IP Advanced Training Course - Exercise

IP Advanced Training Course - Exercises

Advanced Exercise 1a – Cluster Analysis of the Blue Cat wells

Objectives:

Using the Cluster Analysis for Rock Typing module with the log data from the Blue Cat 1 well
develop a model to derive a Facies curve that captures the reservoir heterogeneity. The focus
should be on the rocks and not the fluids, i.e. facies and flow units. Apply this model to the
other Blue Cat wells.

1 Select the data to use in the Cluster Analysis module

1.1 Initially just select the Blue Cat 1 well.

1.2 Select logs, typically GR, RHOB, NPHI, and DT.

• If available (not in this case) can also use PE, K, U, Th, or even NMR curves.

• Consider fluid effects on logs – not related to Facies (Resistivity, N-D gas effect).

• Apply discriminators if required, e.g. badhole, (not required in this case).

1.3 Select the depth interval over which to build the model and over which to apply the model,
in this case the entire well for both.

2 Seed the Clusters and Run Clustering

2.1 Initially just use the default 15 clusters

and then

1 Advanced Exercise 1
IP Advanced Training Course - Exercise

3 Display the Crossplot and check that the results look meaningful

3.1 You can zoom on the entire plot with the zoom tools.

3.2 On an individual crossplot:

• You can scroll with a right click and drag.
• You can zoom by setting a zoom box with a left click and drag.

3.3 If you decide to manually edit any of the cluster means you can do so.

• Check Edit Cluster Means box, then right click and drag cluster mean symbol.

• If you do then you will have to again.

• The results will change, probably to something different than you expect.

3.4 Generally it is better not to manually edit the cluster means, that is a last resort if you
cannot get the desired result through Cluster Consolidation.

2 Advanced Exercise 1
IP Advanced Training Course - Exercise

Zoom Tools

Note – in this plot all of the crossplots on the upper right have been zoomed and scrolled.

4 Cluster Consolidation – different numbers of cluster groups (facies)

4.1 Initially, leave the Clustering Method set to the default as it usually works well:

4.2 You can consolidate to different numbers of final output facies and see which level of
resolution best matches the geology:
• There will be ‘extra’ clusters that are not really facies but are just artefacts of the
different vertical resolutions of the tools and depth matching errors.
• i.e. some clusters will just be bed boundaries.

4.3 The full resolution output, without any consolidation, should always be output,
• in this case 15 clusters (facies) will be output as All Clusters

4.4 Try 13, 11, 9, 7 and 5 clusters, copying to User set 1, 2, 3, 4 and 5 respectively:

3 Advanced Exercise 1
IP Advanced Training Course - Exercise

4.5 The Dendrogram will show you which cluster are joined together when consolidating
• i.e. in going from 15 to 13 clusters 6 and 8 will join first and then 4 and 5 will join

Cluster Grouping Dendrogram

29

28

27

26

25

24

23

22

21

20

19

18

17

16

11 7 9 8 6 3 5 4 2 1 12 10 15 14 13

Number of Cluster Groups : 13 Clustering Method : Minimize the within-cluster sum of squares distance

5 Cluster Output – different numbers cluster groups

5.1 Use default curve naming, which is by User Set, or edit to something more specific such
the number of Facies, as is the case in this example:

5.2 Use default Output Curve Set naming, which is LFacies, or rename to something more
specific such as the consolidation method, as is the case in this example:

5.3 Output the curves.

4 Advanced Exercise 1
IP Advanced Training Course - Exercise

6 View the results

6.1 Either view the Multi-well Plot or the individual well plot selected on the Input tab.

5 Advanced Exercise 1
IP Advanced Training Course - Exercise

7 Cluster Consolidation – 5 different Clustering Methods

7.1 You can output the same five levels of resolution for all 5 Clustering Methods, but you can
see how the different methods affect the result at one resolution, e.g. for 9 facies copying
to User set 1, 2, 3, 4 and 5 respectively:

