More Than Noise: Assessing the Viscosity of Food Products Based on Sound Emission

In the era of Industry 4.0, manufacturing is rapidly shifting towards automation, particularly in processes such as quality control, production lines, and logistics. However, the food industry poses distinctive challenges to automation due to the variability in raw materials and stringent hygiene standards. Sensory analysis is crucial in maintaining consistent quality and safety while manufacturing food products. This paper focuses on the automatic estimation of viscosity, a key parameter in many industry quality aspects of food products. An often overlooked aspect is the potential correlation between viscosity and sound emissions. While conventional methods for determining viscosity require expensive equipment, this research investigates the possibility of analyzing the acoustic emission when a liquid is sucked through a vacuum pump to determine the viscosity. By simulating industry-like food products with varying viscosity through different flour and water mixtures, we aim to investigate the feasibility of developing an automatic, deep-learning-based system for real-time viscosity estimation in manufacturing processes. Our results indicate that our proposed methodology can automatically determine the difference in viscosity, showing the feasibility of using sound emission analysis as a tool for viscosity estimation in manufacturing processes.