Suitable Intensity Measures for Evaluation of Continues Buried Steel Pipelines in Performance-Based Context

Investigation of the correlation between engineering demand parameter (EDP) and intensity measures (IMs) has received substantial attention in performance-based earthquake engineering for prediction of seismic demand of structures.In this study the seismic demands of buried steel pipelines are investigated in a performance-based context. Several nonlinear dynamic analyses of two buried steel pipe models with different D/t, H/D ratios and different soil properties and different pressures, performed under a suite of far-field earthquake ground motion records were scaled to several intensity levels to investigate the behavior of buried pipeline from elastic response to failure. Several scalar ground motion intensity measures (IMs) are used to investigate their correlation with engineering demand parameter (EDP) which is measured in terms of peak axial compressive strain in critical section. Using the regression analysis the efficiency and sufficiency of candidates IMs is investigated.
To investigate the effects of different material and geometrical properties, two buried pipeline of API 5L Grade X65 models with different pipe and soil properties are considered. To simulate soil effects on pipe in axial, transverse and vertical directions bilinear force displacement curves (elastic-perfectly plastic) representation of soil are employed based on suggestions of the American Lifeline Alliance. The FEM was used in the analyses. The buried pipeline and the surrounding soil are modeled using shell, spring and damper elements. Each node of the model was connected to three spring-dashpots.
Before deciding which ground motion IMs correlate well with seismic demand on buried pipes the first question to be answered is: how is the seismic demand measured? Usual failure modes of continuous buried pipelines are tensile ruptures or buckling because of compressive strains. Compressive straines that result in bukcling are smaller than strains induce tensile failure. Therefore, the peak axial compression strain at the critical section would seem the obvious candidate to use for EDP of buried pipe. It is necessary to examine a wide range of potential IMs for determining the best IM for evaluating the buried pipe response. Therefore a total of 16 different IMs are considered.
Using an efficient IM results in smaller variability in the structural response for any given IM. Chosing an efficient IM causes the number of analyses and earthquake records needed to evaluate the probability of exceedance of each value of EDP given the value of IM to be reduced. In this paper, a one-parameter log-log linear regression of peak axial compressive strain on IM is utilized in evaluating the efficiency of each alternative IM. A sufficient IM results in EDP conditionally independent, for a given IM, of earthquake magnitude M and the source to site distance R. Using a sufficient IM, yields ignoring the effects of magnitude and distance in accurately predicting of EDP. Quantifying the sufficiency of IM is done by using the one-parameter regression of peak axial compressive strain on M or R for a given IM. Among the models investigated in this study it was seen that RMSd is the optimal IM for buried steel pipelines based on efficincy and sufficiency conceptions.