最佳化技術於全黑啟動決策支援系統與二次電壓控制之應用

Abstract

本論文旨在探討最佳化技術於全黑啟動(Black Start)規劃與二次電壓控制 (Secondary Voltage Control, SVC)之應用。在全黑啟動規劃方面，本論文提出一全黑啟動決策支援系統(Black Start Decision Supporting System, BSS)以輸出最佳全黑啟動路徑與加壓模式，其目標函數包含輸電線物理特性與加壓過程模擬結果之考量以期提升全黑啟動之效率與安全性；最佳化技術則結合蟻拓最佳化(Ant Colony Optimization, ACO)與圖論(Graph Theory)之演算法以具備適應性(Adaptivity)與可觀測性(Observability)。在二次電壓控制方面，本論文提出一適應性二次電壓控制系統(Adaptive Secondary Voltage Control System, ASVCS)以校正系統電壓與維持系統電壓穩定度(Voltage Stability)，其數學模型以快速解耦合負載潮流(Fast Decoupled Load flow)之方程式為基礎，並考量在系統限制條件下之電壓變動最小化與系統穩定性維持，經轉換構成一組大型線性規劃(Linear Programming, LP)問題；最佳化運算包含運用導引匯流排(Pilot Bus)同步相量量測器(Phasor Measurement Units, PMUs)之量測資訊以估測系統狀態以及計算控制訊號。此系統結合凸規劃(Convex Programming)開源(Open Source)求解工具以及自行開發之數學模型轉換與獨立求解程式以求解最佳控制訊號作為調配發電機輸出虛功之依據。此外，針對導引匯流排最佳設置地點之選擇，本論文亦提出一有效之規畫策略以增進二次電壓控制之效果。透過應用於台灣電網與IEEE標準系統之測試驗證，本論文所提出之全黑啟動決策支援系統與適應性二次電壓控制系統之最佳化技術確實具有合理性且可行性。

The dissertation focuses on the applications of optimization techniques in the black start planning and the secondary voltage control (SVC), respectively. For black start planning, a black start decision supporting system (BSS) has been developed in this study to generate optimal black start energizing paths and modes. The proposed objective function aims at enhancing the security and efficiency of black start, while taking into account the physical characteristics of transmission lines and the simulation results. The optimization approach integrates Graph Theory and Ant Colony optimization (ACO) to meet the requirements of adaptivity and observability. As for SVC, an adaptive secondary voltage control system (ASVCS) has been developed in this study for voltage regulation and voltage stability maintenance. The mathematical model of ASVCS is constructed based on the decoupled load flow equation and transformed into a set of large-scale linear programming (LP) problems considering the minimization of voltage variation and the preservation of system stableness under the system constraints. The optimization process includes the system state estimation using the measured data of the synchronized phasor measurement units (PMUs) at pilot buses and the computation of control actions, which utilizes a self-developed program for mathematical model transformation and a standalone program of interior-point methods modified from open source convex programming solvers. In addition, to enhance the performance of ASVCS, an effective strategy for the selection of pilot buses is also proposed in this dissertation. Through the application tests on Taiwan Power System (TPS) and IEEE standard systems, the validity and feasibility of the optimization approaches of BSS and ASVCS have been both justified.