Jv. Olivervillanueva et al., INFLUENCE OF STRUCTURAL PARAMETERS ON THE NONDESTRUCTIVE EVALUATION OF ASH TIMBER (FRAXINUS-EXCELSIOR L), Holz als Roh- und Werkstoff, 54(2), 1996, pp. 109-112
The aim of this project is to analyse the influence of wood structure
of ash timber products - as a result from different growth rates - on
their performance, if nondestructive evaluation is applied. To do this
, 261 beams have been tested with the method of natural frequency on f
lexural vibration. The following anatomical and structural parameters
have been documented as independent variables on all tested beams: ann
ual ring width, proportion of latewood, ring position, grain orientati
on, presence or absence of heartwood and other discolourations as well
as the positions and dimensions of knots. The values obtained from th
e nondestructive evaluation have been put into relation with the MOR o
btained through final destructive 4-point-bending test of all beams. T
he influence of the wood structure on this relationship has been analy
sed either with univariate regressions between single structural param
eters and MOR, or including more than one of these structural paramete
rs in multivariate regression models. The regressions between nondestr
uctive parameters and MOR present lower levels of determination than t
he usual determination levels reported for softwood timber. For the ri
ngporous species ash the influence of the wood structure could be more
significant than for the more homogeneous wood structure of coniferou
s species and, consequently, should be analysed in detail. The main re
sults of this analysis show only minimal influence of the ring structu
re and of the grain angle on the relationship between nondestructive v
alues and MOR. The presence of heartwood does not influence this relat
ionship. By far the most important parameter is the presence or absenc
e of knots. Different knot parameters were measured and included into
the models: KAR and vectorial models of position and dimension of the
knots. This detailed measurement of the knots and its inclusion into m
ultiple regression models allow a significant increase of the total le
vel of determination and consequently an improved nondestructive quali
ty assessment of ash timber products. Consequences for the practical a
pplication of the results in industrial sawmilling are considered.