Using artificial neural networks and electromagnetic capacitive NDT sensors for wood engineering application

Wood is a sustainable and renewable natural material with a wide variation in its mechanical and physical characteristics. Determining these properties can be challenging due to the complexity involved. Standard methods for measuring wood strength and characteristics are often destructive, expensive, and time-consuming. Therefore, developing a novel nondestructive method for wood property evaluation is crucial. This paper presents the development of an electromagnetic capacitive nondestructive sensor (ECNDTS) to evaluate wood’s physical and mechanical properties and detect defects based on its dielectric characteristics. Wood specimens of 10 mm thickness, representing 20 different wood species, were prepared and tested. Various physical and mechanical properties were investigated, including density, elasticity modulus, compressive and bending strengths, moisture content, and dielectric values. Artificial neural network (ANN) models, including the improved cascade neural network (CFNN) and the feedforward neural network (FFNN), were employed to enhance prediction accuracy. The results demonstrate that the developed ECNDTS can effectively predict wood properties based on measured permittivity values. The models for predicting moisture content, compressive strength, and flexural strength achieved a high coefficient of determination (greater than 90). The measured permittivity value can also be used to detect wood defects and estimate their size. This study’s findings have the potential to improve wood strength grading methods by enabling automation and reducing human error associated with current manual methods or visual grading standards. The research suggests that combining CFNN and FFNN models with ECNDTS can significantly improve wood strength prediction (accuracy >98.5) and enhance preliminary strength grading through artificial intelligence. © 2025 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group.