Energy and reserve scheduling under ambiguity on renewable probability distribution
Energy and reserve scheduling under ambiguity on renewable probability distribution
复制标题
DOI：
10.1016/j.epsr.2018.01.024
复制
发表时间：
2018-07
影响因子：
3.9
通讯作者：
Alexandre Moreira;Bruno Fanzeres;G. Strbac
中科院分区：
文献类型：
--
作者：
Alexandre Moreira;Bruno Fanzeres;G. Strbac
本文提出了一种新的方法来设计在可再生能源和设备停电的不确定情况下最小成本的能源和储备调度。可再生能源生产中的不确定性由外部模拟情景来解释，这是随机规划中的惯例，而发电机和/或输电线路的停电则通过可调整的稳健优化来解决。通过唯一的概率分布精确地表征RE输出是一项具有挑战性的任务。因此，我们提供了一个允许考虑一组“可信的”概率分布的一般公式。以这种方式，系统操作员对可再生生产中的不确定性的模糊厌恶被考虑在内。我们提出的方法通过一个三层模型来确定最低成本的能源和备用调度。在结构上，上层定义了成本最低的调度，而在可再生生产的不确定性下，中间层确定了给定运行点的最坏应急。然后，较低级别利用由较高级别提供的调度来确定最佳重新调度。为了控制系统的均衡，采用风险约束技术来处理系统的不平衡不确定性，确保可靠的运行水平。为了解决多层次问题，我们提出了一种结合Bders分解和列约束生成技术的算法，在调度功率和储备的同时逼近风险度量。以英国电网的实际数据为例，验证了该模型的有效性和考虑模糊性的重要性。
This paper presents a novel methodology to devise a least-cost energy and reserve scheduling under uncertainty in renewable energy sources (RES) and equipment outages. The uncertainty in renewable production is accounted for by exogenously simulated scenarios, as customary in stochastic programming, whereas outages of generators and/or transmission lines are addressed via adjustable robust optimization. The precise characterization of the RES output by means of a unique probability distribution is a challenging task. Hence, we provide a general formulation that allows the consideration of a set of “credible” probability distributions. In this manner, the system operator's ambiguity aversion to uncertainty in renewable production is accounted for. Our proposed methodology determines the least-cost energy and reserve scheduling through a three-level model. Structurally, the upper level defines a least-cost scheduling and, under uncertainty in renewable production, the middle level identifies the worst contingency for the given operating point. The lower level then utilizes the scheduling provided by the upper-level to determine the best redispatch. In order to control the system equilibrium, we adapt risk constraint techniques to handle the system imbalance uncertainty and ensure a reliable operating level. To solve the multi-level problem, we propose an algorithm that combines Benders decomposition and column-and-constraint generation techniques to approximate the risk measure while scheduling power and reserves. The effectiveness of the proposed model and the importance of considering ambiguity are demonstrated through a case study with real data from the Great Britain power system network.