Surrogate model for condition assessment of structures using a dense sensor network

Abstract

Condition assessment of civil infrastructures is difficult due to technical and economic constraints associated with the scaling of sensing solutions. When scaled appropriately, a large sensor network will collect a vast amount of rich data that is difficult to directly link to the existing condition of the structure along with its remaining useful life. This paper presents a methodology to construct a surrogate model enabling diagnostic of structural components equipped with a dense sensor network collecting strain data. The surrogate model, developed as a matrix of discrete stiffness elements, is used to fuse spatial strain data into useful model parameters. Here, strain data is collected from a sensor network that consists of a novel sensing skin fabricated from large area electronics. The surrogate model is constructed by updating the stiffness matrix to minimize the difference between the model’s response and measured data, yielding a 2D map of stiffness reduction parameters. The proposed method is numerically validated on a plate equipped with 40 large area strain sensors. Results demonstrate the suitability of the proposed surrogate model for the condition assessment of structures using a dense sensor network.