This is the second course of the Geostatistics series. The course involves the study of a real data set; examples have been a large N. Sea clastic reservoir with well and seismic data, a lead contamination site, a large polymetallic porphyry deposit in Chile, topographic data in Nevada. The data typically includes sparse hard data and more extensive soft information (seismic/geological interpretation). Student teams independently perform site characterization and share results in class. Reservoir study through maps, variograms, kriging, and stochastic models. Extensive use of GSLIB and 3D visualization software. Flow simulations for recovery forecast and placement of additional wells. Engineering design for site remediation. Prerequisites: 240, FORTRAN/UNIX. Recommended: 246. 3-4 units, Spr (Caers, Mukerji)
Optimization: Deterministic and Stochastic Approaches (PE284)
Class-Website ( login with user ID : guest, login : guest )Deterministic and stochastic methods for optimization in earth sciences and engineering. Linear and non-linear regression, classification and pattern recognition using neural networks, simulated annealing and genetic algorithms. Deterministic optimization using non-gradient-based methods (simplex) and gradient-based methods (conjugated gradient, steepest descent, Levenberg-Marquardt, Gauss-Newton), eigenvalue and singular value decomposition. Applications in petroleum engineering, geostatistics and geophysics. 3 units, Fall (Caers)
The course is intended to give an
overview of the principles of optimization.
Optimization is probably the most important engineering problem. Most optimization
problems also involve some kind of inverse problem.
For example an unknown medium (Earth) is remotely sensed using various
sensing devices, which yields certain responses of the medium (e.g. seismic
surveys, pressure response in well test). The task is then to invert the
medium from the measurements, for example what is porosity at each location
given a particular 3D seismic survey, or what is the permeability distribution
given a particular well test. I cover in the course the mathematical tools
that are used to solve this problem. Some topics covered are
Introduction to Statistical Methods for Earth and Environmental Sciences: general introduction (GES/PE 160)
Class-Website ( login with user ID : guest, login : guest )Data summaries, graphical display of data, measures of association, time series and trends, sampling, quantification of uncertainty, statistical models, statistical testing and prediction. The analysis of spatial information, introduction to geostatistical methods for estimating spatial phenomena. Examples on prediction and uncertainty quantification in geology and environmental monitoring. Issues of statistical computing and software. 4 units, Spring (Caers)
This course is intended as a general
introduction to statistics and to the statistics used for describing variables
distribution is space (Geostatistics). Basic Calculus background is required
but no prior statistical knowledge.
Introduction to Statistical Methods for Earth and Environmental Sciences: Geostatistics (GES/PE161)
Class-Website ( login with user ID : guest, login : guest )
Review of statistical analysis of data, graphical display of
data, common distribution models, sampling and regression. The
variogram as a tool for modeling spatial correlation, variogram
estimation and modeling, introduction to spatial mapping and
prediction with kriging, integration of remote sensing and other
ancillary information using co-kriging models, spatial
uncertainty, introduction to geostatistical software applied to
large environmental, climatological and reservoir engineering
ddatabases, emphasis on practical use of geostatistical tools.
Either of both of GES 160 and GES 161 may be taken.
This course is intended as a general
introduction to geostatistics from a practical point of view. Students work on various datasets
and learn how to use geostatistical tools. The mathemtaical development is limited.