Geological Engineering: Proceedings of the 1st International Conference (ICGE 2007)
150 Study on Monitoring Analysis and Time Forecasting of Landslide Based on Fourier Transform and Wavelet Analysis
Download citation file:
The monitoring data of landslide displacement are usually disturbed by the impact of rainfall or engineering construction. As a result, the curve of displacement monitoring is always characteristic of step-like. In order to achieve a full understanding of the step-like phenomenon, Fourier transform and wavelet analysis methods were introduced to find out the causes in this study. Firstly, Fourier transform method was used to extract frequency component from the original monitoring data of displacement; secondly, the wavelet transform modulus maxima method was adopted to detect the breakpoint between different frequency components. Based on the detected frequency component and the position of the breakpoint, the quasi-periodic noise could be simulated. Finally, the time-displacement curve without the impact of noise was obtained by subtracting the noise component from the original curve of displacement monitoring. It is suggested that periodic rainfall is the main reason that induces the step-like catastrophic break points in the curve of displacement monitoring. These results provide a reasonable explanation to the occurring of step-like phenomenon in the curve of displacement monitoring and propose a new analysis way for the time forecasting of landslides.