rock-physics-avo

Rock Physics & AVO Workflow

End-to-end pipeline for rock physics analysis and AVO feasibility studies, from well log preparation through elastic property calculation, Gassmann fluid substitution, AVO modelling, and synthetic seismogram generation.

Skill Chain

lasio / welly          bruges                    segyio / obspy
[Well Log Prep]     --> [Rock Physics]         --> [Synthetics / Tie]
  |                      |                          |
  Load LAS/DLIS          Elastic moduli             Wavelet extraction
  QC & despike           Gassmann fluid sub         Reflectivity series
  Resample curves        AVO intercept/gradient     Convolve synthetic
  Extract Vp, Vs, rho    Backus averaging           Well-seismic tie

Decision Points

Step-by-Step Orchestration

Stage 1: Well Log Preparation (lasio + welly)

Select and QC a contiguous interval before these calculations. Required curves are DT and RHOB; DTS is optional. This example explicitly permits the Castagna estimate for missing DTS in an appropriate siliciclastic interval. It is an empirical estimate, not a measured shear log. Nulls in unrelated curves must not remove elastic samples, and invalid elastic samples must not be silently bridged.

import lasio
import numpy as np

las = lasio.read('well.las')
df = las.df().copy()

def unit(curve):
    return las.curves[curve].unit.replace('µ', 'u').replace('μ', 'u').upper().replace(' ', '')

def sonic_velocity(curve):
    factors = {'US/FT': 0.3048e6, 'US/M': 1e6}
    values = df[curve].to_numpy(dtype=float)
    if unit(curve) not in factors:
        raise ValueError(f'Confirm sonic units for {curve}: {unit(curve)}')
    if not np.all(np.isfinite(values) & (values > 0)):
        raise ValueError(f'QC {curve}: nonpositive or missing transit times')
    return factors[unit(curve)] / values  # m/s

vp = sonic_velocity('DT')
vs_estimated = 'DTS' not in df.columns
if vs_estimated:
    vs = 0.8621 * vp - 1172.4
else:
    vs = sonic_velocity('DTS')  # Check existence before reading

density_factors = {'G/CC': 1000.0, 'G/CM3': 1000.0, 'G/C3': 1000.0, 'KG/M3': 1.0}
if unit('RHOB') not in density_factors:
    raise ValueError('Confirm RHOB units before converting to kg/m3')
rho = df['RHOB'].to_numpy(dtype=float) * density_factors[unit('RHOB')]
if not np.all(np.isfinite(rho) & (rho > 0) & (vs > 0) & (vp**2 > 4/3 * vs**2)):
    raise ValueError('Invalid elastic inputs: require positive density, Vs and bulk modulus')

depth_factors = {'M': 1.0, 'FT': 0.3048}
depth_unit = unit(las.curves[0].mnemonic)
if depth_unit not in depth_factors:
    raise ValueError('Confirm LAS depth units')
depth_m = df.index.to_numpy(dtype=float) * depth_factors[depth_unit]
if len(depth_m) < 2 or not np.all(np.isfinite(depth_m)) or np.any(np.diff(depth_m) <= 0):
    raise ValueError('Provide at least two samples on an increasing depth basis')