Optimization Method For Power System Restoraion Path Following A Blackout

Although the reliable and secure power supply has to be enhanced in modern power system,the possibility of a large scale of blackout still exists,such as the United States and Canada blackout on 14th Aug and the Brazil blackout on 4th Feb.When a blackout happens,it's essential to develop an effective system restoration scheme to restore the power supply as soon as possible and minimize the adverse impacts.On the stage of network reconfiguration,in order to quickly establish a stable skeleton-network as a basis for subsequent load recovery,it is necessary to optimize the restoration path.The power grid network can be mapped into a topology model,and the operation status of transmission line is set as decision variable.Optimizing the power system restoration path is determining a reasonable network reconfiguration to speed up the restoration process,when the constraint condition can be satisfied.As the optimization is a complex nonlinear programming problem,artificial intelligent algorithms are widely employed to solve this problem.However,because the dimension of restoration path optimization is very high especially for large scale systems,the computation speed of the existing artificial intelligent algorithms is very slow and need to be improved.Considering the randomness of intelligent algorithm with line state encoding,the new individual generated in the process not scatter uniformly over the feasible solution space and most of it is non-connective,this paper improves on it firstly.Furthermore,considering a restriction factor for speeding up the calculating speed is the nonlinearity of the optimization problem.This paper attempts to improve the computation efficiency from the angle of optimization model.Therefore,this paper represents the improvement on the intelligent algorithms and the optimization model for existing method of optimizing power system restoration path,and main achievements are as follows:1.Because of the high dimension of restoration path optimization especially for large scale system,the artificial intelligent algorithms is easy to be trapped in the local optima,which has not been noticed in current research.Therefore,an orthogonal genetic algorithm is employed to generate the initial and offspring population in this paper,which scatters the individuals of initial and offspring populations uniformly over the feasible solution space,so that the algorithm can explore the whole solution space evenly.The solution quality and convergence speed of intelligent algorithms have been improved without changing the optimization process.2.Most new population generated in the process is non-connective because of the randomness of intelligent algorithms.Considering the characteristics that number of operation lines with state“1”and partial connection sections for non-connective individuals generated in iterative process is few,a rectification method based on agglomerative hierarchical clustering algorithm and prim algorithm was proposed to further enhance correcting speed for non-connective individuals.Simulation results demonstrate that the non-connective individuals are corrected more rapidly,in premise of accuracy assurance.3.In order to further improve the optimization efficiency,the problem of optimizing power system restoration path is formulated as a minimum cost maximum flow modelling.The system topology is constructed as a single-source and single-sink network,and maximizing the network flow is used to ensure the connectivity of restoration path.Then the best set of energizing transmission line can be found by minimizing the cost of flow from source to sink.Finally a typical heuristic algorithm named shortest augmenting path method is adopted to solve this problem.Simulation results demonstrate that the proposed method is more efficient than traditional method.4.The optimization of power system restoration paths cannot be solved by the advanced mathematical programming method due to the situation that the connectivity constraint in existing research has not been analyzed.In this paper,the blackout system is translated into a single-source multiple-sinks network.Then bus operation state and amount of network flow in transmission path are set as decision variable,and the flow conservation and capacity constraints for each bus together constitute the analytic expressions of connectivity constraint,according to the fundamental principle that the path between nodes of flow injection and arrival is connective.On this basis,the mixed integer linear programming model for optimization of restoration paths is built through the analytic expression of connectivity constraint,and the optimal restoration path is solved by CPLEX.Simulation results demonstrate that the proposed method is highly efficient.