Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

Sensitivity analysis and model reduction of nonlinear differential-algebraic systems. Final progress report

Technical Report ·
DOI:https://doi.org/10.2172/354988· OSTI ID:354988
Differential-algebraic equations arise in a wide variety of engineering and scientific problems. Relatively little work has been done regarding sensitivity analysis and model reduction for this class of problems. Efficient methods for sensitivity analysis are required in model development and as an intermediate step in design optimization of engineering processes. Reduced order models are needed for modelling complex physical phenomena like turbulent reacting flows, where it is not feasible to use a fully-detailed model. The objective of this work has been to develop numerical methods and software for sensitivity analysis and model reduction of nonlinear differential-algebraic systems, including large-scale systems. In collaboration with Peter Brown and Alan Hindmarsh of LLNL, the authors developed an algorithm for finding consistent initial conditions for several widely occurring classes of differential-algebraic equations (DAEs). The new algorithm is much more robust than the previous algorithm. It is also very easy to use, having been designed to require almost no information about the differential equation, Jacobian matrix, etc. in addition to what is already needed to take the subsequent time steps. The new algorithm has been implemented in a version of the software for solution of large-scale DAEs, DASPK, which has been made available on the internet. The new methods and software have been used to solve a Tokamak edge plasma problem at LLNL which could not be solved with the previous methods and software because of difficulties in finding consistent initial conditions. The capability of finding consistent initial values is also needed for the sensitivity and optimization efforts described in this paper.
Research Organization:
Univ. of Minnesota, Minneapolis, MN (United States)
Sponsoring Organization:
USDOE Office of Energy Research, Washington, DC (United States)
DOE Contract Number:
FG02-92ER25130
OSTI ID:
354988
Report Number(s):
DOE/ER/25130--T1; ON: DE99003199
Country of Publication:
United States
Language:
English

Similar Records

Using Krylov methods in the solution of large-scale differential-algebraic systems
Journal Article · Mon Oct 31 23:00:00 EST 1994 · SIAM Journal on Scientific Computing; (United States) · OSTI ID:6813942

Recent developments in the numerical solution of differential/algebraic systems
Conference · Fri Oct 09 00:00:00 EDT 1987 · OSTI ID:5789902

Consistent initial condition calculation for differential-algebraic systems
Journal Article · Tue Sep 01 00:00:00 EDT 1998 · SIAM Journal on Scientific Computing · OSTI ID:321000