Prognostic Model Selection for Predictive Maintenance and Integrated Production Scheduling Based on Reinforcement Learning in Dynamic Manufacturing Systems
The project is a cooperation between Prof. Dr.-Ing. Michael Freitag (BIBA Bremer Institut für Produktion und Logistik GmbH at the University of Bremen, Germany), Prof. Dr.-Ing. Carlos Eduardo Pereira (Federal University of Rio Grande do Sul, Brazil), Prof. Dr.-Ing. Enzo Morosini Frazzon (Federal University of Santa Catarina, Brazil), and Prof. Dr. Eng. Iracyanne Uhlmann (Institute of Exact Sciences and Technology of Federal University of Amazonas, Brazil) and is related to the Collaborative Research Initiative on Smart Connected Manufacturing (CRI-SCMfg), funded by DFG and CAPES. The project aims to combine a prognostic method selection approach for predictive maintenance and a reinforcement learning-based method for production and maintenance scheduling in dynamic manufacturing systems.
Period: 2025 – current
AHI – Artificial Intelligence Supporting Human Intelligence in Data-Driven Decision-Making
In today’s globalized world, complex processes are the backbone of efficient operations across numerous industries. These processes, however, are subject to dynamic and unpredictable fluctuations due to factors such as market volatility, geopolitical tensions, natural disasters, and changing consumer preferences. Consequently, there is a need for innovative data-driven decision-making approaches that allow for rapid adaptation of organizations and systems in response to these turbulent conditions. Nowadays the application of Artificial Intelligence (AI) for supporting decision-making in organizations and systems is essential for providing a better future, which is efficient, resilient and sustainable. The project aims to investigate the impact of Artificial Intelligence (AI) application in data-driven decision-making in organizations and systems, promoting the scientific exchange and mobility of researchers and students between the Brazilian and German research groups, assisting on the creation of an international research collaboration. The improvement of AI-assisted data-driven decision-making will be first explored on a theoretical level and in use cases. The project will then follow a research-based learning approach, in a way that our research and demonstration efforts will be connected to our educational activities so that, by means of workshops, seminars and lectures, a new generation of innovative thinkers and decision-makers will be trained for the proper application of AI in data-driven decision-making. The project will create fact-based knowledge in the area of organizations and systems on how AI-assisted data-driven decision-making can streamline organizations and systems considering the ongoing digitalization. We will focus on the development of innovative methods, conceptual propositions and best practices towards new application models, accommodating the increasing dynamics and complexity of systems. The dissemination of knowledge will occur by means of events (workshops, seminars and classes), publications in academic journals and the participation in academic and professional conferences. The project is a cooperation between Prof. Dr. Vitor Azevedo (RPTU Kaiserslautern-Landau), Prof. Dr. Christian Cordes (Uni Bremen), Dr. Matheus Eduardo Leusin (Uni Bremen) and Prof. Dr.-Ing. Enzo Morosini Frazzon (UFSC). The joint research and formation efforts of both research groups embody the initiation of a collaboration, including the exchange of researchers and professors.
Period: 2025 – current
Planning and Control of Manufacturing Supply Chains Based on Digital Twins
The digitalization of production and operations, enabled by the application of Industry 4.0 concepts, technologies, and methods, can enhance data-driven management of manufacturing supply chains, thereby improving their efficiency, resilience, and sustainability. Despite the increasing availability of data and the integration of virtual models into the control of physical processes (cyber-physical systems), addressing problems in uncertain environments that require flexibility still demands human participation. Thus, the effectiveness of implementing cyber-physical systems to support decision-making also depends on understanding the behavioral aspects of the actors involved in socio-cyber-physical systems, as well as the contextual variables that influence them. Within the scope of this project, we propose a vision for planning and controlling manufacturing supply chains using digital twins coupled with physical systems and considering behavioral aspects of the agents. Subsequently, an application model of digital twins for manufacturing supply chains will be structured, and use cases highlighting the impact of this application will be developed. The project will result in the development of empirical-normative knowledge aimed at conceptual propositions and oriented towards new application models, as well as in the improvement of practical knowledge directed to manufacturing supply chains, thereby contributing to the ongoing digitalization of production systems.
Period: 2024 – current
EngMind – Supporting the emergence of a new engineering mindset for the Brazilian digital transformation
The training of a new generation of production engineers will contribute to Brazilian industrial competitiveness. Their education depends on research-based learning practices, fostered by collaboration between industry and university. This proposal aims to support the emergence of a new engineering mindset – sustainable, global, and science-oriented – that will contribute to the transformation of Brazilian civil society, economy, and industry. First, an integrated and convergent vision will be proposed, based on a socio-cyber-physical systems perspective. Second, a method will be developed to develop adaptive cognitive digital twins for planning and controlling global manufacturing networks. On these foundations, and considering an exploratory analysis of the research-education relationship in Germany and Italy, a research-based learning practice will be structured, tested, and disseminated in Brazil, supporting the emergence of an engineering mindset willing to a digital future competitive, sustainable, and resilient.
