(149a) Using Production and Image Data for Real-Time Improvement of Process Operations

DCS and PLC systems collect large amounts of data routinely every few seconds. More complex PAT and image based sensors are also in more common use. Knowing what these data are telling you about your process so that the process operations can be improved, and so that operators or control systems can respond in real-time is critical to improved plant performance. This presentation looks at recent advances and applications in the industrial use of on-line multivariate data analysis systems for process improvement, monitoring and control.  These include:

(1)   Real-time monitoring of batch specialty chemical processes.

(2)   Predictive control of final product quality in batch processes using multivariate data based models and mid-course corrections.

(3)   Real-time inspection of multiple defects on synthetic rubber bales via advanced multivariate image analysis and relating the defect severity results to upstream processing conditions to significantly improve the process.