摘要
題 目 ：建立台灣在經典氣象年中的太陽能資料庫
研 究 生 ：嚴偉倫
指導教授 ：張克勤
　　在日本福島第一核電廠核災後，台灣已經掀起了一波反核能的浪潮，又顧慮火力發電所需燃料大多依賴進口，並且有空氣汙染的問題，故台灣極力發展再生能源，其中由於台灣位於北迴歸線附近，日照時數長，所以太陽光電及太陽光熱的應用在台灣再生能源中扮演著很重要的角色。但在台灣太陽能應用評估的數據上，缺乏漫射率資料，並且中央氣象局在2003年4月以前所裝置之日射量測儀器皆有誤差，故本研究為了更新並補齊台灣太陽能資料庫，使用中央氣象局2004年至2013年十年間全台30個氣象觀測站之氣象資料，將其氣象觀測站缺少資料利用三階方程式進行補齊，並按照經典氣象年的規範與步驟整理出三種加權函數下的經典氣象年，此經典氣象年之氣象資料可代表該地區的常態氣候狀況，且應用在不同的領域可選擇不同之加權函數。往後在評估太陽能應用時可做參考本研究所選取之經典氣象年氣象資料。
　　太陽能輻射量通常可分為兩種形式：直射輻射量與漫射輻射量。而漫射率則為兩者之間的比值，其漫射率直接影響某些太陽能應用之效率，故本研究以2014年國立成功大學歸仁校區量測之漫射率資料進行內差補齊並刪去不符合物理意義之量測值，最後迴歸成一條三階經驗方程式， 以利於往後在台灣預估漫射率時可以使用。

Establishment of Taiwan solar energy information
in terms of the typical meteorological year

Author: Yen Wei-Lun
Advisor: Chang Keh-Chin

Department of Aeronautics and Astronautics, National Cheng Kung University
SUMMARY
Taiwan is located in the subtropics with abundant solar irradiation. So, solar energy is the most available green energy for Taiwan. Thus, Taiwan is suitable for developing solar energy applications. The solar irradiation database of Taiwan is incomplete, and the instruments for the measurements of global-radiation used in each station of the Central Weather Bureau (CWB) were inherent to erros before April 2003. This thesis uses the meteorological data between 2004 and 2013 from 30 meteorological stations of CWB, in order to update and repair the Taiwan solar repository. This thesis uses third-order equations to fill the database which were lacked in the report of CWB. Next, I set a database in a so-called typical meteorological year (TMY), which represents long term solar data for Taiwan. In assessing the application of solar energy, the meteorological database of typical meteorological year will be used as a reference.
Global-radiation is usually divided into two parts: beam radiation and diffuse-radiation. The diffuse-fraction is a proportion of diffuse component in global-radiation. The extent of diffuse-fraction can affect the way we design solar energy application systems. This research utilizes the diffuse-fraction data measured at Kuei-Jen in Tainan and develops a diffuse-fraction model.
Key words: typical meteorological year (TMY), solar thermal energy, diffuse
INTRODUCTION
Green energy includes: solar energy, wind power, hydroelectric power, tidal energy, biomass energy, geothermal energy, and marine energy. Taiwan is located in a subtropical region. Thus, Taiwan has sufficient sunlight for solar energy applications throughout the year. Solar energy applications are usually divided into two parts: solar thermal applications and solar photovoltaics (PV). Solar thermal applications utilize sunlight irradiating on a collector. This results in a temperature rise of the working fluid, and then uses the working fluid to achieve specific goals. Some examples are: solar energy water heaters and factory reheating of working fluid. Solar photovoltaics (PV) utilize sunlight irradiating on photovoltaic modules. The photovoltaic materials absorb sunlight and cause electricity generation. Photovoltaic wavelength absorption is limited. Therefore, solar thermal applications are more efficient than solar photovoltaics.

DIFFUSE-FRACTION MODELLING
In order to establish a diffusion-fraction model, several important parameters must be calculated. Important parameter explanations and formulas are as follows:

Clearness Index (k_t):
Clearness index is defined as the ratio of global-radiation and extraterrestrial radiation. Determining how much radiation is absorbed by the atmosphere will also make us aware of weather conditions. The formula is as follows:
k_t=I_global/H_0 (1)
I_global: hourly global-radiation.
H_0: hourly extraterrestrial radiation.
Extraterrestrial Radiation (H_0):
Extraterrestrial radiation is defined as sunlight radiation without loss when having passed through the atmosphere. Also, there are limits on the maximum solar radiation reaching the Earth's surface. The formula is as follows:
H_0=(3600*12*G_sc*[1+0.033*((360*N)/365)])/π
*∫_w1^w2▒(cos∅*cosδ*cosw+sin∅*sinδ)dw (2)
G_sc: solar constant(1367 W⁄m^2 ).
∅: Latitude.
N: N days of the year.
ω: hour angle.
δ: Declination.

Diffuse-Fraction (d):
Diffuse-fraction is defined as the ratio of diffuse-radiation and global-radiation. The formula is as follows:
d=I_diffuse/I_global (3)
I_diffuse: diffuse-radiation.
I_global: global-radiation.
This research utilizes the diffuse-radiation data and the global-radiation data measured at Kuei-Jen in Tainan. The regression equation is used to establish a third-order empirical model for diffusion-fraction in Tainan, Taiwan.

