灰色GM(1,1)模型與灰色馬可夫模型於電力負載之研究

Apply GM(1,1) Model and Grey-Markov in Power Load Analysis

10.6995/JCTU.200801.0113

張簡士琨(Shih-Kuen Changchien)；葉鎮愷(Cheng-Kai Yeh)；林金桂(Chin-Kuei Lin)；溫坤禮(Kun-Li Wen)

電力負載 ； GM(1,1)模型 ； 灰色馬可夫模型 ； 尖峰負載 ； 平均負載 ； Power demand ； Load forecasting ； GM(1,1) ； Grey-Markov model ； Peak loading ； Average loading

建國科大學報

27卷2期（2008 / 01 / 01）

113 - 131

繁體中文

一般而言，預測必須具有時間性，準確性及可靠性，並且預測的方法應該容易暸解與使用，而灰色理論中的灰色6M(1,1)模型恰巧均符合以上之條件，更重要的是只需要四個點就可以預測，而灰色馬可夫模型是灰色6M(1,1)模型與馬可夫模型的結合，爲此一模型的延伸，因此本文使用灰色6M(1,1)模型，並整合馬可夫模型成爲灰色馬可夫模型，使用於電力負載中的尖峰負載與平均負載之預測，並比較兩者之優缺點。首先，在本文中提出電力負載之相關概念及欲分析之課題。接著使用灰色系統理論的6M(1,1)模型和馬可夫的數學模式，並整合成灰色馬可夫模型。基於所提出之數學模式，以台灣的電力負載爲實例加以驗證，期能對電力負載做出貢獻。

In recent years, the power demand of installed capacity, generation, consumption of electricity, and user number are continuously increasing in Taiwan largely. However, due to the ancient power station, may be turn-off at any time, so, it is necessary to plan and to prevent the accident in power demand. Therefore, the load forecasting is essential and importance in power system field. In the prediction field, we know that the prediction require timely, accurate, and reliable, and the prediction method had better easy to understand for user. Therefore, in this thesis, we present GM(1,1) model and Markov method, and combine GM(1,1) model with Markov method, becomes Grey-Markov to predict the peak load and average power load in power system. First of all, this thesis proposes the related power demand ideas, and the willing analyzed topics. Second, the GM(1,1) model, which in grey system theory and the Markov model are introduced, then, we combine GM(1,1) model with Markov method, becomes Grey-Markov. After the mathematics model is presented, a real example of power demand in Taiwan is given to verify our approach. According to the real analysis and comparison between GM(1,1) model and Grey-Markov model, we can find that the Grey-Markov model is proper to predict the large damping, and the GM(1,1) model is proper to predict smooth variation, this is the major contribution in our paper.