Potential of acoustic sensors for real-time monitoring of physicochemical properties of milk protein concentrate during ultrafiltration

Ultrafiltration (UF) is the step for concentrating protein in milk and whey prior to evaporation and drying in dairy ingredient production, e.g. milk protein concentrate (MPC). To optimize UF process, it is important to monitor changes in product/process parameters. Two in-line sensors with outputs: 1. bulk acoustic wave (BAW), acoustic viscosity (AV); 2. surface acoustic wave (SAW), acoustic impedance (AI) and acoustic transmission (AT), were evaluated to measure MPC physicochemical properties (total solids (TS), density, protein and apparent viscosity) during UF. Trials were quintuplicated in a UF membrane pilot-plant, to concentrate feed (TS 11.36-19.10%). Models for predicting MPC physicochemical properties developed by acoustic parameters performed well, especially by AI and AT, with R-2 > 0.963, SEP <1.076 to predict apparent viscosity, and R-2 > 0.980, SEP <0.627 for all other properties' prediction. This study demonstrated the potential of both acoustic sensors for UF process monitoring.