Descripción
|
|
---|---|
The objective of this work is the development of neural network (NN) models to establish a stress grading structural classification methodology by means of nondestructive techniques. Scots pine (Pinus sylvestris L.) wood is the object of study. Considering the suitability of applying NN models to grade timber beam specimens, the following input variables were considered for a series of NN models: wood provenance, real and nominal dimensions, beam length, visual grade, longitudinal vibration velocity, knot proportion on a section, depth of penetration, and screw withdrawal force. Output characteristics were global density, static MOE, modulus of rupture (MOR), and the three of them together. The models show acceptable figures for correlation results (ranging from 0.552 to 0.779), and performance values are good (ranging from 0.701 to 0.887). No model shows an equivalent result for correlation vs. performance values, but except for the global density model where only one good value was found. In a nutsell, the model showing the best correlation (or performance) is not necessarily the model showing the best performance (or correlation). The observed vs. predicted plots follow a 45° pattern, meaning a good approach for prediction. Further studies will include enlarging P. sylvestris data set, considering other species, studying the economic feasibility of the method, in-factory control, and environmental epigenomics and structural stress changes susceptibility. | |
Internacional
|
Si |
Nombre congreso
|
18th International Nondestructive Testing and Evaluation of Wood Symposium |
Tipo de participación
|
960 |
Lugar del congreso
|
Madison, Wisconsin, EEUU |
Revisores
|
Si |
ISBN o ISSN
|
00-0000-000-0 |
DOI
|
|
Fecha inicio congreso
|
24/09/2013 |
Fecha fin congreso
|
27/09/2013 |
Desde la página
|
808 |
Hasta la página
|
808 |
Título de las actas
|
Proceedings 18th International Nondestructive Testing and Evaluation of Wood Symposium. General Technical Report FPL-GTR-226. Madison, WI: U.S. Department of Agriculture, Forest Service, Forest Products Laboratory |