A General Approach to Classify Operational Parameters and Rectify Measurement Errors for Complex Chemical Processes

Abstract

The reliability of the process measurements' data is extremely important for good monitoring, control and optimization of chemical process. On-line rectification of a measurement error is possible be it a random error or a gross bias, if additional information is available. Such information is supplied by the extent to which the material and energy balances are satisfied by the recorded data. These balances are simple, involve parameters usually well known, and they should be satisfied independently of the measurements accuracy. Previous works have attempted to resolve the above problem through the solution of a least squares problem, using linear balance equations. Swenker [1] was among the first who employed this idea, which was further developed by Hoffman [2] and Vaclaveck [3]. Almasy and his coworkers [4] outlined an iterative solution for the second order balance equations. They also applied the algorithm to linear dynamic system (Gertler, Almasy, [5]) and they