Kinematic analysis of structural mechanics based on convolutional neural network

Leye Zhang,Xiangxiang Tian,Hongjun Zhang

2024-05-05

Abstract:Attempt to use convolutional neural network to achieve kinematic analysis of plane bar structure. Through 3dsMax animation software and OpenCV module, self-build image dataset of geometrically stable system and geometrically unstable system. we construct and train convolutional neural network model based on the TensorFlow and Keras deep learning platform framework. The model achieves 100% accuracy on the training set, validation set, and test set. The accuracy on the additional test set is 93.7%, indicating that convolutional neural network can learn and master the relevant knowledge of kinematic analysis of structural mechanics. In the future, the generalization ability of the model can be improved through the diversity of dataset, which has the potential to surpass human experts for complex structures. Convolutional neural network has certain practical value in the field of kinematic analysis of structural mechanics. Using visualization technology, we reveal how convolutional neural network learns and recognizes structural features. Using pre-trained VGG16 model for feature extraction and fine-tuning, we found that the generalization ability is inferior to the self-built model.

Machine Learning,Artificial Intelligence,Computer Vision and Pattern Recognition

What problem does this paper attempt to address?

This paper aims to use Convolutional Neural Networks (CNN) for motion analysis in structural mechanics, particularly for stability analysis of planar beam structures. The researchers created image datasets of geometrically stable and unstable systems using 3dsMax animation software and the OpenCV module. They built and trained a CNN model on the TensorFlow and Keras platforms. The model achieved 100% accuracy on the training, validation, and test sets, with an additional test set accuracy of 93.7%, demonstrating the ability of the CNN to learn and understand relevant knowledge in structural mechanics motion analysis. The paper suggests that future improvements in model generalization could be achieved by increasing the diversity of the dataset, potentially surpassing human experts' performance in handling complex structures. Additionally, the paper reveals how CNN learns and recognizes structural features through visualization techniques and compares the pretrained VGG16 model, finding that the self-built model has better generalization ability for specific tasks. Experimental results show that although the model's accuracy on the new test set did not reach the level of human experts, the 93.7% accuracy demonstrates good generalization ability. The paper emphasizes the importance of expanding the size and diversity of the dataset to improve model generalization, expecting CNN to have practical applications in the field of structural mechanics motion analysis.

Kinematic analysis of structural mechanics based on convolutional neural network

Dynamic displacement estimation of structures using one-dimensional convolutional neural network

DeepISMNet: three-dimensional implicit structural modeling with convolutional neural network

A Multi-channel Input Framework for Structure Displacement Response Prediction Using Convolutional Neural Network

Automated Debonding Characterization in Reinforced Structures Based on Finite Element Analysis and Convolutional Neural Networks

Dynamic Analysis of Nonlinear Civil Engineering Structures using Artificial Neural Network with Adaptive Training

Probabilistic Seismic Response Prediction of Three-Dimensional Structures Based on Bayesian Convolutional Neural Network

GNN-LSTM-based fusion model for structural dynamic responses prediction

Vibration-based structural damage detection via phase-based motion estimation using convolutional neural networks

Civil Engineering Structural Damage Identification by Integrating Benchmark Numerical Model and PCNN Network

Knee Cartilage Defect Assessment by Graph Representation and Surface Convolution

Seismic response prediction method for building structures using convolutional neural network

Novel Physics-Informed Neural Network Approach for Dynamic and Static Displacement Reconstruction Via Strain and Acceleration

A Hierarchical Deep Convolutional Neural Network and Gated Recurrent Unit Framework for Structural Damage Detection

Rapid Structural Safety Assessment Using a Deep Neural Network

CKNet: A Convolutional Neural Network Based on Koopman Operator for Modeling Latent Dynamics from Pixels

Torsional capacity evaluation of RC beams using an improved bird swarm algorithm optimised 2D convolutional neural network

Convolutional neural network-based recognition method for volleyball movements

Accurate Recognition and Simulation of 3D Visual Image of Aerobics Movement

Background Information of Deep Learning for Structural Engineering

Application of Graph Convolutional Neural Networks Combined with Single-Model Decision-Making Fusion Neural Networks in Structural Damage Recognition