Robust Optimization Comprehensive Dispatch Of Microgrid

Crisis of energy and environment have become more and more prominent with the rapid development of global economy,so the distributed power generation technology which is represented by renewable energy has been widespread concern.Microgrid(MG),as one of the most effective ways to integrate distributed powers,has become a research hotspot of relevant researchers in recent years.However,the output power of renewable energy generation,such as wind power generation(WP)and photovoltaic generation(PV),is susceptible to environmental and climatic and therefore has strong intermittent characteristics which will threaten the security and stability of MG if measures are not taken.Robust optimization theory use a set to describe the range of uncertainty parameters in uncertainty optimization problem.No matter what values the uncertainty parameters take in the set,the robust optimal solution is feasible,so the robust optimization theory is very suitable for solving the uncertainty optimization problem which has high safe and stable requirements,such as MG optimization dispatch.In order to ensure the safe and stable operation of MG,this thesis studys on MG optimization dispatch by using the robust optimization theory,the process is as follows: Firstly,by considering the uncertainty of WP,PV's output power and load demand,with the minimized economic environment comprehensive cost as objective,the MG's robust optimization comprehensive dispatch(ROCD)model which is based on realtime frequency adjustment factor method is established;Secondly,In order to solve the model more simply,the ROCD model is transformed into ROCD1 model by using Soyster robust optimization method,and then the ROCD1 model is converted into ROCD2 model by variable substitution;Finally,aiming at the characteristics of multivariate coupling,multiple constraints and nonlinear objective function of ROCD2 model,a modified differential evolution(MDE)algorithm is proposed.Based on traditional differential evolutionary algorithm,perturbed mutation mechanism,dynamic algorithm parameters and appropriate constraint processing method which are related to the specific form of constraints in ROCD2 model are introduced in MDE algorithm.The ROCD model and MDE algorithm are tested by a numerical example.The results show:(1)The ROCD model is robust feasibility;(2)Compared with two kinds of traditional DE algorithms,the MDE algorithm has obvious advantages in convergence rate and optimization depth;(3)The optimal decision of MG's robust optimization comprehensive dispatch is mainly determined by the game between the comprehensive cost of MG's controllable devices and the comprehensive cost of electricity purchasing and selling,and the result of the game will determine the operation status of MG's controllable devices and large power grid;(4)The robustness and optimality of ROCD model are contradictory.Then,based on the Hermite interpolation curve fitting technique,a robustness level selection method,which avoids the maximum cost of robustness growth,is proposed to keep the balance of robustness and optimality of ROCD model.