学术报告第201552期

题    目：Power System  Optimization 

主  讲 人：Dr.  Qiu, Feng（Argonne  National Laboratory）

时    间：2015年12月28,  上午  10:00

地    点：南一楼中311

Abstract:

      Decision problems in power system operations are becoming more and  more challenging  to solve as the heterogenous infrastructures and renewable penetration continue  to grow. In this talk we will study how deterministic and stochastic  optimization techniques are exploited to solve optimization problems in power  systems. The first topic covered is power System Restoration with  Integrated Sectionalization  and Generator Start-up Sequencing. The  restoration process of the bulk power system after a partial or complete  blackout relies on generating units with black- start capabilities. In the  normal build-up restoration process, the system is sectionalized first into a  set of subsystems in which the generators are started afterwards. Due to the  complexity of the restoration process, the sectionalization and generator  sequencing problems have been studied separately in the literature, ignoring  their interlocking connections, which may lead to a sub-optimal restoration plan  for the overall restoration process. In this work, we integrate the two problems  into a single model that minimizes the restoration duration for the overall  system. We propose an integer programming formulation to model the  sectionalization problem as a graph partition problem with connectivity  constraints. A continuous- time representation of the generator start-up  sequencing problem is used based on a semi-infinite programming formulation. Our  case study shows that the proposed model can achieve a global optimization  solution effectively. The  second topic covered is distributionally Robust Congestion Management with  Dynamic Line Ratings. Dynamic  line rating based on real time meteorological data has been shown to be useful  in transmission line capacity management. Based on a binary rating forecast, we  propose a distributionally robust congestion management model that selectively  uses dynamic ratings on critical lines and keep the risk of thermal overloading  below a prescribed level. A case study illustrates that the proposed model can  effectively alleviate transmission congestion with a low error rate. 

In  his talk, he will also introduce some information about Argonne National  Laboratory, which is the biggest research center of Department of Energy of  US.

Qiu  Feng, is a computational engineer in Argonne National Laboratory. He received  his PhD in industrial and systems engineering in Georgia Institute of  Technology, and MS and BS in Automation in Huazhong University of Science and  Technology. Now his main research interest covers stochastic programming,  mixed-integer programming, and their applications in energy systems. He is  Member of IEEE, IIE, INFORMS, Optimization Society, Computing Society,  Mathematical Optimization Society.