Application of Time Series Model Based on Wavelet De-noising in Groundwater Dynamic Variation Regulation Research

摘要

Wavelet transform can decompose the input signal into high and low frequency signal. The low frequency signal includes main performances of the signal. However, the high frequency signal contains many noises. Wavelet de-noising can be realized by reconstructing the rest signal after the high frequency signal is flatted. According to the effect of noise on hydrological time series, this paper studies hydrological time series including measured noise with wavelet de-noising method. Firstly, wavelet de-noising is done with MATLAB, and then trend component, periodic component and random component are extracted from the hydrological time series respectively. At last, the time series model based on wavelet de-noising is established by superposition to simulate and forecast the groundwater dynamic variation of Chahayang irrigation district. The forecasting precision of this model is good. So it is reasonable to research the groundwater dynamic variation of Chahayang irrigation district through the model. At the same time, this paper can provide scientific basis for sustainable utilization of groundwater resources, making reasonable irrigation system and the development of paddy field in Chahayang irrigation district.