臺灣博碩士論文加值系統

本論文將提出一個新的網路靈敏度係數計算推演法則. 這個新的推演方式
將納入最佳電力潮流的架構中並進行測試, 同時在本論文中一個新的線性
增量模型將在求解過程中加以陳述, 更進一步經由線性規劃技術來加以整
合. 而透過本論文所提之新模型, 將可調整電力系統中之控制變量來排除
系統中潛在或既存具危險性之運轉狀況, 並從而將一具危險性之不安全系
統帶入一個能符合系統需求之新安全運轉狀態.傳統上在利用以線性規劃
技術來解決最佳電力潮流規劃問題上已有諸多研究. 事實上傳統之快速解
耦直流模型已經在很多的能源管理系統中進行即時之線性最佳電力潮流規
劃運作. 無論如何, 快速解耦直流模型因係建立於多項假設前提之下, 而
使得應用上遭受諸多限制. 這些假設包括: (1)電抗值遠較大於電阻值.
(2)匯流排電壓大小固定在公稱電壓, 亦即1.0標么 (3)匯流排電壓相角差
不能太大, 一般而言在5度左右, 最大不得超過30度.很顯然地, 上述假設
不可能恆成立, 特別是當系統條件差時. 本論文即針對此一問題加以考
慮, 並發展一新模型來設法避免上述限制. 事實上根據本論文所提出之新
模型, 在不需任何假設前提即可完成整個理論架構之推演, 而僅需電抗值
遠較大於電阻值之假設即可進一步進行解耦化, 卻仍能保有相當之可靠性
及與傳統快速解耦直流模型相當之執行速度與記憶容量 , 此一新的理論
與模型更具有潛力發展為系統即時控制與調度之用.

This research will formulate a new algorithm for the network
senitivity calcultion. This algorithm will be incorporated in
the optimal power flow for testing. A new linear incremental
model will be formulated in the solving process and the linear
programming technique will be applied. There are already many
research works in the LP-OPF area. As a matter of fact, many
energy management systems are running the LP-OPF in real-time
with the fast decoupled DC(FD-DC) model. However, FD-DC model
needs assumptions on (1)negligible line conductances (2)small
angular differences (within a 30 degrees range) (3)constant
voltages (at their nominal values). The above assumptions are
not necessarily true, especially under severe conditions where
the system may undergo severe violations. This research will
try to develop a new model to avoid the above assumptions. The
new model will try to resolve the OPF problem with the
performance similar to the conventional fast decoupled DC
model. However, no assumptions will be needed. A decoupled
version of the new model will be also investigated. The state
variables, power-injections, and current-injections will be
analyzed to investigate the practicality of the new model. The
convergent characteristics, the robustness, storage and
performance will all be tested to see the feasibility of the
new model in handling the real-time energy management functions.