Artificial intelligence - condition monitoring and prediction for sustainable production

Production companies are increasingly confronted with the challenge of mastering two transformations of their own operations simultaneously. On the one hand, there is increasing pressure from customers and the competition to use digital technologies in production in order to optimize production processes, provide information about the production process or even implement digital business models. On the other hand, there are also increasing demands from the customer base and legislators to make production more ecologically sustainable and, for example, to reduce energy consumption or direct emissions of climate-damaging emissions. The approach of using digital technologies such as artificial intelligence (AI) in a targeted way to make production more ecologically sustainable can enable companies to address and shape both of these transformations together.

Through expert knowledge with AI to sustainable production

In February 2022, the ZD.B Digital Production & Engineering topic platform offered an insight into the state of research on AI and ecological sustainability and its use in the industrial environment as part of the webinar series "A bridge between science and industry - from research to practice". Three experts from science and industry presented applications with which the ecological sustainability in production can be increased through a use of AI and provided insights into their industrial implementation.

Prof. Dr.-Ing. Frieder Heieck from Kempten University of Applied Sciences and head of the Technology Transfer Center (TTZ) Sonthofen explained the potential benefits of AI for more sustainable production using three examples. For example, the use of machine learning models based on sensor data from machines can be used to implement a condition-based maintenance strategy for these machines and thus save resources in the context of maintenance. Another possibility is to network all machines and other energy consumers in a production to form an intelligent energy network in which the energy consumption of the entire production is optimized by an AI-based software agent. In addition, an implemented solution was shown that makes it possible to automate the sorting of waste on the basis of a 3D laser scanner and a gripper with the help of AI in such a way that recyclable materials can be separated by type and reused as far as possible in the sense of a circular economy.

Many older machines, different databases and file formats, and high IT security requirements are some of the challenges Jean-Pierre Hacquin of the Kempten-based iron foundry faced when the company set out to digitize its own production in order to produce more sustainably. Thanks to collaboration with external partners such as Kempten University of Applied Sciences, financial support from the German Federal Environmental Foundation (DBU) and a great deal of effort within the company itself, it is now possible to collect and evaluate production data centrally. By analyzing the data, it has been possible to optimize production processes and also adapt machine ramp-up times to network utilization. The Kempten iron foundry was able, for example, to reduce the CO2 emissions of its own production by 15% through the use of digital solutions.

Florian Huber (M.Sc.) from the Allgäu Research Center at Kempten University of Applied Sciences, a member of the research group for digitalization in the metalworking industry led by Prof. Dr.-Ing. Dierk Hartmann, described three digitalization projects. These were implemented in production at the Kempten iron foundry to make production more sustainable. For example, a system was developed to predict the wear of crucibles in order to make the best possible use of their service life. In another project, a system was developed to help predict the tensile strength of castings based on chemical data. In addition, a digital solution was implemented that allows production to be planned in such a way that the molten material (and thus the energy used for it) is used as efficiently as possible. While an AI is used in the first two examples, the planning in the third example is based on an algorithmic approach, since in this case an AI use would not have generated a great advantage and the solution could be implemented faster with the help of the algorithm.

A first step toward greater ecological sustainability through digitization

Finally, the experts have some recommendations for small and medium-sized enterprises (SMEs) that would like to use digital technologies to make their own production more ecologically sustainable. Since every issue and application is unique, the experts recommend that companies first take a close look at the prevailing problem and the existing prerequisites (e.g., with regard to available data). When designing solutions, companies should not immediately jump on the topic of AI, but should also be open to classic algorithm-based approaches, as these are often easier to implement and can achieve similar results in some cases. When selecting initial problem definitions, companies should also make sure to start with small pilot projects and begin "where the shoe pinches the most."

In addition, SMEs are not on their own when it comes to implementing digitization projects for greater environmental sustainability. For example, it is possible, as the Kempten iron foundry has done, to cooperate with a research institution such as the Kempten University of Applied Sciences and thus draw on extensive technical expertise with a strong application focus. In addition, the experts recommend that companies also look into the financial support available for such projects. On the part of the Digital Production & Engineering thematic platform, we will be happy to support you in identifying suitable scientific cooperation partnerships and in selecting suitable funding opportunities for your digitization project.