Transcript
RockSI Tutorial
Tutorial outline Introduction RockSI software overview
Part 1: Log Analysis Part 2: Rock Physics Template (RPT)
Part 3: Creation and calibration of userdefined PEMs Part 4: Uncertainty analysis using Monte Carlo simulations Part 5: LithoSI analysis using simulated PDFs
May 2015
2
Introduction
Rock Physics: the interdisciplinary glue Petrology
Mineralogy Coring
Stratigraphy
Geology
Logging
Petrophysics
Structural Geology
History Matching
Reservoir Modeling May 2015
Rheology
Reservoir Engineering
Flow Simulation
Gravity
Geophysics Electro Magnetism Seismic
4
Rock Physics: Applications Interpretation and Quality Control: Well logs Inversion results
Predictions: Deterministic: Logs prediction, pseudo-well creation Part of petrophysical inversion
Stochastic: Augmentation of litho-classification training sets through Monte Carlo simulations Uncertainty analysis Time-lapse feasibility studies
Trend modeling Inverse Rock Physics Transforms May 2015
5
Rock Physics: some definitions Note that there are a number of terms used in the creation of rock physics models that are almost synonymous, but have slightly different meanings. These terms are: the Rock Physics Model (RPM) the Petro-Elastic Model (PEM) the Rock Physics Template (RPT) In the following slides we will define these three terms. In the RockSI program, and this tuorial, we will mainly be concerned with the creation of Petro-Elastic Models, or PEMs, and their display are Rock Physics Templates ( RPTs). May 2015
6
Rock Physics Model (RPM)
A Rock Physics Model, or RPM, is a model which describes how a rock react to a given stimulus, e.g.: Change of pressure (seismic waves) Change of temperature Electric/magnetic current
May 2015
7
Petro-Elastic Model (PEM)
A Petro-Elastic Model, or PEM, is a model which describes how a rock reacts to seismic waves. The elastic response mainly depends on: Rock composition (minerals and fluids) Rock texture (spatial distribution of minerals and porosity)
May 2015
8
Petro-Elastic Model (PEM)
A PEM is comprised of a set of equations which model the elastic response of a rock from its petrophysical properties. The PEM is an important link between the well and seismic data.
Porosity Sw Lithology
PEM Vp Vs Density SEISMIC
WELLS May 2015
9
Rock Physics Template (RPT)
A Rock Physics Template, or RPT, is the projection of a Rock Physics Model in a particular N-dimensional space: Type of pores
Change of P-Impedance due to gas injection (ΔSG = +0.6)
May 2015
Cracks
Intergranular
Vugs
10
RockSI Overview
RockSI: Overview The RockSI program has two main windows with tabs: on the left is the input menu and on the right is the QC display.
May 2015
Input
QC and display
12
RockSI: Overview The RockSI program has two main windows with tabs: on the left is the input menu and on the right is the QC display.
May 2015
13
RockSI: Overview The Input tab allows the user to: 1.
2. 3.
4.
Select the input data: - Wells (in Depth or TWT) - SIModels - Seismic (Angle Stacks) - Wavelets Select rock types and PEMs. Create tables showing the correspondence between PEM variables and well logs or model attributes. Create tie point for each well in depth (TVD) vs time (TWT).
1
2
3
4
May 2015
14
RockSI: Overview 3
The PEM Library tab allows the user: 1.
2.
A list of available Petro-Elastic Models: - The ability to create new models - The ability to load existing models - Access to standard models
1
2
The definition of each PEM: - Input variables - Output variables with associated equations
3. Three editing modes: - Read-Only to see the content of the PEM - Parameters editing to edit the value of the internal parameters of the PEM - Advanced editing to modify the input/output and equations of the PEM May 2015
15
RockSI: Overview The variables tab gives: 1.
1
2
3
4
A list of variables: any variable with a valid name is available for use in PEMs and QCs Default list
User-defined if more variables are needed. 2.
Variable display limits used in the Log Viewers and CrossPlots tabs
3.
Variable colormap used in the Log Viewers and Cross-Plots tabs
4.
Variable unit used in the Cross-Plots tab
May 2015
16
RockSI: Overview Fluids tab: Fluid property computation: Density Velocity Bulk modulus Based on Batzle & Wang equations Brine, oil and gas properties as a function of pressure, temperature and fluid composition
May 2015
17
RockSI: Overview
1 2
Simulation tab: 1.
Selection of PEM used in Monte Carlo simulation
2.
Definition of the input distributions and correlations
3.
Simulated correlations
4.
Export of simulated data as ASCII file or xsel file used for lithoclassification (LithoSI)
3
4
May 2015
18
RockSI: Overview Log Viewers Options tab: Definition of the display template in the Log Viewers tab:
1
1. Curve display options 2
2. Log table (selection of variables not the actual logs defined in the Input tab) 3
3. Angle Stack table (selection of keywords not the actual stacks defined in the Input tab)
May 2015
19
1
RockSI: Overview 2
Cross-Plots Options tab: 1.
Definition of the display template in the Cross-Plots tab
2.
Graphical options for the display of RPTs from the PEMs defined in the PEM Library tab
3.
Filters to apply on the cross-plots
4.
Graphical options for the display of the data selected in the Input tab
5.
Graphical options for the display of the predicted logs
6.
Graphical options for the display of the data simulated in the Simulation tab
May 2015
3
4
5
6 20
RockSI: Overview Log Viewers tab: Logs and seismic traces display for each well selected in the Input tab (template defined in Log Viewers Options tab)
May 2015
21
RockSI: Overview Cross-Plots tab: Cross-plot display (template defined in Cross-Plots options tab)
May 2015
22
RockSI: Overview Menus/Toolbars: 1. Load session 2. Save session
3. Create a new PEM
1
2
3
4
5
6
4. Load an existing PEM
5. Save a PEM 6. Update all the current QCs
May 2015
23
Part 1 Log Analysis
Data interpretation and QC QC and interpret log data through templates: Rock type, fluid type Petrophysical and elastic properties
May 2015
25
Part 1: Log Analysis
This first exercise will guide the user through the following tasks: Identify the different lithologies encountered at the wells. Identify the main petrophysical variables that impact the rock elastic response. Understand the impact of each of those variables.
