Research On The Grid-Connected/Islanded Control Methods And Black Start Optimization Of Microgrids Based On Differential Evolution

The microgrid provides an effective solution for the efficient access of various distributed renewable energy sources,high-reliability supplication of different loads and multiple energy forms,and efficient implementation of active distribution networks.It is an important part of the future smart grid.In recent years,the optimal grid-connected/islanded control and black-start optimization problems for microgridshave received extensive attentionsboth in the academia and industry fileds.Evolutionary algorithms provide an effective research idea for solving the above problems.As a novel evolutionary algorithm,Differential Evolution?DE?algorithm has been successfully applied in various engineering optimization problems,but it has rarely been reported in the operation optimization and controlissues of microgrids.This thesis focuses on the optimal grid-connected/islanded control and black-start optimization problems based on DE algorithms.The main research work and innovative pointsof this thesis are as follows:?1?In order to deal with the optimalturning problem of multi-controlparameters in multi-variable robust PID controllers,this thesisformulates it as a constrained optimization problem with the mixedH₂/H?robust performance firstly,and then it presents a hybridoptimizaiton method called CEO-DE by combining extremaloptimization and differential evolution.The simulation experiments oftwo typical multi-variable control systems have demonstrated that theCEO-DE algorithm has better control performance than the reportedalgorithms such as IGA and 2LB-MOPSO.?2?For the voltage/frequency optimization optimal control issue ofdistributed invertersunder islanded mode in a microgrid,a DE-basedoptimal droop control method for distributed invertersis proposed byconsidering he weighted sum of the integrated time absolute error?ITAE?for voltage and the total harmonic distortion value?THD?asthe objective function.The simulation experiments have shown thatthe DE has smaller voltage THD values and smaller frequencyfluctuations than genetic algorithm?GA?-based and particle swarmoptimization?PSO?-based methods.Aiming at the optimal P-Qcontrolproblem,the weighted sum of the ITAE for active power andreactive power has been considered as the objective functions,andthen it presents the DE to optimize the P-Q controllers of thedistributed inverter.The simulation experiments have demonstratedthat the DE-optimized inverters hassmoother power fluctuations.?3?Aiming at the optimal design problem of PI parameters for theswitching control of grid-connected and islanded operations in amicrogrid,the weighted sum of ITAE for frequency and voltage THDhas been considered as the objective function,it presents an DE-basedoptimal switching control method of grid-connected and islandedoperations in a microgrid.And then its optimization performance hasbeen revified by establishing a simple simulation model of themicrogrid.The simulation experiments demonstrates that the DE hassmaller voltage THD values and frequency fluctuations than GA andPSO,and there is no sudden change of the power for each distributedinverter,which reduce the transient voltage shock effectively.?4?To solve theoptimization problems of distributed generation?DG?restoration and load restorationin black start process for amicrogrid,a multi-objective optimization model is established firstlyby maximizing DG's power and loads near DG.It presents amulti-objective population extremal optimization?MOPEO?algorithm to obtain the optimal DG restoration.Secondly,asingle-objective optimization model is established by maximizing theactive power of loads,and a binary load evolution?BDE?algorithm isdeveloped to obtain the optimal load restoration.The feasibility of theproposed optimization method is verified by a 30-node microgrid.