| Engineering structures, such as buildings, public works, underground works and slope engineering are destroyed from local damages, which accumulate , and enlarged and wrecked, The effect of structure failures are catastrophic. To preventing safety accidents, nondestructive detection and health monitoring of engineering structures are high meaningful, which could discover early damages and take preventive measures. The dynamic detection technique is one of important nondestructive detection methods. The damages shape of engineering structures are complicated, and their effective detecting is difficult, and the effectiveness of the dynamic detection technique is still not satisfactory. In this paper, the beam with crack damage is investigated and 2 dynamic detection techniques including the modal strain energy method and the virtual flexibility matrix method are presented. The beam with crack is systemically analyzed by the dynamic FEM. Based the similarity principle, dynamic experiments of two modals with different scales and different materials are completed. A large amount of detailed data are obtained from the model experiments, which provide data and means with theoretical research. BP modal of the artificial neural network(ANN) has been built. It realizes dynamic identification of structural damages. Rock engineering and dam foundation. It damages mainly by sliding. Rock mass is natively geologic materials and can't detect by dynamic techniques. Rock safety monitoring is usually detected by sensors. Because standard monitoring technique is not defective for the with geotechnical engineering, the distributed optical sliding sensing technique is developed and 2 model tests are carried out to detecting rock mass sliding and steel-concrete interface damage. The relation equation of sliding distance-optical loss is obtained. There are some creative results in this paper . Firstly, the modal strain energy method is put forward. Its detection index is the modal strain energy. The theoretic idea is that the change of strain field local crack damages and energy field profoundly and directly. Crack location is determined by modal strain energy change before and after structure cracking and crack depth is determined by frequency perturbation change. Results of dynamic model tests and numeric simulation show that the strain energy index is very sensitive to crack damages and the method is predominant. Secondly, the virtual flexibility matrix method is presented for structure's damages. Unit virtual force is exerted on component elements and generalized strain changeable ratios of the undamaged and damaged structures are obtained. The relation is built between flexibility matrix and damaged element, and the damage location and damage severity can be determined. Using few modal parameters could achieved better precision, which coincides with actual engineering conditions. The method do not need numeric model parameters such as mass matrix. Hunce the numeric model errors have no influence on identification results. This physical meaning of the method and its computation speed are simple. Model tests approved this method of highly precision for damage location and damage severity could be estimated. Thirdly, the BP network model is constructed for data processing and damage identification. A intelligent method is built. For tests data processing ,BP model is convenient, high accurate and suitable to engineering practice. Fourth, model tests of distributed optical sensing for rock mass sliding and steel-concrete interface sliding monitoring are realized by the triaxial shear test of soil. The difficulty of optical-mechanics data collection to instantaneous brittleness fracture of the solid material is solved. The data and relation curve for slidingdistance-optical loss are obtained. The probability of typical application of the optical-fiber sensing to geologic disaster monitoring and safety monitoring of slope engineering and dam foundation, and steel-concrete interface damage has been showed.
|
PDF Full Text Request