Q&A, Cristina Thomas, Spring Meeting Program Chair

Since its introduction in 1924 by Walter A. Shewhart, Statistical Process Monitoring (SPM) has evolved and adapted to the many technological transformations and several industrial revolutions, while maintaining the core principles essentially unchanged. Currently, there is a rich portfolio of SPM approaches that cover different application scenarios and can detect a wide variety of faults (on sensors, actuators, or in the process). Still, when moving to the second stage of monitoring, i.e., to fault diagnosis, the performance of the available methods is not so dependable, often leading to ambiguous or imprecise conclusions. Furthermore, fault Prognosis is attracting increasing interest in industry, but research is still in the early stages.

In this tutorial session, an overview is provided of several evolution trends that led to different monitoring solutions available today. Focus is given to solutions that are robust and mature, rather than recent proposals yet to be fully scrutinized and tested. Some common misconceptions in the use of SPM are also discussed and clarified, and good practices are shared. The complementary and always important topic of diagnosis is also discussed as well as the increasing interest in fault prognosis.