A MINLP formulation to optimize sensor allocation using reconciled data in systems with fewer measurements

Abstract

The positioning of new mass flow sensors in an industrial plant is a challenge problem, especially in water networks where measurements are rarely found. The purpose of this paper is to study the application of a methodology to obtain the optimal location for a new sensor set in water networks. To perform such task, the results of a data reconciliation using the concept of Quality of Information, QI, is used. First, the mapping of the network topology and flows was performed to feed a data reconciliation with QI. Second, with the reconciled data, some runs in objective function, with some variation in optimization parameters, were carried out to obtain the optimal positioning of the new sensor set. The optimization was performed respecting all the mass and economic constraints of the measurement system, maximizing the objective function proposed. The proposed method was applied to a real industrial unit: Fafen-BA, a Brazilian nitrogen fertilizer plant. The results of the methodology application show that it is appropriated to choose optimal location of new sensors in an industrial plant.