在台灣未來電力投資日趨困難下，如何準確預估的用電需求避免發電及電力建設浪費為一重要的課題，本論文以時間序列分析方式分別討論台灣各都會區一般住宅用及小型店家用電的影響因素，以台灣各都會區2007年到2012年底之電燈售電量月資料建立OLS模型，此外分別考慮時間序列資料恆定性及序列相關性，以修正模型，最後比較各用電影響因素和用電量的關係及顯著性，以討論各用電影響因素對用電之影響。 本論文最後結論人口數增加對用電影響甚微，溫度則為主要的影響因子，平均溫度上升1℃每人平均增加3.5度電的使用量，且以大台北增加最多，近6年的3次電價調漲對用電量抑制效果亦不大，同樣的在每年6-9月實施夏季電價對抑制用電量的效果亦不明顯。最後本論文建議如欲抑制用電量，可再增加電價或課增能源稅，也可再增加夏季電價並依區域作差別取價，以維使用者付費及環境節能減碳。 Power investment in Taiwan is increasingly difficult, it is a crucial issue to forecast precise electricity consumption and to avoid the waste of construction of power and electricity generation. In this thesis, time series analysis methods were applied to discuss the power consumption of the influencing factors in both residence and small stores in every urban area in Taiwan. OLS model was established by using lighting sales data from January 2007 to December 2012. Also the stationary of time series data and serial correlation were considered to correct the OLS model. Finally, we compare the power consumption of the influencing factors and relationships and significant, to discuss the impact of various factors on the electricity consumption effect. The conclusion of this thesis shows that the increase of population is minor to electricity consumption. The main factor affecting the power consumption is temperature. Among four regional, average temperature rise 1 ℃ average monthly increase of about 3.5 degrees lighting electricity, and the largest increase is in Taipei. Three tariff adjustments in the last six years seems to be little effect. The effect of summer tariff between June and September is not obvious as well. Finally, the thesis recommends to suppress power consumption, can be increased by increasing tariffs or class energy tax, but also to increase the area for the summer tariffs and price discrimination according to user fees and environmental dimension carbon reduction.
