Previous   |   Next

doi:10.2204/iodp.proc.301.104.2005

CORK design summaries

All of the CORK designs described below require some sort of reentry cone and casing to stabilize the upper part of the hole. Designs of the reentry cones and casing systems evolved during DSDP and ODP, and the latest ODP standard is summarized in Graber et al. (2002). In brief, that standard includes the reentry cone, mud skirt, and casing hanger that provide for running up to four nested sizes of casing from the hanger ever deeper into the hole, at diameters of 20, 16, 13⅜, and 10¾ inches. For most of the descriptions below, except for the ACORK, a 10¾ inch diameter casing is assumed to be (1) deployed as the final casing string prior to the CORK installation and (2) in some way sealed into the formation through which it passes. Some installations have also included a smaller diameter liner; a liner is similar in concept to casing but is emplaced into open hole from deep within the inner casing (i.e., it is not hung at the casing hanger immediately below the reentry cone).

Original CORK

The essential elements of the original single-seal CORK design (Davis et al., 1992) (Fig. F5) are (1) a CORK body that seals within the casing hanger system at the top of a reentry hole and (2) a long-term data logger and sensor string in the sealed hole. The CORK body provides an inner bore and landing shoulder that allows deployment (through the drill pipe) and internal sealing of the data logger and sensor string. This design requires prior establishment of a reentry hole, suitably cased; the standard is for the CORK body to seal inside the 10¾ inch casing hanger and extend up ~1.5 m above the rim of the reentry cone, although versions have been constructed for older holes with slightly different casing sizes and hanger systems (e.g., older DSDP reentry holes with 11¾ inch casing). Original CORKs have been deployed in two types of cased reentry holes: (1) those in oceanic crust that are cased through sediments and then cored with a 9⅞ inch bit to leave an open hole in underlying basement and (2) those in subduction settings, normally completely cased (and lined in some instances) with perforated sections through unstable zones of interest. After a suitable cased reentry hole is established (which may take several days to weeks), it has required an additional 24–36 h, on average, for deployment of the CORK body and sensor string, plus a landing platform to support subsequent experimental equipment and submersible operations at the CORK head.

Note that the instrumentation string is limited to a diameter less than ~3.75 inches that (1) allows deployment down the standard 4½ inch internal diameter drill pipe and (2) will pass through the CORK body inner bore until the data logger lands and seals. In theory, deployment of any kind of sensor string is possible if it meets this diameter restriction and it incorporates the necessary seals and landing shoulder. In practice, sensor strings (Fig. F3; Table T2) have typically comprised (1) thermistor cables below the data logger for temperature profiles within the sealed hole and (2) pressure gauges immediately above and below the data logger electronics housing for seafloor reference and sealed-hole pressure measurements, respectively. Where the hole is left filled with seawater, as is normal practice, the single pressure gauge below the data logger averages the fluid pressure signal from any open hole section below the casing or perforated section within casing. Where the hole is left filled with fluid of different density than seawater (e.g., heavy mud, as in the case of Hole 948D), a vertical array of pressure gauges may provide additional information (Foucher et al., 1997). The data logger incorporates an underwater-mateable connector (UMC) accessible at the top of the CORK for submersible-based data transfer and reprogramming, which has usually been attempted at average intervals of ~2 y (see "Submersible operations," below).

The CORK body assembly includes ½ inch stainless steel or titanium tubing that allows fluid pressure or fluid samples from the sealed section of the hole to be brought to a valve on the wellhead that is accessible by submersible. This also permits the hydrologic properties of the isolated zone of the formation to be actively tested using submersible-borne pumps linked to the wellhead valve. The first CORK sensor strings included thin-walled ½ inch diameter Teflon tubing run from the open hole section to the tubing in the CORK body, but this proved unsatisfactory as a fluid sampling method, largely because of damage to the tubing during deployment. Starting in 1994, a majority of sensor strings in CORKs of the original design have instead included self-contained long-term fluid "OsmoSamplers" driven by osmotic pumps (Shipboard Scientific Party, 1995, 1997; Jannasch et al., 2003) and hung on the thermistor cables deep in the hole. These have required recovery of the data logger and sensor string using submersibles some years after original deployment from the drillship (see "Submersible Operations," below, for details).

