Visual design intuition: predicting dynamic properties of beams from raw cross-section images

Abstract

In this work we aim to mimic the human ability to acquire the intuition to estimate the performance of a design from visual inspection and experience alone. We study the ability of convolutional neural networks to predict static and dynamic properties of cantilever beams directly from their raw cross-section images. Using pixels as the only input, the resulting models learn to predict beam properties such as volume maximum deflection and eigenfrequencies with 4.54% and 1.43% mean average percentage error, respectively, compared with the finite-element analysis (FEA) approach. Training these models does not require prior knowledge of theory or relevant geometric properties, but rather relies solely on simulated or empirical data, thereby making predictions based on ‘experience’ as opposed to theoretical knowledge. Since this approach is over 1000 times faster than FEA, it can be adopted to create surrogate models that could speed up the preliminary optimization studies where numerous consecutive evaluations of similar geometries are required. We suggest that this modelling approach would aid in addressing challenging optimization problems involving complex structures and physical phenomena for which theoretical models are unavailable.

In-Situ Classification of Highly Deformed Corrugated Board Using Convolution Neural Networks

“…Sensors 2024, 24, x FOR PEER REVIEW 3 of 16 illumination and pose conditions [26], using a pretrained ResNet-50 network architecture. Wyder and Lipson identified the static and dynamic properties of cantilever beams using the CNNs, basing their classification on raw cross-section images [27]. [30].…”

“…Caputo et al applied support vector machines to classify materials from images [ 25 ] also acquired under various illumination and pose conditions [ 26 ], using a pretrained ResNet-50 network architecture. Wyder and Lipson identified the static and dynamic properties of cantilever beams using the CNNs, basing their classification on raw cross-section images [ 27 ]. Li et al explored different deep learning techniques to analyze the geometric features of self-piercing riveting cross-section, with SOLOv2 and U-Net architectures yielding the best results [ 28 ].…”

Deciphering Double-Walled Corrugated Board Geometry Using Image Analysis and Genetic Algorithms

“…Iqbal Hussain et al used a convolutional neural network, namely the ResNet-50 architecture, to identify and categorize woven materials [ 17 ]. Wyder and Lipson investigated the use of convolutional neural networks to identify the static and dynamic characteristics of cantilever beams using their unprocessed cross-section pictures [ 18 ]. Li et al used a range of deep learning methods to examine the geometric characteristics of a self-piercing riveting cross-section [ 19 ].…”

Identification of Geometric Features of the Corrugated Board Using Images and Genetic Algorithm

“…Iqbal Hussain et al applied the convolutional neural network, based on a pretrained network architecture ResNet-50, for the recognition and classification of woven fabrics [17]. Wyder and Lipson examined convolutional neural networks for the identification of the static and dynamic properties of cantilever beams based on their raw cross-section images [18]. Li et al applied various deep learning techniques for analyzing the geometric features of a self-piercing riveting cross-section [19].…”