DigiFoods - Digital Food Quality

DigiFoods develops smart sensors for food quality assessment directly in the processing lines. The obtained information will be used for optimization of processes and value chains and make the food industry more efficient and sustainable. The inline measurement of food quality is extremely challenging due to the huge biological variation in raw materials. We develop sensor solutions based on a fundamental understanding of food science, processing and modern sensor technology, which will allow the food industry to effectively measure and handle these variations in true time. We also introduce robotics to enable automatic measurements in challenging processes and in the field. Massive assessment of the essential food qualities throughout the value chains will pave the way for a digital transformation of the food production. Based on such quality measures, we develop novel strategies for process control with the aim of reducing food waste and increasing profitability. We are also studying concepts for how large scale quality measurements can be used for 1) improvements in primary production, and 2) product differentiation to different consumer segments, with the aim of reducing food waste in households. The centre of 25 partners will function as an innovation hub for food industry, technology providers and scientific groups. Fourier-transform Infrared spectroscopy (FTIR) is promising for measurement of protein composition, and there exists no industrial solutions for this today. We have shown that FTIR can measure relevant size variation of peptides in industrial peptide mixes. Different bioprocesses require different sample presentation for optimal use. It is very important to be able to measure peptide size. A prototype portable FTIR instrument has been developed and is presently being evaluated in process. It is promising also for monitoring changes in milk proteins during storage. Other IR technology can be miniaturized based on 1) novel LEDs (light emitting diodes) and 2) quantum cascade lasers (QCL) that both generate light in the IR range and are suitable for low-cost measurement systems. These technologies enable small handheld sensor systems that can determine a range of chemical properties of food throughout the value chain. Two prototype systems have been constructed and will be evaluated for measurement of fatty acids on salmon fillets. Raman spectroscopy has the potential to measure different quality properties of foods in process. We have established that it is possible to measure the amount of bone, fat, collagen and protein in minced poultry byproducts. The method is tested in industrial process with promising results. Raman is also suitable for measuring the fatty acids EPA and DHA in salmon fillets at speed on a conveyor belt. There is great interest in the aquaculture industry for such a method. Hyperspectral imaging in the near-infrared (NIR) is also a promising method for this application. Imaging NIR is also developed for determination of fat, pigment, blood and melanin spots in salmon fillets, and this is being tested in industry. This technology is relevant for assessment of numerous quality properties in fish. In-line measurement of salmon fillets with Raman might require robotic control, including machine vision and algorithms to handle the measurement probe optimally. A demonstrator of such a robotic system has been developed and tested in the center. We have tested sensors for measuring sugar and acids in strawberries. The intention is that autonomous robots can use such sensors for precision picking of strawberries. We have achieved promising results, and this will be developed into a dedicated NIR-sensor that can be integrated on robots. An important activity is to do measurements with smart sensors in industrial processes. We have studied the following cases: In-line determination of dry matter in potatoes, fat content in sausages, core temperature in chicken after heat treatment, fat and protein in byproduct of chicken and salmon, and quality features on salmon fillets. All these methods are novel and can potentially contribute to process improvement. We learn more about the actual variation in the processes and this is the basis for new process control with the aims of stable end quality and less food losses. Within data analytics we are developing solutions based on food quality assessments with the purpose of reducing food waste and increasing profitability in food production. We follow three strategies to handle raw material variability: 1) We gather large data sets where we combine food quality data with available farm or production data to identify risk factors and understand underlying causes for variation. 2) We use in-line measurements of food quality to monitor and adjust processes in true time to ensure stable quality. 3) We develop concepts for product differentiation, where products of different qualities can be marketed to different consumer groups.

The goal of SFI Digital Food Quality (DigiFoods) is to develop inline smart, sensor-driven solutions that deliver the essential food quality information required for successful process optimisation and digitalization of the food industry. Food processes are extremely complex and challenging to measure due to the inherent high level of biological variation in raw materials. The development of advanced solutions that are built on a fundamental understanding of food science, will allow the food industry to effectively measure and handle these variations, enabling a ground-breaking digital transformation of the industry. Effective strategies for real-time analysis and utilization of industrial data at a large scale will optimise processes and reduce waste throughout the value chain. New levels of information flow will also significantly increase productivity. DigiFoods will innovatively combine applied and basic research to generate new basic knowledge, evaluate prototype solutions, and create results for innovation in the following core areas 1) novel sensor systems and food application development, 2) robot and sensor integration, 3) integrated inline sensor solutions, and 4) analysis and utilisation of large-scale process and value chain data. A multi-disciplinary team comprising food companies, technology providers and research institutions is highly motivated to collaborate towards this common goal. The industrial partners will participate in all research tasks, committing essential in-kind contributions to the project. A critical element of this centre is that much of the experimental work will be done in the industrial process lines, which the end-users will make available for research. The centre will educate 9 PhD students and 3 post docs. DigiFoods will be governed by a board with an industrial majority, ensuring relevant research that will build the foundation for a future food industry 4.0.