| Numerical calculation is one of the main parts in pseudodynamic test, and the choice of the numerical integration methods directly affects the validity and the reliability of the testing results. The study on the explicit and implicit numerical integration methods for pseudodynamic tests was implemented from the aspects of stability, accuracy, numerical dissipation and error propagation. Furthermore, an optimized generalized explicit method which can overcome some inherent limitation through introducing dissipation parameter based on the duouble-βmethod and the explicit Newmark method was proposed. Then, the substructure pseudodynamic test on a steel frame model was carried out by adopting the explicit-implicit numerical integration method.The investigated results indicate that the modified Newmark explicit method has small numerical damping for the fundamental mode and significant numerical damping for the higher modes, and theα-function and theγ- function dissipation method can have desired numerical dissipation. The proposed optimized generalized explicit method can effectively suppress the spurious growth of high frequency responses, and possesses much better error propagation characteristics. The parameter a which can produce numerical damping can not control the experimental cumulative errors in the HHT—αmethod, and theα—C method and theα—OSM method can effectively control the cumulative errors while theα—OS method does not have this function.The analysis on the results of the substructure pseudodynamictest shows that whether the single-degree of freedom or the multi-degree of freedom tested substructure is adopted, the seismic responses of the whole structure under the certain earthquake wave can be obtained accurately when the elastoplastic responses may only occur in the bottom story, and the computer simulation on elastic stories almost has no effect on the seismic responses of the whole structure in substructure pseudodynamic tests.
|
PDF Full Text Request