An elasto-plastic finite element approach incorporating strain hardening and softening is used for the prediction of borehole stability. The material model is developed and calibrated on the basis of triaxial compression and hollow cylinder collapse tests, and is then used to investigate the stability of vertical and horizontal boreholes. The finite element calculations show localised failure patterns similar to those observed experimentally. The minimum mud pressures necessary to prevent borehole collapse, predicted by the elasto-plastic analyses, are compared with the values predicted by two elastic-brittle models, and, on average, are found to be considerably lower.
Borehole instability problems can dramatically increase the cost of drilling and completing wells. This is apparent from field observations such as stuck drillstring, cavings, excessively enlarged boreholes, logging difficulties, inadvertent side tracking and stuck casing. In the worst case, wells may have to be sidetracked and the troublesome section redrilled or the well may even be abandoned without having reached the planned target and with total loss of the drilling expenditure. Drilling problems are not often experienced in initial vertical exploration and appraisal wells. However, the drilling of (highly) deviated or even horizontal development wells, especially after production when some formations are depleted, is generally more prone to instability problems. The specific gravity of the drilling fluid, i.e. the mudweight, is the major parameter under the control of the operator which governs borehole stability. If the mudweight exceeds an upper limit (new) fracture(s) will be created or existing fractures will be reopened - causing mud losses. A lower mudweight limit is set by two requirements. Firstly, the mudweight must exceed the highest formation fluid pressure gradient in the open-hole section to prevent fluid influx into the wellbore. Secondly, the mudweight has to be sufficient to prevent hole collapse. This paper adresses the latter problem of accurately predicting the minimum mudweight required to prevent borehole collapse. There is an analytical solution to this problem for general borehole orientations, assuming an impermeable borehole wall and elastic-brittle rock behaviour (Bradley 1979). Hole collapse is considered to occur when the stress state at a point on the borehole wall violates a failure criterion. This approach is quick and easy to use. It has been extended to include the effects of a permeable borehole wall (Geertsma 1978, Hsiao 1987), anisotropic rock properties (Aadnoy 1988) and thermal stresses (Kaiser & Malony 1987, Guenot 1987). Experience indicates that mudweights used in the field are often substantially lower than those predicted by elastic-brittle models, especially for highly deviated wells. The EB method has therefore proven to have a limited quantitative predictive power for field applications. An exception might be the field-calibrated version as demonstrated by Fuh & Whitfill (1988), where the material model parameters are tuned to match field observations of hole collapse. Less conservative predictions result when elasto-plastic (EP) material behaviour is assumed as shown by its application to the problem of borehole stability by both Risnes et al. (1982), Bandis & Barton (1986), and Guenot (1988).
