Steady data reconciliation and gross error detection based on the assumption of bounded error distribution

Abstract

Reliable process data are the key to the efficient operation of chemical plants. As a result of random and possibly gross errors, these measurements do not generally satisfy the process constraints. Thus data reconciliation and gross error detection are needed before the measurements can be used successfully. Almost all existing rectification methods are developed on the hypothesis that the measurement errors are normally distributed with zero mean and a known covariance matrix. However the errors are bounded distinctly in nature, whereas normal distribution is unbounded in both sides. A new method for simultaneous steady data reconciliation and gross error detection is presented assuming that the errors are subject to the bounded contaminated normal distribution. The effectiveness of the method is demonstrated on an atmospheric distillation tower.