| IODP Proceedings Volume contents Search | |||
 | |||
| Expedition reports Research results Supplementary material Drilling maps Expedition bibliography | |||
|
doi:10.2204/iodp.proc.314315316.113.2009 Data and log qualityHole C0001AAvailable dataHole C0001A was drilled with the MWD-APWD tools installed in the drill string. All data were sent to the surface by the drilling fluid telemetry system (see Fig. F3 and text in the “Expedition 314 methods” chapter). At the end of the drilling operation, time and depth information were merged and data were processed following the data flow presented in “Onboard data flow and quality check” in the “Expedition 314 methods” chapter. Data included
Depth shiftFor this hole, the mudline (seafloor) was identified from the first break in the GRM1 log found at 2226.4 m DRF (Fig. F7). The GRM1 log is particularly noisy at the seafloor interface because the fast ROP (jet-in) in the unconsolidated formation is incompatible with a reliable statistical count of the radioactive elements of the formation and possible flow of mud around the bit. Despite all sources of uncertainty, the chosen value shows an acceptable discrepancy of –1.6 m with respect to water column height based on cumulative pipe length. The depth-shifted version of the surface and downhole drilling data and downhole ECD, APRS, and GRM1 logs are given in Figure F8. To help correlate time and depth versions of the data, the time-depth relationship for Hole C0001A is given in Figure F9. Logging data qualityExcept for the GRM1 log of the MWD tool, which is directly related to the formation properties (lithology), all other logs are direct surface drilling and downhole measurements. APRS and ECD derived from APRS expectably increase with depth, except between 430 and 520 m LSF, where drillers had to combine backreaming, increased pump pressure (SPPA), and reduced SWOB and ROP to increase circulation. As GRM1 has a high depth of investigation, it is considered reliable despite the lack of hole shape (caliper) data. No repeat data were available in this hole, but we will later show that the GRM1 log correlates well with the gamma ray log of the geoVISION resistivity (GVR) tool in Hole C0001D except for a few depth intervals discussed in the next section. Holes C0001C and C0001DAvailable dataHoles C0001C and C0001D were drilled with LWD-MWD-APWD tools installed in the drill string. Similar to Hole C0001A, all MWD-APWD data were transmitted in real time, including a limited set of LWD data. This data set includes bulk density (RHOB_DH_ADN_RT) and bulk density correction (DRHO_DH_ADN_RT) computed downhole (with a less sophisticated algorithm than the one used at the surface with the memory data); thermal neutron ratio (TNRA_ADN_RT) and thermal neuron porosity (TNPH_ADN_RT); average borehole diameter from the ultrasonic caliper (ADIA_ADN_RT); gamma ray (GR_RAB_RT) and resistivities (bit [RES_BIT_RT], ring [RES_RING_RT], shallow [RES_BS_RT], medium [RES_BM_RT], and deep [RES_BD_RT]) from the GVR tool; hole deviation data such as relative bearing (RB_RT); hole azimuth (HAZI_RT), and hole deviation (DEVI_RT); sonic compressional Δt; and semblance (DTCO and CHCO) from the sonic tool (see Table T2 in the “Expedition 314 methods” chapter). As previously stated in “Operations,” real-time transfer of seismic data failed because of problems with the downhole picking of wave arrivals in a high-noise environment. When LWD tools were recovered on the rig floor, various problems were encountered that required modifying the data flow presented in Figure F8 in the “Expedition 314 methods” chapter. Figure F10 illustrates the data flow used for Holes C0001C and C0001D. Memory data from the geoVISION were successfully downloaded and converted to depth. Sonic data were downloaded, but merging downhole time data with surface time-depth information initially failed. Both files were sent to Schlumberger’s shore base in Shekou (China) for depth conversion preprocessing and then were sent back to the ship for further processing and analysis. Downloading data from the adnVISION tool was impossible on board because of damage to the output port on the tool; therefore, in