Those tasks are not strictly talking part of the rock physics modelling but are a preliminary work that will ease the modelling task.
May 2015
26
Part 1: Log Analysis Start the HRS10 Geoview program by clicking Geoview icon on your desktop:
When you launch Geoview, the first window that you see contains a list of any projects previously opened in Geoview. Your list will be blank if this is the first time you are running Geoview.
May 2015
27
Part 1: Log Analysis For this exercise, we will start with a prepared project. Before doing that, we will set all the data paths to point to the location where we have stored the workshop data. To do that, click the Settings tab: To change all of the directories to the same location, click on the option Set all default directories and then click the button to the right:
Then, in the File Selection Dialog, select the folder which contains the workshop (check with the instructor the folder location): May 2015
28
Part 1: Log Analysis After setting all three paths, the Geoview window will now show the selected directories (note that yours may be different):
When you have finished setting all the paths, click Apply to store these paths:
May 2015
29
Part 1: Log Analysis Go to the Projects tab. Click Find Project:
Select the project RockSI.prj. Click OK to open the project:
May 2015
30
Part 1: Log Analysis Now the GEOVIEW window looks like this: In the project manager, there are four preloaded deviated wells.
May 2015
31
Part 1: Log Analysis First, look at the tabs to the left of the GEOVIEW window. You will see that one of those tabs is called Processes. Click on that tab to see a list of all the operations which are available in Geoview. Each of the processes can be expanded. For example, if you click on the PetroElastic Modeling (RockSI) option and the RockSI option , the following expanded list is seen:
May 2015
32
Part 1: Log Analysis Double click on New RockSI Session:
A new tab called RockSI is created in the GeoView window: On the left of the RockSI window are a series of tabs for data control. On the right side of the window, there are two tabs for displaying the Log viewer and Crossplot window. (Note that depending on the project itself, some data may already be loaded, but we will re-do the loading). May 2015
33
Part 1: Log Analysis Under the Input tab, select the four available wells by clicking Select All Items button. This will select all four wells and create a QC display automatically in the Log Viewers tab for each of the wells, as well as the synthetic trace shown in red at the far right hand side of the plot.
In the seismic table, add two new lines by clicking the + sign button twice.
May 2015
34
Part 1: Log Analysis For this exercise, there are 3 angle stack seismic volumes at 10, 20 and 30 degrees. Select each of the seismic volumes under the column Seismic by left-clicking on the down arrow and selecting the files Amp_10, Amp_20 and Amp_30. Specify the Amplitude as Seismic. Specify the Angle for each seismic volume (10, 20, 30). Finally, select the wavelets that have been previously extracted from the corresponding seismic volumes: Wav_10_1, Wave_20_1, Wav_30_1. Click the Update QCs button at the top of the dialog. Also note that three logs have already been selected, which will be automatically displayed in the Log Viewers window. May 2015
Update QCs
35
Part 1: Log Analysis After selecting the well and seismic data, the Log Viewers window on the right side of the RockSI window will look like this: As mentioned before, three logs are now displayed. On the right of the Log Viewers window, it shows the extracted seismic traces in black from each of the seismic volumes. The red line is the calculated synthetic trace using the logs and wavelet.
Synthetic
Extracted seismic
Note that well-seismic ties were performed prior to running RockSI. Hence, the good match between synthetic and extracted seismic. May 2015
36
Part 1: Log Analysis We will add a few variables for QC purpose. You can do that by clicking on the ‘+’ button a few times to add the following variables. Any extra row can be removed by clicking the ‘-’ button. Proceed to add: Vsh, Phi, Sw, Ip, VpVs, Vqz and Facies.
May 2015
37
Part 1: Log Analysis Next we need to map the variables to the actual log curves. In each of the well tab, there are Input Variable and QC Variable mapping tables. Input Variable table by default list P-wave(Vp), S-wave (Vs) and Bulk Density (Rho) variables. This table will automatically display the input variables required by the Petro-Elastic Model (PEM) selected in subsequent steps. QC Variables are mainly used in various QC plots.
Specify the variables and log curves mapping as shown. Click on the “Replicate selection to other tabs” button to apply the changes to all the well tabs.
May 2015
38
Part 1: Log Analysis Click on Log Viewers Options tab. Highlight the first row and click on the
+ sign to add three rows. Change the logs in these new bottom three rows to Vp, Vs and Rho. Select Shale volume (Vsh), porosity (Phi), and water saturation (Sw) for the top three logs in the well log table. We also want to color-code these logs by selecting them under the Color column.
May 2015
39
Part 1: Log Analysis We can optimize the well log display in the General Options tab: Update the Max value for the porosity log type to 0.35. We can also change the color scheme by clicking on the color Palette and select the desired color scheme. Here, we will keep the default selection. Click on the Update QCs button and see the new display in the next slide.
May 2015
40
Part 1: Log Analysis
Three newly added logs are shown on the left of the Vp track. May 2015
41
Part 1: Log Analysis Click on Cross-Plots Options tab. The Cross-plot window will appear on the right. Select the logs for the X and Y axis. Click the Update QCs button.
May 2015
42
Part 1: Log Analysis
Select Vsh to color code the crossplot, then click on Replicate current value to apply to the other cross plot. One option of the filter is to limit the number of points which by default is 100,000. The number of points are randomly selected from these 4 wells. We can also select the wells of our interest only. Display legends turns on/off of the color keys on the plots.
