本論文針對一微型電網系統,規劃其低頻電驛設定參數,當電力系統遇到嚴重擾動,如發電機組跳機時,以避免配電饋線不必要的跳脫與停電。 本論文使用NEPLAN V5.5.0模擬軟體建立微型電網電力系統,探討其電力潮流分析,並模擬系統遇到各類分散式電源跳機時之暫態穩定度分析,利用基因規劃法及處罰函數之概念,找出最適合系統的低頻電驛之參數設定。 本文使用馬可夫理論中的兩狀態模型來整理分散式電源跳機機率,利用氣象資料進而設定分散式電源發電量,並應用交談式多目標非線性規劃來設計傳統多段低頻電驛參數,求得微型電網系統之卸載量最小值與初始搖擺頻率最大值,以達到降低用戶端損失。 由模擬結果觀察到基因規劃法中,以多目標函數設計低頻電驛參數,電力系統頻率可趨於穩定,並使發生跳機事故時之總卸載量越小及最低頻率越高,因此驗證本論文所提出方法之可行性。
In this thesis, the parameters of the under-frequency relay are designed for a microgrid power system. The unnecessary load shedding can be avoided once the power system is subject to severe disturbances, e.g., generator tripping. The NEPLAN V5.5.0 software package was used to set up the microgrid power system. The power flow and the transient stability analysis were conducted considering distributed generations (DGs) trippings. The optimal parameters of the under-frequency relay are obtained using Genetic Programming and penalty function. The tripping probability of DGs were combined with the Markov two-state model and generations of DGs were set by considering real meteorological data. The interactive multi-objective nonlinear programming was used to design the traditional multi-stage load shedding. The shed load is minimized and the initial swing frequency is maximized for mitigating the loss of costumer’s load. It can be found from the simulation results that the power system can be stable using the proposed multi-objective approach based on the Genetic Programming. This validates the applicability of the proposed method.