close

Вход

Забыли?

вход по аккаунту

?

187100-MS

код для вставкиСкачать
SPE-187100-MS
Modeling of CO2 Leakage from CCS into Overlying Formations - Quest CCS
Monitoring Evaluation
Jeff Duer, Shell Canada Limited
Copyright 2017, Society of Petroleum Engineers
This paper was prepared for presentation at the SPE Annual Technical Conference and Exhibition held in San Antonio, Texas, USA, 9-11 October 2017.
This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents
of the paper have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect
any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written
consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may
not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.
Abstract
Commercial-scale deployment of Carbon Capture and Storage (CCS) as a viable greenhouse gas emissions
reduction technology requires that the CO2 be confidently contained in geological formations with no risk
of groundwater contamination. To ensure containment is adequately monitored, an evaluation of what a
potential leak could look like and how it can be detected is required. This paper is an example from the
Quest CCS operation in Alberta, Canada.
Quest commenced operation in August of 2015 at a rate of 1 MT/year into two injection wells. After
two years of operations, the project?s rigorous monitoring program has demonstrated that the reservoir
is behaving as expected and no leaks have been detected. However, hypothetical leak paths have been
investigated and modeled. Four scales of models have been used to evaluate the risks associated with the
hypothetical leak paths and therefore on containment: (1) Geological structural models - Regional Static
Model, Field Dynamic Model, (2) Legacy well - Well Brine Leak Path Model (3) CO2 Leakage - Injection
Well CO2 Leak Path Model, and (4) Leak detection - Cooking Lake Model. The results of the modelling were
used in the evaluation of the Quest project proposal, the current operating strategy, and the measurement,
monitoring, and verification (MMV) plan.
The results of the leak path models demonstrate that the risk of a CO2 leakage from the Quest storage
operation is very low. Regional modeling of the overburden confirmed that no leak pathways in the project
area could be identified. Field level dynamic modelling demonstrated that injected CO2 is not expected to
reach far field legacy wells, but the potential for elevated reservoir pressure displacing saline brine into
usable ground water could be a risk if insufficient well count. The impact of a brine leak path was modeled
and concluded to be negligible as the overlying under pressured cooking lake formation was concluded to
be an effective pressure sink. As the injection wells have the highest pressures and concentrations of CO2
in the reservoir, despite excellent wellbore integrity, they are the most likely location for a theoretical CO2
leak path. It was both concluded that the buoyancy force was a very slow moving affect and that the cooking
lake formation ultimately acts as a pressure sink. Therefore, the cooking lake was modeled to understand
what pressure response could be expected and whether a leak into those formations could be detected. It
was concluded that material leaks at the injection well would be differentiable from baseline pressure drift
at the monitoring wells.
2
SPE-187100-MS
Introduction
The Quest CO2 sequestration project in Canada has been in operation for two years. The CO2 source is an
oil sands upgrading facility (bitumen to synthetic crude) where 1 million tonnes of CO2 per year is captured
(1/3 of the emissions) and transported to the storage site via a 65km pipeline with 6 block valves (maximum
14km spacing). The compressor includes a dehydration unit to minimise water within the pipeline and
pressure is maintained above 9MPa to ensure dense phase CO2 throughout.
The storage formation is the Basal Cambrian Sands (BCS), a saline aquifer with an average 17% porosity
and 1,000mD permeability at about 2,000m depth. The storage facility consists of three well pads with an
injection well and a deep monitoring well on each pad, and several shallow groundwater wells. Conventional
drilling techniques were used with multiple steel casings for the injection wells, all cemented to the surface.
A comprehensive measurement, monitoring and verification (MMV) program is in place, with technologies
monitoring the atmosphere, biosphere, hydrosphere and geosphere, including considerable pre-injection
baseline data collection.
Quest has delivered safe CCS operations and has exceed production targets. The performance of the
capture facility has been very reliable, resulting in a higher rate of capture than expected. The pressure data
collected from deep monitoring wells shows no indication of any problems or leakage paths. The current
pressure build up in the reservoir is less than expected and reservoir properties are better than expected.
Total pressure build up in the near well-bore region is expected to be less than 2MPa by the end of project
life. After two years of operations, the Quest CCS facility is demonstrating the safe storage of CO2 in a
saline aquifer on a commercial scale.
The Quest modelling strategy was designed to address the specific risk of subsurface leak paths (Figure
1), and as such various scales of models were used in de-risking the project and to influence the operating
and monitoring strategies [1], [2]. This paper discusses the results of five of these modeling studies.
Figure 1?Schematic of potential leak paths at legacy well, injection well, and fault.
