Periodic Reporting for period 4 - COGNIPLANT (COGNITIVE PLATFORM TO ENHANCE 360º PERFORMANCE AND SUSTAINABILITY OF THE EUROPEAN PROCESS INDUSTRY)
Período documentado: 2023-04-01 hasta 2023-09-30
COGNIPLANT project aims to develop and demonstrate an innovative approach for the advanced digitization and intelligent management of the process industries. Based on a novel vision to data monitoring and analysis and coupled with a disruptive use of the Digital Twin concept to improve Production plants’ operation performance by up to 68% in real time control of the productive environment, 65% in quality control of the final products and 70 % in response time to uncontrolled incidents. In doing so, and having in mind industries with high consume of energy, main objectives may be achieved:
- Provide a hierarchical monitoring and supervisory control that will give a comprehensive vision of the plants’ production performance as well as the energy and resource consumption.
- Apply advanced data analytics to extract valuable information from the data collected about the processes and their effect on the production plant’s overall performance.
- Boost quality control of the final products.
- Speed up response time to unplanned incidents and in addition
- Design and simulate operation plans in Digital Twin models.
- Reduce CO2 emissions of cognitive production plants up to 20%.
The project is being performed in four SPIRE industries: A chemical plant in Austria, an aluminum refinery in Ireland, a concrete production plant in Italy and a metal manufacturer in Spain.
- Provide a hierarchical monitoring and supervisory control that will give a comprehensive vision of the plants’ production performance as well as the energy and resource consumption.
- Apply advanced data analytics to extract valuable information from the data collected about the processes and their effect on the production plant’s overall performance.
- Boost quality control of the final products.
- Speed up response time to unplanned incidents and in addition
- Design and simulate operation plans in Digital Twin models.
- Reduce CO2 emissions of cognitive production plants up to 20%.
The project is being performed in four SPIRE industries: A chemical plant in Austria, an aluminum refinery in Ireland, a concrete production plant in Italy and a metal manufacturer in Spain.
The main objective has been defining and developing the basis and framework to suitably run the pilot cases in the four industrial partners facilities. Partners has continued working in WP1 with the analysis of requirements, expert knowledge, architecture framework and definition of the KPI. Expert knowledge collection has been continued and the information was deepened and refined to support the ongoing technical WP. In WP2, where the aim is to design an advanced and effective digitisation strategy for the targeted process industries, main activities have been focused on selecting the sensors and equipment to enhance the plants while the edge processing node and the Cloud based data system were completely validated. WP2 has been finished during the second reporting period. In WP3, work has been carried out on the creation of different ML tools for the exploitation of the data obtained. A Big Data architecture has been designed and this architecture has been deployed at the Savvy site. Mining tool is being used in the different industrial use cases to facilitate the determination of models and their characteristics.
WP4, has worked on the integration of the modules and their joint deployment. Work has also been done on the first approximations to the ML models related to both T4.2 and Co-Decide by creating the first version of 1 Digital twin for production optimisation and within T4.3 on the creation of ML models. In WP5, the End Users continued to perform the modifications in their plant to adapt them to the new digitisation solution. The sensors and equipment have been partially installed while the deployment of some sensors has been slightly delayed due to production peaks of industrial demo cases. In WP6 work around market analysis, exploitation, and replication of COGNIPLANT solution has been progressing. The customer journey of identified buyer personas was developed and the strategy that the consortium should follow to capture and retain these customers throughout the product lifecycle was designed. The exploitation models were further elaborated based on common characteristics of the use cases and a replication methodology to successfully transfer the COGNIPLANT solution in other process industries has been developed.
In WP7, D&C Plan was delivered, and branding created as well as website and social media channels. This plan has been adapted to the new scenario that has appeared due to the pandemic situation, that has limited the possibility of attending to meetings, workshops, conferences, etc. A novel strategy was applied with a focus on the online presence and digital events is being followed. WP8 has performed the project coordination from the administrative, financial, and technical management standpoint. It was established all the management procedures, the quality plan and all the organizational meetings to hold during the project life duration and in WP9 ethics procedure was defined in the previous reporting period and the coordinator is in charge to verify that all the research and innovation activities carried out during the project will comply with ethical principles.
WP4, has worked on the integration of the modules and their joint deployment. Work has also been done on the first approximations to the ML models related to both T4.2 and Co-Decide by creating the first version of 1 Digital twin for production optimisation and within T4.3 on the creation of ML models. In WP5, the End Users continued to perform the modifications in their plant to adapt them to the new digitisation solution. The sensors and equipment have been partially installed while the deployment of some sensors has been slightly delayed due to production peaks of industrial demo cases. In WP6 work around market analysis, exploitation, and replication of COGNIPLANT solution has been progressing. The customer journey of identified buyer personas was developed and the strategy that the consortium should follow to capture and retain these customers throughout the product lifecycle was designed. The exploitation models were further elaborated based on common characteristics of the use cases and a replication methodology to successfully transfer the COGNIPLANT solution in other process industries has been developed.
