Akademik Veri Yönetim Sistemi
Teknik Dergi/Technical Journal of Turkish Chamber of Civil Engineers, cilt.13, sa.3, ss.2709-2726, 2002 (Scopus, TRDizin)
The first three natural frequencies of frames and beams are obtained using multi-layer neural network based back-propagation learning algorithm. The natural frequencies of beams are calculated for five different boundary conditions as the cantilever beam (C-F), Clamped end-Clamped end (C-C), Free end-Free end (F-F), Clamped end-Pinned end (C-P), and Pinned end-Free end (P-F) via direct solution of governing differential equations of beams and Rayleigh's approximate method. The training of the network has been made using these data for (C-C), (F-F), (P-F) and (C-P) and (C-F) support conditions. The trained neural networks, however, had been tested both end simply supported conditions (P-P) in flexural vibration of beams and a frames which were not included in the training set. The results found by using artificial neural network are sufficiently close to the theoretical and numerical solutions. Artificial neural network solutions are shown sufficient sensitive.