In this part, the parameter functions for clay brick masonry appearing in the non-linear model are established for the wall material used in the experiments by means of experimental data and a particular type of optimization. This special optimization makes use of the fact that the wall behaves linearly at the intensity level of each excitation, as described in Part 1, and involves matching in frequency space the experimental and theoretical complex frequency response functions relating the Fourier transforms of the top and base accelerations of the wall.
It is found that the envelope curves for the parameter functions are bilinear and that the dynamic values of mechanical properties of masonry differ greatly from their static values.
The completed model is appraised by comparing how the wall will respond to strong earthquake excitations when predicted using the model and how it actually responded on the shaking table. The predicted response is remarkably close to the experimental.
[1]
M. R. Osborne,et al.
Methods for unconstrained optimization problems
,
1968
.
[2]
Janusz S. Kowalik,et al.
Iterative methods for nonlinear optimization problems
,
1972
.
[3]
Douglass J. Wilde,et al.
Foundations of Optimization.
,
1967
.
[4]
D A Pierre,et al.
Optimization Theory with Applications
,
1986
.
[5]
P. R. Adby,et al.
Introduction to optimization methods
,
1974
.
[6]
E. Polak.
Introduction to linear and nonlinear programming
,
1973
.