6 Advanced Exercise 1
IP Advanced Training Course - Exercise

7.2 Observe the differences in the Dendrograms and consider the implications

Cluster Grouping Dendrogram

29

28

27

26

25

24

23

22

21

20

19

18

17

16

11 10 12 15 14 1 13 2 7 9 8 6 3 5 4

Number of Cluster Groups : 9 Clustering Method : Minimum distance between all objects in clusters

8 Cluster Output – 5 different Clustering Methods

8.1 Assign unique names and curve sets as desired, e.g. as below, and then

7 Advanced Exercise 1
IP Advanced Training Course - Exercise

8 Advanced Exercise 1
IP Advanced Training Course - Exercise

9 Decide which result best models the reservoir

9.1 This will be a manual process using other information about the reservoir, mudlogs,
cores, geology, seismic etc. Use the Copy and Paste Tracks functionality to build a
suitable composite log plot. In this case we have no extra data so compare the facies to
the PhiSw interpretation plot, such as this example comparing the 7 facies ‘minimize the
within-cluster sum of squares distance’ model.

9 Advanced Exercise 1
 IP Advanced Training Course - Exercise

10 Apply the selected model to the other wells

10.1 Select which other wells you wish to apply the model to, check that the selected curves
are correct, unselect the Use well for Model Build option for these new wells.

10.2 on the Cluster Means tab.

10.3 Select the Clustering Method, Number of cluster groups and Copy clusters to User set 1.

10.4 On the Cluster output tab rename the User set 1 output curve to the final name you want
and make sure it is in the curve set you want. Uncheck all the other out curves.

10.5 and then

10.6 Select a common Zone Set, in this case Tops and select Hawkins as the reference (right
click in the grey area between two log plots). Under Options select ‘Add Zone/Tops lines
to all Tracks’ and then select a colour. Resize and rescale accordingly.

10 Advanced Exercise 1
IP Advanced Training Course - Exercise

Advanced Exercise 1b – Cluster Analysis of the Carbonate well with NMR data

Objectives:

Develop separate Facies models using the conventional logs and NMR logs and compare.
Note: this requires inputs previously derived from the NMR interpretation module

1 Select the data to use to build a conventional facies model

1.1 Select the well ‘Carbonate’.

1.2 Select conventional logs for the first four Input Curves, e.g.GR, RHOB, NPHI and PEF in
curve set ‘IN’ and leave ‘Use Curve’ checked.

1.3 Select NMR bin porosities nmrPhiB1, nmrPhiB2, nmrPhiB3 and nmrPhiB4 in curve set
‘IN’ and uncheck ‘Use Curve’, i.e. initially we will just build a model based on the
conventional logs.

• Note that these bin porosities were generated in the NMR Interpretation module with
start-stop times of 0.3-3ms, 3-20ms, 20ms-90ms and 90-3000ms such that bin1
corresponds to the clay bound water and bin4 corresponds to the free fluid (without
any difference due to fluid).

1.4 and then

1.5 Use the default Clustering Method with a suitable number of clusters, e.g. 7, and copy to
User Set

1:

1.6 Rename the Output Curves to something suitable before e.g. :

1.7 View the plot, , and copy and paste the ‘User set 1’ track onto the CPI.

11 Advanced Exercise 2
IP Advanced Training Course - Exercise

2 Select the data to use to build an NMR based model

2.1 Uncheck the ‘Use Curve’ flag for the four conventional logs and check the ‘Use Curve’
flag for the for the four NMR bin porosities on the Input tab.

2.2 and then

• Note: if there are some negative values in one of the cluster means then the first
depth of the input data may need to be edited.

2.3 Again use the default Clustering Method with the same number of clusters but this time
copy to User Set 2:

2.4 Rename the All Clusters curve, uncheck the User set 1 Output, check the User set 2
output and rename it to something suitable e.g. :

2.5 View the plot, , and copy and paste the ‘User set 2’ track onto the CPI and
compare the two models.

12 Advanced Exercise 2