MODELLING WATER INTAKE PROPERTIES OF HEAT-TREATED BEECH AND SPRUCE WOOD TREATED AT DIFFERENT TEMPERATURES USING BY ARTIFICIAL NEURAL NETWORKS

Authors : Ayşenur GÜRGEN, Sibel YILDIZ

Pages : 6-12

View : 36 | Download : 13

Publication Date : 2020-06-30

Article Type : Research Paper

Abstract :The aim of this study is the modelling the water intake rate of heat-treated oriental beech insert ignore into journalissuearticles values(Fagus orientalis Lipsky); and oriental spruce insert ignore into journalissuearticles values(Picea orientalis insert ignore into journalissuearticles values(L); Link); wood samples. For this purpose, all the needed data were obtained from the beech and spruce wood samples which have been subjected to heat treatment with four different temperatures insert ignore into journalissuearticles values(130, 150, 180 and 200 °C); and three different periods insert ignore into journalissuearticles values(2, 6 and 10 hour); and then which have been subjected to the water intake process at certain periods insert ignore into journalissuearticles values(2, 4, 8, 24, 48, 72, 168 and 336 hour);. Data were modeled using artificial neural networks insert ignore into journalissuearticles values(ANN); method for both tree species in terms of water intake rate characteristics, seperately. Two different learning algorithms insert ignore into journalissuearticles values(Levenberg-Marquardt insert ignore into journalissuearticles values(LM); and Scaled Conjugate Gradient insert ignore into journalissuearticles values(SCG);); were used for the modeling process. In order to achieve the best model, all nodes between 1 and 25 were tested as hidden neuron. A total of 100 models were obtained and 2 models were chosen according to the performance of the models. For two wood species, LM learning algorithm had showed better results than SCG learning algorithm. The structures of the best models for beech and spruce were determined as 3-8-1 and 3-13-1 respectively. As a result, it has been concluded that ANN applications can be evaluated within the discipline of wood protection.
Keywords : Heat treatment, beech, spruce, water intake, artificial neural networks

ORIGINAL ARTICLE URL