以集合涵蓋模式為基礎之電力需量反應用戶群代表決策問題

Abstract

近年來，台灣用電需求逐年上升，而新建電廠將對環境造成大量衝擊。對此，台灣電力公司迫切面臨平衡電力供需之壓力。需量反應(Demand Response, DR)係需求面管理之一環，電力公司透過價格機制或提供誘因等方式，引導用戶改變用電行為，解決電力負載危機。本研究以電力需量反應用戶群為決策者建立整數規劃模型以最佳化合約用電戶之抑電組合。在最小化總合約支付價格之情況下達到台灣電力公司之電力抑低目標，並考慮各合約用電戶之抑電限制。本研究亦針對問題設計演算法，首先透過集合涵蓋(Set Covering)模式之建立，以放鬆模型之複雜度，再來，引進對偶理論，在演算法的迭代過程中，根據對偶值作為評估抑電方案之依據，在維持求解品質的情況下，找出優勢方案以減少方案生成。數值測試中，此演算法能在兩分鐘內，達到最多百分之五之求解誤差。

For years, electrical power companies (e.g., Taiwan Power Company) have been facing the challenge of meeting the ever-increasing and fluctuating demand, given the fact that environmental issues have been preventing the quick and low-cost construction of power plants. Demand Response (DR) programs were recently thoughts as an effective mechanism to better match the demand and the supply of electricity. This study develops an integer programming (IP) model to formulate the decision problem of the DR aggregator, who helps the electrical power companies by adjusting the loads of its represented contract users. The objective of the IP model is to minimize the overall payment to the contract users, while meeting the demand response (load reduction) specified by the electrical power company. In particular, the operational limitations for the contract users to perform load reduction must be taken into account. This study also proposes a set-covering problem (SCP) based solution algorithm to solve the decision problem effectively. The iterative algorithm derives the solution by adding the new alternatives of reduction based on the dual price information from the linearly relaxed problem. In the numerical experiment designed based on the practical data from Taiwan Power Company, the algorithm can achieve the solution quality within a gap at most 5%, given a computation time of two minutes.