A PDE-based data reconciliation approach for systems with variations of parameters

Abstract

Data reconciliation is a mathematical approach that improves the quality of measurements by calculating the reconciled data that satisfy the process constraints. The conventional data reconciliation approach relies on the process model that contains the constant parameters. In the industrial applications, however, there are always possible variations of parameters within the system. In this paper, a new data reconciliation approach based on the partial differential equation (PDE) is developed. The proposed data reconciliation approach is experimentally applied to a case study of temperature measurements for a refinery process. The PDE-based model is employed in the formulation of the optimization problem. Unlike the conventional data reconciliation approach in which the system is assumed to be lumped, the PDE-based data reconciliation approach includes in the problem formulation the variations of parameters within the system in order to describe the real system’s behaviour. The reconciled values can be computed within the computational domain so they can be used as the data for troubleshooting, equipment analysis and maintenance.