Advanced CORK

As noted above, the original CORK averages pressure signals from the open hole or perforated interval, so it is not suitable for resolving processes in hydrogeologically layered systems via a single hole. The Advanced CORK, or ACORK (Shipboard Scientific Party, 2002) (Fig. F6), was the first concept developed to achieve the goal of separately isolating multiple zones in a single hole. It achieves the objective by incorporating large-diameter casing-mounted packers at desired depths as integral parts of the final casing string (10¾ inch diameter for the two installations to date, but smaller sizes are feasible). The packers incorporate pressure-tight, lengthwise hydraulic pass-throughs, allowing fluid pressures and/or samples to be transmitted from sampling screens on the outside of the casing to gauges, loggers, and/or samplers mounted on the wellhead via tough, industry-standard hydraulic umbilicals strapped on the outside of the casing. In the two ACORKs deployed to date (deep strings seaward of and at the toe of the Nankai accretionary prism), the ACORK casing itself was entirely solid, although it would be possible to modify the design to provide hydraulic access between the formation and the inside of the casing. After assembly beneath the rig floor, the ACORK casing string was deployed into a predrilled pilot hole with reentry cone (that had been established during earlier logging-while-drilling operations) and run into the hole without rotation using a mud-motor-driven underreamer system, much like any simple (noninstrumented) casing string might be deployed. Once an ACORK casing string is landed into the reentry cone and the packers are inflated using the rig floor pumps, the casing can be reentered with a coring assembly for deeper drilling, as was done in one of the two Nankai installations to penetrate into underlying oceanic basement. After any hole deepening, the bottom of the casing is intended to be sealed with a drillship- and/or wireline-removable bridge plug, which completes the seal of the deepest monitoring interval and leaves the inner bore free for an instrument string, albeit one without direct hydraulic access to the formation. The ambitious Nankai deployments were flawed because of an inadequate underreamer and premature setting of a bridge plug, but they achieved most of their objectives in a difficult setting and demonstrated the utility of the ACORK concept.

CORK-II or sealed borehole instrument hanger with OsmoSamplers

The next planned deployments were to be at the Costa Rica margin, where experience had indicated that any pilot holes for ACORKs could not be expected to remain open to the planned depth of monitoring. Motivated partly by the difficulties with the Nankai ACORK installations, a new approach, dubbed "CORK-II," was taken that would permit multilevel monitoring deep within an otherwise normally prepared, cased reentry hole. It represented an adaptation of the "borehole instrument hanger" system that had been developed for the broadband seismometer/strainmeter installations in deep reentry holes in the western Pacific (Shipboard Scientific Party, 2000). In those installations, the instrument package was attached to the end of a small-diameter (4½ inch) casing that was suspended from a hanger that landed in the reentry cone; the small-diameter casing conveyed the instruments deep into the hole, provided a structural member for the cabling from the instruments to be run to the seafloor, and also provided the conduit by which the instruments were cemented in place once deployed. For the CORK-II (Jannasch et al., 2003) (Fig. F7), the 4½ inch casing string incorporated packers that could be inflated deep in the hole (either in open hole or within 10¾ inch casing) and perforated sampling screens that would allow formation fluids to be sampled by OsmoSamplers deployed down the inside of the drillstring and 4½ inch casing. Like the ACORK, the CORK-II packers incorporate length-wise hydraulic pass-throughs that allow fluid pressures and samples from the isolated zones to be conveyed to the wellhead by umbilicals mounted on the outside of the 4½ inch casing. In the Costa Rica margin CORK-II design (Jannasch et al., 2003) (Fig. F7), the OsmoSamplers also carried long-term self-contained temperature recorders. The OsmoSampler/ temperature-recorder packages were deployed on Spectra rope attached to retrievable plugs that landed deep within and sealed the inner bore of the 4½ inch casing. Great difficulties were experienced during later submersible-based attempts to recover these plugs and samplers from deep within the holes, so for subsequent CORK-II designs (Fisher et al., this volume), the sampling devices were run on Spectra rope extending to sealing plugs accessible directly at the wellhead.

