19-03-2026 - Lecture Hall H

PhD Presentation

Towards efficient multi-dimensional operationally robust grab design optimisation

Mikhail Anisimov

Bulk materials such as iron ore, coal, and limestone are needed worldwide, and the vessels transporting them require fast and efficient unloading to reduce demurrage costs and improve logistics. However, there are limited opportunities when it comes to the optimisation of the operations themselves, leaving the grab design as the main lever affecting performance. Therefore, high-performing grab designs delivering consistently large payloads at each stage of unloading are essential. Still, their design optimisation remains challenging due to high simulation costs and highly variable operational conditions.
This project explores the use of machine learning to guide the optimisation of grab designs more efficiently. The aim is to reduce the number of required simulations while extracting valuable engineering knowledge from the process. Attention is given to material conditions—especially the evolving surface of the bulk material—and how these can be incorporated into the optimisation framework to deliver grab designs with robust performance under real operational conditions.