微電網之決策支援模型 - 以應用於台中工業區為例

Abstract

近年來，由於社會大眾對環境保護及永續的關注及對更穩定的能源供應系統的需求，結合了分佈式能源及再生能源之應用的微電網概念逐漸獲得重視。然而，截至目前為止，微電網所具有的潛在優勢卻尚未能在實際生活中展現出來，這與該系統本身的複雜性及不確定因素造成的進入門檻頗有關連。在過去，有許多種微電網模型陸續被提出以支援微電網規劃工作的最佳化；然而，強健最佳化理論及方法卻至今尚未被應用在微電網的模型建立上以因應消費者能源需求、燃料價格、電價及碳稅稅率等方面的不確定性。本研究提出一個混合型線性整數規劃(MILP)模型，並採用最壞情況下的成本(經濟性強健化方法之一)作為本模型多目標函數的最佳化目標之一，以便及早在微電網的早期規劃設計階段就導入強健最佳化的分析。本模型被設計為可同時進行微電網系統的成本期望值最小化及最壞情況成本的最小化，而最壞情況成本的最小化即是用來因應系統運作的各種不同情況之間的差異性。本模型藉由一套完整的數學公式模擬系統運作，從兼顧經濟、節能及環境保護等觀點為基礎，致力於為一規劃中的微電網系統提供可行的容量設計建議。此模擬的輸出結果尚包含產生一條描述成本期望值及最壞情況成本之間關係的帕累托曲線(Pareto curve)，該曲線可協助本模型的使用者判斷他們在微電網設計上的風險控管程度。在本研究中，此一模型被應用於台灣的台中工業區以驗證其作為微電網決策支援工具的適用性。

The microgrid concept, which encompasses the application of distributed energy resources and renewable energy, has gained arousing interest in recent years due to the increasing concern on environmental sustainability and demand for a more reliable power supply system from the civil society. So far, the pontential benefits of microgrids have not yet been exploited because of the entry barrier caused by the complexity and uncertainty within the microgrid system. A variety of microgrid models have been presented before to optimize the planning of microgrids; however, robust optimization has not been applied to the modeling of microgrids to deal with uncertainties in customer loads, fuel prices, electricity tariff rate, and carbon tax rate, etc. In this study, a mixed-integer linear programming (MILP) model is proposed to adopt an economic robust measure, worst-case cost, as one of the components in the multi-objective function to allow robust optimization of a planned microgrid as early as in the design stage. The model is designed to simultaneously address the issue of expected cost minimization and worst-case cost minimization, which helps in handling the variation among different scenarios of the system operation. With comprehensive mathematical formulation, the model aims at rendering capacity design recommendations for a microgrid from economic, energy-saving, and environmental perspectives. The results of the simulation include the formation of a Pareto curve between the expected cost and worst-case cost of the project, which enables the model users to judge the degree of their risk-taking on microgrid design. The application of the proposed model to Taichung Industrial Park in Taiwan demonstrates the applicability of the model as a decision support tool for microgrid planning.