SPE-187100-MS
3
Regional Static Model
Regional modeling of the overburden confirmed that no leak pathways in the project area could be identified.
A 100 km x100 km regional model from the basement to overburden was constructed to understand and
visualize any possible leak pathways in the project (upper left corner of Figure 2) [2]. The static model
illustrates an absence of any faults or stratigraphic leak paths that would result in CO2 or BCS brine migration
out of BCS storage complex. Furthermore, sufficient thick, mappable seals and baffles over the area provide
a strong safeguard against any migration into ground water protection zone.
At the Quest site, the CO2 is injected into a deep saline aquifer, the Basal Cambrian Sands (BCS), located
at a depth of about 2 km below ground. The BCS directly overlies the PreCambrian granite basement (Figure
2). Above the BCS, a number of regionally extensive thick seals and baffles are present, including the
Middle Cambrian Shale (MCS) and the Upper and Lower Lotsberg Salts. It is worth noting that the Cooking
Lake regional aquifer (CKLK) is at about 1 km below ground and been partially pressure depleted by offset
oil production from the Leduc reef. The CKLK was identified as a key monitoring formation, as it is both a
strong thief zone and the deepest porous formation. The lateral extent of the BCS within the lease area has
been determined sufficient without fluid displacement [3].
Figure 2?Schematic stratigraphic column of the BCS storage complex and an injection well.
Figure 3 is a cross section from the regional static model. The characterization and containment risk
assessment covered the entire stratigraphic column from surface to basement, and includes an assessment
of the risk to containment of legacy wells in the region.
4
SPE-187100-MS
Figure 3?Cross section from the regional regional static geological
model showing the BCS just above the Precambrian basement
The BCS is at the base of the central portion of the Western Canada Sedimentary Basin (WCSB), directly
on top of the Precambrian basement. The BCS storage complex is defined as the series of intervals and
associated formations from the top of the Precambrian basement to the top of the Upper Lotsberg Salt. The
BCS storage complex includes, in ascending stratigraphic order:
?
?
?
?
?
?
?
Precambrian granite basement unconformable underlying the Basal Cambrian Sands
Basal Cambrian Sands (BCS) of the Basal Sandstone Formation ? the CO2 injection storage target
Lower Marine Sand (LMS) of the Earlie Formation ? a transitional heterogeneous clastic interval
between the BCS and overlying Middle Cambrian Shale
Middle Cambrian Shale (MCS) of the Deadwood Formation ? thick shale representing the first
major regional seal above the BCS
Upper Marine Siltstone (UMS) likely Upper Deadwood Formation ? progradational package of
siliciclastic material made up of predominantly green shale with minor silts and sands
Devonian Red Beds ? fine-grained siliciclastics predominantly composed of shale
Lotsberg Salts ? Lower and Upper Lotsberg Salts represent the second and third (ultimate) seals,
respectively, and aquiclude to the BCS storage complex. These salt packages are predominantly
composed of 100% halite with minor shale laminae. They are separated from each other by 50 m
of additional Devonian Red Beds.
SPE-187100-MS
5
Field Dynamic Model
Dynamic modelling was used to determine the number of required injection wells, and also to understand
the CO2 saturation and pressure distributions in the BCS over the injection period. The key risk was the
migration of injection pressure to the legacy well locations causing an uplift of BCS brine in the well to the
groundwater. The modelling demonstrated that injected CO2 is not expected to reach far field legacy wells,
but the potential for elevated reservoir pressure displacing saline brine into usable ground water could be
a risk if insufficient well count.
Results from the full field 3D dynamic modeling (Figure 4) illustrate that the CO2 plumes are isolated
to within 5 km of the injection well during the expected project lifetime. As the injection wells (Figure 2)
have been designed for this service they are low risk. Therefore, the next biggest risk to containment is
pressurized brine displacement through old legacy wells that may not have been abandoned properly.
Figure 4?Maps showing modelled CO2 saturation and pressure footprint at Quest after 25 years of injection. Note
- 5-35, 8-19, and 5-35represent the injection well locations, and the Redwater 3-4 well was a project delineation
well which was properly cemented. The white space on the right represents an erosional high in the BCS.
