Predicting Strength of Matched Sets of Test Specimens

Zhiyong Cai, George P. McCabe, Michael O. Hunt, James P. Bradtmueller, Kenneth J. Fridley, David V. Rosowsky

Abstract


Five different methods for a priori estimating bending strength of wood and wood composite specimens are compared in this paper. They are: (1) edge-matching, (2) matching specimens by normal distribution, (3) matching specimens by log-normal distribution, (4) simple linear regression, and (5) multiple linear regression. It was found that the square root of mean square error (RMSE) of percent difference (PD) of predicted modulus of rupture (MOR) is the key measure in comparing the five methods. Multiple linear regression was found to be the best method to predict MOR of a specimen in an edge-matched set. Finally, how to create the prediction limits for mean MOR of a subgroup of specimens is discussed. The prediction limits for predicting MOR make it possible to quantitatively determine the effect of various treatments of wood materials.

Keywords


Predicting strength;edge-matching;distribution;regression;prediction limits

Full Text:

PDF

Refbacks

  • There are currently no refbacks.