| China has a vast territory and extensive energy distribution,especially the Yalong River Basin in Sichuan.The water system is rich and the climate conditions are suitable,which lays the foundation for the development of wind energy,light energy and hydro energy resources.Because of the distribution characteristics of wind,light and water energy,it is difficult for wind,light and water to operate independently.The randomness and volatility of wind energy and light energy lead to the instability of wind power and photovoltaic power.And the uncertainty of runoff and water resources,water and electricity dispatching in different stages becomes a difficult problem.The multi-energy complementary development of wind energy,light energy and water energy can make up for their own shortcomings and provide new ideas for energy development.However,there are many mutual existence of multiple uncertainties between wind,light and hydro energy,which makes the development of multi-energy complementary to each other.Combined with the above analysis,the research contents of this paper are as follows:(1)Runoff characteristics of watersheds.The five-grade hydropower stations of clean energy planning demonstration base is determined as the research object,and the runoff distribution characteristics of cascade power station are described with the runoff of 20 years.Analysis of the random characteristics of monthly,seasonal and annual radial flow grade 5 hydropower station by using P-Ⅲ empirical accumulated frequency curve.The correlation between ten-day runoff and annual runoff is investigated by using autocorrelation coefficient and the correlation between annual runoff is analyzed.(2)Wind and light energy characteristics of watersheds.To study the characteristics of wind energy and light energy,the randomness and volatility of wind energy and light energy should be studied.Based on the wind speed parameters,a normal distribution model is established to fit the wind speed and frequency distribution.The average wind speed and wind speed probability distribution and cumulative probability distribution of the weather station are analyzed statistically,and the wind energy fluctuation characteristics are investigated.According to the solar radiation and sunshine hours,the daily and seasonal characteristics of light radiation are studied.The solar energy fluctuation is calculated by the solar radiation in 20 years and the solar radiation is converted into sunshine hours.(3)The multi-energy complementary characteristics of landscape water.Based on 20 years of wind,light and runoff data,the generation of wind,photoelectric and hydropower in typical areas is calculated.Based on the variation of mean wind speed,mean irradiation amount and mean runoff curve,the complementary standard of wind power,photovoltaic and hydropower is established.According to the average daily power generation of wind,light and water energy and the load of complementary system,the year of complementary system is divided into four working stages according to the standard,and the complementary characteristics of the system are excavated.(4)Short-term prediction and uncertainty analysis of wind power based on LSTM-NPKDE.A LSTM model is used to forecast wind power on time scale of 4 hours,1 day and 3 days,and compared with the prediction methods of BP neural network,PSO-BP hybrid model and wavelet neural network.And the confidence interval method based on nonparametric kernel density estimation is used to quantitatively calculate the uncertainty of wind power prediction. |
