Improve regression-based models for prediction of internal-bond strength of particleboard using Buckingham’s pi-theorem

Abstract

Internal bond (IB) strength is one of the most important mechanical properties that indicate particleboard quality. The aim of this study was to find a simple regression model that considers the most important parameters that can influence on IB strength. In this study, IB strength was predicted by three kinds of equations (linear, quadratic, and exponential) that were based on the percentage of adhesive (8%, 9.5%, and 11%), particle size (+5, −5 +8, −8 12, and −12 mesh), and density (0.65, 0.7, and 0.75 g/cm3). Our analysis of the results (using SHAZAM 9 software) showed that the exponential function best fitted the experimental data and predicted the IB strength with 18% error. In order decrease the error percentage, the Buckingham Pi theorem was used to build regression models for predicting IB strength based on particle size, density, percentage of adhesive, face-screw withdrawal resistance (SWRf), and edge-screw withdrawal resistance (SWRe). From there, three dimensionless groups were created by Buckingham’s Pi theorem and IB strength was predicted based on multiple regression models. The results showed these models can predict IB strength with 10.68% and 18.17% error, based on face-screw withdrawal resistance and edge-screw withdrawal resistance, respectively.