Field Experiment for CO2 Sequestration
Update on the Frio Brine Pilot Experiment—six months after injection
Susan. D. Hovorka and research team (listed below)
From October 4–14, 2004 the Frio Brine Pilot team injected 1,600 tons of CO2 1,500m below surface into a high permeability brine-bearing sandstone of the Frio Formation beneath the Gulf Coast of Texas, USA. Eight months into the post-injection phase of the study, we have made substantive progress toward accomplishing the four major project objectives:
- Demonstrate to the public and other stakeholders that CO2 can be injected into a brine formation without adverse health, safety, or environmental effects,
- Measure subsurface distribution of injected CO2 using diverse monitoring technologies,
- test the validity of conceptual, hydrologic, and geochemical models,
- develop experience necessary for development of the next generation of larger-scale CO2 injection experiments.
The Frio Brine Pilot experiment is funded by the Department of Energy (DOE) National Energy Technology Laboratory (NETL) and led by the Bureau of Economic Geology (BEG) at the Jackson School of Geosciences, The University of Texas at Austin with major collaboration from GEO-SEQ,a national lab consortium led by Lawrence Berkeley National Lab (LBNL).
The first objective was accomplished through outreach, which included numerous site visits by researchers, local citizens, and environmental groups, major media interviews, an online log of research activities (www.gulfcoastcarbon.org), a technical e-newsletter, and an informal non-technical "neighbor newsletter". These activities continue as results of analysis are obtained. Public and environmental concerns were moderate, practical, and proportional to minimal risks taken by the project and included issues such as traffic and potential of risks to water resources. Press coverage was balanced and positive toward research goals. Safe site operation was managed by Sandia Technologies LLC, Praxair Inc., and Trimeric Corporation.
The second objective, measurement and monitoring of the subsurface CO2 plume, was accomplished using a diverse suite of technologies in both the injection zone and in the shallow near-surface environment. Wireline logging, pressure and temperature measurement, and geochemical sampling were conducted also during injection. In-zone objectives were to measure changes in CO2 saturation through time, in cross section, and areally, and to document accompanying changes in pressure, temperature, and brine chemistry during and in the months following injection. The in-zone measurement strategy was designed to test the effectiveness of a selected suite of monitoring tools in measuring these parameters. The near-surface monitoring program measured soil gas fluxes and concentrations, introduced tracers, and fluid chemistry in the vadose zone and shallow aquifer in an attempt to detect any leaks upward out of the injection zone, especially those rapid enough to cause releases in a short time frame such as behind well casing.
Tools used for in-zone monitoring included five repetitions of logging with the Schlumberger pulsed neutron capture reservoir saturation tool (RST), which under conditions of a maximum 35% porosity and 125,000 ppm salinity was successful in obtaining high-resolution saturation measurements across the injection interval. During the injection, CO2 saturation increased toward a maximum of 60% of pore space filled with CO2 in both the injection and observation well. Saturation declined in the post injection period; a final log run February 23 quantifed the CO2 permanently trapped in-zone by two-phase (residual) trapping. The log analysis team includes researchers from BEG and Schlumberger–Doll Labs.
An innovative geochemical sampling tool, developed and operated by Barry Freifeld and Rob Trautz (LBNL) to support in-zone fluid chemistry sampling, is the U-tube. The U-Tube is composed of a double length stainless steel tubing, with a check valve open to the reservoir at 1500 m. Formation fluid that was collected in the U-Tube was driven at reservoir pressure into evacuated sample cylinders at the surface by high pressure ultra-pure nitrogen. Samples were collected hourly to facilitate accurate delineation of CO2 breakthrough and recover uncontaminated and representative samples of two-phase fluids. Initial CO2 breakthrough to the observation well 30 m updip of the injection well occurred 51 hours after initiation of injection. Steady increases in the ratio of CO2 to brine produced recorded increasing saturation and plume thickness as the front of the plume expanded past the observation well. Free gas in the sample and gases coming out of solution were pumped from the top of the gas separator through a quadrapole mass spectrometer analyzer and a landfill gas analyzer to measure changes in gas composition in the field. During the 12 hours after breakthough, CO2 replaced brine as the fluid in the perforated zone of the wellbore and became the only fluid produced. At the same time that CO2 was detected at the observation well, the pH of produced, partly degassed brine dropped from 6.7 to 5.7, alkalinity increase from 100 to 3,000 mg/L bicarbonate as a result of mineral dissolution, and iron increased from 20 mg/L to 2000 mg/L, changing the fluid from clear to coffee color (Yousif Kharaka [USGS] and Seay Nance[BEG]). Downhole sampling with a Kuster sampler in April 2005 allowed us to assess geochemical changes as CO2 saturated brine react with the mineralogially complex sandstone matrix for 7 months.
