Research And Implementation Of Short-term Power Forecasting Of Grid-Connected PV Plant

Photovoltaic power generation has been favored by many countries because it is clean and pollution-free.At the same time,because of the randomness and fluctuation of output power,it will cause a certain impact on the public power after power plant grid is connected.This is extremely detrimental to the normal operation of the power grid.Therefore,the accurate prediction of the generation power of grid connected is conducive to the rational allocation and planning of the proportion of photovoltaic energy and conventional energy,it is conducive to adjust the schedule timely,and it makes the power system operated in a safe,stable and economical way.In this thesis,a large number of domestic and foreign literatures on the basis of reading,the historical data of a photovoltaic power plant in Gansu province as the research object was used,and a prediction model by using HS-ESN model was established.Then short-term power forecasting of the photovoltaic power station was analyzed.Finally,through the C# and MATLAB mixed programming,the thesis introduced the design of the photovoltaic power short-term prediction system.Based on the above description,this thesis research content mainly includes the following aspects:Firstly,the output characteristics of photovoltaic batteries were combed by simulation model of photovoltaic batteries based on MATLAB,solar irradiation intensity and the temperature influence on the photovoltaic power were analyzed,and the main factors influencing the photovoltaic power were determined ulteriorly.Then the thesis combined each influence factor into feature vector,similar day selection algorithm is utilized to extract similar day and training samples.Secondly,on the basis of deep research on echo state network(ESN)algorithm,this thesis was proposed a hybrid algorithm for HS(Harmony Search)algorithm to optimize echo state network.In this thesis,the HS algorithm was used to optimize the storage pool parameters of ESN algorithm,which could effectively improve the accuracy of ESN algorithm,and the HS-ESN algorithm was applied to the short-term prediction of photovoltaic power generation,the photovoltaic power generation under different weather types was predicted by using different prediction models,it verified the effectiveness of the similar day selection algorithm,at the same time,it showed that the performance of HS-ESN model was better than the single ESN model and other models.Finally,according to the prediction model and method proposed in this thesis,a shortterm prediction system of photovoltaic power generation was designed,the requirement analysis and structure design of the system were also carried out.The forecasting system has the basic module of short-term prediction of photovoltaic power generation,which includes login interface and data import module,short-term power forecasting module,report statistics module and data query and output module.Therefore,it has certain practical value.