From automated to autonomous process operations

Abstract

This paper considers current trends towards a higher degree of automation of process operations. Often referred to as “autonomous” process operations, these developments involve cyber-physical systems that can automate tasks that have hitherto relied extensively on human plant operators and, in particular, on their accurate assessment of the current plant situation based on a multitude of information sources, and on their ability to devise and implement plans of actions for dealing with often novel situations. The paper analyses the main drivers behind the need for a higher level of automation in process operations, and reviews the industrial applications that have been described in the public domain to date. It also presents a review of advances and potential impact of some of the enabling technologies for autonomy; these include sensors, mathematical modelling abstractions, reinforcement learning, knowledge graphs, and large language models.