馃攳
1707 Dimitra Anevlavi, Automated Well Log Correlation Method visualized with real data - YouTube
Channel: unknown
[21]
Let's begin our presentation, by introducing
important terminology.
[25]
As far as Geological Surveys are concerned
a borehole is defined as any hole drilled
[31]
into the sub-surface for the purpose of extracting
or investigating the material at that particular
[36]
point.
[37]
During the process of well logging scientists
record various properties of the rock/fluid
[42]
mixtures along the well-log depth.
[45]
Sustaining accurate well-log records is expensive.
[48]
So why do we need it?
[50]
Boreholes provide one of the most important
sources of information available on the geology
[55]
and structure of Earths sub-surface.
[59]
The prediction of the presence of ore bodies
and hydrocarbons for example, is up to date
[63]
made possible due to the available well-log
data and their geologic interpretation.
[69]
After the process of well logging is finished,
it is up to geologists to determine corresponding
[74]
depths among the well logs, that are geographically
related to each other.
[79]
Geologic interpretation is conducted among
different types of well-logs.
[84]
Most of the times the corresponding depths
among groups of well-logs represent a single
[88]
geologic time in which sediments of similar
properties where deposited over large areas.
[94]
This interpretation is crucial as far as the
identification of basin areas of interest
[98]
are concerned, in the search for petroleum,
gas, minerals or other substances.
[106]
Recent research in well log correlation showed
that an improved method of simultaneous correlation
[111]
of multiple well-logs is possible.
[114]
An example of this research is shown in the
graph, where in a color-coded manner velocity
[118]
logs have been correlated in the means of
relative geologic time.
[123]
In order to facilitate the work of geologists,
in the means of the well-log data interpretation
[128]
we introduce now the main purpose our of project.
[135]
The goal for us was to evaluate the previous
Well-Log Correlation method through visualization
[140]
with real borehole data, from the Netherlands
and Dutch sections repository.
[146]
For that reason a Python code was developed
for the preprocessing of data, that includes:
[152]
Data extraction based on geologic parameters,
Plotting options and Data storage in various
[158]
formatting or in binary form
We also facilitated the data entry for the
[166]
Well-Log Correlation by adding utilities for
the merging of well-log data in our code.
[171]
This way geologists can create case studies
based on real borehole data and as a next
[177]
step the visualization of the correlation
is more flexible.
[182]
An example of porosity logs correlated in
the means of relative geologic time is shown
[187]
in a color-coded manner as well in the graph.
[189]
Well log correlation in the past has been
conducted solely by geologists, but automated
[195]
methods speed up this procedure.
[198]
In any case they should be tested extensively
before put into industrial use.
[203]
The visualization technique, as an output
of our project, helps researchers discover
[207]
limitations and evaluate the accuracy of the
current method.
[211]
Of course only experienced geologists will
provide us with the final conclusions.
[220]
Recent research also conducted in the field
of facies classification, another area of
[225]
geologic expertise, provides us with promising
test cases of deep convolutional networks
[230]
that facilitate geologic interpretation.
[234]
The adaptation of machine-learning algorithms
for the well-log Correlation problem sounds
[238]
promising.
[239]
In that way our utilities could prove to be
suitable generators of test cases for the
[243]
training of the deep neural network algorithms.
[247]
Stay tuned for more exciting projects!!
[250]
Thanks for your attention
Most Recent Videos:
You can go back to the homepage right here: Homepage