Research On Optimization Of Economic Operation For Microgrid Based On SOA-SVM Forecast Of Power Output And Power Load Of Distributed Generation

With the increasing pressure on resources and the environment due to the shortage of fossil energy resources, the deepening of electricity market reforms, and the user' continuous requirements for improvement of power quality, the electric power industry is facing unprecedented opportunities and challenges. Constructing a safe, economic, environmental and green microgrid is increasingly becoming the common goal of the global electric power industry. The economic operation of microgrid is one of the key problems for microgrid research. It can improve the stability and economy of the great power grid operation effectively that optimizing the output of each generating units, the output of energy storage units in microgrid as well as the power exchange between the microgrid and power grid along with an important theoretical value and engineering value.Firstly, the microgrid is introduced briefly, then its structure characteristics and economic operation theory are analyzed. Through the description of some relevant knowledge about microgrid, a microgrid structure is established which consists of wind power, photovoltaic power, diesel generator, battery and power load. Secondly, in the light of the characteristics for the power load of microgrid, the power load of short-term in a certain area for 24 hours is forecasted by seeker optimization algorithm(SOA) combined with support vector machine(SVM), and then the results are compared and analyzed with other prediction models. The wind power generation is choosed as the micro-source of the microgrid, and according to the basic characteristics of the wind power output, its power output series is decomposed into a series of components by using ensemble empirical mode decomposition(EEMD), the series of components is also predicted by SOA combined with SVM, the final prediction results of the wind power output of short-term for 24 hours can be obtained by certain processing, and these results are compared and analyzed with other prediction models, the validity and accuracy of the combined forecasting method are verified. Finally, considering the price difference in different period, the economic operation model of microgrid is established under grid-connected operating state, and the simulated annealing(SA) combined with particle swarm optimization(PSO) is used to optimize the model, and then the output of different micro-sources at different times and the minimized daily power generation cost can be obtained, and the results are compared with ones generated by other different optimization algorithms, the excellent performance of the combined optimization algorithm is further verified.From analyzing the prediction results for the power load and the wind power output as well as the optimization results of the economic operation for microgrid, it is concluded that the content of data prediction and optimization algorithm are greatly enriched with a good theoretical and applied foundation for deepgoing research on microgrid because the combination method of SOA and SVM is applied to the prediction model of power load and wind poweroutput, and EEMD is used to decomposition of the wind power, and the mixed method of SA and PSO is used to solve the model of economic operation in microgrid.