抽蓄電廠抽水調度最佳化之研究―動態規劃法之應用

Abstract

本文主要以台電抽蓄電廠運轉情形為探討對象。由於臺灣電力系統具有獨立系統運轉、電力需求尖離峰差距極大、水力發電比重低及備轉容量低等不同於其他國家電力系統之特色，所以台電電力系統運轉，如何維持可靠、良質之電力，實在是電力調度人員一大課題。因為抽蓄電廠擁有水力發電機組快速起停、發電與抽水運轉快速切換、協助調整系統頻率、系統轉備容量等功能，所以抽蓄電廠在台電電力系統運轉扮演相當重要的角色。電力調度人員在面對電力系統不同調度情境、電源供給變化、電網限制等各種變數情形下，以電力系統安全為前提，如何安排這些變數又兼具考量經濟調度，使抽蓄機組抽水成本，儘可能符合經濟利益，是電力調度人員每日需面對的問題。 本文針對台電電力系統用電負載預測與特性、電源供給預定、抽蓄電廠機組台數、抽蓄電廠上下池容量、發電機組燃料成本表，建構一抽蓄電廠每日抽水量動態規劃決策程式，然後每日依電力系統不同變數決定抽水量及發電機組成本函數，作為執行動態規劃程式輸入資料，求出每日最小抽水成本下之抽水機組使用台數及抽水時機。 由本文實證與結論分析可知，模擬程式求解結果比實際運轉結果可獲得較佳經濟運轉，且求解結果實際可行，因此本研究作為控制管理機制之工具非常適合。

This paper mainly uses operating situation of TaiPower Pumped Storage Plant as object of study. As the power systems in Taiwan have the characteristics of independent system operation, large differences between peak and off-peak power demand, low percentage in hydraulic power generation, and low standby capacity that are different from the power systems of other countries, therefore how to maintain steady, reliable, and good quality electricity is indeed a big issue to people dispatching the power. As pumped storage power plant has the functions of quickly starting and stopping the hydraulic generator, quickly switching between electricity generation and pumping, assisting in adjusting system frequency, and spinning reserve, the pumped storage power plant plays a very important role in the operation of TaiPower system. Electricity dispatch personnel in facing various variables such as the different dispatch situation, power supply variations, and electric network limitations under the pretext of power system safety, how to arrange these variables at the same time when considering economic dispatch so that the cost of operating pumped storage power plant can meet the economical benefit is the problem that power dispatch personnel must face. This paper uses Taipower system power usage loading forecast and characteristics, power supply reservation, number of pumped storage power generators, capacity of upper and lower reservoir for the pumped storage power plant, and table of cost of generator fuel cost to construct a pumped storage daily pumping dynamic programming decision program. Then, using daily system different variables to determine pumping volume and generator cost function as input information to execute dynamic programming program and find out under daily minimum pumping cost the number of pumping machine needed and the timing of pumping. From the empirical data and conclusion we can find out that the answer to the simulation program can obtain better economic operation than actual operation results. And it is possible to implement the result. Therefore this research as tool to control management mechanism is very appropriate.