Summary
Factors and mechanisms leading to sanding are described within an
integrated-rock and soil-mechanics framework. While the conventional
sanding models generally consider a single-mechanism for sanding, namely the
critical depletion resulting in rock disaggregation, the proposed approach
considers the interplay of several mechanisms that can lead to the rock breakup
and sand transport. One important difference is that rock disaggregation
is not seen to represent the onset of sanding, because the sand mass can offer
significant resistance from frictional properties, interlocking of sand grains,
and arching. The approach presented here can be used to explain why
sanding in the field tends to be episodic, and how depletion, which is a major
factor in rock breakup, can be highly effective in holding broken-up sand
grains together and, in fact, become a sand-stabilizing agent.
The proposed approach is used in discussing sanding at several wells in two
different fields. These wells have been in production for several years
and show that sanding cannot be linked to just one unique mechanism (e.g.,
depletion). However, once all mechanisms for sanding are incorporated, a
more consistent analysis can be used by completion and production engineers to
make more objective and pragmatic decisions in managing sanding while
maximizing production over the life of the well.
Introduction
While a great deal of work has been done in the general area of sand
production (Veeken et al. 1991; Weingarten and Perkins 1995; Risnes et al.
1982; Morita 1994; Sanfilippo et al. 1995, 1997; Tronvoll and Halleck 1994;
Tronvoll et al. 1997; Papamichos and Malmanger 1999; Morita and Boyd 1991;
Bradford and Cook 1994; Van den Hoek et al. 1996; Vardoulakis and
Papanastasious 1988; Willson 1996; Morita et al. 1996), most of the approaches
used for practical applications are on based on the assumption that the onset
of sand production is represented by the failure of the perforation-tunnel
wall, which is generally determined using the thick-wall-cylinder (TWC)
strength test. Such an approach is well suited to predicting the maximum
depletion in relatively competent rocks, particularly if they have brittle
behavior. But what about weak-to-totally unconsolidated rocks having an
almost-zero TWC strength, yet remaining stable under reasonably high drawdown
(DD) and, in fact, showing an increase in stability with depletion (field
examples presented later)? How should the DD strategy in terms of rate of
change and magnitude be adjusted as the rock undergoes a structural change from
a cemented formation to a totally disaggregated sand mass?
Strictly speaking, the conventional techniques for sanding prediction, which
are based on Geertsma’s (1985) equations, disclose the increase in confining
pressure required to fail a perforation tunnel or the wellbore cavity. In
practice, the approach provides an indication of the depletion that can be
sustained before the weakest perforation tunnel undergoes a significant
deformation, leading to its disaggregation (normally referred to as the
critical bottomhole reservoir pressure). This single-case-solution
scenario, which is not coupled with fluid flow, does not provide options to
make objective assessment of the risks at different stages in the well’s life
and the most effective contingencies to mitigate such risks [see Vaziri et al.
(2002a) for a full discussion of the past work in this area, formulations used,
and some of the limitations). For the base case of no active sand
control, operators would like to know, at any stage, how much sand will be
produced (rate and duration) for a given production strategy (e.g., maximum DD,
and rate of bean-up/shutdown frequency) and other changes in the reservoir
conditions, such as water cut (WC). By better understanding the roles of
multiple variables, one is enabled to choose the optimal completion method over
the life of the well. For a more comprehensive discussion of these issues
see Vaziri (2004).
©
2006. Society of Petroleum Engineers
View full textPDF
(
5,337 KB
)
History
- Original manuscript received:
6 January 2005
- Meeting paper published:
26 September 2004
- Revised manuscript received:
8 November 2005
- Manuscript approved:
8 January 2006
- Version of record:
20 November 2006
View Cart
