由於近年來大量併接分散式電源於配電系統，造成傳統電力系統莫大影響，其中又以再生能源發電量波動影響最為劇烈，若本地負載無法消耗分散式電源之發電量即可能發生逆送電力情形，因此發展適用含分散式電源之配電系統三相不平衡負載潮流分析格外重要，透過前一日之排程可以做為系統運轉者調度發電機組之參考，使系統運轉具最高經濟效益。 本論文提出一種新的機率負載潮流方法，使用田口方法中的田口直交表對隨機變數之機率密度函數取樣，求取匯流排電壓、線路潮流之平均值及標準差，並透過因子反應，選擇線路潮流變化最大之情境，結合本論文發展的分散式電源模型，設定極端情境於特定末端節點裝設風力機組及太陽能發電系統，導致系統電壓升高超過法規規範值，透過最佳化方法計算分散式電源電壓設定，依據本論文提出的模型，改善當前的不正常情形，同時考慮電力系統負載變動及再生能源間歇性。模擬結果將與蒙地卡羅方法以及點估計法比較，以25 Bus澎湖電力系統、IEEE 118-bus系統及修改的12 Bus台電系統驗證本方法節省機率潮流的運算時間同時兼顧結果的精確性。 ;As the penetration of distributed generation goes up rapidly, the traditional power system is faced with challenges. One of the most severe problems for renewable energy is the fluctuation of the output power. If the output power of distributed generators could not be consumed by local loads, inverse load flow may occur. As a result, developing three-phase unbalanced load flow algorithms for distribution system considering distributed generation seem to be particularly important. By means of day-ahead scheduling, operators could take it as a reference to make an optimal decision. This thesis proposes a novel probabilistic load flow method that is based on Taguchi’s orthogonal arrays from Taguchi method. The proposed method only needs to sample probabilistic variables through orthogonal arrays and utilize few deterministic load flows to obtain the mean and standard deviation of bus voltages. An optimal experiment is also achieved through main effects from Taguchi’s method with the largest deviation from the nominal line-flow result. This thesis also presents a more practical model for distributed generators. To proof the well-behavior of this model, some abnormal operation scenarios are demonstrated. When over-voltage or low-voltage situation occurs caused by excess output power from wind turbine and PV, the model will acquire a voltage set point from optimal algorithm in order to absorb or generate reactive power to improve voltage profile considering the deviation of demands and output power from renewable energies. The simulation results were compared with results attained by Monte Carlo simulation (MCS) and point estimate method (PEM). A 25-bus Penghu system, an IEEE 118-bus system and a modified 12-bus feeder in Taiwan are used as test systems. The results show that the proposed method attains both accurate means and standard deviations of bus voltages and line flows and is time-saving.
