Geology Reference
In-Depth Information
0.7 and P(dhi
nohc) might be 0.2 (note that the two
probabilities do not need to add up to 1 given that they
are essentially unrelated). The chance of success would
be calculated as (0.7 × 0.3)/(0.7 × 0.3 + 0.2 × 0.7)
j
GOOD
N/A
20%
50%
0.6.
This high chance of success indicates that the DHI has
transformed the perception of the prospect. If there
was similar geological risking but a muchmore equivo-
cal DHI, for example a weak flat spot, giving a value of
0.6 for P(dhi
¼
N/A
10%
25%
nohc),
then the chance of success would be (0.6 × 0.3)/
(0.6 × 0.3 + 0.4 × 0.7)
j
hc) and a value of 0.4 for P(dhi
j
¼
0.39. Thus, the equivocal
DHI has not changed the perception of the prospect
very much. A similar calculation can deal with the risk
of low (non-commercial) gas saturations; in this case
nohc would include sub-commercial saturations.
Bright soft amplitudes on a flank, consistent with the
presence of gas but not showing conformance to struc-
ture or evidence of a gas column from a flat spot, might
then have P(dhi
N/A
N/A
12%
POOR
LOW
HIGH
Confidence in DHI
Figure 10.36 An example of a simple DHI risking matrix. Note that
the numbers in the cells are play specific; in this case they are
modelled on the Yegua example presented by Allen and Peddy
( 1993 ).
nohc) at say 0.6, and
the chance of success is calculated as (0.6 × 0.3)/
(0.6 × 0.3 + 0.6 × 0.7)
j
hc) equal to P(dhi
j
0.3. In this case the amplitudes
are making no modification to the geological risk.
To see what happens if a DHI is predicted
from modelling but it is absent in the seismic data,
then the same equation would be used but now
with nodhi
¼
but where one would be expected if significant quan-
tities of hydrocarbons were present. Combining DHI
information with general geological risking can effect-
ively be handled using Bayes
'
theorem, using the
hc) might
be 0.2, and the chance of no DHI given no hydro-
carbons might be 0.9; then the chance of hydrocar-
bons
instead of dhi.SoP(nodhi
j
following form:
P
ð
dhi
j
hc
Þ
P
ð
hc
Þ
ð
j
Þ¼
,
P
hc
dhi
given no DHI would be
(0.2 × 0.3)/
P
ð
dhi
j
hc
Þ
P
ð
hc
Þ
+P
ð
dhi
j
nohc
Þ
P
ð
nohc
Þ
(0.2 × 0.3 + 0.9 × 0.7)
0.09. The absence of the
DHI has substantially reduced the perception of the
chance of success.
The a priori risk, based on traditional geological
reasoning, has its influence on the outcome. Suppose
the geological chance of success is only 0.1 and a
convincing DHI is present as in the first example, so
that P(dhi
¼
where
P(hc
the probability of hydrocarbons given
the observed DHI,
P(dhi
j
dhi)
¼
the probability of occurrence of the
observed DHI if hydrocarbons are present,
P(dhi
j
hc)
¼
the probability of occurrence of the
observed DHI if no hydrocarbons are present,
P(hc)
j
nohc)
¼
nohc) is 0.2. The
chance of hydrocarbons would then be calculated as
(0.7 × 0.1)/(0.7 × 0.1 + 0.2 × 0.9)
j
hc) is 0.7 and P(dhi
j
the a priori probability of hydrocarbons
ignoring the DHI (i.e.
¼
the
'
geological
'
chance of
0.28. Even a
convincing DHI will therefore be allocated a modest
chance of success if the geological model is sceptical
of hydrocarbon presence.
The Bayesian approach outlined above is a useful
sense check, particularly for comparing with prob-
abilities derived from subjective compromises
between geological and DHI risking. Stratigraphic
prospects which are entirely driven by seismic ampli-
tudes present a difficulty for this approach, however,
as it would be difficult to assign a trap risk independ-
ent of the amplitude information.
¼
success),
P(nohc)
¼
(1
P(hc)) is the a priori probability of
no hydrocarbons.
To illustrate how this works, suppose that there is
a drillable prospect in a well understood play where
the chance of finding hydrocarbons using geological
risking in a particular trap is 0.3 (i.e. a typical risk for a
drillable prospect in a relatively mature basin).
A convincing DHI is present, such that it is very
unlikely it would be seen if no hydrocarbons are
present; thus P(dhi
249
j
hc) might be assigned a value of
Search WWH ::




Custom Search