Optimal Sensor Placement for a Truss Structure Using Particle Swarm Optimisation Algorithm

Abstract

The problem of placing sensors plays a significant role in the domains of structural health monitoring (SHM) applications and parameter estimation in structural dynamics. In this paper, the particle swarm optimisation (PSO) algorithm is introduced firstly and utilised to place sensors optimally on a truss structure for the purpose of modal identification. Then, two different types of fitness functions are constructed as to be the optimal criteria, which are based on modal assurance criterion (MAC) and maximising measures of the observability gramian matrix, respectively. The former one is mainly used to distinguish between two sets of mode shapes with the aim of maximising the off-diagonal elements in MAC matrix, and the latter one ensures proper observability of the structure. Finally, a truss structure model is considered as an example to demonstrate the efficiency and validity of the proposed method. Numerical results show that the two approaches using the PSO algorithm can find the optimal location of sensors successfully and identify the modal frequencies of the truss structure accurately by acceleration FRF method.