Thesis--M. Phil., University of Hong Kong.

Thesis--Ph. D., University of Hong Kong. / Mimeographed.

No description available.

No description available.

No description available.

No description available.

An experimental investigation into the direct shear behaviour of steel fibre reinforced composites utilising discrete fibres at pre-defined angles and fibres randomly distributed is described. The direct shear tests encompassed the complete range of loading from its initial application to failure of the double L-shaped push-off specimens. Hooked-ended and straight steel fibres were used in the tests with the fibres oriented at angles of ??75??, ??60??, ??45??, ??30??, ??15?? and 0?? with respect to a plane normal to the loading direction. The embedment lengths of the fibres, related to the total fibre length lf , each side of the shear plane were 0.5lf :0.5lf and 0.25lf :0.75lf . In addition to the single fibre tests, tests were conducted on randomly oriented steel fibre reinforced composites with fibre volumes of 0.005, 0.010, 0.015, and 0.020 with hooked-ended and straight steel fibres. In addition to the tests outlined above, a series of non-destructive tests employing radiographic techniques was carried out to produce photographic images of events taking place of fibres pulling out from a cementitious element. The tests consisted of hooked-ended steel fibres oriented at angles of -60??, -30??, 0??, +30?? and +60?? to the cracking plane and straight fibres oriented at angles of -60??, 0?? and +60??. The non destructive technique allowed the internal behaviour occurring within the specimen along the shear plane to be investigated without impacting on the direct shear tests. The angle of the fibre to the interface plane is an important parameter in determining the behaviour of the fibres under load and for the mode of failure; viz fibre pullout or fibre fracture. The effect of the end hook on behavioural aspects becomes increasingly less significant for more acute fibre angles where bending and snubbing effects become increasingly influential on the load versus displacement behaviour and mode of failure. Contrary to expectations, the fibre embedment length had little influence over the peak loads attained for the discrete fibre tests and, in a number of specimens, fibres pulled out from the longer embedded side. This observation is contrary to the generally accepted assumption that a fibre remains rigidly embedded on the long side and pulls out from short side. The traditional role that uniform bond stresses along a fibre length and friction have played in the description of fibre behaviour is not as significant as previously reported, other effects such as snubbing are more important in anchoring a fibre. Various models need to be revised with this observation in mind. The experimental results and observations from the discrete hooked-ended and straight steel fibres investigation are incorporated in the development of a behavioural model, the Variable Engagement Model II (VEMII). The VEMII describes the behaviour of randomly oriented discontinuous steel fibre reinforced composites loaded in shear. The model is verified against a series of randomly distributed fibre reinforced mortar specimens carried out in this study. Two forms of models are analysed: 1) a model based on the observation of lumped shear stresses at the fibre hook and in the snubbing zone; and 2) a uniform fibre bond stress applied along the embedded part of the fibre. The lumped bond stress approach and the uniform approach were found to give reasonable comparisons with the test data for the hooked-ended fibres but were conservative for the straight fibres. The VEMII confirms the applicability of the uniform bond approach adopted by previous researchers even though it does not correspond to the observations of fibre pullout behaviour of single fibres. The VEMII model provides a versatile approach that can also be applied to hybrid fibre combinations.

The first section of this thesis (Chapter 2) describes the creation and analysis of a database on concrete and masonry dam incidents known as CONGDATA. The aim was to carry out as complete a study of concrete and masonry dam incidents as was practicable, with a greater emphasis than in other studies on the geology, mode of failure, and the warning signs that were observed. This analysis was used to develop a method of very approximately assessing probabilities of failure. This can be used in initial risk assessments of large concrete and masonry dams along with analysis of stability for various annual exceedance probability floods. The second and main section of this thesis (Chapters 3-6) had its origins in the results of Chapter 2 and the general interests of the author. It was found that failure through the foundation was common in the list of dams analysed and that information on how to assess the strength of the foundations of dams on rock masses was limited. This section applies to all applications of rock mass strength such as the stability of rock slopes. Methods used for assessing the shear strength of jointed rock masses are based on empirical criteria. As a general rule such criteria are based on laboratory scale specimens with very little, and often no, field validation. The Hoek-Brown empirical rock mass failure criterion was developed in 1980 for hard rock masses. Since its development it has become virtually universally accepted and is now used for all types of rock masses and in all stress regimes. This thesis uses case studies and databases of intact rock and rockfill triaxial tests collated by the author to review the current Hoek-Brown criterion. The results highlight the inability of the criterion to fit all types of intact rock and poor quality rock masses. This arose predominately due to the exponent a being restrained to approximately 0.5 to 0.62 and using rock type as a predictor of mi. Modifications to the equations for determining the Hoek-Brown parameters are provided that overcome these problems. In the course of reviewing the Hoek-Brown criterion new equations were derived for estimating the shear strength of intact rock and rockfill. Empirical slope design curves have also been developed for use as a preliminary tool for slope design.

Thesis (Ph. D.)--University of New South Wales, 2002. / Also available online.

Thesis (M.S.)--Virginia Polytechnic Institute and State University, 1991. / Vita. Abstract. Includes bibliographical references (leaves 46-49). Also available via the Internet.