Search

Recent trends in digitization and automation of information systems have led to the Industry 4.0 revolution in manufacturing systems. With the emergence of integrated "smart" systems that communicate through the cloud, collecting and manipulating the system data became a key yet, challenging component for developing optimal control strategies for these complex systems. In this work, we propose a strategy to address this problem with the case study on an air separation unit (ASU). Our approach involves developing an ASU's controllers via high‐fidelity modeling, studies in data‐driven reduced‐order models, and providing implementable control policies for the high‐fidelity model. Connecting the high‐fidelity model to a smart manufacturing platform allows integration into other smart manufacturing tools and applications. Since the high‐fidelity model is computationally challenging for online optimization tasks, such as model predictive control, surrogate models are generated that represent the high‐fidelity model's behavior. The derived reduced‐order models are then embedded into a model predictive control formulation for the optimal control of the whole process through multiparametric programming. A multiparametric approach based on solving a small portion of the multiparametric program is proposed to reduce the computational overhead. We then close the loop by deploying the developed controllers on the high‐fidelity model for tuning with prospects of employing them on the real industrial plant. [ABSTRACT FROM AUTHOR]

As markets become more competitive and dynamic, manufacturing plants are undergoing transitions toward flexible, agile, and low costs operations. Appropriate coordination within the supply chain is an important factor in manufacturing systems' performance. In this study, the impact of preemptive control action in advance of an upcoming demand change on the economic performance of a cryogenic air separation unit is investigated. The effects of various factors are explored through optimization formulations utilizing a high-fidelity collocation based dynamic process model. This includes the amount of lead time, choice of manipulated inputs, direction of demand change, and liquid product market conditions. Plant performance is evaluated and analyzed through a comprehensive multipart case study. © 2017 American Institute of Chemical Engineers AIChE J, 63: 3845-3859, 2017 [ABSTRACT FROM AUTHOR]

Reformer boxes are complex, integrated, and high-temperature units, subject to various failures during continuous operations for extended time periods. Challenges in the development of high-fidelity first principle models, despite easy availability of process measurements motivated the development of data-driven, automated fault detection (FD) systems. Paucity of plant-wide implementation of FD technologies in the chemical industry, accentuates the absence of relevant practical guidelines and best practices. In this paper, a trivially replicable FD system has been developed for large-scale industrial reformer boxes of hydrogen manufacturing units. Actual process data from plant historian has been used for training and validation of a novel model, developed using a combination of partial least squares regression and principal components analysis. Abnormalities based on several important measurements around the reformer were identified. Explicit algorithmic details and insights obtained during development of the expert system have been provided for ease of replication and adaptability. [ABSTRACT FROM AUTHOR]