Study Of Scenario Generation And Reduction In The Stochastic Unit Commitment Problem

Unit commitment problem is an important problem in the economic dispatch ofelectric power system. It is a large scale mixed integer programming problem withlots of constraints. The research of unit commitment problem can bring significanteconomic benefits, so it has received wide attention from researchers. In recent years,with the development of the computer technology and the stochastic optimizationtheory, the stochastic unit commitment problem with consideration of price and loaduncertainties is receiving increasing attention.In the mathematical model of the stochastic unit commitment, it generallyrequires modeling uncertainties into a series of scenarios based on the scenarioanalysis method, therefore, the quality and quantity of scenarios will directly affectthe optimal solution and computation burden. During the process of generatingscenarios, in order to truly reflect the uncertainty, it usually generates a large numberof scenarios. But with the increasing number of scenarios, the computationalefficiency of solving of the stochastic unit commitment will be greatly reduced,leading to its restrictions on the practical applications.In order to provide appropriate number of scenarios for the stochastic unitcommitment problem, this paper focuses on the study of scenario generation andreduction in the stochastic unit commitment problem.Firstly, this paper introduces the research background and significance ofscenario generation and reduction in the stochastic unit commitment problem, anddomestic and abroad research status; Secondly, this paper constructs the mathematicalmodel of the traditional unit commitment problem, the stochastic unit commitmentproblem with consideration of load uncertainties and the stochastic unit commitmentproblem with consideration of electric power price uncertainties; Thirdly, based onthe introduction to the structure of scenario and scenario generation methods, thispaper uses Monte Carlo simulation method to generate the initial price scenarios forthe stochastic unit commitment problem with consideration of electric power priceuncertainties; Finally, based on introduction to the basic principle of scenarioreduction, the special case of scenario reduction and calculation process, this paperadopts the forward selection method and backward reduction method to reduce the initial price scenarios, and uses the relative distance coefficient to validate theeffectiveness of these two algorithms, and compares their computation efficiency. Testresults show that, two algorithms can reduce scenarios effectively, and obtain theappropriate number of scenarios.