Summary
We studied the stability of multilateral (ML) junctions in a combined
experimental and numerical modeling program. The experiments were carried out
in a true triaxial machine on large cubic blocks (40 cm) of weak triassic
sandstone with two holes intersecting. Six tests have been performed with two
different geometrical configurations and three different stress states. The
experimental results are presented and compared with numerical modeling
obtained with finite-element software developed for assessing the integrity of
rock surrounding an ML junction.
Introduction
Drilling inclined wells through producing strata can greatly improve
reservoir drainage and hydrocarbon recovery. The horizontal sections are
accessed through multiple inclined wells drilled from a relatively small
footprint in many or all directions, something that allows better exploitation
of offshore platforms and land rigs that are under economic and environmental
restrictions. Drilling inclined and horizontal wells, though, is more difficult
and more expensive, because of wellbore instabilities. A particular area of
concern is the integrity of the rock near an ML junction. The junction is the
region in which a second wellbore (lateral) takes off from the main wellbore
(parent). In the terminology used for ML junctions, different levels are
defined according to whether there is mechanical and/or hydraulic integration
between lateral and parent holes. By the ML Levels 1 and 2 definitions, the
rock at the junction is not supported mechanically with cemented casing, so the
integrity of the rock around the area of two intersecting tubes becomes very
important in terms of stability.
We performed physical tests on large cubic blocks (40 cm) with two holes
intersecting, in the true triaxial cell of the U. of Lille. The tested rock is
weak Triassic sandstone called “Grés des Vosges.” Six tests have been performed
with two different geometrical configurations (lateral differently oriented
with respect to the main bore) and three different stress states (two blocks
with a hydrostatic stress state and three blocks with anisotropic stress
state). The blocks were loaded to generate breakouts of the borehole wall in
various directions. The deformation of the borehole walls and the development
of breakouts are monitored in real time with a video camera placed in the main
bore. The image is then analyzed by image-processing software. Graphs of the
relative diametric decrease (convergence of the borehole wall) in various
directions can then be plotted against loading. After testing, the blocks were
cut in cross sections perpendicular to the parent-hole axis at different
distances from the junction. The experimental technique and results will be
presented in the Rock Characterization and Experimental Procedure and the
Experimental Results sections, later in this paper.
We compared the experimental results with numerical modeling obtained with
software developed for assessing the integrity of rock surrounding an ML
junction. The tool was developed using finite-element analysis and a graphical
user interface for providing the input data and visualizing the results. The
analysis is based on a generalized plane-strain formulation that is carried out
in cross sections in succession, perpendicular to the parent-hole axis. We
compared the load level at which breakouts are initiated. Results are
presented for two different junction geometries and three different loading
paths (in isotropic loading, all three applied stresses on the block were the
same, and in anisotropic loading, two of the applied stresses were the same,
and the third stress had a different magnitude). The results, based on the
elastic/brittle analysis, qualitatively reproduce those obtained during the
experimental tests.
© 2006. Society of Petroleum Engineers
View full textPDF
(
1,011 KB
)
History
- Original manuscript received:
28 July 2004
- Revised manuscript received:
27 October 2005
- Manuscript approved:
10 November 2005
- Version of record:
20 March 2006
View Cart
