Machine Learning for Process Monitoring

TOPICAL COLLECTIONS

Guest Editors

Marcelo Ferreira de Souza Alves

Programa de Pós-Graduação em Engenharia de Processos Químicos e Bioquímicos, Escola de Química, Universidade Federal do Rio de Janeiro, Rio de Janeiro, RJ, 21941-909 Brazil

Marcelo Ferreira de Souza Alves received the B.Sc. degree in Chemical Engineering and the M.Sc. degree in Chemical and Biochemical Process Technology from Universidade Federal do Rio de Janeiro in 2020 and 2021 respectively. He currently is a D.Sc. candidate for the degree of Chemical and Biochemical Process Engineering at Universidade Federal do Rio de Janeiro. His main research topics are Machine Learning applications, especially regarding industrial processes monitoring, data reconciliation, and uses of ultrasound spectroscopy, with an emphasis on applications to dispersed systems and signal processing. Marcelo worries not only in performing high quality research, but also in communicating with both academia and society. Throughout these first few years of his academic career, Marcelo has integrated four distinct research projects, co-advised undergraduate students, attended national and international conferences, published several technical works and papers, taught minicourses and participated in events focused on publicizing scientific research to society.

Gildeir Lima Rabello

Programa de Pós-Graduação em Engenharia de Processos Químicos e Bioquímicos, Escola de Química, Universidade Federal do Rio de Janeiro, Rio de Janeiro, RJ, 21941-909 Brazil

Gildeir Lima Rabello holds a B.Sc. degree in Chemical Engineering from the Universidade Federal do Espírito Santo (UFES), completed between 2013 and 2018. He earned his M.Sc. degree in Chemical Engineering from the Postgraduate Program in Chemical Engineering at the Universidade Federal Rural do Rio de Janeiro (UFRRJ), with a specialization in Chemical Technology (2018–2021). Currently, he is pursuing a D.Sc. degree in Chemical and Biochemical Process Engineering at the Universidade Federal do Rio de Janeiro (EPQB - UFRJ). His research focuses on advanced topics such as data reconciliation, parameter estimation, machine learning, process monitoring, real-time monitoring, mathematical modeling, simulation, and optimization.

José Carlos Pinto

Programa de Pós-Graduação em Engenharia de Processos Químicos e Bioquímicos, Escola de Química, Universidade Federal do Rio de Janeiro, Rio de Janeiro, RJ, 21941-909 Brazil

José Carlos Pinto holds a B.Sc. degree in Chemical Engineering from Universidade Federal da Bahia (1985) and M.Sc. and D.Sc. degrees in Chemical Engineering from Universidade Federal do Rio de Janeiro (UFRJ) (1987, 1991). He is a full professor in COPPE's Chemical Engineering Program (UFRJ), a permanent member of Graduate Program in Chemical and Biochemical Process Engineering(UFRJ), and member of the editorial boards of the journals Polímeros: Ciência e Tecnologia (ABPol), Macromolecular Reaction Engineering (Wiley) and Processes (MDPI). He has been a full member of the Brazilian Academy of Sciences since 2010 and the National Academy of Engineering since 2014. He has experience in Chemical Engineering, with a general emphasis on chemical reactors and a particular focus on the modeling, simulation, and control of polymerization systems. He has published over 500 papers, with over 12,000 citations in WOS, Scopus, and Google Scholar databases, with an h-index of 54 (February 24, 2024).

Collection Overview

Industrial processes are conceptualized as a sequence of steps under specified conditions that can guarantee the manufacturing of products with properties of interest. Is this context, accurate process knowledge is crucial for guaranteeing product specification, production optimization, safety and enhanced environmental performance.

Measuring relevant variables is essential, but knowledge goes beyond. Process monitoring may be understood as the set of tasks and procedures adopted to assess the state of the process in order to guarantee its conformity to best and desired practices both at a present moment and that may also allow estimations about future states.

Machine Learning approaches, a branch of artificial intelligence, constitute emerging techniques of increasing interest among researchers by both their capability of mapping complex relationships among variables and by providing fast responses.

In this collection, the main interest is the use of machine learning procedures to assist process monitoring. Topics of interest include, but are not limited to:

- use of machine learning algorithms as process models

- fault detection, predictive maintenance diagnostics and / or remaining useful life prediction using machine learning

- development of digital twins using machine learning

- development of data-handling strategies, which may include preprocessing, feature selection and instance selection

- development of key performance indicators to help process monitoring