Placing sensors optimally in base-isolated steel structures using a multi-objective optimization approach

Subject Areas : structural design

Mehdi Firoozbakht ¹ , Hamidreza Vosoughifar ² , Alireza Ghari Ghoran ³

1 - Department of Civil Engineering, Islamic Azad University, Khorasgan branch, Isfahan, Iran
2 - Department of Civil and Environment, Hawaii University at Mano, Hawaii, USA
3 - Department of Civil and Transportation Engineering, Isfahan University, Isfahan, Iran

Received: 2023-06-14 Accepted : 2023-06-22 Published : 2023-02-01

Keywords: Structural Health Monitoring (SHM), Optimal Sensor Placement (OSP), Base Isolated (BI) structures, Nonlinear Time-History Analysis (NTH),

Abstract :

Adequate selection of sensors placement plays a key role in structural health monitoring (SHM) in base isolated (BI) structures. This critical issue is usually done by past experience and knowledge on the force and vibration situations of a structure. During recent decades, techniques have received increasing attention as a tool for determining an arrangement of sensors suitable for SHM. In this paper, a multi- objective numerical method for optimal sensor placement (OSP) in BI structures based on the combination of traditional OSP algorithms and nonlinear time-history analysis (NTH) has been proposed. Next, genetic algorithm (GA) was employed to determine the location of sensors on the structure based on the structural dynamic response of the BI system. To show the efficiency of the proposed method, a BI building was modeled using finite element method (FEM) in which NTH were undertaken using the seismic scaled records of near-fault earthquakes (NF). The novel numerical approach called transformed time history to frequency domain (TTFD) was evaluated to transform NTH results to frequency domain and then the effective frequencies according the maximum range of Fourier amplitude were selected. The modified type of modal assurance criteria (MAC) values can be achieved from MAC with the exact seismic displacement. Results show that the proposed method can provide the optimal sensor locations and remarkably reduce the number of required sensors and also improve their optimum location.