Meet Smartness: The Future of Production Management Systems in the Metals Industry
15 Mar 2023 - Production, Technology
Today's production management systems (PMS) include a high level of process automation, approaching 100 percent digital transparency of plant production. However, experienced human operators still make most critical production decisions. These operators initiate fully automated procedures that have a predefined, mostly rule-based, non-adaptive behavior. If this automation behavior needs to be adapted, it must be "re-programmed" or "re-configured". Unfortunately, production conditions evolve and require regular adjustments. This adaptability is one of the key challenges for the next generation of PMS solutions. The metals industry needs production automation that is able to self-adapt to changing production conditions.
7 Key Features That Drive the Future of Smart PMS Automation
Waves of Production Management Automation in the Metals Industry
© PSI Metals
Digitalization of business and production process emerged following the industrial and digital revolution of the last century. This led to successful introduction of Level 1, Level 2, Level 3 and Level 4 automation systems in the metals industry.
Previously, the steel and aluminum producer’s IT and automation team designed their own legacy systems based on in-house database schemas. User applications were developed as terminal-based data screens by mainly Cobol and database query developers. These systems were highly plant and process-specific and hard to adapt to changes. They were running on expensive hardware and were expensive to maintain, thus questioning their value. Gradually, they were replaced by standard automation systems delivered by external specialized and product-based software suppliers.
In a second wave, external software providers tailored solutions for the metals industry. These new standardized software applications contain a set of generic business and process functionalities that can be customized, configured and embedded into a plant’s specific solution easily. Such standard applications are easier to maintain, as they are highly configurable and designed to run on any hardware platform.
Today, industry oriented B2B software providers deliver state-of-the-art technology and focus on specific levels of the automation pyramid. For example, SAP for Enterprise Resource Planning (ERP) systems, PSI Metals for Manufacturing Execution Systems (MES) and planning systems and Primetals for the L2 systems. These new solutions are adaptable to production process and production management.
The suppliers possess business models from previous installations, allowing the customer to learn how other producers work from the generic solutions they are buying.
The icing on the cake is that the total lifecycle cost of such third-party production software is remarkably low.
With today’s L1, L2, L3 and L4 system integrations and plant automation, the metals industry is close to a full digital twin of the real production. It has all this production data readily available and can track and evaluate every plant activity.
All this production data are stored in one consistent production supply chain and material genealogy model. This data is a gold mine for data analytics and artificial intelligence that would drive the future of automation.
7 Key Features That Drive the Future of Smart PMS Automation
© OleCNX / Adobe Stock Edited by PSI Metals
With the abundant digital production data, the metals industry will be able to evolve to the next generation of self-adapting PMS. We identify these 7 key features that will support this evolution.
1. Plug and play software
Self-adapting software is only possible if it is very easy to plug in or out software modules without any interruption. Virtualization and cloud computing, including seamless updating, contain already the concepts to make that possible. Service-oriented bus architectures such as the PSI Metals Service Platform, allow to you modify “services” according to the current needs of production.
2. Model-driven configurability
Self-adaption often means changing the decision flow chart. This is only possible if the decision flow is not “hard coded” but configurable in an easy-to-adapt model. PSI Metals already manages business processes, rule bases, constraints and decision tables in editable and configurable models that can be adapted at run-time without any interruption.
3. KPI-based production management
To measure and identify changes, we need to monitor key performance indicators (KPI) of the production and business processes. PSI Qualicision for example monitors huge amounts of production data, and uses its data mining technology to concisely generate key production analysis reports and decisions.
4. Machine learning and self-adaption
Currently, production data can be extended with data labels generated by machine learning models. Such prediction models can be re-trained regularly with new data. The resulting up-to-date decision model will auto-adapt to the new reality.
5. Expert-augmented machine learning
This is an automated way to automatically extract problem-specific human expert knowledge and integrate it with machine learning to build adaptive decisions systems.
For example, they systematically learn from the decisions of human experts in scheduling or quality, and automatically adapt the knowledge rule base or optimization goals.
6. Explainable AI technology
It will allow production managers to trust “black box” AI models. Only 20 percent of AI models successfully reach operational production. Mostly because of the “black box” nature of AI decisions. Explainable AI will close that gap between data scientists and production managers.
7. Multi-agent systems (MAS)
MAS allow smart or intelligent software agents to work together to solve complex problems. A multi-agent system consists of multiple decision-making agents, which interact, in a shared environment to achieve common or conflicting goals. This collaboration can further be extended with human agent’s interactions and decisions.
Research in Smart Production Management Automation
PSI Metals is conducting research in the field of smart production management automation and is working together with three partners from Berlin on a future-oriented solution for the steel industry.
For example, the joint project, The Hybrid Multi-Agent System (HYMAS) is developing a hybrid multi-agent system for self-adapting steel production. This system will be able to autonomously, collaboratively and automatically create and monitor production plans and dynamically adjust them as needed. Data analytics methods will be used to enable experts to analyze and, if necessary, influence the decisions of the production management system. The project brings together the combined expertise of partners from the fields of information technology (TU Berlin, DAI Laboratory), production management (TU Berlin, Department of Industrial Production and Service Management), data analytics (Pumacy) and metals production (PSI Metals). Practical testing will take place at a steel mill in Rhineland-Palatinate.
Finally, data analytics methods are used to enable human experts to analyze and, if needed, influence the control decisions of the PMS.
What is the future of automation in production management systems?
What features do you expect from smart production managmement systems?
Luc Van Nerom
Innovation Manager, PSI Metals
lvannerom@psi.de
After studying mathematics and computer science, he founded Artificial Intelligence Systems in 1986. The mission was to bring AI technology to the energy and process industry. Especially in the metals industry this crystalized into a number of production management optimizers today fully embedded into the PSI Metals portfolio. Nowadays at PSI, he focusses on innovation and product management. He is also managing director of PSI Metals Belgium.
Follow Us on
