Soft sensors based on nonlinear steady-state data reconciliation in the process industry

In this article, recent practical experiences of soft sensor projects based on data reconciliation in the process industry are reported. The discussed items comprise the steady-state detection, the gross error detection, the initial value generation, and remarks on the implementation of this method. Two practical examples are presented from the area of process industry. The first is a representative soft sensor application to provide consistent measurements for data monitoring. The second is a soft sensor application which uses a rigorous model to estimate concentrations in a distillation column.