風機發電量推估之不確定性分析-以中屯風力發電廠為例

Abstract

一般在推估風機發電量時，通常係以定值的風速資料或定值的風速分佈參數值推估發電量，但實際上在風速資料的觀測、風速分佈的套配與發電量的估算皆有不確定性存在，以定值無法完全表示其實際情況。故本研究利用蒙地卡羅模擬方法，結合各項影響風機發電量的因子，進行發電量不確定性的分析與驗證。本研究選擇了中屯風力發電廠與鄰近的澎湖氣象站評估模擬發電量是否符合實際發電概況，並將發電資料依時間長度之分為全年、半年強風期、半年弱風期與各月份，討論其模擬結果的不確定性。本研究同時也探討模擬次數對模擬結果的影響。而前人的研究尚未考慮平均風速的不確定性對發電量推估之影響，故本研究將之加入模式當中，並比較其差異。研究結果發現，當模擬次數達100000次時，會有較穩定之模擬結果。此外，考慮平均風速之不確定性會比不考慮平均風速之不確定性更為合理。各時期比較方面，全年、半年強風期、半年弱風期的發電量不確定性較低，強風期各月份的發電量不確定性次之，弱風期各月份的發電量不確定性最高。

Estimation of the wind power generation requires the knowledge of velocity distribution fitting, velocity prediction, and power output prediction. There parameters are treated as a specific value, but there exist various uncertainties in modeling and data acquisition. The traditional approach may not fully express the actual energy output. As a result, this study presents an integrated assessment of the uncertainties in energy output estimation using the Monte-Carlo approach to provide a comprehensive understanding of the variance in the energy output estimation. We choose Jhongtun wind power station and Penghu weather station as our study case to investigate the effect of various time scales, namely, whole-year period, strong-wind period, weak-wind period, and monthly period on the estimated energy output accuracy. It is more reasonable to incorporate the uncertainty of mean velocity, which is not considered by Kwon’s work (2010). According to our results, the prediction will converge for simulations with over 100,000 trials. whole-year period, strong-wind period and weak-wind period have relatively low uncertainties, separate monthly periods of strong wind have relatively high ones, and separate monthly periods of weak wind have the highest ones.