Wireline CORK or wireline instrumented multipacker system

The wireline CORK (Fig. F8) followed, in many ways, an independent approach that utilized the capabilities of the Control Vehicle of the Marine Physical Laboratory (MPL), Scripps Institution of Oceanography (La Jolla, California, USA), for wireline reentry of existing cased holes from an oceanographic research vessel. The Control Vehicle had been developed partly with U.S. Science Support Program support as a facility for wireline reentry, logging, and emplacement of instruments within stable reentry holes without requiring a drillship (Spiess et al., 1992). The concept for the "wireline CORK" included an in-cone platform from which was suspended a bundled sensor string that included electrical leads, a mechanical strength member, hydraulic tubing, and inflatable packers to seal the hole at the desired depths. As with the ACORK and CORK-II, the packers incorporated hydraulic pass-throughs to allow fluid pressure signals and samples to be transmitted to gauges and valves on the in-cone platform via tubing from the zones isolated by the packers. In this case, the packers also incorporated electrical feed-throughs to bring thermistor signals from the isolated zones up to a data logger on the in-cone package. The sensors and data loggers were quite similar to those used for contemporary drillship CORKs. Two such installations were deployed in 2001 from the Roger Revelle in a pair of deep crustal reentry holes on the Costa Rica Rift (Becker et al., 2004). One of these worked well with only one packer to seal at the base of the casing and isolate the open hole section below. The other incorporated two packers, one intended to seal in casing and the other to seal in open hole, thus isolating two zones in the formation; that installation failed when the deeper packer became stuck in open hole ~20 m above its intended inflation position, so that the upper packer was not pulled the final ~20 m to its intended seat within the upper casing.

CORK instrumentation packages

Table T2 provides a summary of the specific instrumentation originally installed in each ODP CORK. The scientific measurement objectives for the original CORK design were actually modest: long-term records of temperature profiles and pressure in the sealed hole, sampled hourly. The geometry of the original CORK and down-the-pipe deployment method for the instrumentation defined a small-diameter form factor and the basic geometry of the instrumentation package. This included an elongated, small-diameter data logger housing, above which was mounted a seafloor reference pressure gauge and UMC for communication with the data logger, and below which was suspended in the sealed hole a thermistor cable and single absolute pressure gauge (multiple gauges not being useful in the normal case of the hole left filled with seawater). As is described below, various other options were ruled out early because of the basic form factor. The instrumentation for the majority of single- and multiseal CORKs were provided by our collaborative group supported by the U.S. National Science Foundation (NSF) and the Geological Survey of Canada, so we focus on the evolution of that instrumentation in this review. However, we note that the Institut Francais de Researche pour l'Exploration de la Mer (IFREMER) provided a quite successful sensor string of independent design for one of the single-seal CORKs installed in 1994 (Table T2) and refer the reader to Foucher et al. (1997) for details.

Pressure gauges

With the exception of the IFREMER string noted above, all the ODP-era CORKs utilized Paroscientific Digiquartz depth sensors (4000 or 7000 m models, as appropriate) to provide absolute pressure measurements. The associated "Paroscientific Intelligent Module" analog-to-digital converters (ADCs) have been incorporated within the data loggers for all installations except the IFREMER string. The narrow form factor required by the original CORK design essentially precluded consideration of a differential pressure gauge to assess formation pressure relative to hydrostatic because passage of one or more fluid line(s) through or into the electronics housing would have been required. The Paroscientific gauges have proven to be sufficiently accurate and very reliable over the long term, so they have also been used in the multilevel ACORK and CORK-II, even though the newer configurations would allow use of differential gauges. Recorded pressures have been accurate to ~0.01% of the full-scale range, and pressure variations have been resolved to 1 ppm of full scale (~40 cm and ~4 mm, respectively, for the Juan de Fuca deployments). Precise differential pressure determinations are facilitated by hydrostatic reference checks before installation and at the time of service or recovery operations.

