Summary
Field data from StatoilHydro on hydrate-plug formation in the Tommeliten
gas/condensate field are compared to predictions of the hydrate-growth model
(CSMHyK-OLGA) for four typical operating scenarios: steady-state operation with
failure of inhibitor injection, restart of an uninhibited line, restart of an
underinhibited line, and restart of a depressurized line.
Although the CSMHyK model was designed for oil flowlines, the model is able
to predict the correct time scale for hydrate-plug formation in this
gas/condensate tieback. The predicted locations of the plugs are often farther
upstream than observed in the field trials. This is mainly because of the
assumption of a "hydrate/oil slip factor" of zero, which forces the
hydrate to accumulate where it initially formed. In reality, hydrate
agglomerates would be carried further downstream before eventually jamming in
dips.
Predicting where and when hydrate plugs will form in subsea tiebacks is of
increasing importance as the industry strives to manage the risk of plugging in
oil and gas flowlines while minimizing the use of costly and environmentally
harmful chemicals for hydrate inhibition. The Colorado School of Mines has been
developing the CSMHyK model for the past 5 years, in collaboration with the SPT
Group and several leading energy companies.
Introduction
Hydrates continue to be the most prevalent flow-assurance problem in
offshore oil and gas operations: an order of magnitude worse than waxes and two
orders of magnitude worse than asphaltenes (Sloan and Koh 2008). The risk of
hydrate plugging increases as the industry moves into deeper water with
corresponding higher pressures from the additional liquid head and to longer
tiebacks in which the production fluids cool deep into the hydrate-stability
zone. A recent review of hydrate-plug-prevention strategies is provided by
Mokhatab et al. (2007).
The cost of thermodynamically inhibiting such tiebacks under steady-state
and transient operations can be prohibitive. It is often not possible for the
flow-assurance engineer to avoid the hydrate-stability zone in all foreseeable
operating scenarios. Instead, a risk-management approach is often adopted to
prevent hydrate-plug formation (Kinnari et al. 2006; Pausche et al. 2002).
Because of the potentially severe economic impact of forming a hydrate plug, it
is critical to develop models that the flow-assurance engineer can confidently
apply when making a risk assessment of a new field design or restart
procedure.
A number of models have been proposed previously for hydrate-formation rates
in laboratory-scale systems. The models are based either on chemical-kinetics
equations (Vysniauskas and Bishnoi 1983; Englezos et al. 1987; Christiansen and
Sloan 1995; Lee et al. 2005) or interfacial-mass-transfer resistances to
hydrate formation (Skovborg and Rasmussen 1994). The applicability of these
models is generally limited to apparatuses of a geometry similar to that of the
apparatus in which the measurements were taken. The Colorado School of Mines
has been developing a model for hydrate formation in industrial-scale flowlines
in conjunction with SPT Group since 2003. The model, CSMHyK, is incorporated as
a plug-in module in the transient multiphase-flow simulator, OLGA.
CSMHyK was developed initially for oil flowlines, but has also proved a
valuable tool for Chevron when making design decisions for new field
developments. Key to the development of CSMHyK has been the extensive testing
against industrial-flowloop data from ExxonMobil and the University of Tulsa
(Boxall et al. 2008) and against industrial field data as described in this
paper.
No published prior method exists that enables predictions similar to this
work. Recently, Calsep has been developing an alternative model for hydrate
formation in industrial systems. The model, Flowasta, is based on mass-transfer
rates between the hydrocarbon liquid and the water phase. The model currently
relies on fitted paramerers for the mass-transfer resistances, and work is in
progress to validate the model against industrial-scale flowloops.
© 2009. Society of Petroleum Engineers
View full textPDF
(
754 KB
)
History
- Original manuscript received:
1 July 2008
- Meeting paper published:
21 September 2008
- Revised manuscript received:
12 December 2008
- Manuscript approved:
19 December 2008
- Published online:
27 August 2009
- Version of record:
25 November 2009
View Cart
