(17b) Towards a Discipline of Industrial Process Analytics: From Case-Based Problem Solving to a Systematic Body of Knowledge | AIChE

(17b) Towards a Discipline of Industrial Process Analytics: From Case-Based Problem Solving to a Systematic Body of Knowledge


Reis, M. - Presenter, University of Coimbra
The use of data for supporting inductive reasoning, operational management, and process improvement, has been a driver for progress in modern industry. Many success stories have been shared on the successful application of data-driven methods to address different open challenges, across different industrial sectors. The current and future importance of Industrial Process Analytics is therefore undeniable and is becoming widely considered as a competency that should be internalized and integrated into the company’s knowledge lake and skills set. But isolated success stories, even when abundant, are of limited value if the only way to make use of them is by replication on closely related scenarios. It is the mechanisms that have conducted to such successes, not the final solutions themselves, that really matters when it comes to creating sustainable value from Industrial Process Analytics. The recent advances on AI/ML technology in the fields of image& video analysis and natural language have generated many new methods and tools, spiking the interest of the research community to explore their application outside these domains, namely in the chemical, food, biotechnological, semiconductor, and pharmaceutical industries, among others. But this boost in activity has also increased the difficulty of understanding the multiple underlying rationales for applying them, other than the mere curiosity of “to see what comes out” (still valid, but arguably inefficient). Furthermore, it is often difficult to assess the added value of using these new methods, as many times they are not rigorously compared with conventional solutions presenting state-of-the-art performances. Therefore, it is now (the) time, I believe, to work on the formal structuring of a discipline of Industrial Process Analytics, and establish a solid methodological ground for research and teaching in this field. This is essential for undergraduate engineering students to learn basic inductive skills and to carry them to industry, where they become naturally part of the company’s knowledge lake. When we achieve that, then the transformative power of Industrial Process Analytics will be more than a set of isolated successes, but the result of implementing the principles and methodologies from a consolidated and structured discipline.