Thermistor cables

The initial CORK deployments provided the greatest temperature-measuring challenge of all the installations because of high formation temperatures in excess of 270°C. (The use of thermocouples on the sensor string for such high temperature settings was ruled out because of the complication in providing reference junctions at the tops of long cables.) We standardized on a Thermometrics "SP100" thermistor of high nominal resistance so that line resistances could be ignored and specially aged at high temperature for >4 months to achieve acceptable long-term stability. For the initial installations, two cable manufacturers were contracted to mold the thermistors into cables specified to be able to withstand the expected temperatures; unfortunately, one (Vector Schlumberger) could not deliver in time and the readings from the cables from the other (Cortland Cable Co.) displayed problems indicative of pervasive seawater leakage at the high temperatures within days of deployment. For subsequent installations, we took attachment of the thermistors to the multiconductor cables into our own hands, with significantly better results (in formations at lower temperatures), using three different methods:

1. Bringing each conductor pair and thermistor into a grease-filled Teflon capsule of our design;

2. A proprietary epoxy encapsulation of the thermistors by Ocean Design Inc., with leads brought out for crimp pins and slip-on rubber boots sealing over the crimps to conductors; and

3. Molding by Branter/SeaCon of the thermistors into MAW-2 connectors, with the mating MAW-2 then molded onto each conductor pair at the appropriate depth.

Early conductor cable designs by Cortland Cable Co. incorporated 10 twisted pairs around a ¼ inch Kevlar center strength member; insulation of the 20 gauge wires was Teflon of the grade appropriate for the expected temperatures. There were quality control problems with the insulation on these cables, and for 1996–1997 deployments at moderate temperatures (20°–65°C), we changed to a design by Neptune Technologies (now unfortunately out of business). These included outer Kevlar and polyester braiding as strength member and abrasion cover, respectively, a conductor core made by South Bay Cable of 12 twisted pairs of conductors with extra-thick insulation of polypropylene, and MAW-2 conductors molded onto the conductor pairs with a proprietary Neptune Technologies technique. These worked very well in multiyear deployments on the Juan de Fuca and Mid-Atlantic Ridge flanks, although the two units at highest temperatures (60°–65°C) displayed some degradation after 2 years that seemed to originate where the MAW-2 conductor pairs were molded to the cable conductor pairs.

Our experience with thermistor strings seems to be consistent with industry experience in long-term reservoir monitoring using more sophisticated cable assemblies. In those efforts, the most significant long-term failure rates are with conductors and connectors, not with sensors or electronics (M. Kamata, presentation at IODP interim Technical Advisory Panel, pers. comm., 2003). Similarly, our worst problems have been with the conductors and thermistor-conductor connection. These problems were wholesale with the first cables at very high temperatures; even with the subsequent, better quality designs, problems in general increase with both long time and in situ temperature, as insulation degrades and/or seawater penetrates insulating materials.

Finally, an important aspect of our experience is that about half of the thermistor cables deployed to date have had to be field-shortened, often under tight time constraints, when the realized open hole depths fell short of planned depths. We anticipated this likelihood and over the years have employed two methods when shortening was necessary: folding the cable or reterminating the top of the cable assembly. The former was necessary for the original cable designs with a central Kevlar strength member that could not be easily reterminated, and it worked reasonably well as long as proper thimbles were used at the folds to avoid crimping the conductors. The latter was made possible when we changed to the Neptune Technologies cable design, in which the strength termination was a Kevlar cable grip applied to the outer braid cable strength member. It was also made possible by the success of our original design for bringing the thermistor and borehole pressure signals into the electronics pressure case, given the restriction to a single bulkhead connector feasible with the narrow-diameter form factor. We made these connections by bringing the thermistor conductors into an oil-filled boot, where they were mated individually, using standard single-pin connectors with slip-on rubber boots, to leads on a custom-molded multilead "pigtail" that brought all the signals to the connector that mated to the bulkhead connector on the electronics package. For all the single-seal CORK deployments except the IFREMER string in Hole 948D, the custom pigtail included 20 thermistor leads plus the standard four-pin connector to the sealed-hole pressure gauge, which was also made up within the compliant oil-filled boot. The boot and its conductor feed-through bulkhead were made by our group; the connector system was made by Branter/SeaCon, based on the MINM-25 bulkhead and mating cable connector models. The design proved to be reliable for all the installations.

Data loggers

Functionality of the data loggers for the original CORK design (ODP Legs 139 through 195) was constrained largely by the form factor dictated by the deployment scheme and pressure case diameter. All components in the reentry cone assemblies (pressure sensors, cable terminations, and electrical connectors including the UMCs, batteries, and electronics) were required to fit within the 64 mm (2.5 inch) inside diameter of the pressure-case and strength-member sections of the assembly, which totaled ~13 feet in length.

