|
Descripción
|
|
|---|---|
| Non-destructive testing is becoming increasingly common as the method of choice for studying wooden structures. However, findings using this method, based on statistical models in which the mechanical properties are related to variables measured in place, are not always as reliable as desired. Artificial neural networks, developed in the last forty years, are mathematical structures which have demonstrated their use in modelling complex relations between variables, particularly when obtaining a reliable result is more important than determining how the variables interact. These networks imitate the functioning of biological systems in such a way that they are capable of extracting knowledge from a series of sampling data and applying it to unknown data. They have shown enormous potential in several fields, ranging from medicine to finances, including engineering, in which timber engineering is no exception. In this study the modulus of elasticity of Abies pinsapo Boiss. wood was obtained using the data of test piece dimensions, moisture content, density and ultrasonic wave propagation velocity. The artificial neural network used was a feedforward multilayer perceptron, which improved on results obtained using traditional statistical methods by 65%. | |
|
Internacional
|
Si |
|
Nombre congreso
|
World Conference on Timber Engineering 2010 |
|
Tipo de participación
|
960 |
|
Lugar del congreso
|
Trento - Italia |
|
Revisores
|
Si |
|
ISBN o ISSN
|
0000-0000 |
|
DOI
|
|
|
Fecha inicio congreso
|
20/06/2010 |
|
Fecha fin congreso
|
24/06/2010 |
|
Desde la página
|
1 |
|
Hasta la página
|
7 |
|
Título de las actas
|
2010 World Conference on Timber Engineering |