光伏发电系统发电功率预测研究-电气工程专业论文.docxVIP

ABSTRACT As Chinas vigorous development of renewable energy, the installed capacity of grid-connected PV systems is increasing. Grid-connected Photovoltaic system is affected by meteorological factors, its output is intermittent and volatility. A large number of grid-connected PV systems will increase the complexity of the power system, changing the existing grid system margin and power generation plan, causing the system to face the risk of collapse. Improving the prediction accuracy can help power sector develop a detailed scheduling program and improve the power system stability. In this paper, a lot of reading at home and abroad on the basis of the literature had been conducted and three different aspects of photovoltaic power generation projections were studied. By analyzing the impact of grid-connected PV power related factors, we select the main factors as neural network input variables to build photovoltaic power generation based on neural network prediction model. Through data mining techniques, we screened from large amounts of data and forecast periods with similar meteorological characteristics data sequence, using gray-correlation theory to predict the photovoltaic power generation. By using theoretical weight calculation techniques of the combined forecast, neural network prediction and grey prediction are given different weights for the combination of photovoltaic power forecasting technology. In this paper, all the data is based on the grid-connected PV monitoring system of Tianjin University. Neural network-based prediction of photovoltaic power generation technology, gray relational analysis of photovoltaic PV power forecasting techniques and combined forecast techniques are analysed under the same weather types for simulation and data validation. And these three predictions prediction techniques were assessed through the use of five different error indexes. Assessment indicates three prediction methods have reached a higher prediction accuracy. K