The Analyses And Solutions Of The Power System Controlled Separated Strategies

Considering the.demand for electricity increases rapidly the.development of the society and the power supply companies unceasingly pursue their economic benefits:the power system is approaching its extremity for stable operation.When encounteriing mutual oscillation of system units caused by a major fault:stability control devices are required to remove some load so as to restore the svnchronization.Once the above means failed.the system has to be splitting to stop the disturbance from influencing the whole grid and causing widespread blackout.In traditional out-of-step splitting:splitting device is adjusted off-line according to expected scenes,and then operated on-line based on local measurement and criteria.This method has some weakness.For example,the adjustment is complex,it’s diffcult for devices to cooperate,and expected scenes are very different from practical conditions.Now the communication and monitoring means of power system are constantly improved:people tend to study online active splitting control which has the characteristics of centralization and coordination.The paper put forward a soluion scheme of "searching and verifying" based on the power flow tracing.The solution combined the graph theory with characteristics of the power system:and realized quick searching of splitting surface through initial partition and optimization.Finding the optimal splitting surface is changed to finding the feasible one:thus reducing the complexity of the problem and increasing the efficiency of finding splitting surface.In the searching phase:initial strategy space is narrowed down to sub space with smaller scale through basic constraints that are easy to describe(usually synchronization constraint and power balance constraint).In the verifying phase:enumeeration method is used to verify all the strategies in subspace one by one.If the disconnected island can satisfy other basic constraints of stable power system by the candidate strategy,then we can detemine the proper splitting point in energency through verifying the degree of unbalance.If the system is very complicated:it is hard to divide a large number of nodes The workload will be tremendous if only the power flow tracing method and enumeration methods are used.So:this piper introduced the greedy algorithm and genetic algorithm and anticipated to select the best algorithm for splitting through comparative study on these two algorithms.Based on the simplified topological graph of power flow tracing method firstly,it initialized the node commumity,searched the adjacent groups of each group,and based on the principle of assuring the minimum imbalance of all groups after incorporation:selected two groups each time to incorporate and formed a new group,confimed the locally optimum solution:finished the one-step search and continued to the next round of incorporation after finishing previous round,till all nodes had been ivided and formed a given number of islands.Through actual operation,the splitting results of greedy algorithm were found to be not ideal and the running time of program was long In the genetic algorithm,firstly,it made high-dimension mapping on the complicated power grid topology by the embedded mapping method with double-layer crossed counting to avoid the two-dimensional intersection,thus solving the problem that the cross wire which shall not be cut was cut by accident.Then,based on simplification of the complicated topology by power flow tracing method,the optimum cutting plane was found out by using the genetic algorithm,namely the minimum cut of the complicated topology was found out by using the genetic algorithm.According to the results,it was relatively reasonable to use this division method,the minimum cut finally acquired and the imbalanced power of each island were both smaller In the meanwhile,the whole calculation process of genetic algorithm was quicker and could solve the large calculated amount in the emumeration method and traditional algorithms.