Distributed Optimization-Based Hourly Coordination for V2G and G2V
Distributed Optimization-Based Hourly Coordination for V2G and G2V
复制标题
DOI：
10.1109/tpec.2019.8662148
复制
发表时间：
2019-02
期刊：
影响因子：
--
通讯作者：
F. Safdarian;Logan Lamonte;A. Kargarian;M. Farasat
中科院分区：
文献类型：
--
作者：
F. Safdarian;Logan Lamonte;A. Kargarian;M. Farasat
网络约束经济调度（NCED）问题是一个规模大、复杂且计算成本高的问题，它考虑了电动汽车（EV）的随机移动和每辆电动汽车充放电周期对应的附加变量。为了减少与该优化问题相关的计算负担，引入了分布式优化。由于电动汽车随机地从一辆公交车移动到另一辆公交车，本文提出了一种时间分解方法，而不是地理分解方法来划分坡道约束的NCED。为了利用并行计算的优势，减少求解时间，在调度范围内考虑了数千辆电动汽车。每个子层次都有一个斜坡约束的NCED，子问题之间的联系被建模为共享变量/约束。为了协调各子问题并找出整个运行视界的最优解，提出了分布式辅助问题原理（APP）。在此基础上，提出了一种有效的初始化策略，提高了算法的收敛速度。采用该方法在一个6总线和IEEE 118总线测试系统上解决了一周前的NCED问题。结果与集中式方法的结果进行了比较，验证了该方法在缩短求解时间方面的有效性。
Network-constrained economic dispatch (NCED) problem, which takes into account the random mobility of electric vehicles (EV) and additional variables corresponding to each EV’s charge/discharge cycles, is large scale, complex and computationally expensive. To reduce the computational burden associated with this optimization problem, distributed optimization is introduced. Since EVs randomly move from one bus to another bus, this paper proposes a temporal, rather than a geographical, decomposition approach to divide a ramp-constrained NCED. Thousands of EVs are considered over the scheduling horizon in order to take advantage of parallel computing and achieve reduced solution time. A ramp-constrained NCED is formulated for each sub-horizon while the connections between subproblems are modeled as shared variables/constraints. In order to coordinate the subproblems and find the optimal solution for the entire operation horizon, distributed auxiliary problem principle (APP) is proposed. Further, an efficient initialization strategy is presented to enhance the convergence time of the solution algorithm. The proposed method is employed to solve a week-ahead NCED on a 6-bus and IEEE 118-bus test systems. The results are compared with those of a centralized approach and effectiveness of the proposed method in reducing the solution time is verified.