Collo’s technology is based on RF signals that can penetrate any liquid, machine learning, and sophisticated edge computing analysis. The system is described as adapting to any liquid process automatically and resulting in an eight-dimensional multi-parameter, real-time analysis that makes it possible to adjust the process immediately when there is a quality issue.
“The technology can detect nearly any thinkable change in the liquid properties in real-time”, says Matti Järveläinen, CEO and founder of ColloidTek. “The result is a continuous quality control, contrary to the normal, time-consuming quality control process based on blind samples and laboratory analyses.”
As an example, it can monitor fermentation processes widely used in food, cosmetics, and pharmaceutical industries and many other bioprocesses, to ensure that the delicate microbiological process is progressing as intended.
Fast response to liquid quality changes can be critical in other applications as well. Collo recently conducted a study to find out how unsupervised machine learning could be used to detect abnormal qualities in sewage waste. The study was conducted together with the City of Oulu waterworks in Northern Finland. In the study, Collo collected data from the sewage well for a couple of weeks and then compressed the information into a couple of models showing how the process should look when it was working well. After that, the analyzer monitored the sewage quality according to these models.
“With the aid of machine learning, our system was able to pinpoint the liquid quality deviations in the process,” Matti Järveläinen says. “The study showed that our analyser could be used on a large scale to monitor industrial wastewater streams for early detection of anomalies like chemical spillage, product losses or other abnormal behaviour.”