🌐 中文版 / Chinese version
Worksheet for full-waveform inversion of zero-offset DAS-VSP data. Nine main steps; each row lists the operation, the data it requires, and its product. Data providers can match deliveries against the “Data required” column. Primary target is near-well Vp; optional extensions are Q and anisotropy.
| Data | Key requirements | Format |
|---|
| Raw DAS shot gathers | Per shot per channel, unstacked, undenoised; plus a preprocessed version with its flow documented; headers must carry shot coordinates, channel number, MD, offset, sample interval, start time; header byte map attached | SEG-Y preferred; HDF5 / TDMS / SEG-D accepted |
| DAS acquisition parameters | Sample rate, record length, channel spacing, gauge length, pulse rate, interrogator model and firmware, physical quantity and units, gain and polarity, cable installation and fiber type, known coupling issues | Document |
| Source information | Source type; per-shot coordinates and offsets; sweep parameters; pilot or correlated wavelet; shot times and synchronization; number of sweeps per point, single-sweep records preferred | Document + SEG-Y/ASCII |
| Well information | Trajectory MD/TVD/x/y/inclination/azimuth; wellhead coordinates, elevation, KB; casing program; cable top/bottom depth; channel-depth calibration table and method | CSV or Excel |
| Well logs | DT, DTS, density, gamma, caliper, temperature | LAS |
| Initial-model data | Checkshot time-depth; interval velocities; velocity volume through the well; formation tops | SEG-Y / ASCII |
| Metadata | CRS and projection; unit system; time datum; polarity convention; record of prior processing | Document |
| Data | Purpose |
|---|
| Geometry files SPS or UKOOA P1/90 | Geometry cross-check |
| Statics | Near-surface correction |
| Ambient-noise records without shooting | Step 0 SNR assessment |
| Acquisition report and observer logs | Tracing anomalous channels |
| Legacy VSP products: corridor stack, up/downgoing fields | Steps 7A/8 reference |
| Surface seismic section through the well, prior velocity or Q results | Step 8 validation |
| Walkaway / multi-offset / multi-azimuth shots | Step 9 extensions, high value |
| # | Step | Operation | Data required | Product |
|---|
| 0 | Data reception & QC | Verify trace headers, CRS, units, polarity; SNR and usable bandwidth statistics per shot and per depth interval | Raw DAS shot gathers; header byte map; metadata; ambient-noise records | QC report; usable shot/channel list |
| 1 | Channel-depth calibration | Map channel numbers to MD/TVD using tap tests, casing collars, known markers | Calibration table; well trajectory; casing program; cable installed depth | Channel-depth table |
| 2 | DAS preprocessing | Remove bad channels and poorly coupled intervals; suppress common-mode noise and cable ringing; unify polarity; strain-rate/strain conversion as needed | Full acquisition parameters: gauge length, channel spacing, physical quantity and units, gain and polarity, cable installation, coupling notes | Clean shot gathers |
| 3 | First-break picking | Pick direct-arrival first breaks on every channel | Clean gathers; shot times and synchronization scheme | First-break table |
| 4 | Initial model & tie | Compare first-break traveltimes with checkshot and sonic; build 1-D starting Vp | Checkshot; sonic DT/DTS; density; interval velocities or velocity volume; formation tops | Initial Vp model |
| 5 | Wavefield separation | Flatten on first breaks -> extract downgoing field by median/FK/SVD -> subtract to obtain upgoing field | Clean gathers; first-break table | Downgoing and upgoing wavefields |
| 6 | Wavelet extraction | Window the downgoing direct arrival to extract the measured source wavelet; cross-check against the pilot-correlated wavelet | Downgoing field; pilot sweep; single-sweep records; sweep parameters | Measured source wavelet |
| 7A | Conventional branch | Deterministic deconvolution of upgoing by downgoing - compresses the wavelet and removes downgoing multiples; corridor stack | Up/downgoing wavefields | Corridor stack as reference |
| 7B | FWI main line | Forward modeling = Green function of the current model convolved with the measured wavelet; DAS strain operator in the objective; iterate near-well Vp, optionally invert the source term | Initial model; measured wavelet; clean gathers or upgoing field; channel-depth table; trajectory | Near-well Vp model |
| 8 | Validation | Compare inverted Vp with sonic, checkshot, corridor stack; residual analysis | Products of steps 4 and 7A; legacy VSP results | Validation report |
| 9 | Optional extensions | Q estimation if bandwidth allows; VTI parameters if walkaway or multi-offset shots exist | Multi-offset gathers and their geometry files | Q profile; VTI parameters |
| Data type | Preferred | Alternative |
|---|
| DAS gathers | SEG-Y | HDF5, TDMS, SEG-D |
| Wavelet / pilot | SEG-Y | ASCII |
| Well logs | LAS | CSV |
| Trajectory / calibration | CSV | Excel |
| Geometry | SPS | UKOOA |
| Velocity / horizons | SEG-Y | ASCII |
| Parameters & notes | PDF or Markdown | Word |
If time-lapse VSP/DAS observations are later coupled to the reservoir model for seismic-constrained history matching, the following reservoir-side data will be needed. Not required now; listed in advance so the data channel can be planned.
| Category | Content | Format |
|---|
| Static model | Corner-point grid, porosity/permeability/saturation properties, faults, zonation | GRDECL / EGRID |
| Dynamic model | Full simulation deck: PVT, relative permeability, initialization, schedule; or at minimum per-timestep pressure and saturation restarts | Eclipse-style DATA / INIT / UNRST |
| Production history | Rates and injection by well, bottom-hole pressures, RFT/MDT points, well-test interpretations | CSV / Excel |
| Rock-physics calibration | Core elastic measurements, dry-frame moduli; DT/DTS/density logs shared with Section 2 | Report + LAS |
| Time-lapse alignment | Restart file of the reservoir state at each VSP/DAS acquisition date | UNRST |
Coupling route: two simulated states pass through Gassmann fluid substitution to give delta-Vp, compared against time-lapse VSP/DAS observations; the residual enters ensemble history matching as an observation term.