There are three legacy wells within the Quest sequestration lease area which are forecasted to encounter
pressurized saline brine in the project lifetime. Given the BCS reservoir pressure (20,036 kPa) and insitu
fluid gradient (11.7 kPa/m), it was calculated that a minimum incremental pressure of 3.3 ? 4.5 MPa in the
BCS would be required to lift BCS brine into the Base of Ground Water Protection zone (BGWP) through
an open hole at hydrostatic conditions (Table 1.) The subsequent field dynamic modeling demonstrated
that the expected pressure increases at the legacy wells will be less than half of the calculated incremental
pressure. This significantly decreases the containment risks of the
Well Brine Leak Path Model
The impact of a brine leakage at a legacy well was concluded to be negligible as the overlying
underpressured cooking lake formation was concluded to be an effective pressure sink. A 1 km
compositional radial well model was constructed from surface to basement for a legacy well to evaluate the
potential for cross flow between formations. 10 000 MD was applied to the first cell surrounding the well to
generate a vertical flow path. Each of the 50 discriminate layers was initialized with its own permeability,
porosity, initial pressure, and salinity. For details of the model inputs see Table 2. The BCS was initialized
with an 8 MPa delta pressure (28 MPa) with an infinite aquifer attached.
6
SPE-187100-MS
Table 1?Pore Pressure increase required to lift BCS brine to the Base Groundwater Protection (BGWP).
Table 2?Legacy Well Leakage Model Inputs for evaluating brine leak paths.
SPE-187100-MS
7
The resulting invasion profile seen in Figure 5 is illustrates the lack of any invasion profile. As most
of these layers have tight permeability they do not readily receive water injection. It should be noted that
the Cooking Lake is under pressured and no scenario investiaged had brine move higher than the Cooking
Lake as it is a pressure sink.
Figure 5?Permeability X-Section and 30 year forecast of the invasion profile as depicted in the Salinity X-section.
Injection Well CO2 Leak Path Model
As the injection wells have the highest pressures and concentrations of CO2 in the reservoir, despite excellent
wellbore integrity, they are the most likely location for a theoretical CO2 leak path. It was both concluded
that the bouancy force was a very slow moving affect and that the cooking lake formation ultimately acts as a
pressure sink. Vertical migration of CO2 from the BCS reservoir through a potential micro annulus crossing
the overlaying seals into the near-surface environment is considered as a very low risk as the cement bond
logs (CBL) in the injection wells confirm isolation. To mitigate this risk in any legacy wells (Figure 2), none
are expected to be exposed to CO2 or acidic brine as evaluated by the field dynamic modelling. However, in
the unlikely event that a micro-annulus was to be created at one of the injection wells, the vertical migration
of CO2 from the BCS into overlaying layers though was modelled [2] with an assumed vertical permeability
of 100mD. In addition to the CBLs, hydraulic isolation logs were run post injection and indicate that vertical
permeability is ? 100mD [4].
Such vertical migration through a permeable injector wellbore can reach 700m above the BCS at 25 years
(the end of the injection), 800m above BCS at 200 years. The driving force of buoyancy is marginal at these
depths due to the fact that at the insitu conditions CO2 is in dense, supercritical state.
In overburden formations, the scenario demonstrates that CO2 can reach horizontally about 30 meters
away from the injector wellbore into the formations above the BCS. The invasion of leaking fluids into the
overburden will remain a very localized event of very slow dynamics.
8
SPE-187100-MS
Figure 6?Modeled vertical migration of CO2 from the BCS into
overlaying layers though an injector well assumed permeable (100mD).
Cooking Lake Model
In the event that CO2 managed to leak past the seals, an observation well was drilled at each injection
site to monitor pressures above the storage complex. A number of overlying formations were evaluated
as potential monitoring targets. Figure 7 illustrates the identified formations and the potential of the
Winnipegosis, Contact Rapids/Lower Winnipegosis, Moberly, and Cooking Lake Formations to be used as a
deep monitoring formation based on XPT/MDT tests. The Cooking Lake was determined to be the preferred
leak detection monitoring formation and baseline pressure data was collect prior to any injection of CO2.
Figure 7?Monitoring Zone Assessment Summary. DMW stands for deep
monitoring well. Stratigraphic thicknesses are indicated in blue font.
The Cooking Lake formation benefits from proven geological extension well beyond the sequestration
lease area and it has a well-documented history of pressure communication across great distances from
production activities occurring in the Leduc reef. The Leduc reef is connected to the Cooking Lake as
illustrated in Figure 8 [5].
SPE-187100-MS
9
Figure 8?Schematic Cross Section illustrating Cooking Lake to Leduc Connection
The offset production from the Leduc reef induces pressure transients which must be accounted for in
order to interpret any observed pressure anomalies in the deep monitoring wells. Figure 9 is map illustrating
the offset of the Quest injection wells relative to the Redwater Leduc Reef. Figure 10 is a pressure plot from
the deep monitoring wells that illustrates the slow, smooth pressure response in the CKLK as the aquifer
re-pressures from offset production in the Leduc reef.
Figure 9?Aerial view of wells in the Quest to Leduc Reef area
10
SPE-187100-MS
Figure 10?Pressure in the Cooking Lake as recoreded at the observation wells.
