| In this paper,the technical analysis theory and discriminant methods of the stress wave reflection method in measuring the signal pile were firstly discussed and compared; Based on the dynamic finite element method,the measuring pile process of the stress wave reflection method was simulated; Combined with the wavelet packet technology in Signal Processing and neural network technology,the identification of defects types of pile signal was simulated.This paper mainly worked as follows:(1) Numerical simulation of signal curve of stress wave reflection method in measuring pile. With the dynamic finite element method, the stress wave propagation in the concrete pile under the pulse load and reflection properties were numerically simulated, and the pile top-speed curves were resulted.The certification of theoretical curves and measured curves achieving better, clarified the authenticity of the model.(2) The wavelet packet eigenvalue extraction on simulated signal of pile. Used wavelet analysis technology to dear with signal noises,with the wavelet packet dynamic test signal, extracted eigenvalues of typical defects of several piles, and realized the accurate position on defects of piles.(3) The use of neural network detecting defect types of piles.Features were extracted from the three-layer wavelet packet on the frequency domain of measuring piles, finally the features were trained and classified under the neural network, so as to realize the accurate certification of pile defects.In this paper, researches on the stress wave reflection method in the field of detecting piles provided a certain degree of guiding significance,and played a role in promoting development of detecting technologies.I am sure that after continuous improvement of the stress wave reflection method and detecting pile technology, we can be more integrated intuitive to get information of internal defects of piles,and also have good prospects of promotion on the effective detection and improving the quality of concrete structural engineering. |
