Geometric Process Control (GPC) Takes Over where Statistical Process Control (SPC) Leaves Off

Statistical Process Control (SPC) has been successful over many years in improving quality management, particularly in manufacturing applications. This approach separates process variables and quality achievement to fit univariate statistical methods and misleadingly tracks qualities individually. An example is a manufacturer of a complex assembly tracking 25 quality parameters individually with SPC Charts. Imagine the reaction when a few minutes’ work with GPC showed none of the over 100 assemblies made simultaneously conformed to all quality constraints.  SPC is not as common in the process industries where most quality results are related to many process variables. Attempts to extend the application into multi-variate statistical methods have not been as successful.

The answer to this need in industries with common multi-variable relationships has come from a different branch of mathematics. Geometric Process Control (GPC) is based on multi-dimensional geometry. GPC solves the challenges of relating product qualities to each other and to process variables with self-service analytics that requires good knowledge but only basic mathematics to discover cause-and-effect relationships.

GPC can also be used in real-time applications for advisory process control and early detection of equipment issues. GPC is visual due to its foundation in geometry, making it easily understood with a brief explanation by anyone in an organization.

In our webinar, we will demonstrate how the analytics tool is used to find better Operating Windows and continue to the modeller tool to generate a real-time machine-learning geometric model of an Operating Envelope where derived variable relationships are intriguingly represented in the shape of the geometric object rather than in equations. The modeller includes an Operator Display and OPC Client and can handle models of a single mode or phase as well as multi-mode (e.g. grade processes such as polymer production) and multi-phase models (as many batch processes).