Process Systems Engineering
Process systems engineering (PSE) is the application of systems engineering to biophysicochemical processes, where systems engineering is the methodical approach for the design of systems via mathematical modelling, data analytics, optimization, and control, and asystem is a set of units that interact. Numerous journal papers have been published in the last decade in which authors provide various perspectives on process systems engineering.16 Some papers argue that the discipline of process systems engineering was founded in the 1950s, whereas others argue for the 1960s or even later. The process systems engineering field actually can be traced back to the 1930s to early 1940s, when the use of mathematical modeling to design controlled process systems was already practiced in the chemical industry.17 18 1920 21 Process systems engineering was an active research area by the 1950s, with pioneering work by Rutherford Aris, Neal Amundson, and Roger Sargent in various types of chemical reactors and separations.2223 24 2526 27 The use of numerical computing in process systems engineering, which had already been applied in industry in the 1950s, became widely studied by the 1960s, with additional research teams including that of Ernst Gilles, Lowell Koppel, Leon Lapidus, Dale Rudd, W. Harmon Ray, Larry Evans, Reuel Shinnar, and Art Westerberg.28 2930 31 3233 34 3536 37 3839 40 4142
The 1970s saw the development of many research groups in process systems engineering, including by G.V. Reklaitis, George Stephanopoulos, Iori Hashimoto, John Perkins, Thomas Edgar, Ignacio Grossmann, Dale Seborg, Michael Doherty, Manfred Morari, Charles Cutler (at Shell), and Jeffrey Siirola (at Eastman Chemical).43 44 4546 47 4849 50 Much of the research was in developing PSE solutions for specific classes of processes. Other research considered processes more generally, and precise mathematical formulations were derived for much of the PSE technologies used in today’s industry, including for model predictive control, data reconciliation, and process scheduling. Methods were developed for better handling of practical considerations such as uncertainties and mixed continuous-discrete operations. At the same time, several research groups were making an impact outside of the field of chemical engineering, alongside their contributions within the field. As in earlier years, many individuals within the PSE community were widely known outside of PSE, including in reaction engineering and separations.
The 1980s to mid-1990s saw the founding of a large number of PSE research groups worldwide, with a large proportion focusing their attention only within the PSE community, with very little interaction outside of PSE, either within or outside of chemical engineering. A small proportion became highly engaged outside of PSE, in some cases becoming very well known outside of PSE, including in the areas of optimization and control theory and in applications such as to energy systems, pharmaceutical manufacturing, and biomedical devices. Individuals at such interfaces included James Rawlings, Lorenz Biegler, Nikolaos Sahinidis, Babatunde Ogunnaike, and Frank Doyle.
Data analytics, which was an active area of research within a subset of the PSE community and applied in industry for decades, has become of increased activity with the meteoric rise (and some rebranding) of machine learning. Recent years have seen PSE faculty increasingly collaborating with non-PSE faculty to apply PSE methodologies to their research problems, which has resulted in an increase in the number of publications by PSE faculty in highly scientific venues includingNature , Science , Cell , and PNAS . These collaborations include applications of machine learning, systems analysis, and process design, operations, and control. The “PSE community” is largely two distinct groups today, with one group being very inward-looking and focused on publishing papers in traditional PSE publication venues, and another group whose primary goal is to make an impact in other communities. The interactions between the two groups are rather limited, although some individuals have received some level of acceptance by both groups. The outward-looking community is likely to continue to grow, as automation, high-throughput experimentation, sensor technologies, and computing continue to favor methods that make best use of these trends.
With the growth of computing and data, process systems engineering continued to grow in methods, software, and applications to the point of becoming indispensable in the design and operation of modern chemical and biotechnological processes. Today software for computer-aided process systems engineering such as Aspen Plus, gPROMS, DMCplus, and DeltaV is ubiquitous in the process industries. While systematic methods for addressing such characteristics as time delays, nonlinearities, disturbances, and uncertainties have been developed since the early days of the process systems engineering discipline,51algorithms and software have become increasingly powerful since then, enabling increasingly complex systems and types of design problems to be addressed. The engineering designs are only as good as the mathematical formulation of the process systems engineering problem to be solved, and chemical engineers need to continue to be trained in the language of mathematics and in chemical and process systems engineering fundamentals to be able to address the increasingly complex chemical and biological systems that arise in today’s and tomorrow’s technological problems.