Period: 2022 – current
ManDigital – Manufacturing networks planning and control in the era of digital production and operations
Economies of scale and scope are favored by distributed production systems. This type of productive arrangement imposes challenges from the number of actors and contexts involved. The digitalization of production and operations, enabled by the application of Industry 4.0 concepts, technologies and methods, can improve the structuring and monitoring, based on data, of manufacturing networks, which contributes to improve their efficiency, effectiveness and scalability. Despite the increasing availability of data and the integration of virtual models into the planning and control of physical processes (cyber-physical systems), dealing with problem solving in uncertain and flexible environments requires human participation. Therefore, the outcome of implementing cyber-physical systems to support decision making also depends on understanding the behavioral aspects of the participants involved and the context-related variables that influence them. In this context, through the implementation of this proposal, it will be possible to advance the frontier of scientific and applied knowledge about the planning and control of manufacturing networks, considered from the perspective of socio-physical systems. In the scope of this proposal, the planning and control of manufacturing networks will be improved through: (i) a new concept that considers a socio-physical perspective, (ii) a hybrid simulation-optimization-analysis method to support operational decisions, and (iii) an integrative data-driven approach. The proposed project will result in the development of scientific knowledge, focused on conceptual propositions and oriented to new hybrid methods of simulation-optimization-analysis; and practical knowledge, directed to applied approaches and empirical studies, in the scope of planning and control of manufacturing networks considering the increasing digitalization of production and operations.
Period: 2021 – 2024
MetaMaintain – A meta-learning approach to select appropriate prognostic methods for the predictive maintenance of digital manufacturing systems
The project aims to use data from cyber-physical systems to feed a meta-learning method that selects appropriate failure prediction models for predictive maintenance of advanced manufacturing systems. The international cooperation project is coordinated by Prof. Enzo Frazzon and involves 8 student-researchers and 6 professors in Brazil, as well as 4 student-researchers and professors in Germany.
Period: 2019 – 2023 | Resources: R$ 822,576.00 covering capital, funding, and scholarships financed by CAPES/PIPC. Process CAPES88887.207652/2018-00.
INCANTO/PrINT – Analysis of distributed and automated manufacturing lines for manufacturing customized medical treatment items
The project aims to analyze the operational, economic, and social feasibility of implementing an IoT platform for distributed and highly automated manufacturing lines aimed at manufacturing customized items for personalized medical treatment. This is a collaborative project with other institutions in Brazil and abroad. The technologies and methods developed in this project have great scientific and technological appeal and will place UFSC, in the medium and long term, in a scenario of data and prototype exchange with other groups interested in the project’s target theme. The project is technically coordinated by Prof. Enzo Frazzon (together with Prof. Bornia/UFSC), involves 6 student-researchers and 9 professors in Brazil, as well as researchers and professors in Italy, USA, Canada, Germany, and Portugal.
Period: 2018 – 2022 | Resources: R$ 1,439,300.00 covering capital, funds expenses, and Visiting Professor scholarships in Brazil, as well as sandwich doctorate scholarships and Visiting Professor Abroad, financed by CAPES PrINT/UFSC. Process 88887.310575/2018-00.
FASTEN manufacturing
FASTEN project aimed to develop, demonstrate, validate and disseminate an integrated and modular framework for efficient production of highly customized products. The technology demonstration covered two pilot cases (TRL 6): one specified by Thyssen Krupp (TSK) in Brazil and other by Embraer Portugal (EMBPT) in Portugal. The FASTEN project was coordinated by INESC Brasil and involved student-researchers and researchers in Brazil and abroad. The project includes capital resources, as well as funding and scholarships of different modalities.
Period: 2018 – 2021 Resources: R$ 6,144,718.00 covering capital, funding, and scholarships financed by NATIONAL NETWORK FOR EDUCATION AND RESEARCH – RNP – (BR-EU Coordinated).
AdaptiveSBO – An adaptive simulation-based optimization approach for the scheduling and control of dynamic manufacturing systems
Efficient management of manufacturing systems requires reliable methods for planning and controlling resource allocation, scheduling tasks, machines, and personnel, as well as predicting delivery times for each order. AdaptiveSBO project developed a new simulation-based and data-driven adaptive optimization method for scheduling and controlling dynamic manufacturing systems, with complex structures and subject to stochastic behavior. The proposed method uses available real-time data to characterize the state of equipment, production orders, and stocks in the manufacturing system, especially those from IoT sensors. It was applied in three partner companies, which were integrated with the decision-making systems responsible for scheduling and controlling the manufacturing systems. The international cooperation project was coordinated in Brazil by Prof. Enzo Frazzon, involved 13 student-researchers and 4 professors in Brazil, as well as 4 student-researchers and professors in Germany.
Period: 2016 – 2021 | Resources: R$ 1,264,631.00 covering capital, funding, and sandwich masters and doctoral scholarships abroad financed by CAPES and DFG (BRAGECRIM Program). Process CAPES 99999.006033/2015-06.
I2MS2C – Integrating Intelligent Maintenance Systems and Spare Parts Supply Chains
The objective of the I2MS2C project was to integrate the spare parts supply chain with intelligent maintenance systems (IMS), using a distributed approach to asset management and supply chain planning. To this end, intelligent production and logistics systems were proposed, fully considering the concepts, approaches, and technologies of Industry 4.0. The international cooperation project was coordinated in Brazil by Prof. Dr. Carlos Eduardo Pereira from UFRGS. At UFSC, this project was coordinated by Prof. Enzo Frazzon.
Period: 2012 – 2017 Resources: R$ 412,000.80 covering capital, funding, and scholarships financed by CAPES and DFG (Program BRAGECRIM).