A model study of rock foundation problems underneath a concrete gravity dam

Abstract

In recent years photoelastic analysis has been a most valuable aid to the study of the stress distribution around mining excavations, as simulated by models constructed of birefringent materials. More recently the technique has been used to study the deformation of rock specimens under load and to investigate apparent anomalies in their compressive strength. An interesting recent project employing this technique was a model study of rock foundation problems underneath a concrete gravity dam. The object of the investigation was to study the effect of various alternative structural details in the dam foundations. Several models were therefore constructed, each representing a cross section of the dam and the foundation beds to a scale of 1:200, the dam being represented by a concrete mortar cast three inches thick. The foundation beds at the site consisted of beds of limestone and friable brown lignite together with marl which became unconsolidated when wet. The preliminary tests indicated that there was a difficulty in making a reliable measurement of the compressive strength of the lignite. However, an investigation on the crushing strength of irregularly shaped fragments gave results from which a compressive strength of the order of 2,000 lb./sq.inch was deduced. In the selection of the model materials, those with similar mechanical and rheological characteristics to those of the prototype were used and these had elastic properties similar to those of the terrain and the structure so far as available information allowed. Cement mortar was chosen to represent the dam material and the limestone, and a gypsum plaster plus a porous weakener, such as wood, as the material to represent the lignite. Research was carried out on the behaviour of the model materials by means of uniaxial and triaxial compression tests. A two-dimensional study mainly of a qualitative nature was made first on one side of the models constructed using photoelastic sheet and observation by reflection polariscope at various conditions of load. Having determined the principal stress directions these were transferred to the rear face so that strain gauges could be fixed in suitable alignment. The models were also tested to failure after removing the photoelastic coatings and soaking the foundation beds in water. The effects of reinforcement by an added foundation key, by vertical rock bolts, and pre-stressed tendons were also investigated. The rock bolts were found to act as passive reinforcement during normal loading conditions and only became active strengthening elements when slip between the dam and foundation beds actually began. On the other hand, the tendons were found to be a very effective form of reinforcement giving a higher “coefficient of safety” than either the extension key or the rock bolts. The general conclusions are that the results of the model study described are primarily of a qualitative nature. Such a study is, however, able to pattern the behaviour of the situation more realistically than idealised models described in mathematical terms or photoelastic bench techniques. The combination of the use of photoelastic coatings and electrical strain gauges, is a very powerful one and may have considerable potential in many other fields of application.