Shady Adi
civil engineering and Digital Twin researcher, Global AI in Science Fellow at Imperial College London, and Visiting Lecturer at the University of Hertfordshire.Dr Shady Adib is a civil engineering and Digital Twin researcher, Global AI in Science Fellow at Imperial College London, and Visiting Lecturer at the University of Hertfordshire. He holds a PhD in Civil Engineering from Newcastle University, where he developed a hybrid Digital Twin framework for real-time structural damage identification. His research focuses on advancing intelligent, adaptive and resilient Digital Twin systems for civil infrastructure, with particular emphasis on Self-Generating and Cognitive Digital Twins. His work integrates artificial intelligence, structural health monitoring, physics-based modelling and data assimilation to enable continuous structural assessment, autonomous adaptation and reliable engineering decision-making under uncertain and changing conditions.
At Imperial College London, his research explores the application of Generative Artificial Intelligence (Gen-AI) to reconstruct missing structural monitoring data and enhance the reliability and resilience of cognitive digital twins. His ongoing research explores the application of AI-driven data reconstruction to real-world bridge monitoring, aiming to improve structural assessment, monitoring reliability and infrastructure resilience under both normal operating conditions and extreme events."
My Sessions at Day 2
Thursday 11 March 2027
When Bridge Sensors Go Silent: Using AI to Fill the Gaps in Your Monitoring Data.
Structural health monitoring and digital twins rely on continuous sensor data, but sensor faults and communication outages leave gaps, often when information matters most, such as during floods and earthquakes. Drawing on his I-X Global Early-Career AI Fellowship research at Imperial College London, Dr Shady Adib compares generative AI with conventional techniques for reconstructing missing measurements, and explains how the work is being extended to five years of monitoring data from the Forth Road Bridge.
- How diffusion-based and generative adversarial models compare with linear interpolation and Gaussian Process Regression, tested on shaking-table data from a three-storey timber structure under real earthquake records
- What generative AI can and cannot reliably recover, from short random gaps to prolonged sensor outages
- Applying the approach to the Forth Road Bridge, with uncertainty-aware assessment to support maintenance and decision-making
Panel discussion: Faster, Smarter, Safer? Putting AI to Work on Our Bridges
Much of the talk about AI in engineering is about abstract possibilities, so this panel focuses on what engineers are actually doing with it. Panellists share real results, where AI is helping and where it needs careful handling.
- Practical examples of AI in bridge design, assessment and monitoring
- The risks of inappropriate use and how to keep engineering judgement central
- Skills, cost and environmental questions that come with wider adoption
