Periodic Reporting for period 2 - COGNITWIN (COGNITIVE PLANTS THROUGH PROACTIVE SELF-LEARNING HYBRID DIGITAL TWINS)
Reporting period: 2021-03-01 to 2022-02-28
COGNITWIN will enhance the potential of the process industry in Europe by creating and validating a new approach for cognitive digital twins affordable for all process industries.
One aspect to the digitalisation vision is the "cognitive element", where the process plants can learn from pattern recognition in historical data and adapt to changes in the process, simultaneously being able to predict unwanted events in the operation before they happen
The COGNITWIN toolbox will be validated in the following SPIRE process industry sectors: Non-Ferrous (Aluminium and Silicon), Steel and Engineering.
COGNITWIN will demonstrate an approach for the design, development and operation of the European process industry by introducing a platform for virtual component-based architecture that integrates IoT, Big data, AI, smart sensors, machine learning and communication technologies, all connected to a novel paradigm of self-learning hybrid models with proactive cognitive capabilities.
The overall objectives includes the following:
• COGNITWIN for Industry Process Excellence: Show improved performance in cognitive production plants by a technology demonstration of fully digitalized pilots.
• Cognitive Digital Twins for Cognitive Retrofitting: Enabling an efficient and well-defined approach for “cognitive augmentation” of physical assets, processes and systems for Cognitive Digital Transformation in Process industry.
• Hybrid Twins for Optimised Process Performance by hybrid models that combines first principle and data-driven models and use machine learning, AI and the connected data bases to pro-active forecast and communication, as well as self-learning by recognition of patterns in the data.
• COGNITWIN Interoperability Toolbox as a Service: A reference architecture for the cognitive elements including of Big Data, Databases, IoT, Smart Sensors, Machine Learning, and AI technologies that realizes hybrid modelling, self-adaptivity and cognitive recognition, leveraging/extending the existing work into relevant communities.
• COGNITWIN for increasing European Technology Dominance: Ensure the dominance of the Europe in technologies related to cognitive plants, thereby influencing the further development of Big Data, Databases, IoT, Smart Sensors, Hybrid Modelling, Machine Learning and AI technologies in relevant communities, focusing on the capabilities of the developed technologies for creating new generations of self-adaptive and cognitive algorithms and models.
• COGNITWIN for SPIRE: Ensure the knowledge transfer of results and experiences from the COGNITWIN project to the SPIRE Process Industry community, focusing on active participation in the new SPIRE DG7 Digitalisation group and in SPIRE organized events.
• COGNITWIN for boosting European Industry: Provide competitive advantage to the European industry, esp. SMEs in the global market, through better exploitation of the synergies between Big Data, Databases, IoT, Smart Sensors, Hybrid Modelling, Machine Learning and AI technologies for an efficient resolution of complex process industrial challenges.
• Effective dissemination and ensuring transfer of knowledge and experience generated in the pilots to the wide (European) audience in different industrial sectors by providing practical experiences from large-scale pilots to hundreds of companies through associated DIHs.
One aspect to the digitalisation vision is the "cognitive element", where the process plants can learn from pattern recognition in historical data and adapt to changes in the process, simultaneously being able to predict unwanted events in the operation before they happen
The COGNITWIN toolbox will be validated in the following SPIRE process industry sectors: Non-Ferrous (Aluminium and Silicon), Steel and Engineering.
COGNITWIN will demonstrate an approach for the design, development and operation of the European process industry by introducing a platform for virtual component-based architecture that integrates IoT, Big data, AI, smart sensors, machine learning and communication technologies, all connected to a novel paradigm of self-learning hybrid models with proactive cognitive capabilities.
The overall objectives includes the following:
• COGNITWIN for Industry Process Excellence: Show improved performance in cognitive production plants by a technology demonstration of fully digitalized pilots.
• Cognitive Digital Twins for Cognitive Retrofitting: Enabling an efficient and well-defined approach for “cognitive augmentation” of physical assets, processes and systems for Cognitive Digital Transformation in Process industry.
• Hybrid Twins for Optimised Process Performance by hybrid models that combines first principle and data-driven models and use machine learning, AI and the connected data bases to pro-active forecast and communication, as well as self-learning by recognition of patterns in the data.
• COGNITWIN Interoperability Toolbox as a Service: A reference architecture for the cognitive elements including of Big Data, Databases, IoT, Smart Sensors, Machine Learning, and AI technologies that realizes hybrid modelling, self-adaptivity and cognitive recognition, leveraging/extending the existing work into relevant communities.
• COGNITWIN for increasing European Technology Dominance: Ensure the dominance of the Europe in technologies related to cognitive plants, thereby influencing the further development of Big Data, Databases, IoT, Smart Sensors, Hybrid Modelling, Machine Learning and AI technologies in relevant communities, focusing on the capabilities of the developed technologies for creating new generations of self-adaptive and cognitive algorithms and models.
• COGNITWIN for SPIRE: Ensure the knowledge transfer of results and experiences from the COGNITWIN project to the SPIRE Process Industry community, focusing on active participation in the new SPIRE DG7 Digitalisation group and in SPIRE organized events.
• COGNITWIN for boosting European Industry: Provide competitive advantage to the European industry, esp. SMEs in the global market, through better exploitation of the synergies between Big Data, Databases, IoT, Smart Sensors, Hybrid Modelling, Machine Learning and AI technologies for an efficient resolution of complex process industrial challenges.
• Effective dissemination and ensuring transfer of knowledge and experience generated in the pilots to the wide (European) audience in different industrial sectors by providing practical experiences from large-scale pilots to hundreds of companies through associated DIHs.
