Practical application of uncertainty-based validation assessment
- Mark C.
- Jeffrey E.
- Ryan D.
- Amanda C.
Validation of simulation results by comparison with experimental data is certainly not a new idea. However, as the capability to simulate complex physical phenomena has increased over the last few decades, the need for a systematic approach to validation assessment has become evident. Organizations such as the American Society of Mechanical Engineers (ASME) and the National Laboratories are in the process of formulating validation requirements and approaches. A typical depiction of the validation process is given in Figure 1, derived from current ASME efforts regarding computational solid mechanics. Note that uncertainty quantification plays an integral role in the validation comparison step of the process defined in the figure. This is a natural consequence of the need for verification and validation to facilitate decision-making by the customers of simulation results. Since very little is exactly known about real systems, questions of economy, reliability, and safety are best answered in the language of uncertainty. The process illustrated in the figure above is very logical, but very general. Examples of concrete applications of this are still rare. Engineers at Los Alamos National Laboratory have been applying a systematic verification and validation process, much like that of Figure 1, to structural dynamic simulations for the past several years. These applications have resulted in the realizations that there are many details not mentioned in general process descriptions that can complicate a validation assessment. Such details include the following: (1) The need for a hierarchical approach in which the interactions between components within/between assemblies are considered in addition to the overall input/output behavior of the entire system; (2) The need for system state and response data within the important elements of the hierarchy in addition to the observed characteristics at the system level; (3) Selection of appropriate response features for comparison between analytical and experimental data; (4) Selection of a comprehensive, but tenable set of parameters for uncertainty propagation; and (5) Limitations of modeling capabilities and the finite element method for approximating high frequency dynamic behavior of real systems. This paper illustrates these issues by describing the details of the validation assessment for an example system. The system considered is referred to as the 'threaded assembly'. It consists of a titanium mount to which a lower mass is attached by a tape joint, an upper mass is connected via bolted joints, and a pair of aluminum shells is attached via a complex threaded joint. The system is excited impulsively by an explosive load applied over a small area of the aluminum shells. The validation assessment of the threaded assembly is described systematically so that the reader can see the logic behind the process. The simulation model is described to provide context. The feature and parameter selection processes are discussed in detail because they determine not only a large measure of the efficacy of the process, but its cost as well. The choice of uncertainty propagation method for the simulation is covered in some detail and results are presented. Validation experiments are described and results are presented along with experimental uncertainties. Finally, simulation results are compared with experimental data, and conclusions about the validity of these results are drawn within the context of the estimated uncertainties.
- Research Organization:
- Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)
- Sponsoring Organization:
- USDOE
- OSTI ID:
- 977856
- Report Number(s):
- LA-UR-04-6444; TRN: US201012%%557
- Resource Relation:
- Conference: Submitted to: 46th AIAA/ASME/ACSE/AHS/ASC Structures, Strucural Dynamics&Materials Conference,18-21 April 2005, Austin, TX
- Country of Publication:
- United States
- Language:
- English
Similar Records
Comparative Assessment of Experimental Testing of Instrument with an Embedded Digital Device Using Model-Based and Conventional Methods
Systematic approach to verification and validation: High explosive burn models