Endentação dinâmica aplicada na predição de propriedades físico-mecânicas da madeira serrada e de painéis de madeira
Carregando...
Data
2020-05-27
Autores
Assis, Albert Augusto de
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
Este trabalho avaliou o potencial de utilização de parâmetros característicos do ensaio de endentação dinâmica para predição de propriedades físico-mecânicas da madeira, de painéis de madeira de média densidade (MDP) e de pisos laminados. A técnica apresentada consistiu em identificar diversos parâmetros de seis ciclos do ensaio dinâmico, selecioná-los por meio de método estatístico e utilizá-los para predição das propriedades de interesse com o uso de regressão múltipla. No programa experimental foram estudadas nove espécies de madeira de diferentes densidades, 40 painéis MDP convencionais, 10 painéis de alta performance (HPP) e 60 peças de pisos laminados. Os ensaios de endentação dinâmica foram realizados com o uso de um durômetro portátil automatizado e processados em um programa computacional desenvolvido especificamente para análise do fenômeno de endentação. Foram obtidos 144 parâmetros característicos do ensaio de endentação dinâmica, os quais foram selecionados por meio do método Stepwise e utilizados na calibração dos modelos de regressão múltipla. A capacidade de generalização dos modelos foi avaliada por meio de validação cruzada k-fold. Os resultados de calibração revelaram bons coeficientes de determinação na predição de quatro propriedades da madeira (0,84 ≤ R² ≤ 0,98), cinco propriedades dos painéis MDP (0,84 ≤ R² ≤ 0,89) e sete propriedades dos painéis HPP (0,78 ≤ R² ≤ 0,98). Os modelos apresentaram baixos erros de predição (RMSE) e boa performance na etapa de validação. Os valores preditos do ensaio de resistência à deformação estática dos pisos laminados foram categorizados e utilizados na classificação das peças obtendo acurácia de 90% e coeficiente kappa de classificação de 0,8. O ensaio de endentação demonstrou potencial para ser utilizado como método alternativo para predição de diversas propriedades físico-mecânicas da madeira e de seus produtos, podendo ser aplicado com o uso de um equipamento portátil com procedimento de operação simples e expedito.
This study aimed to investigate the use of characteristics of the dynamic indentation test to predict physical-mechanical properties of sawn wood, particleboard and laminate flooring. The technique here introduced consists of identifying many characteristics from six cycles of the dynamic test, selecting them by means of a statistical method and using them to predict the desired properties by means of multiple regression. The experimental program studied 9 wood species varying in densities, 40 three-layer standard particleboards, 10 high-performance particleboards (HPP) and 60 pieces of laminate flooring. The dynamic indentation tests were performed using an automated portable tester and were processed by a software specifically designed for dynamic indentation analysis. Predictor variables were selected by Stepwise method from a set of 144 characteristics of the dynamic test and used for calibration of multiple regression models. The generalization performance of the models was assessed by means of k-fold cross validation. The calibration tests revealed good coefficients of determination in the prediction of four properties of sawn wood (0.84 ≤ R² ≤ 0.98), five properties of particleboard (0.84 ≤ R² ≤ 0.89) and seven properties of high-performance particleboards (0.78 ≤ R² ≤ 0.98). The models reached low prediction errors (RMSE) and good performance in the validation process. The pieces of laminate flooring were classified using the results of the predicted residual deformation to static load, reaching accuracy of 90% and a kappa coefficient of classification of 0.8. The dynamic indentation test revealed potential to be used as an alternative method for estimating various physical-mechanical properties of the wood and wood products. It can be applied with a portable tester using a simple and quick procedure.
This study aimed to investigate the use of characteristics of the dynamic indentation test to predict physical-mechanical properties of sawn wood, particleboard and laminate flooring. The technique here introduced consists of identifying many characteristics from six cycles of the dynamic test, selecting them by means of a statistical method and using them to predict the desired properties by means of multiple regression. The experimental program studied 9 wood species varying in densities, 40 three-layer standard particleboards, 10 high-performance particleboards (HPP) and 60 pieces of laminate flooring. The dynamic indentation tests were performed using an automated portable tester and were processed by a software specifically designed for dynamic indentation analysis. Predictor variables were selected by Stepwise method from a set of 144 characteristics of the dynamic test and used for calibration of multiple regression models. The generalization performance of the models was assessed by means of k-fold cross validation. The calibration tests revealed good coefficients of determination in the prediction of four properties of sawn wood (0.84 ≤ R² ≤ 0.98), five properties of particleboard (0.84 ≤ R² ≤ 0.89) and seven properties of high-performance particleboards (0.78 ≤ R² ≤ 0.98). The models reached low prediction errors (RMSE) and good performance in the validation process. The pieces of laminate flooring were classified using the results of the predicted residual deformation to static load, reaching accuracy of 90% and a kappa coefficient of classification of 0.8. The dynamic indentation test revealed potential to be used as an alternative method for estimating various physical-mechanical properties of the wood and wood products. It can be applied with a portable tester using a simple and quick procedure.
Descrição
Palavras-chave
Madeira, Testes dinâmicos, Matéria - Propriedades, Wood, Dynamic tests, Physicalmechanical properties