0.15 0.17 0.19 0.21 0.23 0.25 0.27 0.29 0.31 0.33 0 1000 2000 3000 4000 5000 6000 Predicting accuracy for non-FP facies Major range of variogram, m INTRODUCTION DATA AND METHODOLOGY Fig. 4. Locations of the study area (a) wells (b). RESULTS Can Decompaction Increase the Predicting Accuracy for Depositional Facies in Geological Modelling? Fengde Zhou ab ; Daren Shields ab ; Alistair W. Buchanan a ; Steven Tyson ab ; Joan Esterle ab a School of Earth Sciences , The University of Queensland b Centre for Coal Seam Gas, The University of Queensland CCSG Research Review, 9 December 2016 PROJECT: SURAT SUPERMODEL II – AN INTEGRATED GEOLOGICAL FRAMEWORK FOR CSG Wells used for modelling (19) 1.4 km Wells used for checking (28) N 2.5 km Moonie Dalby Chinchilla Wandoan Injune Taroom SURAT BASIN GALILEE BASIN BOWEN BASIN Taroom Trough NEW ENGLAND OROGEN NAMBOUR BASIN MARYBOROUGH BASIN BRISBANE Rosewood Walloon Coalfield 28˚S 26˚S 148˚E 152˚E 150˚E EROMANGA BASIN Limit of Surat and associated Mesozoic basins Limit of Permo-Triassic Bowen and Galilee basins Fault Walloon subgroup Great Artesian Basin Syncline Anticline Local model area 0 100km B A 1500*1500m*5m Depositional Facies from logs 2000*2000m*5m 3000*3000m*5m Fig. 1. Peat/coal compaction changes thickness hence structure (after Titheridge, 2016?) Fig. 5. An example showing depositional facies comparison between log and model predicted depositional facies at one borehole. A workflow in Petrel was used to automatically calculate the prediction accuracy by using different variogram ranges, major direction, and vertical range. Major range varies from 0 to 5 km; major-range/minor-range ratio varies from 1 to 2; vertical range varies from 1 to 5; major range direction varies from 0 to 180˚. Each case includes 200 realisations. Well correlation and/or seismic interpretation based stratigraphic framework is basic for geological modelling. However, peat/coal compaction may change the sedimentary structure (Fig. 1). Also, there is big debate about the peat/coal compaction ratio as shown in Fig. 2. A local area, about 6×8 sq. km with 47 wells’ wireline logs is selected for this study. The well spacing ranges from 0.8-1.5 km. Among those wells, 19 wells were used in modelling and 28 wells were used as monitors for prediction accuracy assessment. Two sets of models were generated, one with un-decompacted wireline logs and the other with de-compacted wireline logs. Fig. 3 shows the compaction ratio for different lithologies. A workflow was generated and used to optimise the variogram geometry. peat lignite Sub- bituminous bituminous anthracite Peat : coal compaction ratio 0.1 1.0 10 100 60(h) 31(f) 11(e) 1.4(c) 5(b) 7.7(a) 2.3(a) 1.2(b) 1.1(c) 1.4(d) 1.7(f) 1.2(g) 7(i) 20(j) Fig. 2. Variations of peat:coal compaction ratio in the literature (after Widera, 2015). a-Bloom (1964); b-Bird et al. (2004); c- Widera (2013a); d-Hurnik (1990); e-Smith and Clymo (1984); f-White (1986); Nadon (1998); h-Winston (1986); i-cited by Ryer and Langer (1980); j-Elliot (1985). Fig. 7. One each realisation of depositional facies by truncated Gaussian method with (a) un- decompacted wireline logs and (b) decompacted wireline logs (decompaction ratios of Sandstone, Dirty Sst, Siltstone, mudstone, carb coal, and coal are 1.7, 1.7, 1.9, 2.5, 4.5, and 18, respectively). (a) (b) 0.15 0.16 0.17 0.18 0.19 0.2 0.21 0.22 0.23 0.24 0 1000 2000 3000 4000 5000 6000 Predicting accuracy for non-FP facies Major range of variogram, m Fig. 8. Predicting accuracy against (a) major range of variogram; (b) major range direction; (c) vertical range of variogram for coal with un-decompacted wireline logs. Major range at about 2 km yields highest predicting accuracy of about 23 %. Major range direction and vertical range have less impact on predicting accuracy for coal. 0.15 0.16 0.17 0.18 0.19 0.2 0.21 0.22 0.23 0.24 0 50 100 150 Predicting accuracy for non-FP facies Major range direction, ˚ 0.15 0.16 0.17 0.18 0.19 0.2 0.21 0.22 0.23 0.24 0 1 2 3 4 5 6 Predicting accuracy for non-FP facies Verticle range of variogram, m ? (a) (b) Fig. 9. Predicting accuracy against major range of variogram with decompacted wireline logs. Major range at about 3 km yields highest predicting accuracy of about 0.31 %. Note that the different decompaction ratio will affect the grid numbers for different facies. 