Vulnerability Assessement And Differentiated Planning On Complex Power Grid To Prevent Blackouts

As an important part and basic guarantee of energy supply security,electric power security is one of the key scientific problems of the basic research for national major strategic needs.The large scale interconnection of power grids and the development of permeability of stochastic volatile new energy bring great social and economic benefits along with more complexity of system topology and dynamic performance,which resulted in new challenges of preventing and reducing the risk of blackouts.Meanwhile,the future power systems are required to provide more secure,reliable,high quality and self-healing power supply service,as well as lower blackout risk.Due to the new challenges and requirement,,it is necessary to research on power grid vulnerability of complex power grids and propose a strong backbone grid differentiated planning method to prevent blackouts.Blackout events are mainly caused by the following three aspects:some vulnerable sources triggered resulting in abnormal operation state,cascading failure process from cause to effect of vulnerable sources,and vulnerabilities of physical structure in the power grid itself.Therefore,supported by natural science foundation of China,the vulnerability and method on differentiated planning of the power grid considering the causes,processes and physical nature resulted in blackouts are studied.The main work and of contributions of the thesis are summarized as follows.Firstly,to find out the vulnerable sources of the blackout accidents,the risk entropy assessment method on the vulnerable sources of large blackout event is proposed based on time serial of blackout accidents,which is combined with the pressure and release of the PAR risk model and information Entropy model.The resilience of power system is considered,which reflects that the system risk is relatively low when it has high self-healing recovery performance.The detection degree of interference is also considered,which reflects that the interference with poor detectability often leads to larger risk.And through the entropy weight enhancement coefficient,the model overcomes the shortcoming that small probability events resulting from probabilistic direct multiplication are easily overlooked.Furthermore,a power law distribution parameter estimation method based on logarithmic partition-rank sum is proposed.By applying the risk entropy method,taking time serial of blackouts of UCTE from year of 2002 to 2015 as example,it is found that the main three types of vulnerable sources,in turn,are the other and unknown reasons?RE₆?,extreme weather,natural disaster?RE₅?and failure of protective equipment or other components?RE₃?,which can provide a reference for system planning,design and decision making.Secondly,to explore the vulnerable regions of blackout incidents spatial distribution,the spatial-temporal evolution model considering the regional differences is proposed,and the fractal dimension characteristic interconnection of power grids is discovered.The differences of load growth,line reconstruction,geography and energy distribution,topology change caused by planning and management in different region are considered in the regional spatial-temporal evolution model.The method on similarity of time serial of blackouts between the large grid and its sub-grids is described.From the model and the actual grid,the conclusion on relationship between the fractal dimension and the evolution of the grid can be summarized to get the vulnerable regions of the power grid.This demonstrates that the similarity in the fractal dimension of various sub-grids and between large power grid and its regional sub-grids reflects the similarity in topology development and evolution,which provide theory basis for complex power system hierarchical structure and self-similar model.Thirdly,to identify the vulnerable lines that cause the blackouts,two kinds of electric PageRank?E-PageRank?algorithms based on the power gird dual model are proposed:TE-PageRank algorithm based on topology dual model and FE-PageRank algorithm considering the interaction of line faults.And then the importance of lines can be assessed by using their PR values.The topology dual model of the power grid is constructed by using the nodes to represent the branches of power grid and using the lines to represent the relationship among the branches.Considering the line voltage level,the transmission power flow and the contribution of the power supply to the load nodes,TE-PageRank algorithm is proposed to identify the important lines in charge of energy transmission.Based on the frequency relationship between interactions of branches after failure,the grid interaction dual model is proposed.Considering the pre-fault load and the line failure rate,the FE-PageRank algorithm for the grid interaction model is proposed to identify the vulnerable branches that cause the cascading failures.The proposed dual model can extend the existing methods from nodes to branches.Compared with the literature results,E-PageRank algorithms are effective,the convergence speed is fast and the recognition accuracy is not sensitive to parameters.Fourthly,to construct the backbone network of transmission grid,a model based on important nodes and branches for the differentiated planning of grid backbone is proposed.The important components in power grid for the differentiated planning are evaluated,which include the key nodes with classification characteristics and the key branches with dual model and electrical PR algorithm.The purpose of the core backbone differentiated planning model is to select components with minimum econometric cost with maximum sum of the PR values of the branches and nodes in the backbone network.The binary particle swarm algorithm is used to solve the model.The results show that the model can not only cover the key nodes and branches with high PR value,but also satisfy the economy and power balance of power grid operation.Fifth,to improve the power supply capacity of important loads in disaster,the procedure,model and method of differentiated planning in distribution system are proposed.The specific procedure of differentiated planning in distribution system includes identification and prediction of special loads,analysis of current status,measures of power supply for special loads,and high-standard power grid planning.Then the model is established to select the subsystem components with minimum cost increment over general planning considering disaster scenario and general scenario.To solve the constrained optimization problem,genetic algorithm combined with minimum spanning tree?MST?and improved radial topology maintaining method is adopted.The case study and analysis show that the proposed model and method can effectively achieve disaster reduction benefits.Focusing on preventing the blackouts,the vulnerable sources of blackout incidents,temporal-spatial evolution model considering the regional characteristics,the fractal dimension characteristic research method which can get the vulnerable regions and the identification of vulnerable branches,and the differentiated planning model and method are proposed.The research shows that grasping the major vulnerable sources that cause blackouts,utilizing the evolutionary similarity among the regional power grids,focusing on the vulnerable regions,grasping the key elements of the topology and realizing the differentiated planning of key components are important to improve the vulnerability of the power grid and prevent the blackouts.This can provide reference for improving the safety of the power grid and ensuring the national energy security.