Program flatten reads in an input seismic or attribute volume as well as a picked horizon and generates a flattened output volume:

Transcription

1 FLATTENING A SCALAR DATA VOLUME PROGRAM flatten Computation Flow Chart Program flatten reads in an input seismic or attribute volume as well as a picked horizon and generates a flattened output volume: Input volume Horizon flatten Flattened volume Output file naming convention Program convolutional_modeling will always generate the following output files: Output file description Flattened data volume program log information program error/completion information File name syntax flattened_unique_project_name_suffix.h convolutional_modeling_unique_project_name_suffix.log convolutional_modeling_unique_project_name_suffix.err where the values in red are defined by the program GUI. The errors we anticipated will be written to the *.err file and be displayed in a pop-up window upon program termination. These errors, much of the input information, a description of intermediate variables, and any software trace-back errors will be contained in the *.log file. Overview Attribute-Assisted Seismic Processing and Interpretation March 8, 2019 Page 1

2 Extracting phantom horizon slices and stratal (or proportional) slices two some of the more common interpretation activities performed in interpretation workstation software. Since the interpretation workstation is the place where you picked your horizons, it is the obvious place to do such slicing and subsequent analysis. Nevertheless, there are reasons to create flattened or stratal sliced subvolumes in the AASPI software. Generating flattened volumes has value if your commercial software does not have a state-of-the-art spectral decomposition algorithm and you wish to generate a suite of volumes about a target horizon. Similarly, AASPI provides horizon-based clustering (also called classification) algorithms. AASPI programs flatten, complex_spectral_flatten, and vector_flatten flatten a user-defined window of input data defined about a picked horizon. AASPI program unflatten complex_spectral_unflatten, and vector_unflatten reverse this process. AASPI program stratal_slice, complex_spectral_stratal_slice, and vector_stratal_slice generate a suite of stratal (proportional slices) between two user-defined horizons. Flattened slices are computed by interpolating the input data using a φ=2πfδt Fourier phase shift of each Fourier component. Because the distance between stratal slices varies from trace to trace, it is more efficient to compute stratal slices using a simple six-point interpolation in the time (or depth) domain. Program flatten, unflatten, and stratal_slice work on scalar data volumes. Because there is a discontinuity about ±180, phase and azimuth cannot be interpolated along the vertical axes. For this reason, complex spectra and vector quantities such as dip magnitude and dip azimuth should be flattened as pairs using programs complex_spectral_flatten, and vector_flatten. The same argument holds for interpolating to generate complex and vector stratal slices. Program flatten reads in a single data volume and a horizon file and outputs a flattened volume. Program flatten is launched from the Formation Attributes within in the main aaspi_util GUI: The following GUI appears: Attribute-Assisted Seismic Processing and Interpretation March 8, 2019 Page 2

3 Begin by (1) entering the input data volume file name and (2) the input horizon. Click (3) to view the horizon. If the file is generated using Windows-based software (e.g. Petrel), they will have the annoying carriage return (^M) at the end of each line (Shown in Figure 1). Use this button to delete those carriage returns if you prefer. Their presence will not negatively impact the software. Then (5) choose the horizon format. Currently gridded (e.g. EarthVision in Petrel) and interpolated (ASCII free format, e.g. SeisX) formats are supported. Gridded horizons are x, y location based and often define the location of knots in a B-spline surface in the North and East directions. In general, a gridded surface is independent of the location of inlines and crosslines. In contrast, interpolated horizons are inline- and crossline-based, where each trace has an x and y value. Examples of both formats are shown in Figure 1. If interpolated is selected, the user needs to manually define each column in the file. The only other value needed for a gridded horizon is (11) the value of the znull identifying a missing portion of the horizon. The interpolated horizon needs other parameters to define its format including (6) the number header lines to skip, (7) the total number of columns, (8) the column number containing the inline indices, (9) the column number containing the crossline (or CDP) indices, and (10) the column number containing the two-way travel time (or depth) value. Next, (12 and 13) define the limits of the output data about the picked horizon (in this case ±0.100 s). Finally (14) choose the direction of the vertical axis defined in the horizon file to be positive down or negative down (e.g. Petrel uses negative down) and (15) the time unit used in the horizon files (usually s or ms, depending on the software used). Attribute-Assisted Seismic Processing and Interpretation March 8, 2019 Page 3

4 Selector 15: The input data looks like this while the flattened volume looks like this The phantom horizon slice s above the picked horizon shows some turbidite channels: Attribute-Assisted Seismic Processing and Interpretation March 8, 2019 Page 4

5 Attribute-Assisted Seismic Processing and Interpretation March 8, 2019 Page 5