Modeling to yield estimation in deploying logs of Araucaria angustifolia (Bertol.) Kuntze.

Abstract

The objective of this study was to test and select mathematical models to predict the yield in deploying logs process of Araucaria angustifolia (Bertol.) Kuntze. The volume was determined by Huber. Eleven models were tested using as dependent variable the volume yield of wasting wood (R) and as independent variable the smaller diameter (d), the length (ℓ) and the total volume of log (v). All the dimensional relations were analyzed through correlation matrix. The choice of the best model was based on the highest of adjusted coefficient determination, smallest residual standard error and most uniform behavior and homogeneous of residuals values. The average yield observed was 49,2% varying from 42,5 to 57,4%. The highest correlations were obtained for total volume and yield, as well as the diameter and yield, 97,2 and 96,9%, respectively. Among the tested models, the best result was obtained by the equation R=-0,0040596673+0,0000903028*d²+0,0000146008*(d²*ℓ) with (R²_aj.) 96,7% and standard error (S_yx) = 12,8%. With this equation the yield prediction becomes possible in deploying process, in advance.