Process analyzer location and performance assessment for optimal process monitoring

The influence of process analyzer location and performance on plant-wide pr ocess monitoring was evaluated using five uncertainty contributions to the estimation error: measurement error/uncertainty, analysis frequency, sample size/grab error, analyzer memory effect/response correlation, and delay ti me. The choice of both the location and performance characteristics of diff erent process analyzers can be evaluated using a measurability factor M, ra nging from zero to one, where one indicates perfect monitoring capabilities . Due to the unifying nature of the measurability factor, this factor can b e used to make a rational decision between very different process analyzers . This allows for finding optimal process analyzer configurations for exist ing processes or for processes in the design phase. A tubular reactor simul ation model was used for the partial oxidation of benzene to maleic anhydri de to demonstrate the use of the measurability factor.