Some thoughts on model verification
The most powerful attribute of numerical models for simulating fluid flow in partially saturated fractured media is that they can handle arbitrarily complex problems in terms of geometry, heterogeneity and forcing functions. In contrast, analytical solutions can be only obtained for relatively simple problems. Therefore, the very power of the numerical models that is to be exploited is unnecessarily abridged if the numerical method is viewed merely as a tool for approximately solving the partial differential equation. A broader definition of verification is needed in order to exploit the power and generality of numerical models. In the ultimate, verification should be dictated by axiomatic testability. The development of self-verifying numerical models is a challenge that deserves pursuit.
- Research Organization:
- Lawrence Berkeley Lab., CA (USA)
- DOE Contract Number:
- AC03-76SF00098
- OSTI ID:
- 6933641
- Report Number(s):
- LBL-23021; CONF-861220-5; ON: DE87007615
- Resource Relation:
- Conference: American Geophysical Union fall meeting, San Francisco, CA, USA, 8 Dec 1986; Other Information: Portions of this document are illegible in microfiche products
- Country of Publication:
- United States
- Language:
- English
Similar Records
Numerical modeling in hydrology
Compiler verification meets cross-language linking via data abstraction