Using Correlation in MFM Model Design and Validation

Industrial process signals do affect one other. However, causal dependencies, or correlation, and thereby consequences of control actions cannot always be understood by the operator, which may lead to accidents. Models explain dependencies and are mostly built on process knowledge and blueprints, while causalities are visible in the measurements and ought to be used to increase process understanding. In this paper we will discuss why the cross-correlation technique is insufficient for causality detection in industrial processes and we will present a new method that, for instance, can be used to validate the design of multilevel flow models.