Adaptive neural network feedback control for uncertain fractional-order building structure vibration systems

• Tipo: Artigo em Revista Científica Internacional
• Data: 2024

Autores

Xu, K (Autor) Outra; Chen, LP (Autor) Outra; Gu, PP (Autor) Outra; António Mendes Lopes (Autor) FEUP; Wang, MW (Autor) Outra; Du, WX (Autor) Outra

Revista

A Revista está pendente de validação pelos Serviços Administrativos.

Título: ALEXANDRIA ENGINEERING JOURNAL

Vol. 104

Páginas: 627-635

ISSN: 1110-0168

Indexação

ISI Web of Knowledge - 0 Citações

Scopus - 0 Citações

Outras Informações

ID Authenticus: P-016-RP5

DOI: 10.1016/j.aej.2024.08.018

Abstract (EN): In previous studies, structure vibration control mainly focused on integer-order systems or commensurate fractional-order (FO) dynamics systems. In this paper, we propose an adaptive radial basis function (RBF) neural network feedback (ARBF-FK) controller for FO building structure vibration systems with viscoelastic (VE) dampers, uncertain structure parameters and unknown seismic waves. Firstly, use a FO multi-order state space description for FO building structural vibration systems. Then, the design of the ideal feedback controller is based on the stability theory of FO multi-order systems. Moreover, to reduce the cost and facilitate the practical implementation of the control, unknown earthquake seismic waves is approximated by a RBF neural network, the ARBF-FK controller is proposed. In addition, to guarantee the stability of the closed-loop control system and avoid falling into local optimum, network weights are adapted by the FO Lyapunov stability theory instead of gradient descent algorithms. Finally, the convergence rate of the system is analyzed and perform various tests on the ARBF-FK controller. The simulation results demonstrate that the ARBF-FK controller has superior performance and is very robust to control FO building structure vibration systems with uncertain structure parameters and unknown external earthquake excitation.

Idioma: Inglês

Tipo (Avaliação Docente): Científica

Nº de páginas: 9