Based On The Photovoltaic Power Generation Forecasting Grid Inverter Control Strategy Research

With the development and utilization of new energy, renewable energy will replacefuel energy become the dominant force in power generation industry in the future. As therepresentative of new energy, Photovoltaic is more and more attention by researchscholars at domestic and abroad in recent years. However, photovoltaic power generationis greatly influenced by environmental climate. PV grid can bring impact to power grid,and affect the safety of power grid, stable and economic operation. So, grid photovoltaicpower generation forecasting has become a staff of decision-making basis and guarantee.Based on D-S evidence theory for generating initial data preprocessing, in this paper,forecast photovoltaic power plant generating capacity by support vector machine (SVM)algorithm, and complete photovoltaic (pv) grid inverter based on power generationforecasting model of the design and implementation. This paper mainly for the followingwork:First of all, researching theory of support vector machine (SVM) algorithm based onD-S evidence. Through the theory of evidence, get with similar characteristics to predictmonthly training data highly value, and support vector machine forecasting model isestablished to forecast.In the second, using the improved variable step size perturbation method ofobservation, on the basis of building a mathematical model of pv cells, combined with Sfunction to build PV maximum power tracking module, being used to simulate pv cellmodules under different light intensity and temperature of the I-V, P-V characteristics andto achieve Maximum Power Point Tracking.In the third, PQ control method is used as a photovoltaic grid inverter control strategy.The control strategy of research and discussion has carried on through PLL (phase-lockedloop) and dq transform, power and power factor control, current control for photovoltaic(pv) grid inverter in three stages.Finally, setting up the grid inverter experimental platform, complete two parts of thegrid inverter design of hardware and software. Bring about grid control with inverter through the DSP digital control technique. Put the forecast through the D-S evidencetheory combined with support vector machine prediction algorithm as a reference ofinverter output power. Then I have carried on a closed-loop text of the grid-connectedoperation of the inverter.