Time filter

Source Type

Lin J.,Tianjin University | Sun L.,Tianjin University | Pu T.,China Electric Power Research Institute | Yu T.,China Electric Power Research Institute | And 5 more authors.
Zhongguo Dianji Gongcheng Xuebao/Proceedings of the Chinese Society of Electrical Engineering | Year: 2014

Complete splitting model of the controlled partitioning strategy is a large-scale nonlinear mixed integer programming problem and mostly solved by approximate decomposition algorithms. Only the approximate solution can be gained. To solve this problem, a new alternate optimization method based on master-slave problem is proposed to search for optimal splitting strategy. The complete model was converted into master-slave problems based on connected graph constrained knapsack problem (CGKP). The master problem is graph partitioning problem, which is solved by CGKP algorithm, and the slave problem is direct current optimal power flow (OPF). The coupling between master problem and slave problem is achieved through load adjustment. A better splitting strategy, which is closer to the optimal solution of the complete model, can be obtained through the alternating iteration between the master problem and the salve problem. The results of the examples show that the proposed method can obtain better splitting strategy with less shed load than other approximate algorithms, which verifies the feasibility and effectiveness of the proposed method. © 2014 Chin. Soc. for Elec. Eng.


Lin J.,Tianjin University | Wang X.,Cooperation Technology | Li S.,Tianjin University | Wu P.,Tianjin University | And 3 more authors.
Zhongguo Dianji Gongcheng Xuebao/Proceedings of the Chinese Society of Electrical Engineering | Year: 2012

Graph partition technology is significant in transportation, power grid fault diagnosis, power system partition, etc. In this paper, a new graph partition problem that is connected graph constrained knapsack problem (CGKP) was first built, and an efficient approximation algorithm was designed to solve it in the paper. Four node sets relating to graph connectivity were defined, the characteristics and searching methods of which were proved and provided in the form of propositions. On the basis of that, the algorithm of graph constrained knapsack problem (GKP) was extended, and the approximation algorithm for CGKP was achieved. The computation complexity of the algorithm was also thoroughly discussed. The test result of the sample demonstrates the validity of the proposed method. Since the optimal controlled partition of power system could be treated as a CGKP at some condition, all the above new node sets, propositions and algorithms for CGKP provide theory foundation for solving the optimal controlled partition of power system. © 2012 Chin. Soc. for Elec. Eng.


Lin J.,Tianjin University | Li S.,Tianjin University | Wu P.,Tianjin University | Wang X.,Cooperation Technology | And 3 more authors.
Zhongguo Dianji Gongcheng Xuebao/Proceedings of the Chinese Society of Electrical Engineering | Year: 2012

Power system partitioning is to split interconnected grid into two or more stable subsystems when the structure of power systems can not be hold. Based on the related graph theory notions and the practical physical process, the complete model for optimal controlled partitioning of power systems which synthetically considers all kinds of security constraints was constructed, and a two-stage strategy-"searching and regulation" based on connected graph constrained knapsack problem (CGKP) was designed to solve this model. In order to reduce the computation complexity, the model was decomposed into an optimal balance graph partition problem and a regulation problem based on optimal power flow, both of which were solved successively in two stages. The optimal graph balance partition problem was further decomposed into multiple CGKPs, which were solved by the graph partitioning method based on CGKP. The algorithm proposed owns the advantages of strong theoretical foundation and low computational complexity. Results of samples demonstrate that the model and algorithm proposed are valid. © 2012 Chinese Society for Electrical Engineering.

Loading Northeast Electrical Power Dispatching Center collaborators
Loading Northeast Electrical Power Dispatching Center collaborators