THE EFFECT OF GROUND MOTION UNCERTAINTY ON THE SEISMIC RELIABILITY OF GEOTECHNICAL SYSTEMS

Abstract

The ground motion variability is typically not included in earthquake design. The main goal of this research was to quantify the effect of the ground motion intensity variability on the seismic reliability of geotechnical systems. This is to calibrate the seismic design of geotechnical systems included in codes and to assess if the standard of practice is adequate to estimate the seismic resiliency of geotechnical systems not regulated by design codes. Due to the complexity of the problem, a simplified probabilistic model was developed, which is based on the execution of few numerical analyses and the use of simple functions correlating the random variables of the problem. The simplified probabilistic model still requires numerical integration, however integration can be performed using commercial spreadsheets. The simplified probabilistic model overcomes the distinction between epistemic and aleatoric uncertainty. It is believed that this distinction is not real and that the total estimated variability of the ground motion is necessary for design and performance assessment. The simplified probabilistic model was applied to three problems described below. For the problems where no target performance is available, target reliability indexes were proposed based on available literature and engineering judgement. First, geotechnical resistance factors were determined for single piles subject to vertical and horizontal seismic loads. The results of the calibration produced geotechnical resistance factors which are smaller than those indicated in the Canadian Highway Bridge Design Code (2019). During this effort, an algebraic approximate solution of the probabilistic model was developed. The algebraic expression can be used to check future calibration executed using either the simplified probabilistic model or more complex methodologies. It was found that, when considering different sites having the same load and soil distributions, the calculated probability of failure depends on the local ground motion hazard. As a consequence, to achieve the same reliability level at different sites, either different geotechnical resistance factor for different seismic regions should be selected, or the national ground motion hazard should be adjusted to achieve a uniform risk level. The author of this research believes that the second approach is the only one able to achieve similar levels of geotechnical risk throughout Canada. Second, the effect of the ground motion variability on the probability of failure of tailing dams was assessed using the same simplified probabilistic model used for the pile problem. There is no official dam “design code” in Canada. The design standard used in practice is the Canadian Dam Association Manual (CDA, 2013), which is in reality a code as it is referred to by provincial regulation of Canada. The performance targets given by CDA (2013) are in terms of loss of life, which are conditional on the occurrence of a dam seismic failure. As result, an accurate estimation of the dam probability of seismic failure is essential. It was found that the inclusion of the ground motion variability has a profound impact on the calculated probability of failure and on the dam classification and as such it should be always included in the calculation of the probability of failure. The third problem investigated in this research concerns how the ground motion uncertainty affects the calculated probability of damage caused by induced seismicity in an area located in the northeast of British Columbia. The simplified probabilistic model was applied to this problem, including a conditional probability of failure related to structural damage. Since damage states and associated acceptable reliability levels for these problems are not defined, reliability indexes were derived from existing codes or literature and proposed for the area. In addition, a probabilistic seismic hazard model was developed to describe the non-Poissonian nature of the problem.