Research On Time Series Modeling Of Long-term Meteorological Data

Time series analysis is not only an important branch of statistics,but also an indispensable part of data analysis.Precipitation and wind speed are two kinds of meteorological data with obvious non-stationary and periodic fluctuation characteristics.To a certain extent,the variation and fluctuation of precipitation can reflect the drought and flood in the area.At present,the world is in a shortage of water resources,and precipitation is random and uncertain in time Therefore,the accurate prediction of precipitation can be used to deal with drought and waterlogging disasters in a timely manner.It also has certain guiding significance for the growth and irrigation of crops.Natural resources are constantly being excavated.Wind energy belongs to renewable energy,and the use of wind energy to generate electricity is currently the need of sustainable development strategy in China's energy construction.Wind speed has the randomness and uncertainty of random variation.Accurate prediction of wind speed can reduce the cost of wind farm,improve the utilization rate of wind,and at the same time,it is of great significance to the structural adjustment of power industry.Therefore,in order to realize the efficient utilization of meteorological energy and establish the prediction model of long-term meteorological data time series,this thesis mainly carries on the following three aspects:(1)Based on the threshold denoising algorithm of wavelet,an improved denoising algorithm for soft threshold is proposed,and the signal denoising performance is optimized and simulated.The results show that the signal curve after denoising not only has good smoothing performance,but also preserves the basic characteristics of the original signal.(2)Based on the ARMA model in time series,the seasonal model of ARIMA is constructed Through the simulation of noisy sinusoidal series,the results show that the ARIMA seasonal model is superior to the ARMA model.Finally,The simulation results of precipitation and wind speed of meteorological data show that the performance of ARIMA seasonal model is obviously better than that of ARMA model.(3)Based on the seasonal model of ARIMA,a ARIMA seasonal model based on soft threshold denoising is proposed,which is verified by sinusoidal sequence simulation.The performance of ARIMA seasonal model based on soft threshold denoising is significantly better than that of ARMA and ARIMA seasonal models.The simulation results of meteorological data further show that the performance of ARIMA seasonal model based on soft threshold denoising is obviously better than that of the first two models.