0.31 0.23 Section location Fig. 6. Correlation based on un- decompacted (a) and decompacted (b) wire logs and depositional facies. W E 14 8 15 9 60 47 25 8 3 11 PC CH FP MS CS 0 10 20 30 40 50 60 Decompacted Un- decompacted Percentage, % Fig. 7. Histogram of facies. y = 252.13x -0.527 R² = 0.9843(Coal) y = 81.296x -0.167 R² = 0.863(Carb Coal) y = 100.1x -0.131 R² = 0.9564(MSst) y = 156.55x -0.145 R² = 0.95689(SSst) y = 166.69x -0.139 R² = 0.9535(DSst) y = 177.35x -0.145 R² = 0.9597(Sst) 0 10 20 30 40 50 60 70 80 90 100 0 500 1000 1500 Compaction ratio, % Burial depth, m CONCLUSIONS A workflow has been generated in Petrel to compare the predicting accuracy for depositional facies with different variogram geometry. Major range has a strong relationship with predicting accuracy compared with variogram direction and vertical variogram range. The incremental predicting accuracy is about 8% with decompacted logs. More decompaction ratios will be assessed in future. Fig. 3. Variation of compaction ratio of thickness for different facies with burial depth (after Buchanan, 2015). Acknowledgments: We thank Renate Sliwa and OGIA for provision of well markers and a normalised wireline data set. We also acknowledge all the coal seam gas companies who provide data to the DNRM, Centre for Coal Seam Gas for the support, and Schlumberger for providing the license of Petrel. (a) W E (b) W E
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
0.15
0.17
0.19
0.21
0.23
0.25
0.27
0.29
0.31
0.33
0 1000 2000 3000 4000 5000 6000
Pre
dic
tin
g ac
cura
cy f
or
no
n-F
P
faci
es
Major range of variogram, m
INTRODUCTION
DATA AND METHODOLOGY
Fig. 4. Locations of the study area (a) wells (b).
RESULTS
Can Decompaction Increase the Predicting Accuracy for Depositional Facies in Geological Modelling?
Fengde Zhou ab; Daren Shields ab; Alistair W. Buchanan a; Steven Tyson ab; Joan Esterle ab
a School of Earth Sciences , The University of Queenslandb Centre for Coal Seam Gas, The University of Queensland
CCSG Research Review, 9 December 2016
PROJECT: SURAT SUPERMODEL II –AN INTEGRATED GEOLOGICAL FRAMEWORK FOR CSG
Wells used for modelling (19)
1.4 km
Wells used for checking (28)
N
2.5 km
Moonie
Dalby
Chinchilla
WandoanInjune
Taroom
SURAT BASIN
GALILEE BASIN BOWEN BASIN
Taroom Trough
NEW
ENGLAND
OROGEN
NAMBOURBASIN
MARYBOROUGHBASIN
BRISBANE
RosewoodWalloonCoalfield
28˚S
26˚S
148˚E 152˚E150˚E
EROMANGA BASIN
Limit of Surat and associated Mesozoic basins
Limit of Permo-Triassic Bowen and Galilee basins
Fault
Walloon subgroup
Great Artesian Basin
Syncline Anticline
Local model area
0 100km
B
A
1500*1500m*5m
Depositional Facies from logs
2000*2000m*5m3000*3000m*5m
Fig. 1. Peat/coal compaction changes thickness hence structure (after Titheridge, 2016?)
Fig. 5. An example showing depositional facies comparison
between log and model predicted depositional facies at one borehole. A
workflow in Petrel was used to automatically calculate the prediction accuracy by using different variogram ranges, major direction, and vertical
range. Major range varies from 0 to 5 km; major-range/minor-range ratio
varies from 1 to 2; vertical range varies from 1 to 5; major range
direction varies from 0 to 180˚. Each case includes 200 realisations.
