Page 344 - Standard Handbook Petroleum Natural Gas Engineering VOLUME2

Estimation of Waterflood Residual Oil Saturation 311

Interpretation. Computer simulation is used to model the injection, reaction,
partitioning, production of the tracers and to correct for overall drift of fluids
past the wellbore [340]. Fluid drift is mainly caused by injection or production
of fluids in the vicinity of the test well. A best fit is obtained for the injected
and produced ester and methanol, and parameters given by this fit are used to
model the alcohol production for a range of assigned values of residual satura-
tion. The measured alcohol production curve is then compared with the
simulated results in order to estimate the residual saturation in the formation.
Reservoir heterogeneity and loss of tracer can present problems in interpretation.
Reservoir Heterogeneity. Although the tracer test samples a relatively large pore
volume, results will be weighted towards the higher permeability zones. However,
this may not be a disadvantage because these zones will normally be swept
preferentially by tertiary processes.
The single-well tracer method is not recommended for fractured reservoirs.
Fractures cause nonradial flow, which results in tracer profiles that are almost
impossible to simulate [341]. Severe permeability variations are also difficult to
interpret. Where gross permeability variations exist, it may be necessary to
conduct frequent injection profile tests to determine the intervals that are
experiencing high-rate fluid flow [342].
Loss of Tracer. When the chemical tracer is injected into zones that do not
subsequently produce fluids, tracer will be lost to the reservoir. In one field
test, loss of tracer was estimated at 15% of the injected amount [339]. Conversely,
dilution of produced fluids by water from zones that did not receive tracer
injection will also present interpretation problems. If high drift-rates cannot be
controlled, the slug of water containing chemicals can move so far from the
wellbore that tracer profiles may not be well defined. Drift-rate should be less
than 1 ft/day [342].
In some cases, it may not be practical to perform tests in wells with large
intervals open or in wells with large holdup volumes in the wellbore. In such
instances, well workovers may be required [340].
With wells produced by gas lift, corrections are required to account for the
loss of part of the more volatile tracers to the gas by stripping action in the
wellbore. Tests conducted [341] with ethyl acetate in gas lift wells indicate a
loss of about 30%. Appreciable gas accompanying crude oil lifted by other
production methods may cause loss of tracer by stripping.

Accuracy. Success in application of the tracer test depends to a considerable
extent on the skill and experience of those conducting the test. While there is
no absolute measure of success, comparison of measurements with simulated
results provides a useful guide as to the reliability and tolerance of the results.
In about 10% of the early tests conducted, residual oil saturation could not be
determined from results. Oil saturations of 10% to 20% pore volume have been
measured with reported accuracies of +2% to +4%. Although conventional
methods of estimating residual oil cannot be considered an acceptable standard
for determining absolute accuracy of the tracer technique, measurements
obtained from tracer tests have generally agreed with values obtained from
pressure core tests [319,339-3421.
Field Application. Logistic considerations require adequate preplanning which
means it may be difficult to schedule tests on quick notice [339]. As compared
to well logging techniques, considerable time is required to obtain and interpret
the data.