Flexural Performance of Cross-Laminated Timber Panels Using Evolutionary Artificial Neural Networks

Cross-laminated timber (CLT) is a panelized engineering wood product known for its strong and lightweight properties. Construction with CLT panels has been a growing trend to meet the low-carbon building alternatives. This study generates a reliable artificial neural networks (ANNs)-based model for estimating the flexural performance of CLT panels. Genetic algorithm (GA) with multilayer perceptron (MLP) was implemented on a dataset of CLT panels considering width, span length, thickness, bending, and shearing strength variables as input parameters to determine the flexural strength of the panels. 70% of the data were used for training and 30% for testing phases. The accuracy of GA-based MLP model was evaluated by comparing the results with multiple linear regression (MLR) and a variety of feed-forward (FF) models. The results revealed that the GA-optimized MLP model could estimate the flexural strength of CLT panels with the highest accuracy.