Large resource investments are necessary in order to minimize the limiting problems arising from food industrial intensive productivity. One of the most challenging concerns is the cleaning status uncertainty among heat transfer areas in dairy heat exchangers, since the effectiveness of this process cannot be easily validated. The present study aimed to develop a low-power ultrasound sensing method for monitoring the removal of milk fouling deposits along cleaning processes inside an experimental plate heat exchanger structure, connected to a milk piping unit. For that purpose, signal processing, namely acoustic feature extraction, over different wave patterns combined with artificial neural network techniques was used. Measurements were taken in pulse-echo mode with a handmade 4 MHz ultrasound transducer. While fouling deposits having initial average thickness values of 250 μm (34.5 ± 4.5 mg/cm²) were removed, the acoustic transmissivity increased. Results showed that the signal features follow the expected trends in both, clean and fouled cases, within right guess detection accuracies above 80%. Therefore, when calibrated well, this could be a very sensitive and noninvasive technique for material characterization, as well as a suitable validation method for industrial cleaning cycle operation optimization that could significantly reduce the associated costs.     «

Large resource investments are necessary in order to minimize the limiting problems arising from food industrial intensive productivity. One of the most challenging concerns is the cleaning status uncertainty among heat transfer areas in dairy heat exchangers, since the effectiveness of this process cannot be easily validated. The present study aimed to develop a low-power ultrasound sensing method for monitoring the removal of milk fouling deposits along cleaning processes inside an experimenta...     »