影像視覺追日偏差量測技術開發與追日太陽光電系統之實測;Development of Measuring Sun-Tracking Deviation Technology with CCD Image and Field Test of Photovoltaic System

NCU Institutional Repository > 工學院 > 能源工程研究所 > 博碩士論文 > Item 987654321/53779

請使用永久網址來引用或連結此文件: http://ir.lib.ncu.edu.tw/handle/987654321/53779

題名:	影像視覺追日偏差量測技術開發與追日太陽光電系統之實測;Development of Measuring Sun-Tracking Deviation Technology with CCD Image and Field Test of Photovoltaic System
作者:	林武君;Lin,Wu-chun
貢獻者:	能源工程研究所
關鍵詞:	CCD;追日偏差角度量測;PSD;閉迴路追日控制、混合式追日控制;影像處理;Closed-loop sun-tracking control;Image processing;PSD;Hybird sun-tracking control;CCD;Offset angle measurment
日期:	2012-07-27
上傳時間:	2012-09-11 18:14:24 (UTC+8)
出版者:	國立中央大學
摘要:	本文開發追日偏差角度量測方法，採用CCD攝影機做為量測設備，並以此設備分別在戶外實測二套閉迴路追日控制（億芳光感測器）與混合式追日控制（短路電流搭配太陽軌跡公式）的雙軸追日PV系統，監控數據包括追日偏差和PV電力性能。本文以LabVIEW開發影像處理演算法，影像處理目的在獲取目標物的輪廓，再根據輪廓尋找出目標物的質心位置，處理過程分為四個步驟：太陽影像擷取、影像二值化處理、去雜點處理與計算目標物質心位置。實測顯示分群二值化比固定閥值二值化更適合在光源多變的情況使用；本文亦比較重心法與邊緣檢測法對太陽影像形狀判讀的差異，晴天時兩者平均偏差角差異範圍介於0.008o~0.01o，多雲時則介於0.02o~0.05o，顯示邊緣檢測法即使太陽被雲層遮蔽仍不會影響其準確性，因此邊緣檢測法優於重心法。接著以PSD量測技術驗證CCD量測技術的可行性，前者具有高精度以及高傳輸速度特性，後者則可藉由影像處理修正太陽被雲層局部遮蔽的影像。以CCD實地測試閉迴路追日控制與混合式追日控制，前者6天平均追日偏差為0.147o，平均輸出功率為1226 W（統計6天累計發電量為73.73 kWh），後者混合式追日控制5天平均追日偏差為1.291o，平均輸出功率為1291 W（統計6天累計發電量為68.9 kWh）。前者偏差角大都小於0.3o，後者的追日則主要依據短路電流偵測策略採用擾動方式，短路電流受到日照分布影響甚大、加上追蹤器受重力影響產生變形，導致以短路電流尋找到最大功率點的同時，追日精度不夠小以及實測天數有限（僅有完整六天實驗），因此偏差角分布範圍較廣達3.5o。還需再進一步長期監控PV電力表現才能全面評估短路電流追日方法的特性。雖然短路電流的追日精度不如億芳光感測器，但對PV系統而言，發電量才是決定系統性能的最終指標，因此混合式追日搭配短路電流方法仍為一實用且有效的追日方法。This thesis developed an apparatus for measuring the offset angle of sun-tracking system with CCD camera and performed the outdoor tests of two photovoltaic (PV) systems which are controled by closed-loop and hybird sun-tracking (combine the detection of short-circuit current and sun-position calculation), respectivily. Monitoring data including offset angles and PV power performance.CCD sensor is a discrete element structure based on the matrix elements to record images. This thesis used LabVIEW to develop an image processing algorithm to acquire the image profile, and then find out the centroid of image. Image processing consists of four parts: acquiring of sun images, image threshold processing, target filtering, and calculating centroid position. Clustering-threshold method is better than the fixed-threshold binary approach in condition of varying light attenuation/scattering on outdoor testing. Comparing with centroid approach and edge-detection method for distinguishing sun images was carried out. On clear days, the average of offset angles ranging from 0.008o-0.01o, and values of 0.02o-0.05o was obtained on cloudy days. Results with the edge-detection method shown that in the case of cloud covering on sunlight wouldn’t degrade its accuracy. Thus the edge-detected method is better than the centroid method. Position sensitive device (PSD) technique is adopted to validate CCD technique for its applicability. The former has high accuracy and transmission, and the latter can correct local cloud covering on sunlight by image processing.Field testing with CCD for monitoring the close-loop and hybrid sun-tracking found that the former has an average of offset angle of 0.147o on a six-day period and the average power output of 1226 W (total electricity on six days is 73.73 kWh). The hybrid sun-tracking has an average of offset angle of 1.291o on a five-day period, and the average power output of 1291 W (total electricity on six days is 68.9 kWh). Most of offset angles of the former are less than 0.3o. While the latter mainly using perturbation to detect the short-circuit current of PV module, which significantly being affected by solar radiation, deformed by gravity when finding maximum power point where offset angle is not small and limited test period (only full 6 days measurement). Thus the hybrid sun-tracking has large offset angles ranges up to 3.5o. Further long time monitoring of PV electrical performance is needed to fully assess the characteristics of sun-tracking based on short-circuit current. Although the short-circuit current tracking accuracy is less than the sun-tracking sensor, yet for assessing the performance of PV system, electricity is the ultimate factor. Thus, the hybrid sun-tracking with the short-circuit current algorithm is a practical and effective sun-tracking method.
顯示於類別:	[能源工程研究所 ] 博碩士論文

文件中的檔案:

檔案	描述	大小	格式	瀏覽次數
index.html		0Kb	HTML	746	檢視/開啟

在NCUIR中所有的資料項目都受到原著作權保護.

社群 sharing

資料載入中.....