I am a computational mathematician at Sandia National
Laboratories. My research interests include numerical analysis,
applied mathematics and computational science, with particular focus on
 Compatible and alternative discretizations of PDEs,
including mimetic, DG, and stabilized methods.
 Finite element methods of leastsquares type
 Finite element methods for PDE constrained optimization
 Multiscale methods: atomistictocontinuum coupling, domain bridging
I earned my PhD at Virginia Tech
under the direction of Max Gunzburger .
My current research is in the following areas:
Compatible discretizations
Discretization is a model reduction process which converts infinite dimensional mathematical models into finite dimensional algebraic equations that can be solved on a computer. Consequently, discretization is accompanied by inevitable losses of information which can adversely affect the predictiveness of the discrete models. Compatible and mimetic discretizations control these losses directly by using discrete fields and operators which inherit the key structural properties of their continuum counterparts. As a result, compatible discretizations transform partial differential
equations into discrete algebraic problems that mimic fundamental
properties of the continuum equations. I work on a common framework for mimetic
discretizations using algebraic topology to guide the analysis. This framework provides the basis for the development of the Intrepid package.
Optimizationbased modeling
This research is outgrowth of previous work on compatible discretizations and extends it in an important complementary direction. Compatible and regularized methods excel in controlling "structural" information losses responsible for the stability and wellposedness of the discrete equations. However, they encounter considerable difficulties in at least two cases: multiphysics models which combine constituent components with fundamentally different mathematical properties, and loss of "qualitative" properties such as maximum principles, monotonicity and local bounds preservation.
The main objective of this research is to develop a formal approach which uses optimization ideas to control externally information losses which are difficult (or impractical) to manage directly in the discretization process. Ultimately, the goal is to use optimization and control ideas in order to improve predictiveness of computational models, increase robustness and accuracy of solvers, and enable efficient reuse of code.
Software tools for compatible discretizations
Intrepid is a library of interoperable tools for compatible discretizations of Partial Differential Equations (PDEs). Design of Intrepid is motivated by the common framework for compatible discretizations. A key feature of the design philosophy is separation of cell topology from the reconstruction process. As a result, Intrepid can be easily extened to include new cell shapes and/or function spaces. Development of Intrepid started in 2007 and the package was publicly released in 2009 with Trilinos 10.0. Intrepid is part of the Discretization Capability Area (DCA) in Trilinos. The goal of DCA is to provide, over time, a collection of libraries and interfaces that enable rapid development of application codes for applications that require numerical solution of Partial Differential Equations (PDE).
Atomistictocontinuum coupling
AtomistictoContinuum (AtC) coupling is a critical
component in computational materials science and other
applications of interest to the DOE Office of Science. Our
goal is to understand and quantify mathematically the limits in AtC
coupling and the resulting impact on multiscale simulations.
Multiscale methods
Part of this research deals with application of Variational
Multiscale Analysis (VMS)to develop new finite element methods
for PDEs. Another research direction is domain bridging where we
look at novel Lagrangemultiplier formulations for mesh
tying
Top of page

Pavel Bochev — Distinguished Member of the Technical Staff.
Contact
Email: Pavel Bochev
(505)8441990 (Phone)
(505)8457442 (Fax)
Mailing address
Sandia National Laboratories
P.O. Box 5800, MS 1320
Albuquerque, NM 871851320
Delivery address
FedEX/UPS/DHL
Sandia National Laboratories
1515 Eubank SE,
Albuquerque, NM 871231320
Useful Links
SINUM
IJNAM
LANL
