Inscrição na biblioteca: Guest

ISSN Online: 2377-424X

ISBN Print: 1-56032-797-9

International Heat Transfer Conference 11
August, 23-28, 1998, Kyongju, Korea

CONFIDENCE INTERVALS FOR ESTIMATION OF THERMAL PROPERTIES - SIMULATIONS

Get access (open in a dialog) DOI: 10.1615/IHTC11.2920
pages 59-63

Resumo

The present work investigates the effects of correlated measurement errors on parameter estimates and confidence intervals in transient estimation experiments. The more difficult, but important case of serially-and-spatially correlated errors in multiple sensors experiments is addressed. A first-order autoregressive is used to specify a model for autocorrelated measurement errors in time. Spatial error correlation between sensors is due to non-constant variances for each sensor. The approximate Bonferroni method is used to provide simultaneous 95 percent confidence intervals for estimated parameters. Test cases for simultaneous estimation of thermal conductivity and volumetric specific heat, and the associated 95 percent confidence intervals are presented. A Monte Carlo study is used to investigate the reliability of parameter estimators and confidence intervals. The results of simulations show that correlated errors tend to increase the variances of the estimated parameters and hence enlarge confidence intervals.