Gross Error Detection Strategies when Constraints are Bilinear

Abstract In chemical processes, accurate process data is necessary in order to achieve optimal operations. Thus it is important to identify and correct process measurement error. This paper evaluates two gross error detection techniques developed by (Rollins and Roelfs, 1992) for the case when constraints are bilinear. In a Monte Carlo simulation study, the size of the bias, the location of the bias, the sample size, and the level of significance were varied. Both techniques were shown to capable of highly accurate performance in identifying biased measurements and estimating process variables.