Health monitoring of lightweight structures, like thin flexible plates, is of interest in several engineering fields. In this paper, a recursive Bayesian procedure is proposed to monitor the health of such structures through data collected by a network of optimally placed inertial sensors. As a main drawback of standard monitoring procedures is linked to the computational costs, two remedies are jointly considered: first, an order-reduction of the numerical model used to track the structural dynamics, enforced with proper orthogonal decomposition; and, second, an improved particle filter, which features an extended Kalman updating of each evolving particle before the resampling stage. The former remedy can reduce the number of effective degrees-of-freedom of the structural model to a few only (depending on the excitation), whereas the latter one allows to track the evolution of damage and to locate it thanks to an intricate formulation. To assess the effectiveness of the proposed procedure, the case of a plate subject to bending is investigated; it is shown that, when the procedure is appropriately fed by measurements, damage is efficiently and accurately estimated.
Structural health monitoring (SHM) allows the acquisition of information on the structural integrity of any mechanical system by processing data, measured through a set of sensors, in order to estimate relevant mechanical parameters and indicators of performance. Herein we present a method to perform the cost–benefit optimization of a sensor network by defining the density, type, and positioning of the sensors to be deployed. The effectiveness (benefit) of an SHM system may be quantified by means of information theory, namely through the expected Shannon information gain provided by the measured data, which allows the inherent uncertainties of the experimental process (i.e., those associated with the prediction error and the parameters to be estimated) to be accounted for. In order to evaluate the computationally expensive Monte Carlo estimator of the objective function, a framework comprising surrogate models (polynomial chaos expansion), model order reduction methods (principal component analysis), and stochastic optimization methods is introduced. Two optimization strategies are proposed: the maximization of the information provided by the measured data, given the technological, identifiability, and budgetary constraints; and the maximization of the information–cost ratio. The application of the framework to a large-scale structural problem, the Pirelli tower in Milan, is presented, and the two comprehensive optimization methods are compared.
Abstract. We present an optimal sensor placement methodology for structural health monitoring (SHM) purposes, relying on a Bayesian experimental design approach. The unknown structural properties, e.g. the residual strength and stiffness, are inferred from data collected through a network of sensors, whose architecture, i.e., type and position may largely affect the accuracy of the monitoring system. In tackling this issue, an optimal network configuration is herein sought by maximizing the expected information gain between prior and posterior probability distributions of the parameters to be estimated. Since the objective function linked to the network topology cannot be analytically computed, a numerical approximation is provided by means of a Monte Carlo analysis, wherein each realization is obtained via finite element modeling. Since the computational burden linked to this procedure often grows infeasible, a Polynomial Chaos Expansion (PCE) approach is adopted for accelerating the computation of the forward problem. The analysis expands over joint samples covering both structural state and design variables, i.e., sensor locations. Via increase of the number of deployed sensors in the network, the optimization procedure soon turns computationally costly due to the curse of dimensionality. To this end, a stochastic optimization method is adopted for accelerating the convergence of the optimization process and thereby the damage detection capability of the SHM system. The proposed method is applied to thin flexible structures, and the resulting optimal sensor configuration is shown. The effects of the number of training samples, the polynomial degree of the approximation expansion and the optimization settings are also discussed.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.