Fail-safe sensor distributions for impact detection in composite materials

Abstract

This paper studies the problem of optimal sensor placement for impact detection and location in composite materials. The study involves a simple impact experiment on a composite box panel. The time-varying strain data are measured using piezoceramic sensors. An effective impact detection procedure is established using a neural network approach. The procedure determines the location and amplitude of impacts. A genetic algorithm is used to determine the optimum sensor positions for a diagnostic system. The main object of the paper is to study fail-safe distributions, i.e. those whose sub-distributions also have a low probability of detection error. The results are validated against an exhaustive search. The study shows that genetic algorithms combined with neural networks can be effectively used to find near-optimal sensor distributions for damage detection. The methods presented are generic and can be used in similar sensor position problems.

Impact Location and Quantification on an Aluminum Sandwich Panel Using Principal Component Analysis and Linear Approximation with Maximum Entropy

“…Comparison between Equations (6) and (7b) clearly reveals the analogy between linear approximation and the constraints to which the maximum entropy is subjected. Thus, on replacing the possibility p i by the weighting function w i , the function g r (x i ) by the feature vector X i , and the known expected value < g r (x) > by the test vector X, we have a method to compute the weighting functions that are needed to completely determine the impact estimateŶ in Equation (5).…”

“…Within data-driven algorithms, methodologies based on classification, pattern recognition and regression have been proposed, with artificial neural networks (ANNs) being most frequently used. Worden and Staszewski [4] and Staszewski et al [5] used two feed-forward multi-layer perceptron (MLP) networks to identify impacts on a composite plate. The first network was trained to detect the impact location, whereas the second network quantifies the impact magnitude.…”

Structural Health Monitoring and Impact Detection Using Neural Networks for Damage Characterization

“…11 Neural networks transform measured inputs into expected outputs using a set of weights and are computationally efficient after training. 12 The outputs are a linear combination of these inputs and the weights.…”

Identification of impact force for smart composite stiffened panels

“…Artifitial Neural Networks (ANNs) have also been used to predict the peak impact force using different features of the sensor signal (Jones et al 1997;Staszewski et al 2000;Worden et al 2000;Haywood et al 2005). ANNs are mathematical models that can be trained to model complex nonlinear relationships between the inputs and outputs.…”