Research On Theory And Application Of Intelligent Non-destructive Testing Of Geotechnical Anchorage Systems Based On Wavelet Neural Network

The development and application of geotechnical anchorage technology is an important mark of modern geotechnical engineering. At present, geotechnical anchorage technology has been widely used in practice. Because of various factors, there are usually many deficiencies in geotechnical anchorage systems. If there are internal defects in them, they would be very dangerous, even to cause huge loss of people's life and properties. To avoid the situation, a high-performance, accurate, reliable, convenient and practical testing technology is needed. The defect identification, quality diagnosis and real-time testing to anchorage systems have been extensively concerned in the field of geotechnical engineering. Traditional static load testing method with difficulty in meeting the need of testing the large-area practical bolts is a testing method of destroying anchors, but the dynamic testing method of stress wave reflection, as a convenient, economical, rapid and reliable testing method, can not only meet the need of testing the large-area practical bolts but also don't destroy themselves. The dynamic testing method of stress wave reflection makes up for the deficiency of static load testing method, so it is very necessary and significant to research and apply it.By researching the energy leakage characteristics of stress wave reflection in anchor body and spreading rule of elastic wave in anchor body, rock and soil medium, and their coupling systems, this paper find the attenuation rule and spreading mechanism of elastic wave, which provides an important theoretical basis for non-destructively testing anchors' quality. According to one-dimension pole's vibration theory, it establishes mathematical and mechanical models of a complete anchor and a defective anchor, derives the formula of time-domain response and frequency-domain response, and makes a conclusion that there are two approaches to analyzing the signal of non-destructive testing anchors, namely time-domain analysis and frequency-domain analysis. Taking management buildings and museum of navigation sport school of general administration of sport of China as an example, the article compares the practical frequency curve of a complete engineering anchor with that of a defective engineering anchor and analyze them. Studying the theory of wavelet analysis, comparing the characteristics of various methods of wavelet analysis, researching how to choose the wavelet basis function suitable for defect diagnosis, using MATLAB software to deal with the practical engineering signal by wiping off yawp, and applying frequency-band energy and wavelet packet coefficients of crunodes into neural network to train by combining wavelet analysis with neural network, are also its research work.Originality of this paper can be shown as follows:(1)By analyzing the characteristic between complete anchors and defective anchors in practical engineering, a conclusion can be made that frequency curve is able to be regarded as important basis of diagnosing the defect of anchors.(2)Applying the training chart of wavelet neural network to the intelligent non-destructive testing in practical engineering can make up for the defects of traditional method not meeting the need of testing large-area anchors.