Exceptional events management applied to continuous pharmaceutical manufacturing

Intan Munirah Hamdan, Purdue University

Abstract

The advent of the process analytical technology (PAT) initiative has been accompanied by the desire to transition from batch to continuous processes in the pharmaceutical industry. The advantages associated with continuous processing are numerous; however, the complexity of monitoring also increases. Despite the design of sophisticated control systems to maintain the process at or near optimal conditions, exceptional events inevitably occur. Thus, we develop an exceptional events management (EEM) framework that deals with detection, diagnosis, and mitigation of exceptional events that control systems are unable to handle. The EEM framework detects and diagnoses using a combination of signed directed graph (SDG) and qualitative trend analysis (QTA). This framework is applied to the roller compaction process where exceptional events that are inherent to particulate processes are experimentally induced. The framework is found to successfully detect and diagnose these exceptional events. Additionally, fast Fourier transform (FFT) analysis is incorporated into EEM to enhance its diagnostic capabilities. The EEM framework is then modified for real-time and continuous applications. A moving window technique of monitoring is implemented and incipient diagnosis is performed by virtue of SDG and trend analysis. The modified EEM framework is demonstrated on a partially continuous dry granulation line consisting of two feeders, blender, and roller compactor and is shown to be capable of incipient fault diagnosis. In addition, the simultaneous occurrences of exceptional events are considered in this study, and a protocol is developed for multiple fault identification. The improved EEM framework is observed to detect, diagnose, and offer mitigation strategies within 10 s of event inception for the following cases: (1) simultaneous occurrences of exceptional events isolated to a piece of equipment, (2) simultaneous occurrences of exceptional events spanning multiple equipment, and (3) consecutive occurrences of events. Additionally, the EEM framework is capable of limiting the progression of exceptional events originating in an upstream piece of equipment, thus ensuring minimal to no propagation of exceptional events. A fault detection and diagnosis (FDD) framework is also developed using principal component analysis (PCA) and compared with the EEM framework. The diagnostic performance of both frameworks is determined to be comparable; however, the EEM framework is found to be more adaptable to different process conditions and exceptional events. The PAT initiative also advocates the implementation of systems engineering tools to improve status quo pharmaceutical operations. Thus, the scale up of wet granulation is studied using placebo and active formulations and evaluated using a dimensional analysis technique developed by Faure et al. [1]. Experiments have shown that placebo and active formulations require different scale up models. Thus, work is suggested that could elucidate complexities of the wet granulation process with the intent of determining a universal wet granulation scale up model.

Degree

Ph.D.

Advisors

Venkatasubramanian, Purdue University.

Subject Area

Chemical engineering

COinS