The suite of tracers injected with the CO2 include perfluorocarbon tracers (PFTs), the noble gases, krypton, neon, and xenon, along with sulfur hexafluoride. Tracer injection and analysis was performed by researchers from Oak Ridge National Laboratory, Lawrence Berkeley National Laboratory, and Alberta Research Council. The tracer arrival times and elution curves allow assessment of the percentage of CO2 that is trapped by dissolution into the brine, based on partitioning of the tracers from CO2 into the brine, along with facilitating estimation of evolution of CO2 saturation as injection proceeded.
Pressure and temperature histories during injection provided comparative effective permeability under brine- and evolvingCO2+brine conditions. Downhole installation of pressure and temperature gauges proved to be critical for interpretation of complex (gas, supercritical CO2, brine) phases in the wellbore. LBNL and Sandia Technologies designed the hydrologic test program.
Geophysical measurements of plume evolution include cross-well seismic, an azimuthally dependent vertical seismic profile, and cased-hole cross-well electromagnetic (EM) surveys. These surveys were made pre- and post injection and analyses to date show that tools were successful in measuring CO2. The entire test is a proxy for a leak that might escape from a large injection; additional analysis is underway to determine success of geophysical methods in leak detection under these conditions. The geophysical team includes LBNL, Paulsson Geophysical, Schlumberger-EMI Technology Center, and Australian CO2CRC/CSIRO.
Near-surface monitoring includes soil-gas CO2 flux and concentration measurements, aquifer chemistry monitoring, and tracer detection of PFT with sorbants in the soil and aquifer. Pre-injection baseline surveys for CO2 flux and concentration-depth profiles over a wide area and near existing wells were done in 2004. Minor variability in aquifer pH and gas concentrations have been measured but analyses of tracers needed to determine whether change is related to leakage are still underway. The near-surface research team includes BEG, NETL SEQURE, Colorado School of Mines, and LBNL.
The third objective is to test the validity of conceptual hydrologic and geochemical models. Reservoir characterization by BEG to provide inputs to the simulations used existing and newly collected wireline logs, existing 3-D seismic survey, baseline geochemical sampling by USGS and Schlumberger, and core analyses by Core Labs. A drawdown interference test and a dipole tracer test conducted by LBNL researchers provided interwell permeability estimates (2.3 Darcys) confirmed that the core-based measurements of the porosity-thickness product (6.2 m thickness with 0.35 porosity) were appropriate at site scale for the Frio C sand targeted for CO2 injection.
Two groups of modelers, LBNL using TOUGH2 and The University of Texas Petroleum Engineering Department using CGM, input geologic and hydrological information along with assumptions concerning CO2 /brine multiphase behavior to predict the evolution of the injected CO2 through time. The observed CO2 breakthough occurred somewhat faster and in a narrower zone than the predicted arrival. Further refinement of the relative permeability and capillary pressure-saturation properties allow the model to better match the acquired data. Geochemical modeling by Lawrence Livermore National Lab predicted elements of brine composition evolution.
As the Frio experiment analysis and modeling continue, it supports the fourth objective, development of the next generation of larger-scale CO2 injection experiments. Confidence in the correctness of conceptual and numerical models and the effectiveness of monitoring tools tested will encourage the next pilots to investigate more complex factors such as stratigraphic and structural heterogeneity and upscaling. The Frio Pilot results provide a model for the U.S. Regional Partnerships Program participants as well as international collaborators to us to design test programs in various settings.
The pilot site is representative of a broad area that is an ultimate target for large-volume storage because it is part of a thick, regionally extensive sandstone trend that underlies a concentration of industrial sources and power plants along the Gulf Coast of the United States. The Gulf Coast Carbon Center, in cooperation with the Southeast Regional Carbon Sequestration Partnership, is proposing one of these ambitious pilots in the Frio or related sandstone to conduct a multi-month injection to "prove- up" the concept of stacked storage in an oil reservoir in decline and the underlying brine-bearing sandstones.
Frio Brine Pilot research team
DOE/NETL project managers
Charles Byrer and Karen Cohen
Bureau of Economic Geology
Susan D. Hovorka (PI)
Texas American Resources
Edward "Spud" Miller
Lawrence Berkeley National Lab
Sally Benson and Larry Myer (GEO_SEQ lead)
Oak Ridge National Lab
National Energy Technology Lab SEQURE group
Lawrence Livermore National Lab
Alberta Research Council
CO2 CRC/CSIRO Australia
T. S. Ramakrishnan Nadja
University of West Virgina