Applications Of Variography To Process Control

Large-scale variability in process plants and process control can be identified and controlled through disaggregating the overall variability in a process stream. Conventional statistics and Statistical Process Control (SPC) fail to address the concept of stream heterogeneity, therefore fail to identify and distinguish between the various sources of variability in a process stream. A modified set of 445 data from a process plant is used as an example to explain the benefits of variography in process control. Large-scale variability, usually in the form of product composition cycles and chemical cyclic variations may be manifest throughout a sampling system or a process plant, but this type of variability is most easily discernable in one-dimensional process streams. The Moving Average is a powerful tool for identifying large and small scale variability and cyclical behaviour in process streams, while the variogram is a custom built tool that allows the components of variability to be identified and resolved. The positions in the variogram at which the different variance can be read and the nomenclature of these types of variability is described. The transfer of variances converted to standard deviations to the Control Charts, allows an operator to appreciate the appropriateness of Upper and Lower control and specification limits. The overall contribution of the various sources of variability is represented in a pie diagram.