The aim of this project is to study the use of a particular non-linear system identification scheme to determine the in situ cyclic properties of soils, using records of the response of such soils to the strong ground shaking induced by large earthquakes. Soils are known to exhibit significant non- linear behavior even under moderate ground shaking. The degree of this non-linearity can have a large effect on the response of soils to strong ground shaking, and hence on the response of structures founded on these soils to the earthquake loading. The use of non- linear system identification methods is difficult and any particular scheme is limited in its application. This study is investigating the use of the Extended Kalman Filter to identify the in-situ cyclic properties of soils existing in shear beam type soil structures. The first phase of the research involves a numerical parametric study aimed at finding guidelines for such issues as: sensitivity of the initial estimates, the required number of parameters, the effects of measurement noise, the impact of ground shaking intensities, and the relationship between global iteration and convergence. The second phase applies these results to the downhole instrument records measured during earthquakes. Emphasis is placed on the consistency of the estimates and the accuracy of the estimates in comparison with laboratory test data.

Project Start
Project End
Budget Start
1990-08-15
Budget End
1993-07-31
Support Year
Fiscal Year
1990
Total Cost
$60,900
Indirect Cost
Name
University of Pittsburgh
Department
Type
DUNS #
City
Pittsburgh
State
PA
Country
United States
Zip Code
15213