The COGNITWIN project aims to develop several enabling technologies that collectively realize the vision of cognitive production plants. Technology components are being provided through the COGNITWIN toolbox.
To realise this vision,a foundation for the concept of digital, hybrid and cognitive twins has been created, with the publication of results from this in 13 scientific papers.
COGNITWIN will empower the process industry with cognitive twins to find new answers to emerging questions by marrying the expert knowledge with the power of hybrid analytics.
The COGNITWIN Toolbox contains a number of components in terms of software and methods suitable to support the realisation of hybrid and cognitive digital twins for the process industry.
The COGNITWIN project is pilot driven and the components in the toolbox have been developed based on the needs and requirements of 6 different process industry pilots. The results have been illustrated by demonstrators for each pilot.'
HYDRO Pilot - Aluminum: Until now, the necessary level of digitalisation to meet the ambitions of the Hydro Pilot has been analysed and defined, followed up with installation of defined sensors and a hybrid digital twin.
ELKEM Pilot - Silicon: Until now, the project period has focused on developing and implementing the hybrid digital twin version of the on-line mass/energy balance model.
SAARSTAHL Pilo - Steel: The focus has been on establishing and providing the requisites for a steel billet tracking system, in particular installing the required cameras, generating a 3D-Modell of the section of the rolling mill with the second demonstrator.
SIDENOR Pilot - Steel: The industrial starting point for ladles information has been assessed. Relevant process data and type of available measurement of the ladle profiles were established and a hybrid digital twin demonstrator created.
Noksel Pilot - Steel: An analysis of the real-world data has been conducted, with a hybrid digital twin demonstrator.
Sumitomo Pilot - Engineering/Boilers: The construction of a hybrid digital twin model for a boiler has been illustrated with a demonstrator.
To realise this vision,a foundation for the concept of digital, hybrid and cognitive twins has been created, with the publication of results from this in 13 scientific papers.
COGNITWIN will empower the process industry with cognitive twins to find new answers to emerging questions by marrying the expert knowledge with the power of hybrid analytics.
The COGNITWIN Toolbox contains a number of components in terms of software and methods suitable to support the realisation of hybrid and cognitive digital twins for the process industry.
The COGNITWIN project is pilot driven and the components in the toolbox have been developed based on the needs and requirements of 6 different process industry pilots. The results have been illustrated by demonstrators for each pilot.'
HYDRO Pilot - Aluminum: Until now, the necessary level of digitalisation to meet the ambitions of the Hydro Pilot has been analysed and defined, followed up with installation of defined sensors and a hybrid digital twin.
ELKEM Pilot - Silicon: Until now, the project period has focused on developing and implementing the hybrid digital twin version of the on-line mass/energy balance model.
SAARSTAHL Pilo - Steel: The focus has been on establishing and providing the requisites for a steel billet tracking system, in particular installing the required cameras, generating a 3D-Modell of the section of the rolling mill with the second demonstrator.
SIDENOR Pilot - Steel: The industrial starting point for ladles information has been assessed. Relevant process data and type of available measurement of the ladle profiles were established and a hybrid digital twin demonstrator created.
Noksel Pilot - Steel: An analysis of the real-world data has been conducted, with a hybrid digital twin demonstrator.
Sumitomo Pilot - Engineering/Boilers: The construction of a hybrid digital twin model for a boiler has been illustrated with a demonstrator.
The progress beyond the state of the art focuses in particular on the development of the concept of Cognitive Digital Twins and the application of this to support Cognitive Plants in the Process industry.
Expected results until the end of the project includes the final version of the COGNITWIN Toolbox with comments that can support the development of Hybrid and Cognitive Digital Twins for the Process industry, supported by the technology partners.
An important part of the expected impact will come inside the process industry partner organisations related to the 6 pilot cases of the process industry partners in the project.
During the initial project phase, each of the process industry partners have provided a pilot case, described their respective pilot cases. fFr the purpose of estimating the impact of the project, they have: provided a number of Key Performance Indicators (KPIs) including how to measure baseline values of the KPIs – based on historical data up to the time of project start-up.
The overall goal of COGNITWIN’s exploitation, dissemination, communication and standardisation activities is to ensure that the COGNITWIN results will have a maximum impact and that the consortium reaches its impact objectives.
The COGNITWIN consortium has active involvement in relevant standardisation activities, in particular in the Digital Twin Consortium, the Industrial Internet Consortium Digital Twin Interoperability group and in ISO SC41 IoT and Digital Twin and ISO SC42 AI and Big Data.
Expected results until the end of the project includes the final version of the COGNITWIN Toolbox with comments that can support the development of Hybrid and Cognitive Digital Twins for the Process industry, supported by the technology partners.
An important part of the expected impact will come inside the process industry partner organisations related to the 6 pilot cases of the process industry partners in the project.
During the initial project phase, each of the process industry partners have provided a pilot case, described their respective pilot cases. fFr the purpose of estimating the impact of the project, they have: provided a number of Key Performance Indicators (KPIs) including how to measure baseline values of the KPIs – based on historical data up to the time of project start-up.
The overall goal of COGNITWIN’s exploitation, dissemination, communication and standardisation activities is to ensure that the COGNITWIN results will have a maximum impact and that the consortium reaches its impact objectives.
The COGNITWIN consortium has active involvement in relevant standardisation activities, in particular in the Digital Twin Consortium, the Industrial Internet Consortium Digital Twin Interoperability group and in ISO SC41 IoT and Digital Twin and ISO SC42 AI and Big Data.