Agreement DE-FC26-02NT15342, Seismic Evaluation of ...

More documents

Recommendations

Info

Impact of pressure on velocity and modulus Fluid saturation effect results in an increase of P-wave velocity and decrease of shear velocity. But more clearly, we see an increasing bulk modulus but maintain a constant for rigidity (shear modulus). The pore pressure effect results in a decreasing bulk modulus and increasing pore pressure. For dry rock the Vp/Vs ratio decreases with decreasing pressure. This means that bulk modulus drops faster than shear modulus with decreasing pressure for consolidated dry rocks. The fluid saturated Vp/Vs ratio is higher than dry bulk modulus due to increasing bulk modulus as well as P-velocity and decreasing S-velocity due to increasing density. Also, the fluid saturated Vp/Vs ratio increases with decreasing pressure due to faster drop of shear modulus than bulk modulus with decreasing pressure. An example of these measurements can be seen in Figure 4. Velocity VS. Pressure (G12677) Modulus VS. Pressure (G12677) 4.00 3.50 3.00 2.50 2.00 1.50 1.00 0 10 20 30 40 50 Pressure (MPa) 17.00 15.00 13.00 11.00 9.00 7.00 5.00 3.00 1.00 0 10 20 30 40 50 Pressure (MPa) Vpw Vsw Vpd Vsd Kw Gw Kd Gd Figure 4. Velocity versus pressure and modulus versus pressure. Water saturation estimation with Support Vector Machine Low saturation gas and commercial gas have close amplitude responses at shallow water reservoir (less than 2000 m) because a small amount of gas can dramatically reduce the bulk modulus of the gas-water mixture. Increasing the depth of reservoir, the bulk modulus of gas increase gradually. Gas saturation can be significant to affect bulk modulus of gas-water mixture. This improves the chances to discriminate low saturation gas from commercial gas with seismic data. we trained a Support Vector Machine (SVM) with a saturation log data from Gulf of Mexico, then apply it to regress the water Agreement DE-FC26-02NT15342, Seismic Evaluation of Hydrocarbon Saturation 18
saturation from seismic data. Validations of SVM regression on the gas andlow saturation gas show it is possible to separate commercial gas and low saturation gas at deep water reservoirs. Low saturation gas or “fizz” refers that the percentage of gas in the pore volume is small (less than 25%). The discrimination between low saturation gas and economical gas attracts much attention because wells are classified as “dry holes” even when small amounts of hydrocarbons exist in the reservoir. Discrimination between fizz and economical gas will reduce exploration cost and improve efficiency. At low-pressure condition (shallow depth < 2000 m), gas modulus is much less than 0.1 GPa. The presence of a small amount of gas can dramatically reduce the P-wave velocity of the reservoir, so “fizz” and economic gas saturations have nearly the same seismic responses. This well known physical phenomena can be modeled by Gassmann’s equation (Domennico, 1976). Under higher pressure condition (depth greater than 2000 m), modulus of gas-water mixture shows progressive decrease with increasing gas saturation (Han and Batzle, 2002). This improves the chances to discriminate low saturation gas from commercial gas with seismic data. Our objective of this study is to estimate the water saturation from seismic volume at deep-water reservoir. We first train a Support Vector Machine (SVM) to learn the water saturation log from seismic trace nearest to the wellbore. With this SVM, we estimate water saturation with all the traces in the seismic volumes from the commercial gas reservoir and low saturation gas reservoir. An SVM is an algorithm using selected subset of data (known as support vectors) in function estimation. We explain this algorithm in detail in the following section. SVM regression Support Vector Machines were introduced to the computer learning community in the mid 1990s (Vapnik, 1995) and are just beginning to applied in geophysical field (Kuzma, 2004; Li and Castagna, 2003; Zhao et al., 2005). They are most commonly used to solve very large classification problems such as handwritten digit recognition and document sorting. However, SVMs can also be used for regression (Smola and Scolkopf, 2004). This part introduces SVM regression based on Smola and Scolkopf’s tutorial. Suppose we are given training data {(x 1 , y 1 ), …, (x l , y l )}, x i ∈ R d is “d” dimension vector, y i ∈ R. In ε–insensitive Support Vector regression, our goal is to find a function f(x) that has at most ε deviation from the actually obtained targets y i for all the training data, and at the same time is as flat as possible. In other words, we do not care about errors as long as they are less thanε, but will not accept any deviation larger than this. Suppose our target function has a linear form f ( x) = 〈 w, x〉 + b with w ∈ R d , b∈ R (1) where 〈⋅,⋅〉 denotes the dot product in R d . Flatness in the case of (1) means that one seeks a small w. One way to ensure this is to minimize the norm, i.e. || w || 2 = 〈 w, w〉. We can write this problem as a convex optimization problem: Agreement DE-FC26-02NT15342, Seismic Evaluation of Hydrocarbon Saturation 19
