Research On The Identification Method Of Vulnerable Parts Of Power Grid Considering Attack Risk

Power system is the important infrastructure which plays an essential role in ensuring the proper functioning of the national economy and social stability and development. The modern power system has developed into a complex coupled network system composed of physical power system and information communication system, which makes the vulnerable parts of the power grid becoming more complex and the components are more vulnerable to attack.Power grid not only needs to ensure its reliable operation in routine environment, but also has to maintain its normal operation instead of widespread collapse. Presently the risk of large power outages caused by malicious attack is receiving more and more attention. Traditional power grid vulnerability research focuses on individual dynamic characteristics of components and evaluates the vulnerability of power grid triggered by its self-factors and natural disasters. In this thesis, the risk of malicious attacks against power grid is considered, and the identification method of the vulnerable parts of power grid which may cause cascading failure is studied. The researching work which has been done is presented as follows.In this thesis the existing research on the power grid vulnerability is summarized and component vulnerability index based static N-k vulnerable parts identification strategy is obtained. Dynamic N-k vulnerable parts identification strategy is proposed based on the static N-k vulnerable parts identification strategy considering the effects of component interaction and the effect of component removal on network structure. The simulation results verified the effectiveness of the dynamic N-k vulnerable parts identification strategy.Considering the diversity and complexity of the N-k vulnerable parts of power grid, an N-k vulnerable parts identification algorithm is proposed based on the space reduced traversal algorithm for the same / multiple kind of components. Space reduction traversal algorithm is based on the results of the lower order N-2 traversal search space to reduce the effective reduction of the computation of the traversal algorithm. The space reduced traversal algorithm based same / multiple kind of N-k vulnerable parts identification algorithm can comprehensively identify N-k vulnerable parts, and the vulnerable parts which cannot be identified by the traditional identification method also can be recognized by the space reduced traversal algorithm. The space reduced traversal algorithm based multiple N-k vulnerable parts identification method provides a new train of thought for identifying vulnerable parts composed of multiple kinds of components. The simulation results verified effectiveness of the algorithm.To overcome the problem of extensive searching space and low search efficiency when conducting higher order N-k analysis in a large-scale power grid, the space conversion reduction algorithm is proposed, furthermore, the binary coding PSO algorithm is improved and the virtual distance coding PSO is proposed. The search space of space conversion reduction strategy is transformed and reduced based on the N-k sampling results, which makes the search space of N-k vulnerable parts significantly reduced and the distribution of vulnerable parts more concentrated. VDPSO algorithm further decreases the search space and improves the search efficiency. The simulation results show that compared with the random chemistry algorithm, the space conversion reduction strategy based virtual distance coding PSO further improves the search efficiency of the N-k vulnerable parts of power grid, which is especially suitable for the higher order N-k vulnerable parts identification in a large-scale power grid.