Abstract

This article presents a novel linear programming (LP) based two-stage stochastic approach for microgrids (MGs) under uncertainties. In this regard, the day-ahead programming of dispatchable resources in MG was modeled, considering the uncertainties in demand loads, and upstream network electricity price. Moreover, the inherent stochastic nature of wind and solar resources as well as the environmental aspects were modeled to find a realistic solution. Also, real-time pricing (RTP), demand response (DR) program considering the energy storage system (ESS) as a DR option was implemented on an MG as a smart customer. The extensive form of the two-stage stochastic recourse model was properly implemented for dispatchable and non-dispatchable resources. Furthermore, scenario generation and reduction procedures were realized with autoregressive and moving average (ARMA) model-based time-series (TS), and backward reduction (BR) method by the Kantorovich distance (KD), respectively. The simulations on a grid-connected MG, including micro-turbine (MT), fuel-cell (FC), wind-turbine (WT), photovoltaic module (PV), and ESS were reported in operational cases based on one month of real recorded data for wind speed, solar irradiance, demand load, and upstream network electricity price. The results for the next day in real-time confirmed the accuracy of the developed optimization methodology.

摘要

本文提出了一种基于线性规划（LP）的不确定性微电网（MG）的两阶段随机方法。在这方面，考虑到需求负荷和上游网络电价的不确定性，对MG中可调度资源的提前编程进行了建模。此外，对风能和太阳能的内在随机性以及环境因素进行了建模，以找到切合实际的解决方案。此外，作为智能客户的MG实施了将能源存储系统（ESS）作为DR选项的实时定价（RTP），需求响应（DR）程序。对于可调度和不可调度资源，已正确实施了两阶段随机追索模型的扩展形式。此外，场景生成和减少过程分别通过基于自回归和移动平均（ARMA）模型的时间序列（TS）和后向减少（BR）方法（通过Kantorovich距离（KD））实现。在运行案例中，基于一个月的实际运行情况，报告了对并网MG的仿真，其中包括微型涡轮机（MT），燃料电池（FC），风力涡轮机（WT），光伏模块（PV）和ESS记录有关风速，太阳辐照度，需求负荷和上游网络电价的数据。第二天的实时结果证实了所开发优化方法的准确性。根据一个月的实际记录数据（包括风速，太阳辐照度，需求负荷和上游网络电价），在运行情况下报告了光伏模块（PV）和ESS。第二天的实时结果证实了所开发优化方法的准确性。根据一个月的实际记录数据（包括风速，太阳辐照度，需求负荷和上游网络电价），在运行情况下报告了光伏模块（PV）和ESS。第二天的实时结果证实了所开发优化方法的准确性。