Font Size: a A A

Grid Photovoltaic Power Station Power Short-term Forecast Method Research

Posted on:2018-10-18Degree:MasterType:Thesis
Country:ChinaCandidate:J ZhangFull Text:PDF
GTID:2322330512471712Subject:Electrical engineering
Abstract/Summary:PDF Full Text Request
As energy drying up and environment worsening,grid type photovoltaic power generation system is the development trend of power industry in the future.However,due to the photovoltaic easily influenced by the external environment changes,has a strong volatility of output,impact,make its power grid after the impact on the grid of originally stable,stable operation of power system and power grid power quality has a larger effect.Therefore,in order to ensure the system safe and stable operation,overall arrangements for the distribution of coordination development,effectively projections for photovoltaic power generation is of great significance.In view of the research topic,first of all,analyses the various factors influencing the power photovoltaic power station,and uses the Pearson similarity analysis method to find the influential factors,a prediction model of input variables;Then analyzed based on the LS-SVM and markov model to predict the theoretical basis,the optimal combination and the two methods to establish the forecast model of the short-term photovoltaic power;Based on similar day selection method to calculate the prediction model of training data and the optimal parameters of cross validation method,ensure the fitting precision of the data;On the basis of using the error correction of the model,markov correction algorithm to further improve the prediction accuracy;Choose different season in the process of simulation of four types of weather data in combination with the LS-SVM combined with markov theory of the optimal combination forecast model test,the results show that the prediction method has higher prediction accuracy,satisfy the current electric power dispatching of power grid planning and technical requirements.Finally,based on the LS-SVM combined with the practical application system for predicting the output of markov theory,the system can not only realize output prediction function,but also can real-time monitor the power parameters of photovoltaic power station,the system has the function of the photovoltaic power station monitoring and forecasting two parts,which can realize real-time data acquisition,history power generation data collection,data curve,power prediction,etc,have a certain practicality,for dispatcher work provides an efficient platform for the management of the power grid.
Keywords/Search Tags:power prediction, grid type photovoltaic power station, LS-SVM, markov models, similar day
PDF Full Text Request
Related items