ABSTRACT
Optimal sensor placement (OSP) algorithms may require vibration measurements from all candidate sensor positions. A combination of experimental and numerical campaigns is proposed to obtain the full-field data. First, the response under typical excitation is measured with a redundant sensor network. Then, Bayesian virtual sensing is applied to estimate the full-field response that can be further used for OSP. Numerical simulations were performed for a structure subject to unknown random excitation. Noisy response was measured with an initial redundant sensor network, and the full-field response was estimated. Noise was added to the virtual sensors to mimic physical measurements. An OSP algorithm was applied using the virtual sensor data. The resulting optimal sensor network with a given number of sensors was compared to the initial sensor network with the same number of sensors.
