Research On Photovoltaic Power Prediction And New Energy Cluster Division To Promote New Energy Consumption

The prospects for the development of new energy in Yunnan are very promising.The development of new energy sources is of great significance to Yunnan’s economic development and industrial restructuring.But the special plateau mountain terrain in Yunnan has brought a series of problems on new energy consumption.During the operation of Dali new energy dispatching management demonstration project in Yunnan,the following two problems were highlighted: Firstly,the accuracy of wind power prediction is not high under the three-dimensional climatic conditions of the plateau,while the photovoltaic(PV)power is difficult to predict,and the large-scale PV power plants are difficult to connect to the grid.Secondly,the distribution of new energy stations is scattered,which makes the coordination and cooperation among various stations prominent.Improper control will lead to underregulation and overregulation,resulting in idle or wind and light abandoning of the wind power/PV delivery channel.To this end,this paper focus on the way to formulate a precise prediction scheme of PV power generation,and controlling a power plant associated with a dispersed area uniformly by forming clusters according to certain complementary characteristics.The paper collects meteorological factors historical data,PV power historical data and wind power short-term predicted power data in Yunnan Dali new energy dispatching platform.Based on the above data,the short-term nonlinear combination of PV for improving similar sample selection and outputting feature extraction is proposed respectively.The power prediction method and the dynamic division method of the new energy cluster are as follows.Draw four meteorological factors and photovoltaic power generation curve diagrams to visually analyze the influence law of various meteorological factors on photovoltaic power generation,and use the sample distance correlation coefficient to measure the correlation between meteorological factors and PV power generation to find the main forecasting influence factors.excluding non-major influencing factors.A scientific description system is established with the most critical predictor irradiances to command and describe integrally and locally,and its feasibility is verified by the irradiance historical data verification index system of Dali Xicun PV power station.Because the PV power is difficult to be predicted,a new PV power forecasting model is proposed.Step one: Identify and eliminate the wild value data in the historicaldata,and complete the rejected data and missing data of the partial samples by using the interpolation method.Step two: Propose a method to calculate the weight of meteorological factors based on the fuzzy clustering and rough set theory.In the method,reasonable daily samples are selected by the idea of weighted euclidean distance.Step three: The specific meteorological factor is selected as the input of prediction model,and extract the irradiance index feature by the irradiance index description system while the extraction result is used as the prediction model,and the input signal achieves the purpose of scientific dimension reduction.Step four: Propose a nonlinear combination forecasting model based on improved similarity statistic,and testify its feasibility by the historical data of Xicun PV power station based on MATLAB.In this paper,the internal reason why PV power is stable under the model of colony is explained,and the trend inconsistency distance is defined as the distance metric between power stations.To meet the grid topology constraints and cluster sites,take the short-term forecast data sequence of the next day of wind power plant and PV power plant as input.To satisfy the controllable power supply constraints and make the number of clusters is as small as possible,cluster algorithm is used for clustering and the clustering results are evaluated.The optimal cluster number is selected to realize the optimal cluster division of new energy power stations.The feasibility is tested by the new energy field station of 220 kV Haidong substation in Dali area based on MATLAB.