Research On Grid Connected Prediction Of Photovoltaic Power Station Based On Improved Particle Swarm Optimization Algorithm

"Carbon peaking and carbon neutrality" is our commitment to the world.In order to realize the development of low-carbon clean energy and speed up the clean energy transformation in China,we should adapt to local conditions and vigorously develop new energy sources such as wind power photovoltaic,so as to reduce carbon emissions.In view of China’s comparative advantages in geography,climate and technology reserves,vigorously developing wind power and photovoltaic will be the first way to achieve the goal of carbon neutrality in China.However,due to the strong intermittence and randomness of photovoltaic and wind power output,a large proportion of wind power photovoltaic grid connection will bring security and stability risks to the main grid operation.Secondly,due to the relative uncertainty of output,it also brings difficulties to short-term load forecasting of regional power grids.In this paper,firstly,the necessity of integrating photovoltaic power station into Fushun local power grid is expounded,and on this basis,two feasible access system schemes are put forward.Through technical analysis of each grid-connected scheme and referring to relevant economic construction indexes,the problems such as whether the power flow distribution of related substations is reasonable,whether the power grid is unstable,short-circuit current exceeds the standard,etc.are systematically analyzed,and the optimal access system scheme is comprehensively selected from the aspects of safe and economic reliability of power grid operation.To solve the short-term load forecasting problem of regional power grid with photovoltaic grid-connected power generation,this paper uses BP neural network to establish a multi-input BP network short-term load forecasting model by considering various inputs such as weather,temperature,weather conditions and historical load,discusses the shortcomings of the standard BP neural network,improves the particle swarm optimization algorithm through parameter adjustment and other methods,and then optimizes the parameters of the neural network model with improved particle swarm optimization to improve the accuracy of neural network forecasting.According to the geographical and climatic characteristics of Fushun area,the actual situation of power grid operation and the trend of regional load change,the basic BP neural network model and the improved BP neural network model based on improved particle swarm optimization algorithm are used to predict the short-term load of regional power grid after considering the factors that have great influence on photovoltaic output and regional load change.The prediction results of the two models are compared with the actual load value,and compared with the prediction results of the load forecasting module in the smart grid dispatching control system of the regional power dispatching control center of the power supply company to verify the effectiveness of the short-term load forecasting algorithm considering photovoltaic grid connection.