May 2015
43
Part 1: Log Analysis The right of the RockSI window shows the cross plot of the selected wells. The red points have a higher volume of clay (shale), while the blue points have a lower volume. From the figure, we can clearly see that the porosity decreases with increasing Vsh. The shale lithology can be easily identified by the cluster of red points as pointed by the arrow.
May 2015
44
Part 1: Log Analysis Repeat previous steps to display the effects of volume of quartz (Vqz).
The sand is now identified on the opposite side from the Vsh case, as shown by the arrows
May 2015
45
Part 1: Log Analysis Repeat for Phi (Porosity) as the variable used to color-code each of the cross-plots.
Again, the high porosity clusters roughly correspond to the sand as pointed by the arrows.
We can check additional properties like: Sw(Water Saturation), Vcalc (Calcite Volume), Vcoal (Coal Volume) … etc. May 2015
46
Part 1: Facies Logs Next, we will either read in or create a Facies log for each well to specify the associated rock type for each sample. First we will define a series of Facies by clicking on Define Facies Classification from the process list in the project manager.
On the dialogue that appears, click on the Load button:
May 2015
47
Part 1: Facies Logs
Then, in the file chooser that appears, locate the file Facies.xml and select it and click Open. If you can’t find a facies xml file, skip the next slide. May 2015
48
Part 1: Facies Logs When the facies xml file is loaded, it will look like this, with Coal, Shale, Sand, Oil and Calcite facies:
Then, click OK. May 2015
49
Part 1: Facies Logs
Next we want to find out if the four wells contain facies logs. To do this, select Project Data / Well / well A and see if one of the logs is called FACIES, as shown on the right. Check wells B, C, and D also.
May 2015
50
Part 1: Facies Logs Go to the RockSI window and click on the Log Viewers Options tab. Click on the first line in the log viewer table, and click the + sign to create a new row. Select Facies to display and the Facies as color-coding variable.
May 2015
Click here before the + sign
51
Part 1: Facies Logs Go to the General Options tab. Click on the + sign to add a row at the bottom of the table, select Facies for the Property Type. Specify 0-7 for min-max range. Click the color palette and select Facies.
May 2015
52
Part 1: Facies Logs Now the Facies log is displayed on the first track of the log viewer window. From the figure, we can see clearly that the Facies log is closely related to the other color coded logs (Vsh, Phi, Sw and Vp). At depth of 1930 m, it is the Oil sand and Shale interface. By tracking the curser position, we can see a strong reflection at this interface. We can also check the other wells by clicking those well tabs. May 2015
53
Part 1: Facies Logs Click on Cross-Plot Options tab. Select Facies as the variable used to color-code each of the cross-plots.
The cross plot windows now look like this. The color key of the Facies displayed is the color of the five facies we have just defined.
May 2015
54
Part 2 Rock Physics Template (RPT)
Different types of effective medium models
Empirical models: Gardner Greenberg-Castagna
Effective Medium models: Granular models: Pack of spheres with interstitial porosity Used to model high porosity rocks such as unconsolidated and lightly cemented sandstones Inclusion models: Background and inclusions (spheres, needles, penny cracks): Self-consistent model (SC) Differential effective medium (DEM) Used to model low permeability rocks such as tight sandstones and carbonates
May 2015
56
Soft (unconsolidated) sandstone model
Pack of identical smooth spheres Porosity filled by a single phase
May 2015
Real rock
Model
Multiple mineral types Multiple fluids Complex texture
Effective grain Effective fluid Simple texture 57
Soft (unconsolidated) sand model Effective matrix: mix of Quartz and Clay 1 M matrix Vclay M clay 1 Vclay M quartz 2
ρmatrix Vclay ρclay 1 Vclay ρquartz
1
Hill average 1 Vclay
Vclay 2 M clay M quartz
ρ fluid SW ρW SO ρO SG ρG Effective fluid:
K fluid
May 2015
1 SW SO SG KW K O K G
(Wood equation) 58
Soft (unconsolidated) sand model Effective frame: C 1 φC G K HM 2 18π 1 PR 2
GHM
2
2 matrix 2 matrix
Peff
1 3
2 2 5 4 PRmatrix 3C 2 1 φC Gmatrix Peff 2 2 52 PRmatrix 2π 1 PRmatrix
K dry
1 φ φC 1 φ φC 4 4 K HM GHM K matrix GHM 3 3
φ φC G G 9 K 8GHM GHM HM HM 6 K HM 2GHM May 2015
1 3
(Hertz-Mindlin model)
4 GHM 3
1
1 φ φC G 9 K 8GHM Gmatrix HM HM 6 K HM 2GHM
GHM 9 K HM 8GHM 6 K HM 2GHM 59
Soft (unconsolidated) sand model Combination of frame and fluid: 2
K dry 1 K matrix K B K dry K dry 1 K fluid K matrix K matrix2
(Gassmann equation)
Elastic attributes: fluid 1 matrix VP VS May 2015
4 KB G 3
G
60
Unconsolidated sand model D
Modified Voigt
Unconsolidated Sand
PHI
C B
A SG
D May 2015
C
B
A
61
Part 2: Rock Physics Template (RPT) In the previous exercise, we used volume logs to interpret the different lithologies. When the volume logs are not available, we can use the standard Petro Elastic Models (PEMs). You can also create a new PEM with your own equations. In this next exercise, we will learn how to display Unconsolidated Sand RPT using one of the predefined PEMs. Also see how we can display these RPT on standard GEOVIEW crossplots.
Go to the PEM Library tab. Right click on the PEMs item and select the Load PEMs.
May 2015
62
Part 2: Rock Physics Template (RPT) A number of published models are included in RockSI. Click different tabs to review them. Select the Unconsolidated Oil Sandstone model To see the full documentation of this model, click the Information Icon.
May 2015
63
Part 2: Rock Physics Template (RPT) This brings up the documentation for the Unconsolidated Oil Sandstone model as shown on the right.
May 2015
64
Part 2: Rock Physics Template (RPT)
Click Apply to load the Unconsolidated Oil Sandstone PEM, then close the dialog.