Page 1 and 2: June 29, 2006 Seismic Evaluation of
Page 3 and 4: Abstract: During this last quarter
Page 5 and 6: Ursa Data .........................
Page 7 and 8: Figures: Figure 1. Relationship of
Page 9 and 10: Figure 38. Schematic of geometry an
Page 11 and 12: Figure 73. The extracted wavelet an
Page 13 and 14: Executive Summary: In this joint pr
Page 15 and 16: Results and Discussion Introduction
Page 17: Inversion of Sw and porosity from s
Page 21 and 22: The Lagrange function (5) has a sad
Page 23 and 24: ρ ma − ρ + ( ρ w − ρ g ) ×
Page 25 and 26: With a water saturation log calcula
Page 27 and 28: Figure 11. permeability versus poro
Page 29 and 30: Figure 13. Measured dry and brine s
Page 31 and 32: Figure 15. The HP (deep) sands have
Page 33 and 34: pressure-porosity relations. The n
Page 35 and 36: Figure 19. measure dry and brine sa
Page 37 and 38: Relations between Sw, porosity, and
Page 39 and 40: Figure 21. An example realization o
Page 41 and 42: otherwise the seismic properties we
Page 43 and 44: correctly reproduce those found in
Page 45 and 46: study of field data from the Teal S
Page 47 and 48: Figure 30. Spectral decomposed pres
Page 49 and 50: (Adapted from CSM Slope and Basin C
Page 51 and 52: Figure 34. Sequence stratigraphic f
Page 53 and 54: Figure 36. Schematic depiction of c
Page 55 and 56: Figure 38. Schematic of geometry an
Page 57 and 58: Figure 40. Middle slope channel (Ad
Page 59 and 60: Figure 43. Amalgamated sandstone sh
Page 61 and 62: CHANNEL AXIS SINGLE STORY CHANNEL C
Page 63 and 64: MEDIAL LEVEE/OVERBANK 7.62 cm Figur
Page 65 and 66: L1054 L1041 • Ursa Well A4 Green=
Page 67 and 68: 1.9 miles 37 ft Fizz 19 ft Oil 4.0
Page 69 and 70:
esults in shifts followed by subsid
Page 71 and 72:
Figure 55. Schematic of depositiona
Page 73 and 74:
4 miles 1.5 s 2 s 3 s Hydrocarbon I
Page 75 and 76:
In addition to comparing calculated
Page 77 and 78:
Figure 61. Shear velocity trend for
Page 79 and 80:
each reservoir interval modeled. Th
Page 81 and 82:
Figure 65. Histograms of frequency
Page 83 and 84:
the pressure trends were analyzed,
Page 85 and 86:
P Impedance Model Synthetic 2 -D Mo
Page 87 and 88:
Above Magenta reservoir model The A
Page 89 and 90:
Figure 73. The extracted wavelet an
Page 91 and 92:
saturation and 50% gas saturation.
Page 93 and 94:
Figure 77. The amalgamated channel
Page 95 and 96:
Figure 79. Channel complex model 2
Page 97 and 98:
Figure 81. AVO stack and gathers fo
Page 99 and 100:
Figure 82. Log signature includes g
Page 101 and 102:
Figure 85. Extracted wavelet and mo
Page 103 and 104:
Figure 87. Amalgamated sheet sand m
Page 105 and 106:
Figure 89. Pseudo logs, wavelet, sy
Page 107 and 108:
Figure 90. Medial basin floor chann
Page 109 and 110:
Figure 92. Pseudo logs, wavelet, sy
Page 111 and 112:
Figure 94. The post stack seismic e
Page 113 and 114:
Figure 96. Statistically extracted
Page 115 and 116:
Figure 98. The proximal levee overb
Page 117 and 118:
Figure 99. The response of the prox
Page 119 and 120:
Figure 101. The pseudo log, wavelet
Page 121 and 122:
Figure 102. The distal levee overba
Page 123 and 124:
Figure 104. The complex well log ba
Page 125 and 126:
then scaled to the model data by mu
Page 127 and 128:
Figure 107. The Ursa log displays g
Page 129 and 130:
Figure 109. Composite reflection co
Page 131 and 132:
Conclusions One of the main conclus
Page 133 and 134:
Table 5 Summary of the model facies
Page 135 and 136:
REFERENCES Alford, R. M., Kelly, K.
Page 137 and 138:
Gardner, M. H., Anderson, D. A., Bo
Page 139 and 140:
Mahaffie, M.J., 1994, Reservoir cla
Page 141 and 142:
Rothman, D.H., Grotzinger, J.P., Fl
Page 143 and 144:
Vail, P., 1977, Seismic Stratigraph
Page 145 and 146:
downlap (Mtchum, 1977): A base-disc
Page 147 and 148:
direction and downlap onto the tran
Page 149 and 150:
Agreement DE-FC26-02NT15342, Seismi
Page 151 and 152:
Page 153 and 154:
Page 155 and 156:
Page 157 and 158:
Page 159 and 160:
APPENDIX E: VALUES USED FOR RESERVO
show all

Agreement DE-FC26-02NT15342, Seismic Evaluation of ...

Create successful ePaper yourself

Delete template?

Save as template?