|
<< Click to Display Table of Contents >> Navigation: VSP > Convert DAS to VSP |
This module converts raw Distributed Acoustic Sensing (DAS) fiber-optic borehole recordings into conventional Vertical Seismic Profile (VSP) gathers. DAS data are recorded as continuous time-series along optical fiber deployed in a well; each channel corresponds to a depth interval along the fiber. The conversion is performed by cross-correlating each DAS channel with the vibroseis sweep signal, which collapses the swept-frequency record into a zero-phase impulse response equivalent to a conventional VSP shot record.
One or more DAS SEG-Y input files can be provided. When multiple files are supplied, the module processes each file independently and then adaptively stacks the resulting VSP gathers. The stacking weights each trace by the inverse of its pre-event noise power, so noisy records contribute less to the final result. The output is a single stacked VSP gather covering all depth channels from the wellhead channel downward.
The vibroseis sweep used for correlation can either be taken from a connected input gather (measured pilot sweep) or generated synthetically by the module from the specified sweep parameters. Optional sub-procedure chains allow custom preprocessing of the raw DAS channels before correlation and postprocessing of the correlated gather before it is written to the output.
A list of one or more DAS data files in SEG-Y format. Each file should contain the continuous time-domain DAS recording for one vibroseis shot or sweep activation. Add all files that belong to the same acquisition pass to enable adaptive stacking across shots. At least one file must be provided.
Controls whether the vibroseis sweep for correlation is taken from a connected sweep gather or generated synthetically. When set to true, the Input sweep gather connection becomes active and the Vibroseis parameters group is hidden. When set to false, the module synthesizes a linear sweep from the Sweep length, Sweep start frequency, and Sweep end frequency parameters. Using a measured pilot sweep is recommended when one is available, as it captures the actual source signature including non-linearity and harmonic distortion. The default value is true.
A measured pilot sweep gather used as the correlation operator. This input is active only when Use input sweep is enabled. The sweep is automatically resampled to match the sample rate of the DAS input files before correlation. Connect the output of a sweep extraction or import module here to supply the measured vibroseis pilot signal.
Specifies the SEG-Y format parameters used to read the DAS input files, such as sample format (IBM float, IEEE float, integer) and byte order. Set these to match the format in which the DAS acquisition system wrote the SEG-Y files. Incorrect format settings will produce garbled or empty data.
Selects the trace header byte layout used in the DAS SEG-Y files. Different acquisition vendors place geometry and timing information in different byte positions within the 240-byte SEG-Y trace header. Choose the format that corresponds to the DAS acquisition system used. The default is Geomaге Format. If the headers are not being read correctly, switch to New Format or consult the acquisition system documentation to identify the correct byte assignments.
This group controls the geometry and timing of the output VSP gather, including the spatial sampling interval of the DAS fiber, the starting channel index, and the desired record length and sample rate of the converted VSP data.
The spatial sampling interval of the DAS fiber, that is, the physical distance along the well between adjacent DAS channels, measured in metres. This value defines the depth axis spacing of the output VSP gather and should match the gauge length or channel spacing used during DAS acquisition. The default is 2 m. Valid range is 0.01 m to 1000 m.
The one-based channel number in the DAS file that corresponds to the wellhead depth (zero depth). All channels with indices equal to or greater than this value are treated as in-well receivers and are written to the output VSP gather. Channels with indices below this value are ignored, as they correspond to the surface or above-wellhead portion of the fiber. Set this to the first channel index that is inside the borehole. The default is channel 1.
The time length of each output VSP trace, in seconds. This determines how much of the correlated record is retained. Set this to cover the maximum two-way travel time of interest plus any corridor of events you need to analyse. Increasing this value increases output file size but does not affect processing speed significantly. The default is 4 s. Valid range is 0.01 s to 100 s.
The time sampling interval of the output VSP gather, in seconds. The correlated data are resampled from the native DAS sample interval to this output interval using linear interpolation. A typical value for VSP data is 0.001 s (1 ms) or 0.002 s (2 ms). The output sample rate cannot exceed the Nyquist limit of the input DAS data. The default is 0.001 s (1 ms). Valid range is 0.0001 s to 1 s.
This group defines the parameters of the synthetic vibroseis sweep that the module generates when Use input sweep is disabled. The module constructs a linear (Klauder) chirp signal from these parameters and uses it as the cross-correlation operator. These parameters are ignored when a measured pilot sweep is connected.
