Towards Scalable Voltage Control in Smart Grid: A Submodular Optimization Approach

Towards Scalable Voltage Control in Smart Grid: A Submodular Optimization Approach
复制标题

DOI:
10.1109/iccps.2016.7479120
复制
发表时间:
2016-04
期刊:
2016 ACM/IEEE 7th International Conference on Cyber-Physical Systems (ICCPS)
影响因子:
--
通讯作者:
Zhipeng Liu;Andrew Clark;Phillip Lee;L. Bushnell;D. Kirschen;R. Poovendran
Zhipeng Liu;Andrew Clark;Phillip Lee;L. Bushnell;D. Kirschen;R. Poovendran
中科院分区:
其他
文献类型:
--
作者:
Zhipeng Liu;Andrew Clark;Phillip Lee;L. Bushnell;D. Kirschen;R. Poovendran

文献摘要

相似文献

当电力系统不能满足一个或多个母线的无功功率需求时,就会发生电压不稳定。电压不稳定事件已经导致了几次重大停电,并且由于能源需求的增加,电压不稳定事件有可能变得更加频繁。未来的智能电网可以通过快速检测可能的电压不稳定并实施纠正措施来帮助确保电压稳定性。只有以及时、可扩展的方式选择这些纠正措施,才能有效地恢复稳定。然而,用于选择控制动作的当前技术依赖于穷举搜索,并且因此可能选择低效的控制策略。在本文中,我们提出了一个子模块优化方法来设计一个控制策略,以防止电压不稳定在一个或多个总线。我们的关键见解是,与期望电压的偏差是所采用的无功功率注入集合的超模块函数,从而导致具有可证明的最优性保证的计算高效的控制算法。此外,我们表明,我们的方法的最优性界可以提高从1/3到1/2时,电力系统在重负荷条件下运行。我们证明了我们的框架,通过广泛的模拟研究IEEE 30总线测试用例。
Voltage instability occurs when a power system is unable to meet reactive power demand at one or more buses. Voltage instability events have caused several major out- ages and promise to become more frequent due to in- creasing energy demand. The future smart grid may help to ensure voltage stability by enabling rapid detection of possible voltage instability and implementation of corrective action. These corrective actions will only be effective in restoring stability if they are chosen in a timely, scalable manner. Current techniques for select- ing control actions, however, rely on exhaustive search, and hence may choose an inefficient control strategy. In this paper, we propose a submodular optimization approach to designing a control strategy to prevent volt- age instability at one or more buses. Our key insight is that the deviation from the desired voltage is a super- modular function of the set of reactive power injections that are employed, leading to computationally efficient control algorithms with provable optimality guarantees. Furthermore, we show that the optimality bound of our approach can be improved from 1/3 to 1/2 when the power system operates under heavy loading conditions. We demonstrate our framework through extensive simulation study on the IEEE 30 bus test case.