ABSTRACT

INTRODUCTION

Early attempts to predict the creep response of rock salt around underground rooms at the Waste Isolation Pilot Plant (WIPP) produced closure estimates that were one-third to one-fourth of values measured in situ (Morgan et. at., 1985). A subsequent study (Morgan, et. at., 1986) of the WIPP reference elastic secondary creep model (Krieg, 1984) used to make these predictions revealed that room closures and even closure rates could be increased by reducing the elastic constants. This study also indicated that a vertical cylindrical shaft configuration could be substituted for more complicated and expensive rectangular room configurations in studying constitutive parameters for rock salt. Sjaardema and Krieg (1986) used these results to determine how much the W1PP reference value of Yomtg's modulus E had to be reduced to increase the creep closure and closure rate of a hypothetical borehole in rock salt by factors of 3.5 to 4. They found that E had to be divided by 12.5 to produce the desired results.

Motivation for this empirical "fix" to the creep model was to provide realistic room closures in examining the consolidation of crushed salt, a proposed backfill for rootns at the WIPP. The empirical model was not intended to replace more scientifically based models for predicting the creep response of rock salt. Instead it was an expedient that allowed crushed salt investigations to proceed while other creep models were being improved. However, the empirical model was subsequently used to predict the response of several different WIPP room configurations, and the resulting closure estimates were surprisingly good when compared to field measurements (Munson et. al. 1986, Morgan & Krieg, 1988). Consequently, the model has received more use than was originally intended, and various hypotheses have been made as to why it produced such good results. One hypothesis was that damaged or microcracked rock, known to exist near the surfaces of WIPP excavations (Borns & Stormont, 1989), dominated room closure, and the reduced elastic modulus captured its behavior. Other speculation implied that the degraded modulus increased the elastic strains in the salt mass remote frown rotan surfaces to such an extent that the integrated effect produced the increased closures. However, none of these hypotheses explained the observation that the moduli reductions increased closure rates as well as closures.

In this paper, we present the results of shaft calculations, similar to those used by Sjaardema and Krieg, to investigate why dividing E by 12.5 "works." The goal of this investigation is not to justify or promote the empirical model, but instead to explore possible physical phenomena that it captures better than the WIPP reference and other creep models.

This content is only available via PDF.
You can access this article if you purchase or spend a download.