Author(s)

Wang Xiruo & Rong Gang

Abstract

Measured process data are inherently inaccurate and violate process constraints because of their underlying stochastic properties and possible gross errors caused by process disturbances, malfunctioning or miscalibrated instrumentation and even process leaks, etc. Therefore, the theory of gross error detection (GED) and data reconciliation has been developed to solve the contradictions between the measurements and their constraints. A number of gross error detection techniques are widely applied in different chemical processes. However, their performances have not always been so satisfactory: if the probability of type I error is low then that of t

Keywords