Several statistical methods and their corresponding principles of application to structural dynamics problems will be presented. This set was chosen based upon the projects and their corresponding challenges in the Engineering Sciences & Applications (ESA) Division at Los Alamos National Laboratory and focuses on variance-based uncertainty quantification. Our structural dynamics applications are heavily involved in modeling and simulation, often with sparse data availability. In addition to models, heavy reliance is placed upon the use of expertise and experience. Beginning with principles of inference and prediction, some statistical tools for verification and validation are introduced. Among these are the principles of good experimental design for test and model computation planning, and the combination of data, models and knowledge through the use of Bayes Theorem. A brief introduction to multivariate methods and exploratory data analysis will be presented as part of understanding relationships and variation among important parameters, physical quantities of interest, measurements, inputs and outputs. Finally, the use of these methods and principles will be discussed in drawing conclusions from the validation assessment process under uncertainty.
[1]
Sallie Keller-McNulty,et al.
Testing the untestable: reliability in the 21st century
,
2003,
IEEE Trans. Reliab..
[2]
H. Martz.
Bayesian reliability analysis
,
1982
.
[3]
Gwilym M. Jenkins,et al.
Time series analysis, forecasting and control
,
1972
.
[4]
Noel A Cressie,et al.
Statistics for Spatial Data.
,
1992
.
[5]
Paul G. Hoel,et al.
Elementary statistics
,
1971
.
[6]
Mike Rees,et al.
5. Statistics for Spatial Data
,
1993
.
[7]
Jane M. Booker,et al.
Eliciting and analyzing expert judgement - a practical guide
,
2001,
ASA-SIAM series on statistics and applied probability.
[8]
Jane M. Booker,et al.
Ch. 11. Predict: A new approach to product development and lifetime assessment using information integration technology
,
2003
.