Accurate quantitative interpretation of pulsed-neutron-capture (PNC) logs cannot be achieved in the low-salinity formation-water environment of the Gippsland basin because formation fluid responses are small relative to uncertainties introduced by lithologic variation. Time-lapse interpretation is used effectively although formation responses often are masked by larger responses caused by changes in the logging environment. Several easy techniques, such as using the gamma ray log to indicate formation-water movement, are used to aid interpretation.
Esso Australia Ltd. has been running PNC logs in the Gippsland basin since 1971. Although PNC monitoring programs are in place in all the large oil and gas fields, this paper is restricted to a discussion of the problems encountered in the oil fields shown in the shaded area of Fig. 1, where formation-water salinity is about 35,000 ppm NaCl equivalent and is considered marginal for definitive PNC interpretation. Examples have been chosen from the Fortescue, Cobia, Mackerel, Kingfish and West Kingfish oil fields to illustrate common PNC logging and interpretation problems. Although all examples shown are of Schlumberger's Thermal-Decay Time (TDTSM) logs, the problems discussed are fundamental to PNC logging and are experienced to some degree by all logging companies.
The contrast in capture cross sections across an oil/water contact (OWC) is given by
As a general guideline, PNC logging is considered applicable when
In Gippsland, the problem is two-fold: the low salinity of the connate water, which ranges from 25,000 to 40,000 ppm NaCl equivalent throughout much of the basin, and the low crude-oil GOR. The low capture cross section of the connate water (34 c.u.) and high capture cross section of the crude (20 to 21 c.u.), coupled with moderate to good porosities (15% to 25%), give rise to contrasts of only 0.5 to 2.5 c.u. across OWC'S. Maximum contrasts of 2.5 c.u. are achieved only in the cleanest quartz sandstones because of their exceptional saturation changes of 70 to 75 saturation units (s.u.).
Design characteristics of PNC tools will not be dealt with in this paper because they have been detailed in other publications. It is sufficient for our purposes to understand that PNC logging tools have been designed for use in cased holes to evaluate fluid-saturation changes in reservoirs where formation-water salinity is high and that optimal logging conditions occur when the wellbore is filled with high-capture-cross-section fluid, ensuring rapid decay of the borehole signal and hence less interference with the formation signal.
The formation curve generated by PNC tools is the sum of the volumetrically weighted values of each formation component. For a clean formation, can be expressed as the sum of a rock component (first term on the right side of Eq. 3) and a fluid component (second term):
Fig. 2 illustrates the response of a TDT log to a fluid saturation change averaging 70 s.u. across an OWC (3530 m) in a clean, massive sandstone with about 25% porosity. Openhole gamma ray, resistivity, and porosity logs are displayed in the first three tracks, and the near-detector formation curve, SIGM, is displayed in the track to the right. Fig. 3 displays SIGM data 15 m either side of the 1982 OWC vs. effective porosity. Oil and water are clearly discriminated, although there is scatter of about 20 s.u. in each group of points because of minor variations in shaliness that are too large to allow meaningful fluid-saturation calculations. Fig. 4 illustrates the much greater effect that changes in lithology, as opposed to changes in fluid saturation, have on . Openhole logs are displayed in the first three tracks, and effective porosity and SIGM are displayed in the last two tracks. Feldspathic sandstone underlies the clean quartzose sandstone above 3245 m. Neutron-density and resistivity logs indicate that both are clean and permeable. Fig. 5 shows SIGM data from across the matrix (3245 in) and fluid (3239 in) boundaries in Fig. 4 vs. effective porosity. While quantification of fluid saturations in a constant lithology (Figs. 2 and 3) is inaccurate at best, Fig. 5 shows it to be impracticable in variable-matrix reservoirs.
Uncertainties associated with rock effects can be accounted for by use of a time-lapse method (Fig. 6). OWC movements are indicated by differences between time-lapse pairs of SIGM curves acquired in 1981, 1982, and 1984, displayed in the last two tracks. The 2.5-c.u. separation is the maximum fluid effect observed in the Gippsland basin wells. The excellent repeatability of the SIGM curves above and below the swept zones is a function of logging under almost identical conditions. In Gippsland, formation-fluid contract movements affect the and gamma ray curves. Radioactive scale buildup near perforations after formation-water production begins has been well-documented in many areas (e.g., Alaska) and is evident across the 2644- t0 2657-m interval in Fig. 7. Displayed in the left track are the openhole and cased-hole gamma ray curves acquired along with the PNC logs during 1981–89. Once the cased-hole gamma ray logs have been scaled to allow for casing and cement attenuation, progressive radioactivity buildup can be seen in permeable zones away from the perforations, particularly in the lower half of the log. This radioactivity buildup tracks the movement in the OWC, as indicated by the SIGM curves in the last three tracks. This effect has been observed in many wells, is always related to formation-water movement through permeable zones, and is crucial to the interpretation of many of the thin sandstone reservoirs in Gippsland (Fig. 8). In 1989, SIGM curves in the right track were interpreted as having had no fluid saturation change in the sandstone at 2720 m between the 1987 and 1989 logging runs. On the basis of that interpretation, the reservoir was perforated (2719 to 2722 m) but produced only water. A subsequent time-lapse comparison of the cased-hole gamma ray curves that have the openhole log as a base log is in the left track and indicates clearly that this sandstone had watered-out before recording the 1987 base log.