结构损伤自诊断和寿命监测智能系统

Combining the knowledges from several modern techniques such as damage mechanics, wave propogation theory, neural networks, engineering materials and structural enginnering, the study has expoilted the relationship between structural health, structural damage and structural vibration characteristics, and has built the wave-propagation and adaptive neural-networks-based computation model for the global/local damage evaluation in civil engineering tructures. Based on the computation model, a software which is situable to detect the damage of simple structures has been realized and applied successfully in some projects. Theoretical and experimental study of the characteristics of different structural damages and sensitivities of damage to different vibration parameters has been conducted. The method based on wave-propagation, substructure technique, smart materials and structures, modern computation and analysis technique has been proposed to identify the damage of large and complex structures, and the preliminary study in this respect has been carried out. Furthermore, the study has also addressed the relation between the health and damage for concrete structures and the key problems when the intelligent sensors are cast inside the concrete structures by using experimental method and optimal theory. The present research is the basis and important reference for the further study, even the realization of on-line health monitoring of large and complex civil engineering structures. The results from the model experiments and actual engineering application indicate that the proposed approach and way are scientific, and the achievements are of important theoretical and applicant value.

利用模糊神经网络建立结构物理参数与结构模态参数间的非线性映射关系，将损伤力学与神经网络结合起来建立结构物理参数与结构损伤部位和大小间的非线性关系，然后依据这种非线性映射关系，利用智能材料结构获得的结构模态特性，确定结构的损伤状况。随着特殊载荷的多发性以及结构物使用年限的增加，进行结构损伤自诊断和寿命监测具有重要意义。