台灣地區家用冷氣機持有量預測：三種預測模型之比較

Abstract

由於冷氣機需求的成長，連帶使得社會保有冷氣機的持有量亦相對成長。然而冷氣機耗電量相當大，對於台灣地區電力系統夏季尖峰負載造成很大的負擔。復以政府環境保護規定日趨嚴格，致使台電公司在近年來即使積極開發各項能源，仍無法因應負載之快速成長。因此，瞭解台灣地區家用冷氣機持有量其成長的趨勢以及尖峰負載時冷氣機造成的負載影響程度，對於電力事業的負載管理，必有相當之助益。 本研究分別以傳統迴歸、灰色理論及羅吉斯曲線，預測93年前台灣地區冷氣機持有量，並比較各方法之預測結果。其次，推估93年前台灣地區家用冷氣機負載情形其及對尖峰負載的影響。 經研究結果，根據模式本身的預測能力以及冷氣機持有量資料特性，三種模式以羅吉斯曲線顯然最符合實際。 以羅吉斯曲線冷氣機持有量(本研究)為數據，預估93年冷氣機開機率每提高5%時，尖峰負載約增加1,159千瓩。以冷氣機持有量(台電)為數據，預估93年家用冷氣機開機率每提高5%時，尖峰負載約增加1,437千瓩。

The possessing quantity of household air-conditioners（PQHAC）is growing up with the raising tide of the needs for air-conditioners（ACs）. However, during summer peak-load, the heavy duty of ACs is a big burden to the power system in Taiwan. Although Taipower has developed various energy resources in the past few years, power supplies still do not meet the surging need because of limitations from the strict tendency of regulations upon environmental protection. Therefore, it is beneficial to the load-management of power utility to understand the raising trend of PQHAC and the degree of effect of ACs upon power system in peak-load time. This study adopted regression analysis, grey system theory and logistic curve to forecast the PQHAC in Taiwan before 2004 and compared the outcomes of each model. In addition, the estimation of loading of household ACs and its effect upon peak-load of the system followed from the analysis. Among the three models, logistic curve is the most applicable one basing upon the forecasting ability examination and the data property of PQHAC. In conclusion with the data of PQHAC from Taipower, logistic curve shows that peak-load in 2004 increases about 1,437 mega watts with 5% increase of the turn-on rate of ACs; while with the data of modified PQHAC from this study, logistic curve shows that peak-load in 2004 increases about 1,159 mega watts with 5% increase of the turn-on rate of ACs.