Application And Research On Outer Approximation Method And Distributed Alternating Direction Multiplier Method For Power System Optimal Dispatching

Greenhouse gases generated by fossil energy power generation have had a very negative impact on the environment.Therefore,since 2016,the use of non-fossil energy to gradually replace the traditional thermal power generation has become an important measure to prevent and control air pollution,improve environmental quality and adjust energy structure.In order to build a clean,low-carbon,safe and efficient energy system,it is of important theoretical and practical significance to study the operation and dispatch of power system in depth.This thesis aims to study the outer approximation method(OAM)and distributed alternating direction multiplier method,and take the hydrothermal unit commitment(HTUC)and direct current optimal power flow(DC-OPF)in power system as examples for application research.Firstly,this thesis applies the outer approximation method combined penalty function(OAM-PF)to the non-convex HTUC problem.OAM-PF starts first solves the continuous relaxation problem of the primal problem,then alternately solves the mixed integer linear programming(MILP)master problem and the nonlinear programming(NLP)sub problem.Our proposed method employs scaled quadratic cuts(SQCs)as valid underestimation of convex nonlinear function in master problems.Therefore,the MILP master problem is transformed into the mixed integer quadratic constrained programming(MIQCP)master problem.In addition,a heuristic method is proposed to accelerate the solution of continuous relaxation problem of the primal problem.In this paper,the effectiveness of the proposed method is verified by two systems of up to 44 thermal unit,15 hydro unit and 24 hours.Secondly,this thesis applies the consensus alternating direction multiplier method to the direct current dynamic optimal power flow with carbon emission trading(DC-DOPF-CET)problem.As opposed to other ADMM-based distributed approaches,which disclose boundary buses and branches information among adjacent subsystems,our proposed method only disclose boundary branches information among adjacent subsystems.In addition,we improve the convergence speed of our method further by reducing the number of dual multipliers and employing an improved update step of the multiplier.The effectiveness of the proposed method is tested on cases ranging from 6 to 1062 buses.The serial and parallel performances of the distributed alternating direction multiplier method are compared by experiments.