Optimal placement of sensors for structural health monitoring using improved genetic algorithms

Abstract

The global optimization of sensor locations for structural health monitoring systems is studied in this paper. First, the performance function based on damage detection is presented. Then, genetic algorithms (GAs) are adopted to search for the optimal locations of sensors. However, the simple GAs can result in infeasible solutions to the problem. Some improved strategies are presented in this paper, such as crossover based on identification code, mutation based on two gene bits, and improved convergence. The analytical results from the improved genetic algorithm are compared with the penalty function method and the forced mutation method. It is concluded that the convergence speed with the proposed improved genetic algorithm is faster than that with the penalty function method and the forced mutation method, and the result of placement optimization is better.

Sensor Placement Optimization Using Ensemble Kalman Filter and Genetic Algorithm

“…The next step is to evaluate the ensemble mean and the ensemble perturbation matrices as follows, (3) where,…”

“…H. Y. Guo et al proposed to use an improved version of the GA (improved crossover and mutation) to determine the optimal locations of sensors in a SHM system consisting of a 2D truss structure [3]. The SA is an optimization method, named after the physical process of annealing in thermodynamics and used for finding a global minimum or maximum for problems having multiple local minima and maxima [4,5].…”

MONSOON: A Coevolutionary Multiobjective Adaptation Framework for Dynamic Wireless Sensor Networks

“…Several research efforts have applied genetic algorithms to WSNs, for example, to cluster-based routing [14][15][16][17], data processing [18], localization [19] and node placement [20,21]. Every work uses a fitness function that combines multiple objective values as a weighted sum, and uses the function to rank agents/genes in elite selection.…”

“…As a result, BiS-NET/e requires much less configuration cost for application designers. Also, [14,15,17,[19][20][21] do not consider dynamics in the network, but assumes the network is static.…”

Co‐evolutionary algorithm for structural damage identification using minimal physical testing

“…Genetic algorithms have been successfully used for error minimization in various damage identification problems [2,[4][5][6][7][8] and have gained wide acceptance due to their inherent advantages such as convergence to global optima, noise tolerance, handling of multimodal problems (non-unique solutions), and gradient-free search.…”