TYPICAL METEOROLOGICAL YEAR
A typical meteorological year (TMY) represents the long-term observation years of climate in an area, composed by 12 different typical meteorological months, respectively. TMY’s information can be used for the performance assessment of solar system.
This thesis uses a third-order polynomial equation to correlate the database. The daily meteorological data is composed by average the hourly meteorological data. The formula is as follows:
Daily meteorological data=(∑_(i=1)^24▒x_i )/24 (4)
The long-term cumulative distribution function (CDF) and store-term cumulative distribution function are calculated as follows:
CDF(x_i){█(0 ,x<x_1@((i-0.5))/n,x_i≤x<x_(i+1)@1 ,x≥x_n )┤ (5)
Using Finkelstein-Schafer statistics’s (FS statistics) results to compare with the long-term CDF and store-term CDF. The Finkelstein-Schafer statistics (FS statistics) can be used to see which year is closest to the average meteorological datebase. The formula is as follows:
FS statistics=1/N ∑_(i=1)^N▒|〖CDF〗_LT (x_i)-〖CDF〗_ST (x_i)| (6)
Different purposes are used to select a different weighting factor (WF). Weighted statistics is calculated by FS statistics multiplied weighting factor for this year. The formula is as follows:
WS=∑_(i=1)^N▒〖〖(FS statistics)〗_i*〖(WF)〗_i 〗 (7)
Comparing with weighted statistics in different years and picking up the smallest weighted statistics comes out with a typical meteorological month. Typical meteorological year is composed by 12 different typical meteorological months.

CONCLUSION
In this thesis, I utilized the latest and most accurate database of diffuse-radiation in Kuei-Jen.We calculated the absolute error and the relative error of Shadow Band Stand (Model SBS) and AutoMatic Solar Tracker (Model SMT). The absolute error of data obtained between SBS and SMT become larger when the sun is closer to noon. The relative error of data is larger when the sun is closer to sunrise or sunset. The regression equation was used to establish a third-order empirical model for diffusion-fraction. Developed diffusion-fraction model in Taiwan is shown in the section 4.1.1.
This thesis uses the meteorological data from Taiwan’s 30 meteorological stations of CWB, from 2004 to 2013, in order to update and repair the Taiwan solar repository. The typical meteorological years for the 30 CWB’s stations in Taiwan are respectively determined which is presented in the section 4.2.2.

參考文獻
[1] 黃明輝，“台灣日照量高低的爭議”，科技報導，第396期，pp. 10-13，2014年。
[2] 歐文生、何明錦、陳瑞鈴、陳建富、羅時麒，“台灣太陽能設計用標準日射量之研究”，中華民國建築學會建築學報，第64期pp. 103-118，2008年。
[3] 蕭鋒、林博雄、賴彥任，“由MTSAT衛星影像和數位地形資料估算台灣地表太陽輻射量”，大氣科學學報，第39期pp. 53-67，2011年。
[4] 方煒，“台灣的太陽能輻射現況”，台灣大學生物產業機電工程學系，2000年。
[5] 林博雄，“中央氣象局測站之地表全天空輻射量的效驗與分析”，交通部中央氣象局委託研究計畫成果報告，pp. 398-453，2003年。
[6] B. Y. H. Liu, and R. C. Jordan, “The Interrelationship and Characteristic Distribution of Direct, Diffuse and Total Solar Radiation,” Solar Energy vol. 4, pp. 1-19, 1960.
[7] C. W. Kuo, W. C. Chang, and K. C. Chang, “Modeling the hourly solar diffuse fraction in Taiwan,” Renewable Energy vol. 66, pp. 56-61, 2014.
[8] C. W. Kuo, W. C. Chang, and K. C. Chang, “Distribution of solar diffuse fraction in Taiwan, ”Energy Procedia vol. 57, pp. 1120-1129, 2014.
[9] D. T. Reindl, W. A. Beckman, and J. A. Duffie, “Diffuse Fraction Correlations,” Solar Energy vol. 45, pp. 1-7, 1990.
[10] J. L. Torres, M. De. Blas, A. Garcia, A.de. Francisco, “Comparative study of various models in estimating hourly diffuse solar irradiance,” Renewable Energy vol. 35, pp. 1325-1332, 2010.

[11] National Renewable Energy Laboratory (NREL), “User's Manual for TMY2s,” 1995.
[12] A. L. S. Chan, T. T. Chow, K. F. Fong, and J. Z. Lin, “Generation of a typical meteorological year for Hong Kong,” Energy Conversion and Management vol. 47, pp. 87-96, 2006.
[13] 何明錦、歐文生、陳建富，“台灣太陽能設計用標準日射量與相關檢測規範之研究”，內政部建築研究所，2006年。
[14] 廖顯峻，“國內太陽能漫射率預測模式建立”， 國立成功大學碩士論文，2013年。
[15] 李聰盛，“廠區漫射量測及對太陽能集熱器性能影響探討”，太陽熱能南部檢測中心之技術報告，2010年。
[16] J. A. Duffie, and W. A. Beckman, “Solar Engineering of Thermal Processes,”John Wiley & Sons, Inc, pp. 37, 2013.
[17] 郭家瑋，“如何選取經典氣象月”，太陽熱能南部檢測中心之技術報告，2012年。