The new AI-based system for fill level control from Krones is another significant step towards resource-friendly beverage filling with the Dynafill – and it has promptly earned the AI Excellence Award.
Ensuring the correct fill level of a beverage – and thus the quality of the finished product – requires the utmost precision in the filling process. Fill levels can be indirectly controlled by carefully regulating the bowl pressure. However, a number of factors must be taken into account here: Even the smallest changes in product temperature, CO2 content or vacuum pressure can have a significant impact on the process. Therefore, it is extremely difficult for an operator to manually set the proper pressure at just the right time. An AI-based controller within the Krones Dynafill uses a machine learning model to automatically calculate the optimal bowl pressure from the measurements taken at the sensors, thus achieving the desired fill level with a high degree of efficiency and precision.
Small technology – huge progress
“The biggest challenge was to establish the technical basis for the AI-driven fill level control,” recounts Philipp Olenberg, Head of Artificial Intelligence at Krones. “The multidisciplinary expertise available within Krones was crucial to this process. The digital innovation team built the necessary infrastructure and the technical setup for data transmission. They optimised the automatic upload to the cloud and the machine learning of the various parameters. And they tested it all extensively at the customer’s facility.” Another challenge was to ensure the transfer of knowledge between products, bottle types and lines, so that the system could deliver the right model at each changeover, be it from one product to another or simply to a different bottle size.
The flexible, AI-driven control system detects changes in production conditions and responds accordingly, thus ensuring continuous monitoring of fill levels and fast, flexible adjustments to new conditions. What’s more, the AI solution automatically adjusts the fill level during ongoing operations, including the startup phase, which means fewer manual interventions are needed.
And now, the AI solution has not only proven itself in a test run at a customer facility. It has also won over the jury of the Business Intelligence Group, earning the AI Excellence Award in the “Product or Service – Intelligent Control” category.
Ensuring the correct fill level of a beverage – and thus the quality of the finished product – requires the utmost precision in the filling process. Fill levels can be indirectly controlled by carefully regulating the bowl pressure. However, a number of factors must be taken into account here: Even the smallest changes in product temperature, CO2 content or vacuum pressure can have a significant impact on the process. Therefore, it is extremely difficult for an operator to manually set the proper pressure at just the right time. An AI-based controller within the Krones Dynafill uses a machine learning model to automatically calculate the optimal bowl pressure from the measurements taken at the sensors, thus achieving the desired fill level with a high degree of efficiency and precision.
Small technology – huge progress
“The biggest challenge was to establish the technical basis for the AI-driven fill level control,” recounts Philipp Olenberg, Head of Artificial Intelligence at Krones. “The multidisciplinary expertise available within Krones was crucial to this process. The digital innovation team built the necessary infrastructure and the technical setup for data transmission. They optimised the automatic upload to the cloud and the machine learning of the various parameters. And they tested it all extensively at the customer’s facility.” Another challenge was to ensure the transfer of knowledge between products, bottle types and lines, so that the system could deliver the right model at each changeover, be it from one product to another or simply to a different bottle size.
The flexible, AI-driven control system detects changes in production conditions and responds accordingly, thus ensuring continuous monitoring of fill levels and fast, flexible adjustments to new conditions. What’s more, the AI solution automatically adjusts the fill level during ongoing operations, including the startup phase, which means fewer manual interventions are needed.
And now, the AI solution has not only proven itself in a test run at a customer facility. It has also won over the jury of the Business Intelligence Group, earning the AI Excellence Award in the “Product or Service – Intelligent Control” category.