As noted above, the data loggers for all but one of the original CORKs were provided by our collaborative group, and these were manufactured by Richard Brancker Research, Ltd. Power in all of these units was supplied by four 3.6 V lithium thionyl chloride "D" cells, which provided a nominal capacity of 26 A·hr at 7.2 V. This limit, the power consumption dictated by the particular processor and memory used, the number and type of sensors, and the logging rate, defined the monitoring lifetime, which ranged from ~2 y in early units with all channels operational, to several years in some instances (e.g., in Hole 857D, pressures and internal temperature were logged hourly with on-board power from 1996 until 2003, when an auxiliary battery pack was connected). Memory capacity grew as low-power technology advanced, from 0.5 MB in early deployments to 2 MB in later ones, allowing 2 to several years of operation between data downloads (again, depending on the number of sensors employed). Typical CORK instruments provided analog-to-digital conversion for up to 15 analog devices. Typically, these comprised 10 formation thermistors, an on-board thermistor, and two precision resistors that provided a check on logger drift (although none was ever detected). Twelve-bit and, later, thirteen-bit conversion was applied to a temperature range of 0°–100°C and, in some cases, 0°–300°C, to provide a resolution ranging from 0.01 to 0.07 K. Two installations included tilt sensors. Data were typically recorded once per hour, although during certain periods (e.g., hydrostatic sensor checks) sampling intervals were sometimes reduced to the minimum of 10 s allowed by the logger. All data lines have been date/time-tagged with output from quartz oscillators. Drift, determined with periodic checks at the time of submersible visits, has been found to be linear, although often several minutes per year. Clock checks and resets, data downloads, and logger reprogramming were accomplished via a 9600 baud RS-232 serial link utilizing the UMC system described in "Submersible operations."

Despite the highly robust and reliable characteristics of the original CORK data loggers, the relaxation of the diameter constraints permitted by the externally mounted pressure cases of the ACORK configuration stimulated a redesign of the data loggers. Incremental steps were taken on all fronts. The new instruments accommodated greater numbers of pressure and temperature sensors (e.g., 7 pressure sensors were included in the Hole 808I deployment, and 16 thermistor sensors were included in the wireline ACORK in Hole 504B). Much greater temperature sensitivity was achieved with a 24 bit ADC. The limiting factor became the inherent noise of the electronics; resolution realized was significantly better than 1 mK. Space was available for larger battery packs (24 "DD" lithium chloride cells providing 420 A·h at 7.2 V). This, with increases in memory capacity to 8 MB and in serial data transfer rate to 38.4 Kbps, allowed practical logging rates and monitoring lifetimes to be increased substantially. Most of the multilevel CORK installations operating at 10 min sampling rates are limited by battery shelf life and should run for 10 y or more.

Further advances in low-power, high-speed components have prompted improvements in CORK logger technology during the transition from ODP to IODP. These have been applied to the Expedition 301 installations (see Fisher et al., this volume, for further details) and are available for future installations. The new instruments employ flash memory cards for greatly expanded memory capacity, which, along with improvements in power dissipation, allow for significantly better capability for long-term logging at higher rates. The greatest technological advance comes with a new ADC for the Paroscientific pressure sensors designed by Bennest Enterprises, Ltd. A frequency resolution of roughly 2 ppb has been achieved with a 800 ms measurement; applied to the full dynamic range of the pressure-sensitive transducers, this equates to a pressure resolution of 20 ppb, a factor of 50 better than previously attained. This new sensitivity (~2 Pa, or 0.2 mm of water) will permit new studies of oceanographic (infragravity waves, tsunami, and turbulence), seismic (surface waves), and hydrologic phenomena. Another advance involves the portability of the sensor and logging system. Hydraulic connections are provided by lightweight submersible-mateable connectors, and with the reductions in power consumption (and therefore in battery volume), the sensor and logger housings are smaller and more portable. This allows the instruments to be carried to and mated with the wellhead plumbing easily by submersible if they are not mounted on the ACORK instrument frame at the time of drilling, and replaced later if necessary.

Top of page   |   Previous   |   Next