Peter Johansson, who is CEO of the listed technology consulting company Eurocon, sees a great need to optimize the industries of the future with the help of AI.
- Generically speaking, regardless of whether it is a forest industry, mining industry or biorefineries, it is about similar challenges and needs in similar processes, says Peter Johansson, CEO Eurocon.
He believes that AI initially should be used for more realistic simulations in order to train new operators in new systems.
- AI can contribute to developing better training tools for everyone who works in a factory. With good simulators in the form of updated digital twins, which are identical to the current processes, the operators of the future need less training time and higher quality training before they can start working in the real physical industrial environment. If the operators' knowledge and experience then can continuously be refueled and stored with the help of AI, the system will be further sharpened and secured, he says and continues:
- Today, all simulators are statically built and only actual updated a short time after start-up and delivery. Thereafter, the processes gradually change through all the trimming and rebuilding that is carried out and the simulators will quickly become obsolete.
Real-time updated digital twins
The most important task of process tools and simulators is to provide reliable and ongoing support to development, maintenance and operating personnel.
- Unfortunately, no one can possibly update the tools in step with all the changes that regularly take place in both larger and smaller process areas. Doing it afterwards will be a far too big and costly job. Therefore, it would be much easier if you could create real-time updated twins of the facilities from the beginning, says Peter Johansson.
Offline system for decision support
The process industry is capital intensive and due to the large investment amounts forced to be conservative and cautious. Everyone believes in AI's possibilities, but no one wants to be first. Let others make the first costly mistakes.
- Therefore, it is important to initially build offline systems, which do not directly affect the processes if something goes wrong. With the help of effective operator support, we can then simulate and test different events. How is the process run in the most sustainable and efficient way? What can we optimize? What happens to quality if we raise or lower certain values? Which calculation values are most important? What does it cost to produce? AI enables many exciting opportunities. If you can also supplement decision support with reliable predictions, the basis is created for future autonomous control of the processes, Peter Johansson concludes.