Guide for Verification and Validation of Computational Fluid Dynamics Simulation
|Publication Date:||14 January 1998|
The fundamental strategy of V&V is the assessment of error and uncertainty in the computational simulation. The required methodology is a complex process because it must assess errors and uncertainties originating in all three roots of CFD: theory, experiment, and computation. Given these diverse perspectives, it is common to find disagreement and conflict in the terminology of V&V. Furthermore, because fluid dynamics is dominated by nonlinear phenomena, it is common for multiple nonlinearities to be strongly coupled. This introduces significant difficulties in modeling the phenomena and in solving the resulting nonlinear partial differential equations.
This document builds primarily on terminology established by the Society for Computer Simulation and the Defense Modeling and Simulation Office of the Department of Defense [2-4]. Concerning the methodology of V&V, however, there are no publications presenting general and comprehensive procedures in the computational sciences. It is fair to call the present state of the art for V&V methodology ad hoc. The purpose of this document is to promote the establishment of basic terminology and methodology for the V&V of CFD simulations.
It is important to emphasize that this document presents guidelines for V&V of CFD simulations, not standards. The AIAA Standards Procedures are segregated into three levels of state-of-the-art: guides, recommended practices, and standards. This document represents the first level, a guide. The AIAA Computational Fluid Dynamics Committee on Standards unanimously believes that the state of the art in CFD has not developed to the point where standards can be written. The Committee is dedicated to revising this document on a regular basis, following the same approach taken in the preparation of this document. That is, revisions will be made with broad input from other AIAA Technical Committees and any individuals interested in the advancement of CFD.
A few archival journals have developed editorial policies pertaining to the control of numerical accuracy in fluid flow simulations [5-8]. Numerical accuracy is one aspect of V&V, but there are many more aspects as discussed in these guidelines. While it is desired that these guidelines can lead to enhancing the quality of work published in journals, publication-related issues are not specifically discussed within the scope of this document. It should also be made clear that the procedures described in these guidelines are not meant to be necessary conditions for publication of manuscripts in any of the AIAA journals or at any conferences sponsored by AIAA.
This guide is predicated upon the notion that there is no fixed level of credibility or accuracy that is applicable to all CFD simulations. The accuracy level required of simulations depends on the purposes for which the simulations are to be used. In effect, all simulations do not need to demonstrate high accuracy. For example, absolute or high accuracy simulations are not normally required for engineering activities; such simulations only need to be useful, not perfect. The required level of accuracy must be determined for each use of the simulation. Typical practicalities affecting the accuracy obtained are cost, schedule, and safety implications of the simulation.