an attempt to download the data, the tool was sent to the Shekou base. The memory data from the adnVISION tool were returned to the Chikyu shortly before the end of Expedition 314. The seismicVISION tool experienced a similar problem and was sent to the Schlumberger SKK Drilling and Measurement Center (Fuchinobe, Japan), where the data were successfully downloaded and sent back to the ship by file transfer protocol for onboard processing and analysis. Drilling in Hole C0001C was aborted after <9 h of operation (including ~5 h of rig floor equipment repair) (see “Hole C0001C” in “Operations”). MWD, real-time, and memory data from the geoVISION, sonicVISION, and seismicVISION tools cover the depth interval 0–74 m LSF. Because of poor velocity contrast between mud and formation velocity (to ~200 m LSF), sonic data from Hole C0001C have not been processed, though raw waveforms are available at sio7.jamstec.go.jp. Only one check shot above seafloor was acquired at this site (see “Log-seismic correlation”). adnVISION memory is presently unavailable, with only a limited set of real-time data available for Hole C0001C. For Hole C0001D, transfer of adnVISION real-time data to the surface was lost below 510 m LSF, limiting the set of available data to real-time MWD and memory geoVISION, sonicVISION, and seismicVISION data from 0 to 973 m LSF (2193 to 3201 m DRF). adnVISION memory data for 0–510 m LSF were eventually recovered. However, no memory data for the hole below 510 m LSF could be recovered. Depth shiftFor Holes C0001C and C0001D, the mudline (seafloor) was identified from the first break in the gamma ray (GR) and resistivity (RES_RING, RES_BIT, RES_BD, RES_BM, and RES_BS) logs (Figs. F11, F12). In Hole C0001C, the mudline was picked at 2230.5 m DRF, a 4 m discrepancy from drillers depth (2226.5 m DRF). In Hole C0001D, the mudline was picked at 2228 m DRF, again differing from drillers depth, this time by 2 m (2226 m DRF). For both holes, uncertainty in picking the mudline is clearly within ±1 m because the top few meters of the unconsolidated formation was washed out by drilling fluid and resulted in mixing (formation suspension) at the mudline interface, blurring gamma ray and resistivity readings. For Holes C0001C and C0001D, the depth-shifted versions of the main drilling data and geophysical logs are given in Figures F13 and F14, respectively. Figures F15 and F16 present the time-depth relationship linking the time (Figs. F5, F6) and depth (Figs. F13, F14) versions of the data from Holes C0001C and C0001D. Logging data qualityHole C0001CFigure F13 shows the quality control logs for Hole C0001C LWD data. The target ROP of 30 m/h (±5 m/h) was generally achieved until 36.6 m LSF, where rig floor maintenance stopped drilling operations for 5 h (see “Hole C0001C” in “Operations”). Drilling continued at a slightly lower ROP (20–25 m/h) until operations stopped completely at 74 m LSF because of hole deviation exceeding 6°. This ROP was sufficient to record 1 sample per 4 cm over the majority of the hole. WOB was minimal (approximately null in the upper 35 m LSF and ~2 kkgf below). SPPA was maintained at 1.2 MPa for the entire drilling period, and no noticeable change in APRS and ECD was observed. Time after bit (TAB) measurements were taken at time intervals of ~5 min for ring resistivity and 4 min for gamma ray logs, except in a short depth interval that corresponds to the 5 h of rig floor repairs. Theoretical TAB measurements for density and neutron porosity in this interval are ~43 and 46 min, respectively. Note that these values have been computed because (1) the data themselves have not been transmitted in real time (not selected) and (2) the memory data are not available at the time of this writing; however, density and porosity measurements had been selected to be transferred in real time and are thus available. Memory data were recovered late in Expedition 314 and provide actual TAB data. The real-time density caliper log (ADIA), which gives the average diameter of the LWD borehole, is the best indicator of borehole conditions. The density caliper log, which should measure a value of 8.5 inches (21.6 cm) for a perfect