Robust data reconciliation and gross error detection in industrial heat exchanger networks

Industrial measurements are invalidated by the presence of random errors and gross errors. Data reconciliation allows to reduce these effects by solving a constrained optimization problem. Furthermore, it is possible detect and identify the gross errors through gross error detection methods. In this Thesis data reconciliation and gross error detection are exploited to provide reliable measurements for the estimation of fouling model parameters in crude-oil heat exchanger networks.