Security Situation Assessment And Prediction Of Power Grid Orient To Intelligent Scheduling

With the development of power grid,the requirements to grid dispatching intelligence are getting higher and higher.But the existing dispatching system lacks the accurate control to the development and change of the grid operation state,the level of automation and intelligence is low,and it lacks the comprehensive analysis and decision-making advices of the overall operation,which increases the stress of grid operators and makes it harder to assure the secure and economic operation.So the accurate forecasting of the grid operating state and the overall control of the operating process are of great importance to the assessment and forecasting of grid security state.It needs to be carefully studied how to accurately predict the development trend of grid operation state and comprehensive control of grid operation process.Through collecting and control various factors that affect the safety of power grid in the operation scope,and understand and predict accurately and effectively control the security situation of power grid,which enables the grid to recover security and its management from passive to active prevention.Over the years many scholars have made a lot of researches on the assessment and forecasting of power grid security situation,resulting in a lot of optimization algorithms.In this paper,the research status and development trend of grid security situation assessment and forecasting are discussed in detail,including summarizing the proposed forecasting algorithms,discussing their shortcomings and unresolved problems and proposing solutions.Since the research on the assessment and forecasting of power grid security situation is still in its infancy,it is necessary to carry out further research.Therefore,many factors affecting the safety of power grid are collected,comprehensive evaluation index system of power grid security situation is established and analyzed,then the complex index system of high dimension is reduced,finally the grid security situation evaluation model and grid security situation forecasting model are established.Specific works are as follows:1.Study the operation mechanism of power grid,analyze the various factors that affect the security of power grid,and use AHP and Delphi method to construct grid security situation evaluation index system.2.Aiming at the characteristics of strong correlation between the indexes in the power grid security situation evaluation index system and the high dimension of the index system,this paper proposes a method of dimension reduction using the self-coding deep neural network(autoencoder network)method to extract the low-dimensional coding.And the extracted low-dimensional data can represent the characteristics of the highdimensional data.The autoencoder has good dimensionality reduction and reconstruction function for high dimensional data.3.The model of power grid safety situation assessment is built by using Extreme Learning Machine(ELM)algorithm.The machine learning model is introduced and the relationship between the input parameters and the output power grid safety situation values is studied.The simulation experiment is carried out to verify the validity of ELM being applied to the grid security situation assessment.4.A forecasting model based on weighted online sequence extreme learning algorithm is proposed to forecast the grid security situation at a certain point in the future using the historical situation values.The validity of the proposed method being applied to the power grid security situation forecasting is validated by the simulation experiment.