in-gauge hole, instead shows values ranging between 9 and 10 inches. (22.9–25.4 cm) for the entire depth interval (0–74 m LSF) except between 5 and 11 m and between 33.5 and 46 m LSF, where noticeable washouts (ADIA = >10 inches [25.4 cm]; standoff = >1.5 inches [3.8 cm]) were detected. These washouts are associated with major decreases in the bulk density log (RHOB), where bulk density corrections (DRHOB) could not be fully compensated for this major washout and are thus underestimated (Fig. F13). Otherwise, a standoff <1 inch (2.5 cm) between the tool and the borehole wall indicates high-quality density measurements with an accuracy of ±0.015 g/cm3. Comparison between deep button (RES_BD) and shallow button (RES_BS) resistivity values shows that drilling fluid invasion is null or not significant, confirming the short TAB readings. Based on experience gained in Hole C0001D (see next section), sonic data were not processed because the velocity contrast between mud and formation is too low throughout the uppermost 200 m LSF. The quality of GVR images is good, but images suffered from hole ovalization and/or tool eccentricity resulting from hole deviation. No significant resolution loss is observed with variation in ROP except in the shallow section (first couple of meters) where the images were degraded by the rapid ROP and low rotation rate of the bit. Hole C0001DFigure F14 shows the quality control logs for the Hole C0001D LWD data. To avoid deviation of the hole as with Hole C0001C, the hole was started with rapid jetting-in (0–55 m LSF). Except for this shallow depth interval, the target ROP of 25 m/h (±5 m/h) was achieved to 340 m LSF. Effective ROP then slightly decreased with local increases in APRS and ECD, requiring minor backreaming (340–520 m LSF). Concomitant with the major increases in APRS and ECD, an increase in SPPA and a decrease in ROP was necessary to maintain drilling fluid circulation (from 520 to TD = 972 m LSF). For this lower depth interval (520–972 m LSF) increased SWOB was necessary to compensate for increasing stick-slip, which was particularly high (>200 rpm) in the high-ECD zone (520–580 m LSF) with a high level of shocks (740–920 m LSF). At ~575 m LSF, the hole slightly deviated by 2°, but hole deviation never exceeded 4° to TD. Except during pipe connections, backreaming, and wiper trips, TAB measurements are ~5–10 min for ring resistivity and 4–8 min for gamma ray logs. Theoretical TAB measurements are ~40–60 min for density/porosity; however, while real-time density and neutron porosity logs are available, they are limited to the upper section of the hole (0–510 m LSF). adnVISION memory data for 0–510 m LSF were obtained shortly before the end of Expedition 314. All adnVISION memory and real-time data below 510 m LSF were lost because of tool failure. The ADIA log, which is indicative of borehole condition, is also limited to this depth interval. Except for the 40 m where caliper readings are unreliable, the density caliper shows almost in-gauge value (standoff between tool and the formation = ~1 inch [~2.5 cm] to 190 m LSF). From 190 to 480 m LSF (loss of real-time data) the ADIA is highly anticorrelated with the RHOB, especially where ADIA exceeds 10 inches (25.4 cm) and where DRHOB is underestimated (Fig. F14). To fill the gap in ADIA and further assess borehole conditions below 480 m LSF, a comparison between natural gamma ray logs from Holes C0001A (GRM1, MWD tool, and real time) and C0001D (GR, GVR tool, and memory), horizontally separated by ~65 m, is presented in Figure F17. To ~500 m LSF, GRM1 and GR are particularly well correlated, at least at a meter scale, confirming proper reading of both tools. Correlation remains relatively high from 500 to 875 m LSF, except in the following depth intervals: 530–590, 625–650, 730–750, 780–860, and 900–925 m LSF, where GR (Hole C0001D) is lower than GRM1 (Hole C0001A). Careful inspection of the depth and time version of the data reveals that low GR with respect to GRM1 corresponds to periods of SPPA and low ROP in Hole C0001D, suggesting washouts at these depth intervals. SPPA in Hole C0001D is between 15 and 17 MPa, slightly above (1–2 MPa) SPPA