Determination of the Strength of an Ice Sheet From Small Sample Data

Abstract

ABSTRACT The results of various types of tests on small-scale samples of ice are discussed and compared. Some of the results, particularly those associated with varying load rates, are explained by introducing the effects of time dependent plasticity and grain size. A simple model is presented which explains some of the load rate effects in uniaxial specimens, and on the basis of this model, a more refined computer model has been developed which simulates small-scale sample tests, the sample being developed from an assembly of grains. The potential application of small-scale tests to predicting strengths of ice sheets is discussed. INTRODUCTION Due to the great difficulty and expense associated with performing full-scale tests on ice sheets, many investigators have sought to supplement these limited results with the results of tests on small samples. It is assumed that the results of these small-scale tests, when properly interpreted, can help to predict. the strength of an ice sheet and its dependence on the parameters of temperature, salinity, load rate, and thickness. Unfortunately, many of the reported small-scale results fail to agree amongst themselves and with some reported full-scale results, and these inconsistencies have yet to be adequately explained. This paper presents the preliminary results of a study designed to improve the understanding of small-scale test results on ice, with the intention of applying the results to full-scale ice sheets. Of primary interest is to determine just what the stress distribution is, in what amounts to an inhomogeneous inelastic material. This is quite a difficult problem in itself, and must be tackled first before small-scale test results can be applied to the prediction of strengths of ice sheets. PREVIOUS TEST RESULTS The types of tests which have received the greatest attention are uniaxial tensile and compressive tests, ring tensile tests, Brazil tests, and flexural tests. These tests are shown schematically in figure 1. In all cases the maximum stress at failure has been computed assuming that the sample is a linearly elastic, homogeneous solid. With the exception of the uniaxial compression test, maximum tensile stress is generally assumed to govern failure. The general trend for these results has been uniaxial compression +ring tensile + flexural uniaxial tension in order of decreasing stress at failure. The most desirable test for tensile strength is the uniaxial tensile, since it produces the most uniform stress field. In addition the uniaxial tensile data of Peyton2 and Dykins3 seem to produce better agreement with full-scale beams cut out of ice sheets, as reported by Dykins4 and Tabata than do any of the other tests. Unfortunately, the uniaxial tensile test is not suited for field work, whereas the ring tensile test, which has yielded the most unrealistic values for tensile strength, is the simplest to perform.