TY - GEN

T1 - A solution to the stochastic unit commitment problem using chance constrained programming

AU - Mazumdar, Mainak

AU - Norman, Brian

N1 - Copyright:
Copyright 2011 Elsevier B.V., All rights reserved.

PY - 2005

Y1 - 2005

N2 - The paper develops a solution method for scheduling units of a power-generating system to produce electricity by taking into consideration the stochasticity of the hourly load and its correlation structure. The unit commitment problem is initially formulated as a chance-constrained optimization problem in which we require that the load be met with a specified high probability over the entire time horizon. The solution procedure consists of solving a sequence of deterministic versions of the unit commitment problem that converge to the solution of the chance constrained program. For the deterministic unit commitment problems, Lagrangian relaxation is used to separate the dual problem into its subproblems. Each subproblem is solved by a dynamic program. The initial results indicate that accounting for the correlation structure of the hourly loads reduces the value of the objective function when the optimization problem is formulated as a chance constrained program. Monte Carlo simulation is used to verify the accuracy of the solution provided by the algorithm. The relationship that the unit commitment solution found using the chance constrained optimization approach has with that found using conventional optimization spinning reserves is discussed.

AB - The paper develops a solution method for scheduling units of a power-generating system to produce electricity by taking into consideration the stochasticity of the hourly load and its correlation structure. The unit commitment problem is initially formulated as a chance-constrained optimization problem in which we require that the load be met with a specified high probability over the entire time horizon. The solution procedure consists of solving a sequence of deterministic versions of the unit commitment problem that converge to the solution of the chance constrained program. For the deterministic unit commitment problems, Lagrangian relaxation is used to separate the dual problem into its subproblems. Each subproblem is solved by a dynamic program. The initial results indicate that accounting for the correlation structure of the hourly loads reduces the value of the objective function when the optimization problem is formulated as a chance constrained program. Monte Carlo simulation is used to verify the accuracy of the solution provided by the algorithm. The relationship that the unit commitment solution found using the chance constrained optimization approach has with that found using conventional optimization spinning reserves is discussed.

UR - http://www.scopus.com/inward/record.url?scp=27144557303&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:27144557303

SN - 078039156X

SN - 9780780391567

T3 - 2005 IEEE Power Engineering Society General Meeting

SP - 1339

BT - 2005 IEEE Power Engineering Society General Meeting

Y2 - 12 June 2005 through 16 June 2005

ER -