Digital twins optimize processes

Optimization of costs starts in the Primary: tobacco is one of the biggest cost drivers in cigarette production, and using this natural product more efficiently is there-fore key to reducing costs. Taking this as their starting point, an interdisciplinary team of Hauni employees to-gether with colleagues from Körber Digital in Porto and Karlsruhe tackled the issue and have started to develop the solution d.digitaltwin.

Optimal process settings

“Our Process Optimization Manager is based on the concept of a digital process twin: we map the customer’s production line in a virtual environment,” explains Jens Heymann, d.digitaltwin Product Owner at Hauni. “The digital twin is continuously fed with data from green-leaf processing and Primary and Secondary production.” 

By monitoring every step of the production process, the AI-software can predict the current optimal parameters for the individual processing steps. Depending on the customer’s requirements, the system can either adjust these settings automatically or provide the operator with suitable suggestions.  

Higher yield, better quality

d.digitaltwin generates these suggestions by combining relevant plant and environmental data to reduce process fluctuations and stabilize product quality.  
“AI-assisted prediction of optimized process parameters draws on the decades of knowledge that Hauni has accumulated about production, equipment and software to significantly increase tobacco utilization,” explains Heymann. As a side effect, d.digitaltwin also solves a widespread problem – it preserves expertise and experience in the company over the long term. 
d.digitaltwin supports both experienced and lesser experienced employees in manufacturing high-quality products. This is possible because the optimal settings for different blends and products are provided at all times.

Successful project start

Benefits like these ensure that the AI-software is a high priority for British American Tobacco (BAT). The company has introduced d.digitaltwin in its plant in Korea. In the first step, all the machines in its lamina and stem lines as well as the process-relevant machines in the Secondary for a blend were connected to d.digitaltwin. Sensors measure the ambient conditions. As well as analyzing the incoming data and the optimization potential derived from it, Hauni is currently building the customized prototype of the user interface. “Thanks to the close cooperation between the highly motivated team at BAT and Hauni employees, d.digitaltwin has been fully operational since May and we have been able to perform the three month validation test,” says Heymann. “Even at this stage we will start to see the potentially significant savings that d.digitaltwil can deliver. We expect them to be at least 1 to 2 percent – with two-shift production running 6 tons of tobacco per hour, that corresponds to a saving of EUR 1.7 million per year.”