| With the deepening of the reform,China’s social and economic are showing a flourishing scene.People’s living standards and happiness index have been continuously improved,and energy demand has also been on the rise,and reached a new high in recent years.Social and economic development is inseparable from the support of the infrastructure system,including electricity,transportation,natural gas,internet,etc.Among these systems,the importance of the electric power system is self-evident,because the electric power system is the power source of modern society and economic operation.Without electricity,all electrical equipment will be paralyzed.In recent years,drastic changes in the climate and environment have led to the increasing frequency of extreme disaster,which have caused many large-scale power outages and even disassembly of the foreign power systems.In order to reduce the impact of extreme disaster events on the power grid,many countries have put forward the relevant strategies to construct the resilience power grid.Based on the resilience of resilience power grid,this paper studies the available load forecast and recovery strategy design of the resilience power grid under strong typhoon.The main work completed and the results obtained are as follows:(1)First of all,this paper outlines the definition of resilience proposed by experts and scholars in various fields with starting from the general definition of resilience.What’s more,it puts forward the definition of resilience in the power grid,and describes the characteristics of the resilience power grid.Then,simulation is applied to analyze the status information of the actual power grid when it encounters natural disasters,and thus an overall analysis framework for resilience power grid is proposed.Finally,under the framework of the overall analysis of the resilience power grid,the vulnerability of the resilience power grid in each stage of disaster is grasped,and the specific analysis method of the resilience power grid is emerged.(2)A Merge-Google network is proposed to predict the available load supply of the resilience power grid in the disaster bearing stage under the influence of a typhoon.Firstly,motion path of the historical typhoon information,the law of typhoon intensity changes(with the batts model)and the maximum wind speed radius are visualized and showed on the picture;secondly,the topological structure,power flow and tower strength information of the resilience power grid are visualized on the picture.Then,two images and the available load in the disaster bearing stage are input into the Merge-Google network at the same time,and the Merge-Google network is trained to mine the value and information of historical data.Finally,images from the validation set are imported into the Merge-Google network to verify the accuracy of the proposed method.(3)A restoration strategy based on load level,line flow power and component reliability,is Researched and constructed.The neural network based on immune algorithm is proposed to optimize the order of component recovery by considering the time-varying of recovery resources and the difference of component repair time.Among all neural networks,the 3-layer BP neural network is considered to be the most classic neural network.It is not only simple in structure,with only three layers: input layer,hidden layer,and output layer,but also can maximize the fitting of nonlinear functions,which is very useful for optimization of complex problems.In generally,when the traditional BP neural network uses gradient optimization,the initial weight and threshold are set randomly.But when using the gradient descent algorithm for optimization,it is easy to fall into a local minimum and fail to search for the global optimal solution.The initial weight and threshold of the neural network will have a great impact on the performance of the network,so the immune algorithm is used to optimize the initial weight and threshold of the BP neural network to improve the accuracy of the BP neural network in predicting the order of component restoration. |
