Question by John (for Bukovansky and Piercy) i) In comparing two finite element methods, one using a simple straight-forward approach and the other a more sophisticated Goodman-type model, is it worthwhile to use the more sophisticated model? ii) Why not use a routine zero tension program?
Prof. John asked whether the finite element analysis, which included the joint elements, is necessary, in addition to the analysis without the Joint elements. As can be seen from both analyses, the model with Joints seems to provide much more realistic results. The authors believe that only this model should be used for the final engineering design. The described finite element models were used for additional analyses of cuts in deep soils. No-tension analyses, mentioned by Prof. John, could certainly be applied for this problem. Discussion by Robertson In an essentially similar problem area in South Africa a rather different design philosopy was adopted. The authors may care to comment on such an approach. The gorge of Buffelspoort is formed through a simple anticlinal fold in Table Mountain Series Sandstone. Total gorge length is approximately 13 km. Through the gorge flows a major river which in a 50 year flood would flood the gorge to a height of 8 m. The restricted flow conditions require the road be located some 10 m up the gorge walls with little fill being permitted to further restrict flow. Resulting cuts would be high and most unattractive. To minimise cut heights vertical or overhanging cuts were considered where ever possible. Natural overhangs of equivalent dimensions suggested that this might be possible. The Jointing patterns as measured in the gorge conform to the classical patterns anticipated for the simple tectonic stress situation which gave rise to the anticlinal fold. Typical slope cuts were first carefully mapped for structural detail. Fracture data almost invariably were as predicted form the major lineations observed in air photos. Kinematic modes of failure were determined from great circle analyses on stereoplots. Stability analyses were generally made numerically. From such analyses failure situations and design measures were determined. Failure modes included plane failure on bedding, wedge failure and toppling. Friction angle estimates were made from field observations of regions where shear failure had occurred along similar features. A value of 36º was obtained for bedding joints. High overhang cuts would be controlled largely by cohesion on the vertically intersecting wedges formed by Joints. Estimates of such cohesion were made from back analysis of existing overhangs. It is our intention to form overhang cuts by blasting out sections of the cut, leaving temporary support pillars. The overhang will then be instrumented. The support pillars are then to be blasted out and the performance of the overhang evaluated. Should this be satisfactory the overhang cuts will be retained.
In considering the quality control of rock anchors should not corrosion control be stressed very strongly?
This is a very valid point.