X1.3.2.2 To obtain “water-free” samples of crude oil, one
5-gal sample of each of two crudes was distilled over the
X1.1.1 This round-robin testing program has shown that the
distillation method as practiced is somewhat more accurate
than the centrifuge method. The average correction for the
distillation method is about 0.06, whereas the centrifuge
correction is about 0.10. However, this correction is not
constant nor does it correlate well with the measured concentration.
X1.1.2 There is a slight improvement in the precision of the
distillation method over the present Test Method D 95 (API
MPMS Chapter 10.5): 0.08 repeatability versus 0.1, and 0.11
versus 0.2 for reproducibility. These figures are applicable
from 0.1 to 1 % water content, the maximum level studied in
this program.
X1.1.3 The precision of the centrifuge method is worse than
the distillation method: repeatability is 0.12 and the reproducX1.3.2.4 The samples for each cooperator were bottled so
that the entire sample had to be used for a given test. In this
way, any effect due to settling or stratification of water was
eliminated.
X1.3.2.5 Samples were coded to mask the presence of
duplicates and a table of random numbers dictated the running