基於適應式RoCoF參考點之快速反應輔助備轉服務以降低儲能系統運轉成本

Abstract

在再生能源高佔比電力系統中，系統慣量降低帶給系統頻率調節很大的挑戰。為處理這個問題，本文提出了一種基於自適應RoCoF參考控制策略的快速反應備轉服務(Fast Frequency Response, FFR)控制器，應用於電力系統發生事故時，維持系統頻率高於設定的閾值，並避免過多電池儲能系統(Battery Energy Storage System, BESS)電能轉移。為了使電力系統在所有可能的偶發事故下都維持安全運轉，FFR控制器使用離線決定的控制增益，並考量最嚴重事故情境設計控制器增益。為避免在較小的事故下有過多的FFR輸出，我們提出的自適應RoCoF參考值是依據線上的電力缺口預測及發電機升降載能力預測決定。這個方法提供的FFR可以符合不同操作情境的系統需求，包括連鎖故障，同時降低FFR電能轉移及操作成本。雖然此方法僅使用簡單機制且未設計頻率響應估測器，但是其減少FFR電能轉移之效果與使用個別事故最佳增益的控制器相當。本文以修改過的9匯流排測試電力系統和規劃中具有高再生能源滲透率的2025年台灣電力系統中驗證該控制器的效能，模擬結果展示了其在各種情境下的適應性和效率。值得一提的是：在不同的事故規模下，自適應RoCoF參考控制器的FFR平均電能轉移指標(Energy Transfer Index, ETI)為17.37%，相對於傳統參考值為零的控制器為28.59%改善許多，突顯了其在高再生能源滲透電力系統中提升系統效能和可靠性的潛力。

Reduced system inertia in power systems with high renewable energy penetration poses significant challenges to frequency regulation. To address these issues, we propose a Fast Frequency Response (FFR) controller designed to maintain frequency within predefined thresholds while reducing operating costs associated with Battery Energy Storage Systems (BESSs). To ensure frequency security under credible contingencies, the control gain of the FFR controller is determined offline, based on the worst-case scenario. To avoid excessive FFR deployment during minor disturbances, we introduce an adaptive RoCoF reference that dynamically adjusts based on real-time estimations of power deficits and generator ramping capabilities. This approach tailors FFR needs to varying operational scenarios, including cascading failures, thereby effectively reducing FFR energy transfer and operating costs. Although the proposed method employs a simple mechanism that eliminate the need for frequency response estimator, its energy efficiency is comparable to controllers using contingency-specific critical control gains. Validation of the controller in a modified 9-bus test power system and a planned 2025 high-renewable Taiwan power system demonstrates its adaptability across various scenarios. Notably, across different contingencies and operating conditions, the average Energy Transfer Index (ETI) for the proposed controller is significantly lower than that of the zero reference controller, at 17.37% compared to 28.59%, respectively.