PROBABILISTIC CHARACTERIZATION OF SHEAR STRENGTH USING RESULTS OF DIRECT SHEAR TESTS

Laboratory direct shear testing of discontinuities in orientated rock specimens provides a data base that represents the relationship between applied normal stress and resulting shear strength. Using a modified power curve model and Newton's approximate procedure, a non-linear regression analysis can be applied to the date of individual test specimens. The resulting regression parameters and standard errors can be used in a weighted, non-linear regression analysis to determine the modified power curve model that best describes the shear strength of a group of like specimens. This group regression relies on linearization methods and iterative calculations to obtain the model, which serves to define the expected shear strength as a function of normal stress. The variance of the expected shear strength at a specified normal stress can also be calculated. The probability density function of expected shear strength should be of the type that allows only positive values of shear strength, such as gamma probability density. However, a normal density can be used if care is taken to avoid negative shear strengths. Regardless of which function type is selected, its parameters are defined by the mean (expected value) and variance of the shear strength at any specified normal stress. Consequently, once an effective normal stress is calculated in a stability analysis, the statistical distribution of the corresponding shear strength along the discontinuity of interest can be readily determined. Scale effects can be incorporated into the stability analysis using a probability distribution of waviness (roughness) estimated from mapping data. (TRRL)

Media Info

  • Features: Figures; References;
  • Pagination: p. 273-276
  • Serial:
    • GEOTECHNIQUE
    • Volume: 34
    • Issue Number: 2
    • Publisher: Thomas Telford Limited
    • ISSN: 0016-8505

Subject/Index Terms

Filing Info

  • Accession Number: 00392422
  • Record Type: Publication
  • Source Agency: Transport Research Laboratory
  • Files: ITRD, TRIS
  • Created Date: Mar 29 1985 12:00AM