[BANANA] Linear Algebra/Optimization Seminar, Wed Dec 5

[Michael A. Saunders]

   Linear Algebra and Optimization Seminar
   Stanford University
   http://icme.stanford.edu/seminars/seminar.php?seminar_id=2&current=true

   4:15pm Wed Dec 5, 2007
   Rm 326 Wallenberg Hall (Bldg 160)

   Optimizations in the Oil Industry

   Dr David Ciaurri
   Energy Resources Engineering, Stanford University
   David.Echeverria at stanford.edu
   http://pangea.stanford.edu/~echeverr/

In the very first part of this talk we will briefly present the
Smart Fields Consortium at Stanford University and describe the
main lines of research followed there. The Smart Fields
Consortium is a multidisciplinary industrial affiliates program
with participants from several departments from Stanford
University. The research there involves the development and
testing of optimization, risk assessment, decision-making and
real-time monitoring and model calibration techniques, within
the oil industry.

The optimization problems aimed at in the consortium are of
large-scale nature and the cost function typically requires
time-demanding computations (for example, the solution of the
discretized flow equations). In the second part of the talk we
will describe two key strategies for solving these problems
efficiently: kernel principal component analysis (KPCA) and the
method of adjoints for obtaining gradient information. KPCA [1]
helps in decreasing the number of optimization variables by
means of a nonlinear mapping to a high-dimensional space. With
this mapping, the correlation between variables is eliminated
and at the same time some (geological) optimization constraints
are honored. The adjoint equations [2,3] allow derivative
information to be extracted from simulation models without a
significant additional cost.

The concepts dealt with during the presentation will be
illustrated by means of some large-scale optimization problems
of practical relevance carried out within related Smart Field
projects.

[1] B. Schölkopf, A. Smola, and K. Müller,
     Nonlinear component analysis as a kernel eigenvalue problem,
     Neural Computation, 10, 1299-1319, 1998.

[2] O. Pironneau, Optimal shape design for elliptic systems,
     Computational Physics, Springer, Berlin, 1984.

[3] P. Sarma, L.J. Durlofsky, K. Aziz and W.H. Chen,
     Efficient real-time reservoir management using
     adjoint-based optimal control and model updating,
     Computational Geosciences, 10(1), 3-36, 2006.


This is the last LA/Opt seminar of Fall 2007.
We hope to continue Gene's tradition in the Winter and Spring.
Anyone who would be happy to speak, please let us know.
-------------- next part --------------
--++**==--++**==--++**==--++**==--++**==--++**==--++**==
linear_algebra_optimization mailing list
linear_algebra_optimization at lists.stanford.edu
https://mailman.stanford.edu/mailman/listinfo/linear_algebra_optimization