Influence of Masing damping on 1D site response using equivalent linear and non-linear methods
Despite significant developments in one-dimensional (1D) site response methods using non-linear (NL) analysis, the Equivalent Linear (EL) analysis remains the most widely used approach by practitioners to investigate the total stress response of a soil deposit. While the EL method is implemented using both shear modulus reduction curves and damping curves from laboratory test results, NL Cyclic Stress (CS) models are often computed so that they are consistent with the backbone stress-strain characteristics, but accept some inconsistencies in the energy dissipated by the hysteresis loops that are represented using a Masing’s rule approach. As such, EL analysis often provides a conservative site amplification factor and ground response spectrum, while NL models can lead to an over-damping of the ground response spectrum. This paper discusses the impact of discrepancies in these damping formulations on 1D site response using the EL model and a non-linear CS model computed in OpenSees. Traditional EL analysis is performed, as well as a modified EL approach that changes the original formulation in order to capture frequency-dependent soil stiffness and soil damping parameters that are consistent with the frequency content of the vertically propagating shear waves. While it is largely recognized that Masing behaviour introduces a bias in the site response at large strains, it has been found that a moderate strain pattern within the soil profile can also lead to a substantial under-prediction of the ground response.