May 2015
65
Part 2: Rock Physics Template (RPT) After loading the sandstone PEM, we can see a list of reservoir conditions, mineral and fluid properties. Click the Advanced toggle to change to full editor mode.
May 2015
66
Part 2: Rock Physics Template (RPT) In Advanced editing mode, you can modify the relationships, parameters or define new variables as needed.
We will not make any changes to the PEM in this exercise.
May 2015
67
Part 2: Rock Physics Template (RPT) Before display the unconsolidated sandstone RPT, it will be useful to limit the crossplot QCs to show only sandstone facies. It requires the FACIES log curve and Classification table to be specified. In the Input tab, select the option Several rock types and the Facies table imported in previous exercise as the rock classification. Scroll down to well tabs parameters section. May 2015
68
Part 2: Rock Physics Template (RPT) Specify the FACIES log curve for Well A.
Then, click the Replicate Icon to select the same log curve for the other three wells.
May 2015
69
Part 2: Rock Physics Template (RPT) In the Cross-Plots Options tab, modify the first crossplot to show Vp vs Rho. Toggle on Display only rock type and select Sand. Then click the Update QCs icon. The plots now show data points for only the Water Sand.
May 2015
70
Part 2: Rock Physics Template (RPT) To display both Water Sand and Oil Sand facies, we will use the Display only range option. We need to specify the Facies classification and the range 2.5 to 4.5. Also, the Display only Rock Type option must be unchecked.
Range: 2.5-4.5
May 2015
71
Part 2: Rock Physics Template (RPT) We are now ready to display the RPT. Go to the CrossPlots Options tab Select PEM Unconsolidated Oil Sandstone, Main graduation: Phi and check Secondary Graduation and select Vsh. Fill in the PEM parameters as shown in the table. With Sw= 0, we are modelling 100% Oil saturation.
May 2015
72
Part 2: Rock Physics Template (RPT)
The soft sandstone template is as shown in the crossplot window. Meshes are displayed in the plots that represent the Oil sandstones with Phi between 0.1 and 0.35 and Vsh between 0 and 0.4.
May 2015
73
Part 2: Rock Physics Template (RPT)
Set Sw to 1 to model the water sand.
May 2015
74
Part 2: Rock Physics Template (RPT)
Observe that the meshes now match the water sand. You can model any combination of two parameters in the table with ranges.
May 2015
75
Part 2: Rock Physics Template (RPT) The color of the RPT can be changes by clicking on the color selector icon.
You can control how often the increments of the mesh without recalculating the RPT. Click the Palette Icon to bring up the Viewer options. Click Close.
May 2015
76
Part 2: Rock Physics Template (RPT)
Click Save current session to save your work in RockSI. Next we will use the GEOVIEW standard crossplot tool.
May 2015
77
Part 2: Rock Physics Template (RPT) The same RockSI PEM/RPT can also be displayed in the standard Hampson-Russell Crossplot tool. It requires a license of RockSI to enable the option.
We will start by creating a standard VpVs vs P-impedance crossplot in GEOVIEW. - Go to the Processes tab. - Expand the Cross Plotting folder - Double click Cross Plot Logs
May 2015
78
Part 2: Rock Physics Template (RPT) In this example, we will only use Well A to demonstrate the steps.
Make the following selections: -Single Well - well_A - Vp/Vs vs P-impedance We will use all data.
Click OK.
May 2015
79
Part 2: Rock Physics Template (RPT) A crossplot similar to the one on the right will appear.
On the menu bar, click Options and select “RPM and PEM…” Note that this require a RockSI license. May 2015
80
Part 2: Rock Physics Template (RPT)
The PEM parameters dialog will appear as shown on the right. See the picture for various ways to access PEM.
May 2015
81
Part 2: Rock Physics Template (RPT) The Unconsolidated Oil sandstone PEM should readily available for use since we previously loaded it in RockSI. Any PEM created in RockSI can be used in this Crossplot tool. Select Unconsolidated Oil Sandstone.
May 2015
82
Part 2: Rock Physics Template (RPT) Fill in the PEM parameters as shown in the table. With Sw= 0, we are modelling 100% Oil saturation.
May 2015
83
Part 2: Rock Physics Template (RPT) RPT of Unconsolidated Oil Sandstone is now displayed in the crossplot. Any dynamically created models in RockSI can be posted on standard GEOVIEW crossplots to guide seismic and well logs interpretation. Click the Cross Plots icon in the lower right hand corner to dock the window. May 2015
84
Part 3 Creation and calibration of user-defined PEMs
Part 3: Creation of Petro Elastic Models In this exercise, you will create or import a Petro Elastic Model (PEM) for each of the previously determined lithologies.
The quality of the modelling will be tested by comparing the model predictions with the actual logs both in 1D viewers and cross-plots.
May 2015
86
Part 3: Geological context Lithostatic Pressure (MPa) Pore Pressure (MPa) Effective Pressure (MPa) 0
10
20
30
40
Temperature (°C)
50
0
10
20
30
40
50
0
Sea Water ρ = 1g/cc
500
9.8
1000
Overburden ρ = 2.25 g/cc 10.9
11.4
12
May 2015
Gradient = 30°C/km
1500 27
26.4
15.5
24.1
4
Reservoir ρ = 2.5 g/cc
35.5
33
Depth (m)
45
2000
38
49
2500
Depth (m) 87
Part 3: Geological context Fluid Parameters:
Temperature: 35 °C Lithostatic Pressure: 30MPa Pore Pressure: 20 MPa Brine Salinity: 0.01 kg/L Gas-Water ratio: 0 L/L
Oil gravity: 30 API Gas-Oil ratio: 50 L/L Gas gravity: 0.6 May 2015
88
Part 3: Creation of PEMs In this exercise, we will: 1. (Optional) Create a user defined PEM to model Shale. Instead of building a full PEM, we will only build it up to modelling bulk density. That should be enough to demonstrate how it can be done. 2. Then proceed with importing a PEM for each lithofacies and use these PEMS to predict elastic properties: Vp, Vs and Rho. The predicted curves can be visually compared with the measured curves.
