INTERNATIONAL JOURNAL OF ADHESION AND ADHESIVES, cilt.123, 2023 (SCI-Expanded)
Although artificial neural networks (ANN) have been used frequently in engineering applications, it has been determined that they are not preferred much, especially in determining adhesive properties and usage. Hence, this study aimed to determine the effect of urea formaldehyde (UF) resin mixed with valonia tannin obtained from Turkish oak acorns (Quercus aegilops) as a filler on the formaldehyde emission and bonding strength of plywood panels by using ANN modelling. For this purpose, six different valonia tannin contents and three different adhesive usages were chosen, and they were compared to control panels. The formaldehyde emission and bonding strength experimental data obtained according to relevant standards were analysed by using ANN. The prediction models with the best performance and acceptable deviations were determined using statistical and graphical comparisons between the experimental data and the prediction values obtained from the ANN analysis. Using these prediction models, the formaldehyde emission and bonding strength data were obtained according to the amount of valonia tannin and adhesive usage not used in the experimental study. Experimental results have proved that valonia tannin obtained from Turkish oak acorns is a good formaldehyde scavenger for UF adhesives, but it has been determined that it has a negative effect on the bonding strength of plywood. As a result of the ANN analysis, the lowest formaldehyde emission values were obtained from 12% to 15% valonia tannin and 140 g/m2 - 146 g/m2 adhesive usage. The bonding strength values of all the predicted plywood groups exceeded the limit value of the relevant standard except for the 5% valonia tannin and 140 g/m2 - 158 g/ m2 adhesive usage.