Prediction of Meteorological Observation Data Based on Time Series Analysis
A short-term prediction method of wind speed change trend is proposed.Firstly,we use wavelet analysis technology to analyze the wind speed data and study the change trend of the data in each frequency domain.Then,we use long short memory network(LSTM)to predict the data in low frequency region and use Kalman filter filtering technology predict the data in high frequency region.Finally,the data in each frequency domain is reconstructed to achieve short-term prediction of wind speed data.The validation was conducted by using 10 minute wind speed data from a airport in January 2019.The experimental results showed that this method can effectively predict the trend of wind speed,providing an effective method for pre-dicting wind speed in airport.
NETL
NSTL
万方数据
