intelligence controller is suitable for uncertain nonlinear systems, and to those not easy to be realized by classical design methods. In this thesis, a wavelet neural network controller for DC/DC power converter is investigated. The presented approaches are better than traditional design methods. First the DC/DC power converter and the proposed algorithms are described. Finally, settling time and percentage overshoot for startup and step response are compared to traditional controls. From the experimental results of forward DC/DC power converter, it can be demonstrated the possibility of applying intelligence control method in practical DC/DC power converter design.