Experimental and Numerical Characterization of the Stress-Strain Behavior of Weak Sandstone Formations for Sanding Assessment

Abstract

In this study an extensive experimental program has been carried out in order to characterize the mechanical behavior of two weak sandstone formations of an offshore field for application to sand production modeling. The experimental tests included Scratch tests, Triaxial tests and Advanced thick wall cylinder tests (ATWC) where the sand production initiation and the cumulative sand produced were registered. Numerical simulations of experimental tests were then performed using an advanced elasto-plastic constitutive model. Triaxial tests simulations allowed calibrating the constitutive model parameters. These parameters were employed for the numerical simulation of the ATWC in order to determine the equivalent plastic strain threshold required to the onset of sand production observed in laboratory for sanding assessment. Results obtained highlight the importance to use a realistic representation of the rock behavior focusing on post-yield behavior in order to build confidence in model predictions.

Evaluating the Accuracy of Bonded Block Models for Prediction of Rockmass Analog Mechanical Behavior

“…However, this rigorous calibration is required to ensure that the micromechanical failure mechanisms of the model approximate those of the real rock [25][26][27][28][29]44,50]. Many of the previous SRMs developed after the jointed Blanco Mera granite specimens (see Table 1) were unable to capture some of the failure mechanisms seen in the jointed Blanco Mera granite laboratory specimens [17][18][19]31,33], in contrast to the more extensively calibrated models developed in this study. For example, in [18,31], the models were not calibrated to match CI and CD thresholds.…”

“…Another example is the discrete model developed by [33], which failed predominantly through displacement along the pre-existing joints, rather than through intact material damage, as was observed in the jointed specimens of Blano Mera granite [24]. The strength of the intact model was calibrated to be too high compared to the strength of the pre-existing joints, which resulted in incorrect failure mechanisms of the jointed models.…”

“…Numerical models-both BBMs and otherwise-have been previously created and calibrated to the Blanco Mera granite laboratory data from Ref. [24] [17][18][19][31][32][33]. However, each of these studies calibrated their models to both intact and jointed specimen data simultaneously, which does not ensure that the model is predictive under conditions different from which the model was calibrated for [34] (see Section 3.5 for further discussion on this topic).…”