Challenge

In Germany, 10 million tons of food end up in the trash every year – a good half of it only because the best-before date indicated on the packaging has expired. However, this information says little about the actual shelf life because many foods can still be eaten long after the best-before date has expired. The date only represents the time until which the manufacturer guarantees shelf life and consistent quality. On the other hand, food production is subject to many fluctuations, both in the quality of raw materials and in the production process itself. Thus, the shelf life of produced food can also fluctuate depending on the batch. The current best-before date is designed for maximum safety; if the individual shelf life could be determined more precisely, manufacturers could also specify a more accurate best-before date and thus help to reduce food waste.

Goals and procedure

In the DProFood project, real-time data on raw material and product conditions as well as on individual production processes are to be collected using various sensors and condensed into information that will enable accurate shelf life forecasts. The system is to evolve by means of machine learning and thus assemble a digital pool of experience. New sensor solutions and approaches to sensor fusion will also be explored, and the design of various participation and communication scenarios is another important aspect of this project. Using a specific use case – the production of a ready-to-serve pizza – the project team first carries out an analysis from different perspectives and derives sensor concepts from this. In addition, transport and storage are simulated. The data is correlated to identify suitable spoilage indicators. Finally, a model for data analysis is derived from the project results.

Innovations and perspectives

Innovative technologies that can be used to determine the best-before date of foods more accurately not only help to reduce food waste; they also make an important contribution to food safety. For the food industry, which is the third largest industrial sector in Germany, the DProFood project can thus provide effective and sustainable impetus. But the potential applications of the project results go far beyond food production: They can also have an impact on the fields of mechanical engineering, sensor technology, automation, and data analysis. Demonstrators developed in the project can be easily integrated into existing production plants and generate progress at the DProFood partner companies even during the course of the project. Dr. August Oetker Nahrungsmittel KG in particular has numerous production lines to which the findings can probably be transferred over the course of the next five to eight years.