AI-Driven Ultrasound for Materials Evaluation

AI-Driven Ultrasound for Materials Evaluation AI-Driven Ultrasound for Materials Evaluation (2026) In this PhD, you will help develop AI-driven ultrasonic methods for materials evaluation. Depending on your interests, you may focus on building fast AI surrogate models that replace expensive simulations, on inverse models that directly extract material properties or defect information from measurements, or on both. You will work across simulation, experiment, and machine learning, including running large-scale ultrasound simulations, designing modern neural network architectures, and validating your models in our ultrasonic laboratory. What you get For 3.5 years, you will receive a tax-free stipend at a standard rate of £21,805 per year and your fees will be waived (at the UK or International rate). In addition, to a one-off Research and Training Support Grant of £2,000. Type of award Postgraduate Research PhD project Ultrasound is one of the most widely used techniques for non-destructive evaluation (NDE) and materials characterisation, underpinning safety-critical inspections in aerospace, energy, transport, and advanced manufacturing. Its ability to probe the internal state of materials, revealing microstructural features, defects, and degradation, makes it indispensable for both quality assurance and structural health monitoring. However, extracting quantitative information from ultrasonic measurements remains a long-standing challenge. Forward modelling of wave propagation in complex defective media is computationally expensive, and the corresponding inverse problem of inferring material properties or defect characteristics from measured signals is typically ill-posed, non-linear, and highly sensitive to noise. Recent advances in artificial intelligence offer a promising pathway to overcome these barriers. Deep learning models can act as ultra-fast surrogates for physics-based simulations, enabling near-instantaneous prediction of ultrasonic responses. They can also provide powerful inverse mapping capabilities that directly link measurements to underlying material states. This PhD project will develop AI-driven ultrasonic methods for materials evaluation, combining surrogate modelling and inverse characterisation. The student will work across the full research pipeline, including building high-fidelity ultrasound simulations to generate rich training datasets, designing and benchmarking machine learning architectures, and validating the developed models against experimental measurements. Application targets will include metals, layered and composite structures, and additively manufactured components. The successful candidate will join a vibrant and growing research centre at the University of Sussex, with access to state-of-the-art ultrasonic instrumentation, high-performance computing resources, and an active network of academic and industrial collaborators. The student will be trained in advanced ultrasonic theory, large-scale numerical si