Exceptional Events Management Applied to Continuous Pharmaceutical Manufacturing: A Strategy to Detect, Diagnose and Mitigate Multiple Exceptional Events

Conventional regulatory control systems enforce process set points in individual pieces of equipment via implementation of PID and MPC controllers. Supervisory control systems, in contrast, encompass a network of these regulatory control systems, thus resulting in higher levels of process intelligence. The main purpose of the supervisory control layer is to reduce deviations of the entire process from the desired operating conditions. Though the supervisory layer is also able to detect abnormal process behavior, it is unable to diagnose and alleviate the problem. Thus, despite the design and implementation of sophisticated control systems, process anomalies persist. We designate these anomalies that control systems cannot handle as exceptional events, and propose a framework that can handle these occurrences.

The exceptional events management (EEM) framework is designed to detect, diagnose, and mitigate exceptional events as they manifest in process variable measurements. The detection and diagnosis is first performed using signed directed graph, followed by a higher resolution method to pinpoint the exceptional event using trend analysis; the framework is complemented with fast Fourier transform analyses to deal with events that manifest as periodic oscillations in measurements. Once the exceptional event has been determined, a mitigation strategy is either advised to the operator or automatically performed. This framework is also designed to run in conjunction with existing control layers; in essence, the EEM framework provides an added layer of security beyond the control systems to ensure that the plant/process remains in normal operating conditions.

Earlier versions of EEM were introduced and demonstrated on one piece of equipment – the roller compactor [1,2]. The most recent version of EEM incorporates the ability to handle multiple occurrences of abnormality, including simultaneously occurring events [3]. In this talk, we once again explore the EEM functionality by applying it to a continuous process, and compare it to a PCA-driven method to fault detection and diagnosis.

References:

• 1. Hamdan, I.M., Reklaitis, G.V., Venkatasubramanian, V. Exceptional Events Management for Continuous Pharmaceutical Manufacturing: Application and Integration into POPE and DeltaV, AIChE Annual Meeting. 2009
• 2. Hamdan, I.M., Reklaitis, G.V., Venkatasubramanian, V. Exceptional events management applied to roller compaction of pharmaceutical powders. Journal of Pharmaceutical Innovations. 2010; 5:147-160
• 3. Hamdan, I.M., Reklaitis, G.V., Venkatasubramanian, V. Exceptional Events Management for Continuous Pharmaceutical Manufacturing: Feeder, Blender, & Roller Compactor in Series, AIChE Annual Meeting. 2010