Optimal Wireless Sensor Placement in Structural Health Monitoring Emphasizing Information Effectiveness and Network Performance

Abstract

A wireless sensor network (WSN) eliminates physical cables and reduces the difficulty of structural information collection in a structural health monitoring (SHM) system. This paper presents a methodology for placing wireless sensors in SHM systems to obtain the most useful information and the highest WSN performance. The problem of optimal wireless sensor placement (OWSP) first was formulated as a multiobjective optimization in which the information effectiveness is specified as the modal assurance criterion and the network performance is measured by the energy efficiency combined with the network connectivity. Then a multiobjective discrete firefly algorithm based on neighboring searching (MDFA/NS) is proposed to solve the rather difficult combinatorial optimization problem presented by the OWSP. The neighboring information instead of global knowledge is utilized to drive fireflies to move toward the Pareto front at high speed. Elite preservation was adopted to enhance algorithm robustness. Numerical experiments demonstrated that the optimal criterion maintains a trade-off between information effectiveness and network performance, and the MDFA/NS outperforms the popularly used nondominated sorting genetic algorithm II.