May 2015
89
Part 3: Creation of PEMs Click the PEM Library tab. Right click on the PEMs item and select Delete all PEMs.
Click Yes to delete the PEMs.
Toggle the Advanced button. Right click on the PEMs item and select Create New PEM.
May 2015
90
Part 3: Creation of PEMs
Right click on the newly created PEM and select Rename PEM.
Name this PEM as Shale and click Ok.
May 2015
91
Part 3: Creation of PEMs Go to the Cross-Plots Options tab, toggle off Display only Range and toggle on Display only rock type with shale selected.
Set Facies as the color-coded variable for the crossplots if it is not already set.
May 2015
92
Part 3: Creation of PEMs We have seen in Exercise 1 that the main variables affecting the elastic response of the shale are Vsh, Phit (total porosity) and the effective pressure Peff .
Click on PEM Library tab, add three new lines in the input table of the Shale tab by clicking on the Add Line Below button. Select Vsh, Phit, Depth (True Vertical Depth) and Pp (pore pressure) respectively. Pay attention to ensure that proper units are specified.
May 2015
93
Part 3: Creation of PEMs Select Peff as the Output Variable, and MPa as the unit.
In many cases, Peff is derived from Depth when Pp is known. Effective pressure in our case is given by the following equation:
Peff 9.81* (1 2.25*(Depth 1000) / 1000) Pp Input parameters
Click the Arrow sign to see a list of Input parameters, Math operators and precompiled equations, which we will use for writing the equations. May 2015
Math operators Precompiled equations
94
Part 3: Creation of PEMs Type in part of the equation as shown on the right.
Next we will select the variable Depth. Click the arrow sign beside the equation box and select Depth.
Type in the rest of the equation and select the variable Pp.
May 2015
95
Part 3: Creation of PEMs The final equation to calculate Peff looks like this: If you have entered the equation correctly, a green check mark will appear after the PEM name. Add eight new lines by clicking on the Add Line Below button. Select the output variables as shown (notice that the new Variables are Temperature, Salinity, Gas/Water Ratio, and five different densities:
A red cross will now appear after the PEM name, indicating that the equations must be filled in: May 2015
96
Part 3: Creation of PEMs
We will start with the modelling of the shale density as this is the easiest elastic attribute to model. The density is given by the following volumetric averages:
Rho Phit*Rhow 1 Phit *Rho ma Rho ma Vsh*Rhosh 1 Vsh *Rhoqz
We will use 2.65 for the quartz density and the shale mineral density is given as a function of the effective pressure:
Rhosh 2.43 0.02*Peff
The water density can be calculated using Batzle & Wang equation as a function of: Temperature [°C] We will assume a temperature gradient as shown earlier:
May 2015
Pore Pressure [MPa] Salinity [kg/L] Gas-Water ratio [L/L]
T 4 30*( Depth 1000) / 1000
97
Part 3: Lithofacies PEMs Next, we will create PEMs for each of the lithofacies defined earlier. For example, the shale PEM is created using the following equations:
w 1 m where : m Vsh sh 1 Vsh qz and
sh 2.43 0.02*Peff .
May 2015
98
Part 3: Creation of PEMs The Batzle & Wang equation for brine density is quite complex and it has already been precompiled. Click the arrow sign beside the equation box for the fifth row (Rhow), and select Precompiled Equations->Brine density (Batzle Wang). On the menu, we can also see a list of other pre-defined equations.
May 2015
99
Part 3: Creation of PEMs Enter the equations for the rest of the output variables as shown. Make sure the units of the input variables and output variables are correct. If you fill in the equations correctly, the green check mark will appear again. See next slide for a description of the variables involved. May 2015
100
Part 3: Creation of PEMs Vsh – Volume of Shale Phit – Total Porosity Depth – True Vertical Depth Pp – Pore Pressure Peff – Effective Pressure T– Temperature Sal – Salinity GWR – Gas Water Ratio Rhow – Brine Density Rhoqz – Quartz Density Rhosh – Shale Density Rhom – Rock Matrix Density Rho – Bulk Density
May 2015
101
Part 3: Creation of PEMs Click the Input tab, and select Shale for the Petro-Elastic Model for the Rock Type shale.
Specify the mapping of input variables to their corresponding log curves: VCL (clay volume), PHIT (total porosity), TVDSS (true vertical depth), and PP (pore pressure). Click Replicate Icon to select the same logs for other wells.
Click Update PEM predictions Icon to calculate the predicted logs May 2015
102
Part 3: Creation of PEMs You can see the predicted density in the Log Viewer of each well. The predicted density (red) for Shale facies (Brown color facies) is now overlaid on top of the measured density (black). The match of the two logs are quite good.
May 2015
103
Part 3: Creation of PEMs You can see the predicted density overlaid with measured values in the crossplot QCs as well. The predicted density (red) for Shale facies and measured density in brown (colored by facies). You can uncheck the Predictions to see measured density.
May 2015
104
Part 3: Creation of PEMs
Now that the density is correctly modeled, we will need the bulk modulus and shear modulus of the rock to calculate the P-wave and S-wave velocities. The modulus of the mineral can be computed with a simple Hill volumetric average:
Kma 0.5*(Vsh*Ksh 1 Vsh *Kqz) 0.5 /(Vsh/Ksh ( 1-Vsh)/Kqz) Gma 0.5*(Vsh*Gsh 1 Vsh *Gqz) 0.5 /(Vsh/Gsh ( 1-Vsh)/Gqz)
We will use 37 and 45 for the quartz bulk and shear modulus respectively. We will use 25 and 9 for the shale bulk and shear modulus respectively. The modulus of the saturated rock can be computed as a function of the mineral modulus, the effective pressure and the total porosity:
G 0.2*Gma 1.36*Peff 90*Phit/Peff 10 Keff 0.3*Kma 1.8*Peff 210*Phit/Peff 8
May 2015
105
Part 3: Creation of PEMs
Now that we have modelled the bulk, shear modulus and density of the shale formation, we have all the input needed to compute its P-wave and S-wave velocities:
VP
VS
G Rho
Because those equations are used very often, they have been precompiled in the software and can be used through the following keywords:
4 Kb G 3 Rho
VP_WAVE_PROPAGATION_EQ VS_WAVE_PROPAGATION_EQ.
