Research Article

Application of an automatic differentiation tool in the development of a compositional reservoir simulator

  1. Jason Abate,
  2. Peng Wang,
  3. Kamy Sepehrnoori*,
  4. Clint Dawson

Article first published online: 29 JUN 1999

DOI: 10.1002/(SICI)1099-0887(199906/07)15:6<423::AID-CNM256>3.0.CO;2-5

Issue

Communications in Numerical Methods in Engineering

Communications in Numerical Methods in Engineering

Volume 15, Issue 6, pages 423–434, June/July 1999

Additional Information

How to Cite

Abate, J., Wang, P., Sepehrnoori, K. and Dawson, C. (1999), Application of an automatic differentiation tool in the development of a compositional reservoir simulator. Communications in Numerical Methods in Engineering, 15: 423–434. doi: 10.1002/(SICI)1099-0887(199906/07)15:6<423::AID-CNM256>3.0.CO;2-5

Author Information

  1. The University of Texas at Austin, Austin, TX 78712, U.S.A.

*Center for Petroleum and Geosystems Engineering, The University of Texas at Austin, Austin, TX 78712, U.S.A.

Publication History

  1. Issue published online: 29 JUN 1999
  2. Article first published online: 29 JUN 1999
  3. Manuscript Accepted:
  4. Manuscript Received:

Funded by

  • U.S. Department of Energy. Grant Number: W-31-109ENG-38
  • U.S. Department of Defense

SEARCH

SEARCH BY CITATION

ARTICLE TOOLS

Get PDF (260K)

Keywords:

  • automatic differentiation;
  • nonlinear equations;
  • reservoir simulation

Abstract

Automatic differentiation is an approach to numerically compute derivatives both accurately and efficiently. This can be quite valuable, for example, in the fully implicit solution of compositional reservoir equations. We describe the procedures necessary to apply the ADIFOR (Automatic Differentiation of Fortran) tool to the IPARS (Integrated Parallel Accurate Reservoir Simulator) compositional reservoir simulator and the approaches used to exploit the sparsity in the required Jacobian matrix. Values and execution times of the ADIFOR-generated derivatives are compared with a handwritten code that implements analytically derived expressions for the derivatives. Copyright © 1999 John Wiley & Sons, Ltd.

Get PDF (260K)