Data-driven resource efficiency on circular knitting machines

By resource efficiency, we mean the optimum use of energy and operating resources to produce a high-quality product on schedule. Resource efficiency on large circular knitting machines is a multi-layered and complex topic which combines technological and business aspects in textile production plants.

In the research project funded by the AiF, the focus is on the optimum consumption of operating resources of  the machines, in this case the consumption of lubricant and energy. The investigations are being carried out in the technical center of the Technology Center Knitting Technology.

Large diameter circular knitting machines are lubricated in abundance. This is due to high stresses on materials and machine parts caused by increasing production speeds and the high quality standards of technical and non-technical knitted fabrics. In knitting machines, the lubricant has a cleaning function in addition to its lubricating function. The cam area, which is sensitive to the quality of the fabric, is cleaned of fiber fly and avivages by the lubricant. Due to overflow lubrication, oil consumption is high and the machine is not operated in a resource-efficient manner.

The correlations between the consumption of operating fluids are being investigated at the DITF by continuously analyzing and evaluating data from the ongoing production process. To this end, a large diameter circular knitting machine at the DITF has been equipped with comprehensive measuring technology to enable all relevant operating parameters being recorded. This is necessary in order to be able to draw conclusions about the machine condition from the data obtained. In the project, sinker forces, machine vibrations, temperatures in the knitting cam and yarn tensions are correlated with the machine's oil and energy consumption. The measurement system stores the process data as a time series and makes it available for visualization and processing with machine learning algorithms. The Center for Management Research has extensive experience in model building in the Big Data environment and is responsible for processing the data in the project.

Since the runtimes of the  industry cannot be realized in the pilot plant, characteristic scenarios are first examined there before the measurement system is used in the industry under real conditions. The necessary data volumes for data-based models will then be generated in the production environment. In the future, resources can be used more efficiently through the far-reaching and in-depth information and knowledge gained from data. Big Data analytics in the sense of Industry 4.0 enables individual, knowledge-based, automated and resource-efficient production in small and medium-sized enterprises.

Awareness of the ecological footprint is increasing, and the consumption of energy and operating resources is now cost- and image-relevant. Therefore, knowledge of the relationships between characteristics of input materials, machine condition, wear, resource consumption and characteristics of output materials is a prerequisite for modern, resource-efficient production processes on knitting machines as well. Possible applications arise in the field of predictive maintenance and predictive quality.