Identification and characterisation of liquids using ultrasound and machine learning

This project involves further development of an analytical system for identifying and characterising liquids based on the formation of liquid drops and employing interferometric ultrasound methods to investigate them. A prototype system was developed and has provided already a wealth of data for a wide range of different liquids, ranging from soft drinks, wines, beers, methanol and ethanol, aircraft fuels and many other liquids of interest.

The distinctive signals obtained from these signals permit them to be distinguished, as evidenced by earlier work using Principal Component Analysis within a processing pipeline. Such work, apart from its scientific interest, has applications in many wide-ranging industrial areas, including for quality control and process monitoring in the food and beverage industries, for biofuels, perfume manufacture and environmental monitoring.

The proposed project has two primary aims: i) to apply machine learning techniques to identify signals from test liquid samples and ii) to further develop the hardware itself, along with the associated control electronics in order to improve the resolution and reliability of the system.

The ideal candidate for this PhD position should have a strong background in one or more of the following areas: applied physics, analytical chemistry, electronic engineering or computer science. More importantly, they should be willing to adapt to interdisciplinary work in this exciting area. Strong computational skills and manual dexterity in handling samples and equipment would also be advantageous, along with skills in collaborative work as well as the ability to work well independently.