In WP7, D&C Plan was delivered, and branding created as well as website and social media channels. This plan has been adapted to the new scenario that has appeared due to the pandemic situation, that has limited the possibility of attending to meetings, workshops, conferences, etc. A novel strategy was applied with a focus on the online presence and digital events is being followed. WP8 has performed the project coordination from the administrative, financial, and technical management standpoint. It was established all the management procedures, the quality plan and all the organizational meetings to hold during the project life duration and in WP9 ethics procedure was defined in the previous reporting period and the coordinator is in charge to verify that all the research and innovation activities carried out during the project will comply with ethical principles.
1. Digitalization in the European process industry: Develop an innovative solution that will involve a relevant number of breakthroughs in the field of enhanced data acquisition and processing, advanced data analytics and a disruptive vision of the “Digital Twin” for realistic simulations and decision making, not addressed so far in these types of industries.
2. Digital retrofitting: Create a digitization platform able to connect to different existent control systems, sensors and actuators, and collecting and transferring the data to a Big Data Cloud environment. In this way, the data will be processed simultaneously in different analysis groups and profiles.
3. Optimisation methods to distributed targeted process monitoring: The huge amount of data generated at by the IoT connected assets a strong edge node will be set up, that will be capable of processing signal transformation functions and moving the results to the cloud as an aggregated output. Advanced distributed logical controls and a comprehensive suite of processors will be developed for a fast response over recurring problems will be directly done on site and will be transferred to the Cloud platform (Data virtualisation layer). In real time, the Co-Digitise platform will offer enough sampling frequency for a detailed information extraction and solutions.
4. Advanced data analytics: A holistic and advanced data analysis methodology will be developed for the processing of manufacturing processes’ data, starting from big stream data management and processing and ending with the derivation of efficient predictive models for real-time performance prediction and optimization support. This goal passes through the development of methods for a) big data management for data analytics, b) process mining and inference, and c) online learning for prediction and decision support.
5. Digital twin: Digital Twin can be considered as the next generation of modelling and simulation in the industry. In this sense, a digital twin will be created for each of the Demo Cases that allow the simulation and optimization of industrial processes according to the KPIs proposed for each of these scenarios. These simulations will allow study which variables impact and how they impact in the KPIs, giving answer to specific questions of the experts on what would happen if certain parameters of any node are modified in fictitious or simulated scenarios, and how the knowledge of these modifications is inferred and "exploded" throughout the network, to understand which indicators in other nodes are affected.
6. Cognitive Reasoning and Reactive scheduling: The application of advanced Decision support systems (DSS) systems in process industries represents still a challenge, and therefore a strong innovation of COGNIPLANT project.
2. Digital retrofitting: Create a digitization platform able to connect to different existent control systems, sensors and actuators, and collecting and transferring the data to a Big Data Cloud environment. In this way, the data will be processed simultaneously in different analysis groups and profiles.
3. Optimisation methods to distributed targeted process monitoring: The huge amount of data generated at by the IoT connected assets a strong edge node will be set up, that will be capable of processing signal transformation functions and moving the results to the cloud as an aggregated output. Advanced distributed logical controls and a comprehensive suite of processors will be developed for a fast response over recurring problems will be directly done on site and will be transferred to the Cloud platform (Data virtualisation layer). In real time, the Co-Digitise platform will offer enough sampling frequency for a detailed information extraction and solutions.
4. Advanced data analytics: A holistic and advanced data analysis methodology will be developed for the processing of manufacturing processes’ data, starting from big stream data management and processing and ending with the derivation of efficient predictive models for real-time performance prediction and optimization support. This goal passes through the development of methods for a) big data management for data analytics, b) process mining and inference, and c) online learning for prediction and decision support.
5. Digital twin: Digital Twin can be considered as the next generation of modelling and simulation in the industry. In this sense, a digital twin will be created for each of the Demo Cases that allow the simulation and optimization of industrial processes according to the KPIs proposed for each of these scenarios. These simulations will allow study which variables impact and how they impact in the KPIs, giving answer to specific questions of the experts on what would happen if certain parameters of any node are modified in fictitious or simulated scenarios, and how the knowledge of these modifications is inferred and "exploded" throughout the network, to understand which indicators in other nodes are affected.
6. Cognitive Reasoning and Reactive scheduling: The application of advanced Decision support systems (DSS) systems in process industries represents still a challenge, and therefore a strong innovation of COGNIPLANT project.