Speaker
Dr
Brad Schofield
(CERN)
Description
Developing and deploying advanced controllers is complex and time consuming. Often, it is not possible or desirable to perform lengthy tests on the real process before deploying. In this presentation an approach based on a dynamic process simulator is introduced, which has recently enabled the very smooth deployment of a Model Predictive Controller for HVAC systems at CERN. The use of the model as an online ‘Digital Twin’ for monitoring and further tuning after deployment is also discussed.