The CKLK was modeled to understand what pressure response could be expected and whether a leak
into those formations could be detected. It was concluded that material leaks at the injection well would be
differentiable from baseline pressure drift at the monitoring wells. A 60� km model was created using
a 1�m grid with homogenous properties to estimate the pressures that should be slowly equilibrating
between the Leduc reef and the regional aquifer (Figure 11). Any anomalous behavior will then be assessed
with a greater understanding. To date, no anomalous pressure behavior has been observed, as illustrated
in Figure 10. It is further concluded that having 2-5 years of baseline data to more clearly establish the
baseline trend would be useful for discriminating anomalous pressure behavior of monitoring zone in a
pressure transient state.
SPE-187100-MS
11
Figure 11?Pressure transient in the Cooking Lake formation from offset Leduc production.
Conclusions
Quest commenced operation in August of 2015 at a rate of 1 MT/year into two injection wells. After two
years of operations, the project?s rigorous monitoring program has shown that the reservoir is behaving as
expected and no leaks have been detected. The modelling results show that CCS in a basal saline aquifer
below a depleted reservoir is overly sufficiently protected. There are multiple barriers afforded by the
regional geology in the form of seals and barriers and the CKLK depleted reservoir which prevent leak
paths from encountering the protected groundwater zone. Despite these multiple barriers, the monitoring
and operational strategy is designed to address these minimal risks.
The results of the leak path models demonstrate that the risk of a CO2 leakage from the Quest storage
operation is very low. Regional modeling of the overburden confirmed that no leak pathways in the project
area could be identified. Field level dynamic modelling demonstrated that injected CO2 is not expected to
reach far field legacy wells, but the potential for elevated reservoir pressure displacing saline brine into
usable ground water could be a risk if insufficient well count. The impact of a brine leak path was modeled
and concluded to be negligible as the overlying underpressured cooking lake formation was concluded to
be an effective pressure sink. As the injection wells have the highest pressures and concentrations of CO2
in the reservoir, despite excellent wellbore integrity, they are the most likely location for a theoretical CO2
leak path. It was both concluded that the bouancy force was a very slow moving affect and that the cooking
lake formation ultimately acts as a pressure sink. Therefore, the cooking lake was modeled to understand
what pressure response could be expected and whether a leak into those formations could be detected. It
was concluded that material leaks at the injection well would be differentiable from baseline pressure drift
at the monitoring wells.
Acknowledgments
Funding for the Quest project from the Government of Alberta [6] and the Government of Canada [7]
is gratefully acknowledged. Thank you to the JV partners of the Athabasca Oil Sands Project (AOSP),
Canadian Natural Resources Limited and Chevron Canada Limited.
12
SPE-187100-MS
References
1.
2.
3.
4.
5.
6.
7.
Yingqi Zhang, Curtis M. Oldenburg and Sally M. Benson. 2004. Vadose Zone Remediation
of Carbon Dioxide Leakage from Geologic Carbon Dioxide Sequestration Sites. https://
dl.sciencesocieties.org/publications/vzj/abstracts/3/3/0858
Winkler, M., 2011. Generation-4 Integrated Reservoir Modeling Report.
Shell Heavy Oil Controlled Document. http://www.energy.alberta.ca/CCS/
Generation-4IntegratedReservoirModelingReport.pdf#search=quest%20gen%204
Winkler, M, Ross Abernethy, Micah Nicolo, 2010, The Dynamic Aspect of Formation Storage
Use for CO2 Sequestration, SPE International Conference on CO2 Capture, Storage, and
Utilization, New Orleans, Louisiana, USA
Quest Carbon Capture and Storage Project - 4th Annual Status Report 31Mar2016 http://
www.aer.ca/about-aer/contact-us.htm
Schematic Cross Section illustrating Cooking Lake to Leduc Connection http://
www.ags.gov.ab.ca/graphics/atlas/fg12_07.jpg http://www.ags.gov.ab.ca/graphics/atlas/
fg12_07.jpg
CCS FUNDING AGREEMENT ? QUEST PROJECT http://www.energy.alberta.ca/
CCS/3822.asp http://www.energy.alberta.ca/CCS/pdfs/CCS_FA_Quest__consolidated_2016_04_29_final_(2)_for_Posting.pdf
Government of Canada Clean Energy Fund ? QUEST PROJECT http://www.nrcan.gc.ca/energy/
funding/current-funding-programs/18168
Документ
Категория
Без категории
Просмотров
1
Размер файла
2 598 Кб
Теги
187100
1/--страниц
Пожаловаться на содержимое документа