Well correlation and/or seismic interpretation based stratigraphic framework is basic for geological modelling. However, peat/coal compaction may change the sedimentary structure (Fig. 1). Also, there is big debate about the peat/coal compaction ratio as shown in Fig. 2. A local area, about 6×8 sq. km with 47 wells’ wireline logs is selected for this study. The well spacing ranges from 0.8-1.5 km. Among those wells, 19 wells were used in modelling and 28 wells were used as monitors for prediction accuracy assessment. Two sets of models were generated, one with un-decompacted wireline logs and the other with de-compacted wireline logs. Fig. 3 shows the compaction ratio for different lithologies. A workflow was generated and used to optimise the variogram geometry.
peat lignite Sub-bituminous
bituminous anthracite
Peat
: co
al c
om
pac
tio
n r
atio
0.1
1.0
10
10060(h)
31(f)
11(e)
1.4(c)
5(b)7.7(a)
2.3(a)
1.2(b) 1.1(c)
1.4(d) 1.7(f) 1.2(g)
7(i)
20(j)
Fig. 2. Variations of peat:coal compaction ratio in the literature (after Widera, 2015). a-Bloom (1964); b-Bird et al. (2004); c-
Widera (2013a); d-Hurnik (1990); e-Smith and Clymo (1984); f-White (1986); Nadon (1998); h-Winston (1986); i-cited by Ryer and Langer (1980); j-Elliot (1985).
Fig. 7. One each realisation of
depositional facies by truncated
Gaussian method with (a) un-
decompacted wireline logs and (b) decompacted
wireline logs (decompaction ratios
of Sandstone, Dirty Sst, Siltstone,
mudstone, carb coal, and coal are 1.7, 1.7, 1.9, 2.5, 4.5, and 18,
respectively).
(a)
(b)
0.15
0.16
0.17
0.18
0.19
0.2
0.21
0.22
0.23
0.24
0 1000 2000 3000 4000 5000 6000
Pre
dic
tin
g ac
cura
cy f
or
no
n-F
P f
acie
s
Major range of variogram, m
Fig. 8. Predicting accuracy against (a) major range of variogram; (b) major range direction; (c) vertical range of variogram for coal with un-decompacted wireline logs. Major range at about 2 km yields highest predicting accuracy of about 23 %.
Major range direction and vertical range have less impact on predicting accuracy for coal.
0.15
0.16
0.17
0.18
0.19
0.2
0.21
0.22
0.23
0.24
0 50 100 150
Pre
dic
tin
g ac
cura
cy f
or
no
n-F
P f
acie
s
Major range direction, ˚
0.15
0.16
0.17
0.18
0.19
0.2
0.21
0.22
0.23
0.24
0 1 2 3 4 5 6
Pre
dic
tin
g ac
cura
cy f
or
no
n-F
P f
acie
s
Verticle range of variogram, m
?
(a)
(b)
Fig. 9. Predicting accuracy against major range of variogram with decompacted wireline logs. Major range at
about 3 km yields highest predicting accuracy of about 0.31 %. Note that the different decompaction ratio will affect the grid numbers for different facies.
0.31
0.23
Section location
Fig. 6. Correlation
based on un-decompacted
(a) and decompacted (b) wire logs
and depositional
facies.
W
E
148
15
9
60
47
25
83
11
PC CH FP MS CS0
10
20
30
40
50
60
DecompactedUn-decompacted
Perc
enta
ge, %
Fig. 7. Histogram of facies.
y = 252.13x-0.527
R² = 0.9843(Coal)
y = 81.296x-0.167
R² = 0.863(Carb Coal)
y = 100.1x-0.131
R² = 0.9564(MSst)
y = 156.55x-0.145
R² = 0.95689(SSst)
y = 166.69x-0.139
R² = 0.9535(DSst)
y = 177.35x-0.145
R² = 0.9597(Sst)
0
10
20
30
40
50
60
70
80
90
100
0 500 1000 1500
Co
mp
acti
on
rat
io, %
Burial depth, m
CONCLUSIONS
A workflow has been generated in Petrel to compare the predicting accuracy for depositional facies with different variogram geometry.
Major range has a strong relationship with predicting accuracy compared with variogram direction and vertical variogram range.
The incremental predicting accuracy is about 8% with decompacted logs. More decompaction ratios will be assessed in future.
Fig. 3. Variation of compaction ratio of thickness for different facies with burial depth (after Buchanan, 2015).
Acknowledgments: We thank Renate Sliwa and OGIA for provision of well markers and a normalised
wireline data set. We also acknowledge all the coal seam gas companies who provide data to the DNRM, Centre for Coal Seam Gas for the support, and Schlumberger for providing the license of Petrel.
(a)
W E
(b)
W E
Related Documents
