在 FPGA-based系統上實現增強型 RBF網路之太陽能最大功率點追蹤設計與研製

Design and Implementation of the Enhanced RBF Network on a FPGA-based System for PV Maximum Power Point Tracking

鄭建德(Chien-Te Cheng)

指導教授 ： 廖炯州

健行科技大學/電資學院/電子工程所/碩士(2010年)

摘要

本文將使用 DE2-70 多媒體開發平台，設計與實現新型太陽能板最大功率追蹤器。藉由操作太陽能板更接近於最大功率點，則太陽能板發電效率將可有效提升。傳統上，k 階均值演算法(KMA)為最常使用於輻狀基底函數網路演算法的前置階段中，分類輸入樣本的一種方法。雖然 k階均值分類演算法能將資料快速的進行分類，但可能因為隨機挑選初始中心點而陷入局部最佳解。為了解決這個問題，因此本文提出結合 k階均值分類演算法和基因演算法的混合演算法，簡稱 GKA，用來解決上述問題。此方法使用基因演算法之特性，可降低原有方法因隨機挑選初始中心點，而對網路收斂結果產生的不穩定性，並且可以藉著基因演算法採取多點搜尋之特性，能具備跳脫局部最佳分類之特性，解決對網路訓練結果產生的不確定性。除此之外，為獲取輻狀基底函數網路中隱藏層適當的節點個數，本文將使用正交最小平方演算法(OLS)來計算求得。並將其應用在追蹤太陽能發電最大功率點上，使其能更準確的追蹤到太陽能的最大功率點。本文將用 Matlab 套裝軟體模擬所提出之方法，並以架設於本校樓館之太陽能板實際量測資做訓練，經由 DE2-70 開發版進行太陽能最大功率點追蹤器之實際硬體製作與測試，將結果比較於傳統輻狀基底函數網路，依此驗證此法的可行性。

關鍵字

太陽能 ； 最大功率點 ； FPGA

並列摘要

This project is to design and implement a new type of photovoltaic maximum power point (MPP) tracker using DE2-70 FPGA multimedia development platform. By operating PV systems more close to the maximum power point, the output efficiency of PV panels can be improved. Traditionally, the k-means algorithm (KMA) is one of the most popular methods to classify the input patterns of the radial basis function (RBF) network. Although the KMA has an ability to cluster the training patterns rapidly, it usually converges to a local minimum and can be oversensitive to randomly initial partitions. To solve these significant problems, a hybrid skill called Genetic k-Means Algorithm (GKA) is proposed to improve the effectiveness of maximum power point track. Besides, the proposed GKA based clustering approach can overcome the problem of oversensitivity to randomly initial partitions in the existing KMA. In order to determine a suitable number of centers in RBF from the input data, the orthogonal least squares (OLS) learning algorithm was used in this paper. By precisely clustering of the training patterns, the objective to accurately and rapidly approximate the MPP of PV system can be achieved with the least squares criterion in RBF network. Also, this project employed the actual data obtained from the DE2-70 FPGA development broad and the practical PV systems and with which the developed MPP tracker method was proven to be effective.

並列關鍵字

Photovoltaic ； Maximum power point ； FPGA

參考文獻

[1]友晶科技股份有限公司，DE2-70_UserManuall，2007.

[2]友晶科技股份有限公司，TRBD_LTM_UserManuall，2007.

[3]李柏宏, “應用基因演算法結合k階均值分類演算法於輻狀基底函數網路之太陽能最大功率點預測”, 私立清雲科技大學電子工程系碩士論文, 2008.

[4]E. A. Burke, G. P. Summers, J. H. Warner, R. J. Walters, S. R. Messenger, and T. L. Morton, “Effect of Omnidirectional Proton Irradiation On Shielded Solar Cells,” IEEE Trans. on Nuclear Science, Vol. 53, pp. 3771-3778, No. 6, Dec. 2006.

[5]D. Van Gestel, D. Vanhaeren, G. Beaucarne, I. Gordon, J. Poortmans, L. Carnel, and P. Eyben, “Impact of Preferential P-Diffusion Along the Grain Boundaries on Fine-Grained Polysilicon Solar Cells,” IEEE Electron Device Letters, Vol. 28, pp. 899-901, No. 10, Oct. 2007.