The total duration of the synthetic vibroseis sweep, in seconds. This should match the sweep length programmed into the vibrator controller at the time of acquisition. Using an incorrect sweep length will shift the correlated wavelet in time and may reduce resolution. The default is 16 s. Valid range is 0.01 s to 100 s. Active only when Use input sweep is disabled.
The starting frequency of the synthetic linear sweep, in Hz. For an up-sweep this is the lowest frequency; for a down-sweep it is the highest. Set this to match the low-end frequency of the vibrator sweep actually used in the field. The default is 12 Hz. Valid range is 1 Hz to 1000 Hz. Active only when Use input sweep is disabled.
The ending frequency of the synthetic linear sweep, in Hz. For an up-sweep this is the highest frequency. Set this to match the high-end frequency of the vibrator sweep used in the field. Together with the start frequency it determines the usable signal bandwidth of the correlated VSP data. The default is 110 Hz. Valid range is 1 Hz to 1000 Hz. Active only when Use input sweep is disabled.
This group controls the adaptive stacking noise weighting scheme. When multiple DAS files are provided, each converted VSP gather is weighted by the inverse of its pre-event noise power before stacking. The noise level is estimated from a window immediately before the detected first-break arrival on each trace. A short, stable pre-event window gives a reliable noise estimate; a window that is too long may overlap the signal.
The length of the pre-event time window used to estimate the noise power for adaptive stacking weights, in seconds. The module automatically detects the first-break arrival time on each trace and measures the RMS amplitude in a window of this duration ending at the first break. Traces with higher noise power receive a lower stacking weight. Set this value to a duration that is long enough to provide a stable noise estimate but short enough to avoid including signal energy. The default is 0.5 s. Valid range is 0.01 s to 100 s.
The number of DAS channels read and processed at one time from each input file. Because DAS files may contain thousands of channels, the module reads them in batches of this size to manage memory usage. Increasing the chunk size may improve throughput on systems with large amounts of RAM, while reducing it lowers peak memory consumption. The default is 100,000 traces. Valid range is 1 to 1,000,000.
Selects whether the correlation and stacking computations are executed on the CPU or on a GPU accelerator. GPU execution can substantially speed up processing for large DAS datasets with many channels and long record lengths. CPU execution is always available; GPU execution requires a compatible graphics card and driver installation.
Enables distributed processing across multiple compute nodes in a cluster environment. When activated, the workload is split across nodes to reduce wall-clock processing time for large multi-file DAS datasets.
The minimum number of traces assigned to each compute node in a distributed execution job. Setting this value too low creates excessive inter-node communication overhead; setting it too high may result in uneven load distribution.
When enabled, restricts the number of CPU threads used per compute node in distributed mode, allowing other jobs to share the node resources concurrently.
An optional text label appended to the distributed job name. Use this to distinguish parallel runs of the same module in the cluster queue, for example when processing multiple survey lines simultaneously.
When enabled, allows manual specification of CPU core affinity for the processing threads. This is an advanced option for performance tuning on NUMA (Non-Uniform Memory Access) systems and is typically left disabled.
Specifies the CPU core mask or affinity string when Set custom affinity is enabled. Leave at default unless you have a specific hardware binding requirement.
The number of CPU threads used for parallel processing of DAS channels. Channel correlation and noise estimation are parallelised across threads, so increasing this value reduces processing time proportionally on multi-core workstations. Set to the number of available physical cores for best performance. Leaving it at the default allows the system to choose automatically.
Enables execution of custom scripts before and after the module runs. Use this to automate pre- or post-processing steps that are not handled natively within g-Platform, such as file format conversion or external quality control checks.
Path to a script that is executed before the module starts processing. Leave blank if no pre-run script is needed.
Path to a script that is executed after the module completes. Leave blank if no post-run script is needed.
When enabled, the module is bypassed and its output data items are passed through unchanged. Use this to temporarily disable the DAS-to-VSP conversion in a processing flow without removing the module from the sequence.
The synthetic vibroseis sweep signal constructed from the Sweep length, Sweep start frequency, and Sweep end frequency parameters. This output is populated only when Use input sweep is disabled. Connect it to a downstream module to inspect or export the generated sweep wavelet, for example to verify that it matches the acquisition parameters.
The converted and adaptively stacked VSP gather. Each trace in this gather corresponds to one DAS depth channel (from the wellhead channel downward) and contains the cross-correlated, noise-weighted, and normalized VSP record. The time axis runs from zero to the specified Output length at the specified Output sample rate. This gather is the primary result of the module and should be connected to downstream VSP processing steps such as wavefield separation, deconvolution, or depth imaging.