PROJ00326 - Laserfiche WebLink

Home Browse Search

Correlation of porosity data is better between core data and log data from the core holes than between core data from the core holes and log data from well USGS. The standard error of estimate for core porosity data compared to log porosity data from holes CI and CIA (fig. 5) is 0.029 porosity units. A similar correlation between core porosity data from the core holes and log porosity data from well USGS has a stan- dard error of estimate of 0.047 porosity units. Increased variance is introduced in the data by compar- ing core data to log data from a hole other than the core hole. A similar increase in variance can be expected in subsequent regression relations of specific yield derived from core data on log data from well USGS. SPECIFIC-YIELD ESTIMATES FROM GEOPHYSICAL LOGS Although laboratory measurements of specific yield can be made on core samples, and specific yield can be estimated from laboratory grain-size detenninations, a considerable reduction in time, cost, and effort could be achieved if specific yield could be estimated directly from geophysical logs. Logs commonly used in the water-well industry, such as spontaneous potential, natural gamma, and resistivity have little potential for use in estimating specific yield. However, other logs more commonly used in the petroleum industry, such as free-fluid index, effective porosity, and apparent grain density have considerable potential for use in estimating specific yield. Each of these logs is an interpretative log produced as part of a computer-assisted well-evaluation package available from commercial geophysical-logging companies. Schlumberger's NML and Cyberlook log interpretation packages (Schlumberger Well Services, 1987) were used to produce the free-fluid index, effective-porosity, and apparent grain-density logs used in this report. Nuclear Magnetism Log In general, specific yield is a measure of the vol- ume of water that is free to move through and drain from the pore space of a rock. A free-fluid index is a measure of the volume of water that is not bound elec- trically or chemically to the rock matrix and, thus, is free to move within the pore spaces of a rock. The sim- ilarity of these two definitions indicates that a free-fluid index log could provide a means for estimating specific yield. In Schlumberger's nuclear magnetism log, the logging tool measures the precession of the magnetic moment of protons in the Earth's magnetic field (Schlumberger Well Services, 1987). Under natural conditions, the spin of protons in hydrogen nuclei of unbound fluids is approximately aligned with the Earth's magnetic field. When a strong polarizing mag- netic field is applied, the proton spin can be reoriented approximately perpendicular to the Earth's field. In the logging tool, a polarizing field is produced for a few seconds to cause proton polarization; then the field is quickly shut off. The realigned protons quickly begin precessing about the Earth's field as they revert back (relax) to their original orientation. The spin preces- sion induces a sinusoidal signal in a receiver coil in the logging tool. The amplitude of the signal is propor- tiona to the number of affected protons in the material surrounding the tool. Nonhomogeneities in the Earth's magnetic field cause the spins to dephase as they pre- cess, resulting in an exponential decay in signal strength with time. Protons in the nucleus of hydrogen atoms contained in solids or in water molecules bound to the surfaces of solids have very short relaxation times-generally less than a few hundred microsec- onds. Protons in free water in the pore space of a rock have much longer relaxation times-generally hun- dreds of milliseconds. If the measurement of the relax- ation signal is delayed for 25 to 30 milliseconds after the beginning of free precession, only the signal from free or unbound fluid in the pore space is measured by the logging tool. The recorded signal is the free-fluid index log, which is described in greater detail by Her- rick and others (1979). The protons in the water in the borehole produce a relaxation signal that is indistinguishable from that produced by the protons in the formation water. The borehole signal is eliminated by the addition of magne- tite to the drilling fluid before logging. The magnetite shortens the relaxation time of the borehole fluid signal so the signal is largely attenuated during the 25- to 30-millisecond delay period. Unfortunately, the mag- netite also can shorten the signal from the formation waterifthe magnetite invades the formation (the depth of investigation of the logging tool is about I in.) (Herrick and others, 1979). Analyses of the free-fluid index log run in well USGS indicate that the log response is very sub- dued through the entire depth of the hole and only attains specific yields of 0.20 to 0.40 that are measured in the core at a few thin intervals in the lower part of the hole. Analyses of the log response by Schlumberger log analysts indicated that the logging tools and inter- pretative software apparently functioned correctly, and that the lack of signal shown on the log was due to the invasion of magnetite into the formation. Analyses of the micronormal and microinverse logs indicate that mudcake on the borehole wall is thicker in the lower SPECIFIC-YIELD ESTIMATES FROM GEOPHYSICAL LOGS 13