The Study Of Microgrid Multi-objective Optimization Operation

To effectively solve the increasingly serious energy depletion and environmentalpollution problems, distributed power generation technology rapidly developments andwidely spreads all over the world in recent years, microgrid thus arises at the historic moment.And how to optimize the allocation of resources, reduce the running cost and pollutionemissions in the network under the premise of safe and reliable operation also become one ofthe hot topics in the study of micro network operation, in order to improve the micro economyand environmental protection of the network operation. Actually microgrid optimal operationis a multi-objective optimization problem. Simply considering the operation costenvironmental benefits cannot completely describe the microgrid running status. Therefore,this paper will study microgrid multi-objective optimization operation, establish itsmathematical model and explore its effective solving method.At first, this paper gives the energy consumption and output model of distributedgenerations, and then proposes the micro-grid multi-objective optimal operation model as theobjective of micro network operation cost minimum and carbon dioxide emissions minimumaccording to the principle of operation of micro-grid. It is a complex multi-objectiveoptimization problem with a variety of equality constraints and inequality constraints. Thereare several deficiencies using conventional multi-objective optimization algorithm to solvesuch problems. One is that traditional constraint handling methods dealing withmulti-constraints problems is difficult and the efficiency is not high. The second is that thespecies diversity in the evolution process based on single algorithm is not high and it’s easyfor precocious. Therefore, this paper proposes a new kind of α-constraint dominant sortinghybrid evolution algorithm to solve the model.The algorithm treats all the constraint conditions as α-constraint levelness by usingα-constraint dominant sorting mechanism, and then uses α-constraint levelness forevolutionary selection index to control all individuals transform into feasible solution quicklywhich could significantly improve the efficiency of constraint processing. At the same time, akind of hybrid multi-objective evolution algorithm based on non-dominated sorting wasproposed in order to effectively combine the advantages of differential evolution algorithmand estimation of distribution algorithm, so that the single algorithm’s defects of low speciesdiversity and premature convergence could be overcome. In addition, with technique for orderpreference by similarity to an ideal solution, this paper realizes the multi-attribute decisionmaking and obtains the optimal compromise solution.Finally, this paper simulates the algorithm with an example of microgrid and compares itwith non-dominated sorting differential evolution algorithm using conventional penaltyfunction constraint processing and single objective genetic algorithm toolbox. The example results show that the convergence rate, set quality and operating time of the algorithm in thispaper is better than the other two. And compared with the single objective optimization model,using the multi-objective optimization model is more in line with the actual microgrid runningstatus. It can better reflect the comprehensive benefit of microgrid both in economy andenvironmental protection.