Influence of measurement and model errors in optimal sensor placement for SHM purposes

In this work, the influence of measurement and model errors in optimal sensor placement is investigated. The sensor placement procedure is based on the Information Entropy theory and the solution of the optimization problem is obtained maximizing the determinant of the so called Fisher Information Matrix (FIM). Results of placement process considerably depend on the so called covariance matrix of prediction error as well as on the definition of the correlation function. The paper thus investigates the role of the covariance matrix and the correlation function in optimal sensor placement. Different proposals on their definition are compared. A constant and an exponential correlation function depending on the distance between sensors are firstly assumed; then a new proposal depending from both the distance and modal vectors is presented. The method is finally applied to a benchmark case study and the effect of model and measurement error on results is described.