| 수록사항 |
한국공간구조학회지 , Vol. 25, No. 1 (통권 99호)(2025-03) |
| 주제어 |
; Quantum Neural Network (QNN); Variational Quantum Circuit (VQC); Truss System; Force Method; Entanglement; Quantum Gates; Surrogate Model |
| 요약2 |
Truss structures, widely used in engineering, consist of straight members transferring axial forces. Traditional analysis methods like FEM and the Force Method become computationally expensive for large-scale and nonlinear problems. Surrogate models using Artificial Neural Networks (ANNs), particularly Physics-Informed Neural Networks (PINNs), offer alternatives but require extensive training data and computational resources. Variational Quantum Algorithms (VQAs) address these challenges by leveraging quantum circuits for optimization with fewer parameters. Variational Quantum Circuits (VQCs) based on Quantum Neural Networks (QNNs) utilize quantum entanglement and superposition to approximate high-dimensional data efficiently, making them suitable for computationally intensive tasks like surrogate modeling in structural analysis. This study applies QNNs to truss analysis using 6-bar and 10-bar planar trusses, assessing their feasibility. Results indicate that residual-based loss functions enable QNNs to make reliable predictions, with increased layers improving accuracy and a higher Q-bit count contributing to performance, albeit marginally. |