You can continue to enter equations and variables to build a full PEM to predict Vp and Vs for shale. Now that you have some experience on building a PEM, we will stop and simply import a number of prebuilt PEMs instead.
May 2015
106
Part 3: Creation of PEMs Click the PEM Library tab. Right click on the PEMs item and select Delete all PEMs.
Click Yes to delete the PEMs.
Toggle on the Advanced button. Right click on the PEMs item and select Load PEM(s).
May 2015
107
Part 3: Creation of PEMs
On the Load PEM(s) dialog, click the Folder Icon.
May 2015
108
Part 3: Creation of PEMs Go to the same directory where the RockSI project resides. You should find five files with “.pem” extension. Select Calcite, Coal, Sand (Water sand), Oil (Oil sand) and Shale. Note that water and oil sand use the same model. We separated them so that it will be easier for comparing simulation results later. Click Open to select them. On the Load PEM(s) dialog, Click Apply to import them. May 2015
109
Part 3: Creation of PEMs
The PEM Library tab should now show four PEMs available for use. The Shale PEM is named Shale #1 since we had one defined previously. You can review the equations and parameters defined in each PEM.
May 2015
110
Part 3: Creation of PEMs With all the necessary PEMs available, we are ready to use them for predicting Vp, Vs and Rho for all facies. Go to the Input tab. Specify the associated PEM for each Rock Type.
Make sure the Output Variable table has Vp, Vs, Rho, Ip and Vpvs selected. Use the + sign to add new variables. Scroll down to the Well Tabs. May 2015
111
Part 3: Creation of PEMs For Well A, specify the log curves as shown for each table. Make sure you click the Replicate Icon for each table to ensure the selections are applied to the other three wells. Click through all the wells to confirm the selections. Lastly, click on Update PEM predictions Icon to generate the log curves.
May 2015
112
Part 3: Creation of PEMs In the Log Viewers tab, you should see the predicted Vp, Vs and Rho (red) overlay with their associated measured curves (black). The angle synthetics are also generated based on the predicted curves. You may need to use the horizontal scroll bar to scroll to the right in order to see the results.
May 2015
113
Part 3: Creation of PEMs Click on the Cross-Plot Options tab. Uncheck Display only rock type filter. Go to the Cross-plot tabs to see the crossplot QCs. Predicted values are in red and measured values are colored by facies.
May 2015
114
Part 3: Creation of PEMs
Click the Save session button to save the current parameters and QC templates. This concludes this exercise.
May 2015
115
Part 4 Uncertainty analysis using Monte Carlo simulations
Uncertainty analysis Sw, Sh, Sg
Pore Pressure
Overburden Pressure
PEM
V p f (inputs)
f
Porosity
Clay Volume
Rock Model Parameters May 2015
117
Uncertainty analysis Sw, Sh, Sg
Pore Pressure
Overburden Pressure
PEM
f
Porosity
V p f (inputs)
Output PDF
Output Uncertainty
Clay Volume Input PDFs
Input Uncertainties May 2015
Rock Model Parameters 118
Input Distributions Constant
Uniform: Min and Max
Normal: Mean and standard deviation
Beta (PERT): Min, Mode, Max, Lambda (see Annex C for charts)
Discrete: Min, Max, Step May 2015
119
Beta (PERT) distribution Parameterization suited to rock physics properties: Min/Max Most likely = mode Shape factor reciprocal of “standard deviation” at mode =1 1.5
mode
2 2.5 3 5
lambda
7 9
120
May 2015
Input Correlations Correlations between the input variables are simulated through the ImanConover technique (see Annex C for more details).
PX , Y
PX , Y
PY | X
PY | X
PY
PY
Y
Y|X mY
Y
mY | X
Y|X
mY
mY | X
CC = -0.7
mX
mX
X
X PX
May 2015
Y X
PX
121
Simulation of scenarios Production effect on elastic attributes:
Is water flooding gas out of solution due to depletion shale
water-sand oil-sand gas sand
May 2015
Ip 122
Part 4: Monte Carlo simulations
In this exercise the user will run Monte Carlo simulations using the PEMs defined in the previous exercise.
The objective is to make sure that the PEMs reproduce the variability observed in the log data.
May 2015
123
Part 4: Monte Carlo simulations We will start the simulation of the shale. Set the following cross-plots parameters in the Cross-Plots Options tab. Filter the data to display shale only. Turn off the display of Predictions result.
May 2015
124
Part 4: Monte Carlo simulations
Select the Shale PEM in the simulation tab. Enter the input distributions and correlations as shown. Total porosity (Phit) is negatively correlated to volume of Shale (Vsh). Launch the simulation by clicking Launch Simulation button. Compare the simulated points with the actual logs.
May 2015
125
Part 4: Monte Carlo simulations The Shale simulation result looks like this:
Measured Data
May 2015
Simulated Data With P10, P50, and P90 PDFs contours
126
Part 4: Monte Carlo simulations After the simulation, we can select to output a list of petro-elastic attributes as LithoSI training set. LithoSI is a Bayesian Classification software. Remove all output and add two new output attributes. Select attribute Ip and VpVs. For the training set name, it has to be one of the facies we have defined; Enter Shale. Name this training set as Training_Set and click Save.
