用於配電最佳電壓分佈之電容器配置及電壓預測

Capacitor Installation and Voltage Forecast for Distribution Optimal Voltage Profile

白義和(Yih-Her Bair)

指導教授 ： 高文秀 ； 曾國雄

國立臺北科技大學/電資學院/電機工程系所/碩士(2007年)

https://doi.org/10.6841/NTUT.2007.00374

摘要

電力系統無效電力與電壓之運轉是電力公司營運的重要課題之一，為使電力系統之各節點電壓維持在系統電壓控制之目標容許範圍內，並確保配電線路之損失成本最小化。
台灣地區電力系統電壓變動之容許幅度，在電業法規定電燈為±5%，電力為±10%，台灣地區電力系統配電電壓之容許變動幅度須維持在標準值±3%範圍內，配電電容器之配置地點及其容量大小，對配電系統之配電電壓品質極為重要，而台電公司目前僅採內部規定、配電技術準則及技術經驗等來推導估算，欲使配電電壓品質最佳化，尚有一段須努力之空間，故仍無法滿足用戶之期待。
本論文提出符合三相不平衡線路模型之電力潮流計算軟體分析、基因演算法、類神經網路及配合現有作業模式，組成一套新改善作業模式，藉由停限電管理系統搜尋負載特性資料與配電調度控制系統之電壓、電流、有效功率及無效功率，透過電力潮流計算及基因演算法，獲得電容器最佳配置位置、容量大小及操作時間，並求解線路各節點之最佳電壓值；應用倒傳遞類神經網路透過數據資料進行自我學習、平行處理等能力，使值班人員能隨時掌握配電線各節點電壓值之運轉情形，並適時作電壓之控制調整。新改善作業模式之電容器最佳配置及電壓預測等功能可納入配電管理系統，俾提升配電線供電品質及供電電壓穩定使配電線運轉最佳化。應用在台電公司配電線路之實際模擬結果，印證本論文建議之方
法，確實有效且屬實用。

關鍵字

電力潮流計算軟體 ； 基因演算法 ； 類神經網路 ； 停限電管理系統

並列摘要

The reactive power and the operation voltage of the power system are important to TPC’s Utilities operates. To maintain each node voltage of the distribution feeder as the system voltage which is also permitted in the range of the control goal, and to make the loss cost of the distribution line be minimized, lots of research recently probe about optimization.
The Taiwan electricity industry law stipulates the permission range of the voltage deviation in TPC power system which is ±5% to the electric light and ±10% to the electricity. The voltage profile is maintained within ±3% of nominal voltage for the distribution systems in TPC. Managing the voltage deviation of the power system in a minimum level is absolutely important. Nowadays, the TPC’s regulation and past experiences are involved to estimate the locations and sizes of the capacitor bank. But the voltage profile of distribution can't satisfy users' expectation; there is still a lot of work to do.
The voltage at each node can be obtained by applying Power Flow Calculation Program (PFCP). This paper used PFCP which is about the three phase unbalanced system model, Genetic Algorithm (GA), the Artificial Neural Network (ANN) and some present mode of operations to compose a new approach. We use Distribution Dispatch Control System (DDCS), searching line current, substation bus voltage, real power and reactive power, and Outage Management System (OMS), searching load characteristics. Then, by PFCP, we can get the node voltages. And then voltage profile improvement can be obtained from GA. So we’ll obtain the locations and sizes of the capacitor bank and switching time. The improvement result inputs to ANN, the short term node voltages which can be predicted. In order to improve the voltage profile of a distribution feeder, allocations of supervised capacitor bank with adequate energized time needs to be calculated. The power loss on distribution feeders can be minimized when the optimal locations and switching time of capacitor bank are obtained. Using the voltages predicted by ANN, incorporating with the Distribution Management System (DMS), the operation of capacitor bank can be automated. Thus, the power loss and voltage deviation of distribution feeders can be minimized. Application of the actual simulation result of the TPC’s distribution line confirms that the propose method of thesis do use effectively and practically.

並列關鍵字

Power Flow Calculation Program (PFCP) ； Genetic Algorithm (GA) ； Artificial Neural Network (ANN) ； Outage Management System (OMS) ； Distribution Management System (DMS) ； Taiwan Power Company (TPC)

參考文獻

[18] 蔡孟伸、劉清華，以JAVA為基礎的三相不平衡配電潮流，碩士論文，中原大學，中壢，民國92年。

[25] 洪穎怡、黃偉芬，以基因演算法進行多目標模糊參數之被動式濾波器規劃，碩士論文，中原大學，中壢，民國90年。

[26] 蔡孟伸、楊宏澤、許富淵，改良式基因演算法應用於配電系統重構問題之探討，碩士論文，中原大學，中壢，民國93年。

[1] B. V. Vibhute, H. P. Inamdar and S. A. Deokar, "Maximum Loss Reduction by Optimal Placement of Capacitors on a Distribution Systems," Power India Conference, April. 10-12, 2006, pp.692-694.

[2] C. T. Su and C. S. Lee, "Modified Differential Evolution Method for Capacitor Placement of Distribution Systems," Proceesings of IEEE Power Engineering Society Transmission and Distribution Conference, Vol. 1, Oct. 6-10, 2002, pp.208-213.