應用時間序列模式與灰模式進行電力需求預測與節能改善評估 以朝陽科技大學為例

Prediction and Improvement of Electricity Need by Time Series Models and Grey Models– An Example of Chaoyang University of Technology

王新春(Hsin-Chun Wang)

指導教授 ： 羅煌木

朝陽科技大學/理工學院/環境工程與管理系/碩士(2010年)

摘要

本論文針對朝陽科技大學所建置的校園電力監控系統介紹基本架構及功能。利用所收集的資料加以分析，讓用電保持在契約容量以下，並以模式預測其用電趨勢，俾有助電量消耗及電費支出控管，達到節約能源的目標。 分析運用的總用電量數據資料共計60 筆數據，經指數平滑法（Exponential Smoothing）、時間序列分析（Time Series Analysis Model）、灰模式（Grey Model）進行電力預測，以準確度比（Accuracy Ratio，AR）、平均殘差值百分比(Mean Absolute Percentage Error，MAPE)、均方根誤差百分比(Root Mean-Square Percentage Error，RMSPE)、相關係數R（Correlation Coefficient）來檢測預測值之效力。 經研究顯示電力監控系統可以提昇電力的穩定度及降低不必要的浪費，並以時間序列相乘模式的預測效力其準確度91.87％、平均殘差百分比8.13％，屬於高精確度預測，供管理者作為未來電力需量管理指標及參考依據。

關鍵字

灰模式 ； 電力預測 ； 指數平滑法 ； 時間序列分析

並列摘要

This study aims at investigating the campus power monitoring system including its basic structure and functions installed in Chaoyang University of Technology. The consumption of electricity was monitored and controlled by adjusting the operating conditions. Operating data were employed to predict the trend of electricity use by several models. Data analysis and model prediction could provide the baseline information for the control of electricity use and the improvement of energy savings.
Sixty data sets were used to models prediction. Predicting models include Exponential Smoothing、Time Series Analysis Model and Grey Model. The accuracy and precision of modeling was verified by Accuracy Ratio (AR), Mean Absolute Percentage Error, (MAPE), Root Mean-Square Percentage Error (RMSPE) and Correlation Coefficient R.
Results showed that the power monitoring system is capable of enhancing the stability of power and reducing the unnecessary waste. The prediction by Multiplicative Model of Time Series Analysis Module could reach 91.87％ and the MAPE was 8.13％ showing the highest accuracy among the predicted models. These results could be used for the references of decision makers.

並列關鍵字

Exponential Smoothing ； Time Series Analysis Model ； Grey Model ； Electricity prediction

參考文獻

Deng J.L., “Introduction to Grey System Theory,” The Journal of Grey System, Vol.1, pp.1-24（1989）

Liang, Ruey-Hsun, “Grey Linear Programming Based Redispatch Technique to Dynamic Generation Allocation,” Science and Technology periodical, Vol. 349-355.（1998）

Morita, H., T. Kase, Y. Tamura, and S. Iwamoto, “Interval Prediction of Annual Maximum Demand Using Grey Dynamic Model,” Electrical Power & Energy Systems, Vol.18, No.7,409-413.（1996）

C. Chartfield, M. Yar, “Holt-Winters forecasting: some practical issues”, The Statistician, Vol. 37, No. 2, pp. 129-140, 1988.

史光榮，「電力系統動態狀態估計與負載預測之研究」，博士論文，國立成功大學機工程研究所，台南（2002）