May 2015
127
Part 4: Monte Carlo simulations Now the simulation has been saved. We can view the simulation results in the saved file. Right click on Training_Set and select View saved session:
In the training set file, we can see that we have three columns including the two output volumes Ip and VpVs and the class of the facies Shale.
May 2015
128
Part 4: Monte Carlo simulations In the Cross-Plots Options tab, Select Phi instead of Phit in the cross-plots. Set the rock type filter to display the water sand only.
Check off Shale simulation results.
May 2015
129
Part 4: Monte Carlo simulations Select the PEM Sandstone in the Simulation tab. Enter the input distributions and input correlations as shown:
Launch the Water Sand simulation by clicking Launch Simulation button.
May 2015
130
Part 4: Monte Carlo simulations The Water Sand simulation result looks like this. Compare the simulated points with the actual logs Simulated Water Sand Data Measured Data With P10, P50, and P90 PDFs contours
May 2015
131
Part 4: Monte Carlo simulations Save the simulation result for Water Sand. For the training set name, it has to be one of the facies we have defined; Enter Sand. Name this training set as Training_Set and click Save.
May 2015
132
Part 4: Monte Carlo simulations Next, we will simulate the Oil Sand. Modify the input distributions as shown in the table. The only difference here is that we are varying Sw with a Beta distribution between 10%-70%. (i.e. So ranges 30%-90%) Launch the simulation by clicking Launch Simulation button. Compare the simulated points with the actual logs.
Set the range filter to display only the oil sand.
May 2015
133
Part 4: Monte Carlo simulations The Oil Sand simulation result looks like this. Compare the simulated points with the actual logs Simulated Oil Sand Data Measured Data With P10, P50, and P90 PDFs contours
May 2015
134
Part 4: Monte Carlo simulations Save the simulation result for Oil Sand. Change the Training set name to Oil.
Select the training set Training_Set and click Save:
On the pop-up message, click Ok.
May 2015
135
Part 4: Monte Carlo simulations We can also simulate what would happen if the reservoir was filled by gas instead of oil. Modify the input distributions for Sw and Sg as shown in the table: Launch the simulation by clicking Launch Simulation button. Compare the simulated points with the actual logs.
Set the range filter to display only the water and oil reservoirs. Compare with the simulated gas reservoir. Click to redraw. May 2015
136
Part 4: Monte Carlo simulations The Gas Sand simulation result looks like this. We will not save this result to Training_Set. This is just for comparison purpose.
May 2015
137
Part 4: Monte Carlo simulations
Reset simulation back to the Oil case by modifying Sw and Sg as shown on the right. Click on Launch Simulation again. There is no need to save this result since we did it previously.
May 2015
138
Part 4: Monte Carlo simulations Select the PEM Calcite in the simulation tab. Enter the input distributions and input correlations as shown. Launch the simulation by clicking Launch Simulation button.
Change the Training set name to Calcite. Select the training set Training_Set and click Save.
May 2015
139
Part 4: Monte Carlo simulations The Calcite simulation result looks like this. Compare the simulated points with the actual logs Simulated Calcite Data Measured Data With P10, P50, and P90 PDFs contours
May 2015
140
Part 4: Monte Carlo simulations Select the PEM Coal in the simulation tab. Enter the input distributions and input correlations as shown. Launch the simulation by clicking Launch Simulation button.
After the simulation, Select the output attribute Ip and VpVs: Change the Training set name to Coal. Select the training set Training_Set and click Save: May 2015
141
Part 4: Monte Carlo simulations The Coal simulation result looks like this. Compare the simulated points with the actual logs Simulated Coal Data Measured Data With P10, P50, and P90 PDFs contours
May 2015
142
Part 4: Monte Carlo simulations To see all five simulated facies data. Go to Cross-Plot Options tab. Reduce to number of crossplots to only showing Ip vs VpVs Set the display parameters as shown on the right: - Uncheck any filters - Uncheck Data and Predictions - Uncheck outline for each Simulation - Uncheck pdf for each Simulation - Specify corresponding color for each facies May 2015
143
Part 4: Monte Carlo simulations This the simulated data for the five facies. In the next excerise we will use these simulated data points as training data set for LithoSI (Bayesian Classification).
May 2015
144
Part 4: Creation of PEMs
Click the Save session button to save the current parameters and QC templates. This concludes Part 4 exercise.
May 2015
145
Part 5 LithoSI Analysis using Simulated PDFs
Litho-classification
Ip
1600 2000 24001.8
2.2
Vp/V
2.6
1200
2.5
2.5
2.4
2.4
2.3
2.3
1350
2.2
2.2
1400
2.1
2.1
1250
Vp/Vs
1600 1700 1800 1900 2000 2100 2200 2300 2400
s
1300
Extract training sets
1450 1500 1550
2.0
2.0
1.9
1.9
1.8
1.8
1600 1600 1700 1800 1900 2000 2100 2200 2300 2400
Seismic inversion results
Well logs
I p
Compute multivariate PDFs
Vp/V 1.0
Oil
s
Water Gas
0.0
Compute probability litho-cubes May 2015
P(Class i I p , V p / Vs )
1600 1700 1800 1900 2000 2100 2200 2300 2400
2.5
2.5
2.4
2.4
2.3
2.3
2.2
2.2
2.1
2.1
2.0
2.0
1.9
1.9
1.8
1.8 1600 1700 1800 1900 2000 2100 2200 2300 2400
I p
147
Litho-classification Training Set Issues
1600 2000 24001.8
2.2
Vp/V
2.6
1200
Too sparse and not representative of natural reservoir variability
s
1300
May not include all “litho-classes” May not capture expected depth trends
2.3
2.4
Gas sand
2.3
2.2
2.2
1400
2.1
2.1
Extract training sets
1500 1550
2.4
1350
1450
2.5
2.5
1250
1600 1700 1800 1900 2000 2100 2200 2300 2400
2.0 1.9 1.8
Cemented sand?
