In many industrial processes, foaming is an unavoidable side-effect of mixing, shearing and powder incorporation. Too much foam, however, causes a number of operational issues, such as reduced equipment efficiencies, impaired heat transfer, product quality issues stemming from ingredients being trapped in bubbles, and the malfunctioning of pumps and valves. Excessive foam also occupies tank volume, decreasing capacity for processing materials or even leading to tank overflow.
Mikrotron EoSens MC1362. Image Credit: Mikrotron GmbH
Over the years, a number of foam management techniques have been implemented with the most common being the addition of antifoam chemicals into the system. While effective, antifoam chemicals have the potential to introduce extra compounds that can disrupt reaction kinetics and transport mechanisms. Antifoam chemical use in pharmaceuticals and food production also raises regulatory concerns.
A study by the Université de Toulouse, Laboratoire de Génie Chimique (Toulouse, France) offers fresh perspectives on how process operating variables affect the development of foam in gas-liquid stirred tanks. The impact of gas flow rate, rotational speed, and impeller type on foam generation in an air-surfactant/water system was investigated. Specifically, the scientists were measuring foam height, that is, the vertical thickness or level of foam formed on the liquid surface during the mixing or agitation processes.
To conduct their study, researchers employed a cylindrical, dished-bottom Perspex tank equipped with four transparent baffles driven by a 120 W variable speed overhead stirrer. Two different impellers were tested, a 6-blade and a 3-blade design, that were run in various pump modes. An aqueous surfactant solution was poured into the tank during the liquid phase, while compressed air was introduced in the gas phase through a ring sparger. A calibrated rotameter regulated the gas flow rate, while liquid viscosity in the tank was measured by a rheometer and surface tension by a tensiometer.
Each set-up was tested three times to ensure accuracy and corresponded to a specific operating condition and impeller type. After 20 minutes of agitation to reach equilibrium, the foam height in the tank was measured via image analysis. Images of the foam were captured with a Mikrotron EoSens MC1362, a CameraLink™ 1.3 Megapixel CMOS color digital camera. At full resolution, the EoSens MC1362 is capable of 1362 x 1024 pixel images at 500 frames-per-second (fps), or as high as 35,000 fps at low resolutions. In this case, the EoSens MC1362 captured 50 images at 10 fps over 5 seconds.
Images of the foam were analyzed with a deep learning method based on Convolutional Neural Network U-Net architecture. This tool automatically detects the foam generated in the stirred tank to determine its height. An existing dataset of foam images collected under various lighting and operating conditions trained U-Net to ensure generalization across all experiments.
Scientists at the Université de Toulouse found that the process conditions that produce smaller bubbles lead to larger quantities of foam and higher gas holdup, while larger bubbles result in lower gas holdup and reduced foam formation. Different impeller designs affect the degree of gas dispersion and liquid mixing, which in turn impacts the amount and stability of foam generated. They also concluded that strong liquid recirculation at high impeller speed decreases foam height via bubble drawdown. Their findings underscore the significance of impeller design and operating conditions for foaming in stirred tank processes.
Future research will focus on correlating foam height with the various influencing parameters and to take into account the role of fluid properties. Scale-up guidance for industrial applications will also be developed to improve foam control strategies across different industries.