Pressure monitoring inside hermetically sealed containers generally involves the use of devices in direct contact with the product inside them. However, this requires a hole to be made in the wall of the container, with the risk of causing leaks or degradation of the product itself.
In this context, the aim of a recent study carried out by a group of US researchers (Prisbrey et al., 2024) was to develop a non-invasive technology for measuring pressure inside closed containers based on the use of acoustic resonance spectroscopy (ARS) and machine learning (ML) algorithms. In particular, during the experiment, the KNN (k-nearest neighbour) regression model was trained using ARS spectra as input. The results show that the proposed system allows non-invasive pressure measurement inside closed containers without the use of sensors in permanent contact with the product.
It was also observed that resonance at lower frequencies correlates more accurately with pressure variation. Further investigation is needed to assess the measurement limitations resulting from the geometry of the containers, defining the robustness level of the proposed system. The acoustic technique presented in the study has wide applications in various industrial sectors for monitoring pressure in systems where the use of permanent sensors is not desirable, such as vacuum-packed food.
Development of a new non-invasive level monitoring technology for the food industry.
Non-invasive technologies for measuring the level of liquids in containers and vessels are of great interest to various industrial sectors, including the food industry. In fact, most commercial solutions use invasive techniques, which are not always compatible with the conditions and/or quality of the products contained within the equipment. In this context, the aim of a recent study carried out by a group of US researchers (Palanisamy et al., 2024) was to develop a new portable and non-invasive technology to perform this measurement.
This technology is based on the analysis of the diffusion of acoustic waves through the walls of the container and their interactions with the liquid content to determine the optimal separation distance of the sensors and the characteristics of the excitation signal, thus obtaining a highly accurate and precise level measurement. In particular, the approach involves a detailed analysis of the wave propagation dispersion properties, on the basis of which a general-purpose liquid level detection methodology is developed.
The latter is then validated using numerical wave propagation simulations and experiments on various containers with different physical characteristics, demonstrating the versatility of the method. In conclusion, the authors argue that, compared to existing techniques optimised for specific applications, the proposed method is potentially universally applicable to any container, regardless of shape, size, wall thickness and the presence or absence of internal components.
References: Prisbrey et al., Machine Learning with Applications, 18, 2024, 100589; R.P. Palanisamy et al., Applied Acoustics, 220, 2024, 109978.
