Research On Ultra-short-term Prediction Method Of Distributed Photovoltaic Output

With the rapid development of the global economy,the imbalance between global energy consumption and energy supply is gradually becoming an important factor restricting the development of human society.In order to alleviate the global energy supply and the global ecological environment,the development ofclean and renewable energy such as solar energy and wind energy has become an important way for human society.As the most abundant renewable energy in the world,solar energy has various development and utilization methods.Among them,photovoltaic power generation has developed rapidly in recent years due to its advantages of no harmful gas emissions,safe and reliable operation,and easy maintenance.However,distributed photovoltaic power generation is affected by factors such as irradiance and temperature of photovoltaic modules.Its output power has obvious randomness and volatility,which will bring huge hidden dangers to the power balance and safe economic operation of the power system.Although there are many modeling methods for the uncertainty of photovoltaic output power,these methods cannot essentially solve the random fluctuation of photovoltaic output power.Accurate photovoltaic power generation prediction can reduce the uncertainty of photovoltaic power generation in essence,and is one of the key technologies to improve the ability of power grid to accept stochastic volatility power supply.In the aspect of irradiance prediction,an irradiance ultra-short-term prediction technique based on geographic information and meteorological information is proposed.Based on geographical location,atmospheric attenuation and other factors,a clear sky horizontal slope irradiance model is established.At the same time,the typical characteristics of various cloud layers are analyzed.Combined with historical meteorological data,the correlation matrix of different cloud types and different weather types is established.Combined with the historical irradiance,the conversion method of the clear sky index to the cloud occlusion is further proposed,and the cloud occlusion area at different times is determined by the clear sky index and the cloud transmittance of different types of clouds under different weather conditions.The trend of the occlusion area is predicted by the moving speed and direction of the cloud layer,and the irradiance change trend is further predicted.In the aspect of single PV output power prediction,the output power probability modeling method of photovoltaic power plant based on general distribution is proposed.Based on this,an ultra-short-term prediction method of output power of single photovoltaic power station based on irradiance prediction value is proposed.By preprocessing the historical number,the irradiance is binned.For different irradiance intervals,the versatile distribution is used to fit the actual PV output power value and the probability distribution of the actual PV output power value is established.For any irradiance prediction value,determine which interval it belongs to and find its corresponding distribution function,use inverse transform sampling to obtain the possible scene set of PV forecast output power,and then weight-average it to obtain the most likely PV forecast output power under the irradiance condition.Subsequently,the non-uniform binning and temperature correction were proposed to improve the above model,which further improved the forecasting accuracy of the PV output power power prediction.In the area of distributed PV output power forecasting,a distributed PV output power ultra-short-term forecasting method considering space-time correlation is proposed.The output power characteristics of regional distributed photovoltaic power plants and the correlation of output power of regional distributed photovoltaic power plants are analyzed.According to the theoretical value of the irradiance of the photovoltaic power station,the theoretical value of the output of the photovoltaic power station is obtained,and the equivalent clear-sky index of the photovoltaic power station is proposed based on the actual value of the photovoltaic output,thereby characterizing the degree of occlusion of the photovoltaic power station by the cloud layer..According to the influence of cloud layer with time and space on different photovoltaic power plants,the expression of occlusion area of regional photovoltaic power plant for time and space is proposed,and then a short-term prediction method of distributed photovoltaic output considering space-time correlation is proposed.