in Hole C0001A (note the significant shift in SPPA by ~4 MPa between 522 and 556 m LSF), but most importantly the ROP in Hole C0001D is significantly lower (mostly below 15 m/h) than in Hole C0001A (mostly above 15 m/h). Major decreases in sonic P -wave velocity (VP) and resistivity for the two largest intervals (530–590 and 780–860 m LSF) favor this interpretation. Major potential washouts are identified in orange and minor potential washouts in yellow. Scalar logging data in these depth intervals must be interpreted with caution. As for Hole C0001C, comparison between RES_BD and RES_BS resistivity values shows that drilling fluid invasion is null or not significant (even with a slightly lower ROP than in Hole C0001C) and confirm short TAB readings. The sonicVISION data for Hole C0001D were processed on board the Chikyu by the Schlumberger Data Consulting Services (DCS) specialist, using two primary filtering sequences. The first “wide” sequence uses a broad bandwidth filter and the second “leaky-P” sequence uses a narrow filter designed to pass leaky P-wave mode arrivals. The composite sonic velocity curve that the DCS specialist prepared for this site includes data from both processed logs (Table T4). In the upper half of the hole from 0 to 524 m LSF and below 874 m LSF, the wide data were superior and were used to assemble the composite log. From 524 to 874 m LSF, the data from the wide and leaky-P processing were used in the composite. Quality control analysis of the sonic data is based on examination of the plots showing sonic waveforms and slowness coherence images for the common receiver data and common source data (Figs. F18, F19, F20, F21, F22, F23). The sonic data from 0 to 175 m LSF show a strong arrival with slowness expected for waves traveling in drilling fluid (an example of data in this interval is shown as “mud arrival” in Fig. F18). This arrival is expected to be large when the formation velocity is very low. We see no sign of a distinct arrival from the formation. We conclude that the formation velocity must be near that of the mud because if it were higher it would be more distinct. The interval 175–325 m LSF is a region in which there is a clear formation arrival distinct from the mud arrival (examples of data in this interval are shown in Figs. F19, F20). The sonic data seem quite continuous and reliable. There is a narrow zone from 192 to 202 m LSF in which the formation slowness drops back into the mud arrival. It is unclear whether this signal is an increase in slowness or a washout in the hole. The interval between 325 and 476 m LSF (Fig. F21 shows an example of data from this interval) is characterized by reasonably continuous slowness coherence broken by occasional gaps one to several meters thick. The picks in these gaps are usually higher slowness than the neighboring reliable picks. This creates apparent fluctuations in velocity that would be detrimental to the creation of synthetic seismograms. From 476 to 874 m LSF the gaps between reliable picks become larger, until the gaps dominate (an example of data in this interval is shown in Fig. F22). At this point, the sonic log is unreliable if used as is, although there are probably good values to be identified by using the coherence plots. From 874 m LSF to the deepest point reached by the sonicVISION tool at 964 m LSF the data improve. The picks are clear and fairly continuous (an example of data in this interval is shown in Fig. F23). The following comments on the processed sonicVISION data were provided by the Schlumberger DCS specialist:
Overall, the quality of the resistivity image data is good in Hole C0001D (Fig. F14; Table T5). Except for two short intervals (541–543 and 602–603 m LSF) showing a stick-slip vertical line indicative of nonrotation of the GVR tool/stick-slip, consistency in image quality between the shallow, medium, and deep data is indicative of good quality. From 529 to 629 m LSF, where hole problems occurred, data losses are more common, but even in this section the general structural patterns are apparent. Interpretation of resistivity image data is further discussed in “Structural geology and geomechanics.” |
|||
