In this course, we will focus its uses in seismic interpretation.
Seismic data have been successfully used in subsurface characterization. Synthetic seismograms are used to interpret recorded seismic data in term of geology. Synthetics are calculated seismic response over an earth model given a wavelet. Convolutional synthetic modeling is the most common method, and it is performed by convolving a reflectivity series with a wavelet. Some of the key uses of synthetics are: 1) seismic-well-tie of stacked seismic data that links seismic to geology and provides time-depth relationship, 2) estimation of wavelet/phase of seismic, 3) aid structural and stratigraphic interpretation, 4) identification of noise in seismic, 5) calibration of seismic AVO data before AVO attributes estimation and seismic inversion, 6) trace matching and for model updates in seismic inversion process, 7) understanding resolution/tuning effect and seismic net pay estimation, 8) evaluate reservoir models with 3D and 4D seismic, 9) evaluating new technology/algorithm, and 10) generating training data for Machine Learning project.
Duration
4 Hours (half day)
Intended Audience
Students and professional involved in analyzing seismic data
Prerequisites (Knowledge/Experience/Education Required)
BS/MS, experience/interest with exploration seismic data
Who should attend?
Graduate students, researchers, professional geoscientists (interpreters, reservoir characterization) – using seismic data
Course Outline
- What is synthetic seismogram and why we need it?
- Computation methods and challenges
- Uses of synthetics, including well-tie, wavelet/phase estimation, inversion, interpretation, and building training data for Machine Learning projects
Learner Outcomes
- Describe inputs and steps in synthetic seismogram computation
- Using seismic-well-tie to link wiggle to geology in seismic interpretation
- Synthetics to estimate wavelet/phase and resolution of seismic
- Value of synthetic in amplitude calibration, including seismic inversion
- Role of synthetics in building training data for Machine Learning projects