1600 1600 1700 1800 1900 2000 2100 2200 2300 2400
Well logs
2.0 1.9 1.8
I p
Compute multivariate PDFs
Vp/V 1.0
Oil
s
Water Gas
0.0
Compute probability litho-cubes May 2015
P(Class i I p ,V p / Vs )
1600 1700 1800 1900 2000 2100 2200 2300 2400
2.5
2.5
2.4
2.4
2.3
2.3
2.2
2.2
2.1
2.1
2.0
2.0
1.9
1.9
1.8
1.8 1600 1700 1800 1900 2000 2100 2200 2300 2400
I p
148
Litho-classification
Sw F Vclay
Vp/V
Monte Carlo Simulation
s Simulate training sets
PEM
1600 1700 1800 1900 2000 2100 2200 2300 2400
2.5
2.5
2.4
2.4
2.3
2.3
2.2
2.2
2.1
2.1
2.0
2.0
1.9
1.9
1.8
1.8
Additional simulated litho-class
1600 1700 1800 1900 2000 2100 2200 2300 2400
Peff
Augmented training set for existing class
Compute multivariate PDFs
Vp/V 1.0
Oil
s
Water Gas
0.0
Compute probability litho-cubes May 2015
P(Class i I p ,V p / Vs )
1600 1700 1800 1900 2000 2100 2200 2300 2400
2.5
2.5
2.4
2.4
2.3
2.3
2.2
2.2
2.1
2.1
2.0
2.0
1.9
1.9
1.8
1.8 1600 1700 1800 1900 2000 2100 2200 2300 2400
I p
149
Part 5: LithoSI analysis In this exercise, we will use the previously defined PEMs to interpret an elastic inversion result.
We will also perform a litho-classification of that inverted model using both the well data and the synthetic training sets created in the previous exercise.
May 2015
150
Part 5: LithoSI analysis
To initiate the LithoSI program, go back to the Processes tab and under Facies Classification (LithoSI) , double-click .
May 2015
151
Part 5: LithoSI analysis The dialog which appears is multi-tab, which will help you set up the LithoSI process. The first tab of the dialog defines which seismic volumes will be used in the process. These seismic volumes are collected in a “Super Volume”. In this example we select both the Zp and VpVs volumes as shown on the left (either click once on each volume and click Select or double-click each volume).
May 2015
152
Part 5: LithoSI analysis When you have changed the first page as shown above, click the Wells tab to see the next page:
This page specifies which wells to use in the analysis. We want to use three wells (well_B, well_C and well_D), so Select these wells:
May 2015
153
Part 5: LithoSI analysis Then, click the Facies Classification tab:
The Facies Classification page is used to define the set of facies classes in our training data. First we select the lithology log, which was created in each of the wells in the previous section. This was called FACIES.
In the rest of the page, we define the classes, telling the program how many classes and what their names are. We will select the classes we just defined in the previous section. Click on Facies: May 2015
154
Part 5: LithoSI analysis Now click the last tab, Seismic Attributes-Well Log Mapping: On this page, we specify how many seismic attributes are used, and how the seismic volumes are related to the well log curves. The default number of attributes is 2, and we will accept that: The mapping of volumes is as on the right. We associate the seismic inversion volume, Inverted_Ip, with the curve IP. Similarly, we associate the seismic inversion volume Inverted_VpVs with the well log curve VpVs. May 2015
155
Part 5: LithoSI analysis
When you have filled in the dialog as shown above, click OK to create the LithoSI window:
May 2015
156
Part 5: LithoSI analysis The LithoSI window contains a series of options for creating the probability distributions associated with each class. The first thing we can do is create more space on the computer screen by temporarily removing the Project Manager. Click the “x” as shown:
Note that we can restore the Project Manager any time later by clicking the rotated name Project Manager on the left side:
May 2015
157
Part 5: LithoSI analysis Go to the Data Selection tab We want to use the simulated data created in RockSI so select “Load training-set from RockSI”.
Select RockSI Session and the training set. Note that the session name will be different in your case since by default the name is created with date and time. Specify the mapping of the properties used in RockSI with the log curves. The a-posterior proportions on the rightmost column in the bottom table are calculated from all the wells based on facies curves.
May 2015
158
Part 5: LithoSI analysis To display the external training set from RockSI in a crossplot. Turn off Well logs data points.
Turn on External Training-set.
May 2015
159
Part 5: LithoSI analysis LithoSI Kernel Analysis now shows the data points simulated in RockSI. Next step is to generate the pdfs for each of these facies and fine-tune by adjusting the smoother parameter.
May 2015
160
Part 5: LithoSI analysis Reduce the smoother length to 2. Click on Compute in Common Parameters panel. In Kernel Analysis tab, turn off Coal distribution. The pdfs between these facies overlap quite a bit. Reducing the smoother length also reduces these overlaps.
May 2015
161
Part 5: LithoSI analysis Use the slider to adjust the smoother to 1. Click Compute button again. The pdfs now fit the facies data points tighter.
May 2015
162
Part 5: LithoSI analysis Click on QC at Wells. Select only Well Logs toggle. The Confusion Matrix gives an overall success and misclassification rates of the facies at all three well locations.
May 2015
163
Part 5: LithoSI analysis Next we QC with an automatically generated arbitrary line that passes through the three wells. Go to QC on Sections. Select “Passing through wells” and click on the Select all wells icon.
May 2015
164
Part 5: LithoSI analysis You can also QC constant time map of the classification by going to the QC on Maps tab.
Finally, you can generate the classification results for the entire volume by running it in “Run on 3D volume” tab. It is not necessary to run the process for this exercise.
May 2015
165
Part 5: LithoSI analysis
This final exercise demonstrate how RockSI simulated data points can be used in the Bayesian Classification process (LithoSI). This concludes our RockSI tutorial.
May 2015
166
