Nonlinear Behavior Identification Of Structures With Dynamic Response Measurements

Due to integration of corrosion, overloading, structural member defect, unavoidable material aging, as well as lacking of proper maintenance, damages accumulate gradually during the lifetime of engineering structures, which reduce the load carrying capability and the reliability of structures. In particular, strong dynamic loadings such as earthquake and typhoon will induce more severe damages. To a great extent, the accumulation of damage which is a typically nonlinear process usually causes the failure of structure. Meanwhile, lots of structural components and materials, lighter and more flexible, are used as the needs of engineering development. These components and materials are characteristic of nonlinearity even more. Therefore, it is of great significance to identify the damages and assess the structural health state through analyzing nonlinearities. Unfortunately, nonlinear system identification (NSI) methods which can effectively identify a wide range of nonlinear behaviors are still few. Based on the above background, the study includes theoretical research, numerical simulation and experiment validation.(1) A data-based identification approach for general nonlinear system is developed. The method starts from equation of motion. Through Least-square method and equation of motion, the nonlinear restoring force can be identified only using dynamic response and excitation measurement time series.(2) In the numerical simulation, take a single-degree-of-freedom system as an example, several different nonlinear models were introduced into this system. Thereinto, the double flag model was used to mimic the macroscopical mechanical behavior of shape memory alloy wires, while a piecewise linear model used to mimic the behavior of the deformable displacement restraint device. And research that performs the inverse analysis using double flag model is rare. The identified effect was reflected mainly by identified parameters and figures of restoring force versus displacement. Results show that the methodology developed in the study is feasible and effective qualitatively and quantitively.(3) Under the guarantee of numerical simulation, a one-story frame model was built in laboratory. Several nonlinear elements such as SMA wires, two dampers based on SMA wires and deformable displacement restraint device, were incorporated into this model to mimic kinds of nonlinearities. Using the external excitation and corresponding acceleration time series, nonlinear restoring force can be identified utilizing the same method and then employed to compare with the measured one. Results validate the performance of the methodology from the perspectives of experiment. And experimental study like these is not very common among the field of nonlinear system identification. Moreover, two retrofitted SMA dampers developed in this study can not only dissipate seismic energy, but also can monitor the performance of damper in real-time. Both have good application prospect in civil and mechanical engineering.(4) Take a3-degree-of-freedom numerical model as an example, the physical and other parameters of a nonlinear system are identified utilizing the advantage of extended Kalman filter method under the condition of limited outputs. Only are known these parameters, which lay a foundation to calculate the total and nonlinear restoring forces.