Interview: Kathleen De Badrihaye on the importance of Big Data
Big data, machine learning, recipe optimisation and digitisation of specifications management are the main trends identified by software company Adifo. “It would be nice if, ultimately, machines could reset themselves after each product changeover, eliminating the need for test batches altogether.”
The most important trend observed by Adifo is the use of big data. "Foodstuffs companies could get more out of big data than is currently the case", says Product Manager Kathleen De Badrihaye.
"They could establish a database of quality and process parameters. By combining the measuring data, machines can be set correctly immediately after a product changeover. For example, we implemented this for a manufacturer of extrusion products.
Every half hour, we registered the moisture content. We then combined this with data on the machine settings, such as temperature and pressure. With this information, after each product changeover the manufacturer is able to set the machine so well and accurately that test runs are barely required any longer.” The product manager expects that, in the future, machines will even be capable of adjusting their own settings based on data they themselves have collected. That’s when machine learning will become reality.
Another trend observed by Adifo is recipe optimisation. “This is already being applied in the meat and dairy industries,” she says. “We see opportunities here for other sectors as well.” For batch optimisation, software programme Bestmix collects all analysis data for incoming raw materials. Then, the programme calculates the optimal combination of the various batches in order to achieve a previously-specified final quality.
One example here is milk standardisation within the dairy sector. “Dairy processors receive various batches of milk with different levels of fat and dry matter, among other things. Our software programme calculates the best combination of these batches, for example for cheese-making. The programme then determines the most economical solution which will still meet the specifications.” De Badrihaye expects that manufacturers from other sectors can also benefit from this functionality, for example to get rid of stock nearing the ‘best before’ date. This of course reduces waste. Additionally, the software can support product developers in creating the most optimal recipes.
Finally, she notes that the digitisation of specifications management is on the rise. “We see that manufacturers are still being supplied with hardcopy specifications for their raw materials. They then manually copy the data into our software programme in order to calculate their recipes. Increasingly, companies are asking for a data sourcing portal where suppliers can enter their own specifications. That way, manual copying is no longer necessary. An additional advantage is that the data can be checked for completeness immediately, and that this can be further automated and structured by means of workflows and authentication.”