Low strain detection, an indirect and nondestructive testing method, is one of the main pile integrity testing methods. We propose low strain testing analysis based on a synchrosqueezing wavelet transformation (SST). Through a typical model pile test, the SST is applied to identify pile bottom signal reflection time and to separate signal from noise. It is also compared with the conventional wavelet de-noising and the empirical mode decomposition (EMD) de-noising method. Results show that the SST technique can be used to identify the reflected signal of the pile bottom, achieve signal and noise separation, and improve signal-to-noise ratio. The method has significant advantage in low strain detection signal processing compared to other methods.
Compared with conventional time-frequency analysis method, synchrosqueezing wavelet transformation (SST) exhibits high resolution capability and good application effect. In this study, SST is introduced to ground-penetrating radar (GPR) processing. This method is applied to analyze a continuous electromagnetic signal. SST can obtain a higher resolution and a better processing effect than conventional wavelet transform and short-time Fourier analysis. In the application of GPR forward analysis data, the transform can correctly distinguish different interfaces and objects. Its resolution increases as frequency increases. However, compression wavelet modulus gradually decays as frequency increases. The proposed method is applied to detect tunnel lining under actual conditions and in a strong noise background. Indeed, the method can efficiently identify interfaces and abnormalities.