Selection of optimal sensor position in a tubular reactor using robust degree of observability criteria

Abstract

Robust selection criteria for the optimal location for in-process concentration or temperature sensors along the length of a tubular reactor for the partial oxidation of Benzene to Maleic Anhydride are developed. A model of the reactor is constructed by rewriting the Pde's describing the mass and heat balances into a set of Ode's through the method of lines on a grid defined over the reactor length. The linearized model is described as a continuous, time-invariant state-space model where the state is formed by temperature and concentration profiles on the grid points. The best sensor location for the reactor is found by specifying scalar measures on the observability Gramian integral from the linear least-squares state estimation problem. New robust criteria for a degree of observability are specified. The scores on these criteria are determined by the amount of signal received by a sensor for a specific system configuration. These new selection criteria are compared with known measures for degree of observability for the optimal sensor location problem from the literature.