We are excited to share our latest research on advancing the design of shell structures under challenging loading conditions. Traditionally, form-finding methods for masonry vaults focus on vertical loads, but extreme wind or seismic forces introduce significant design and safety complexities.
Our new paper explores an alternative to computationally intensive optimisation steps in Membrane Equilibrium Analysis. Using machine learning regression techniques—XGBoost, Random Forests, and k-Nearest Neighbours—we identify optimal Airy Stress Function parameters to improve efficiency and maintain structural integrity.
Case study results show that these methods can reduce computational demands while achieving material-efficient designs, with k-Nearest Neighbours delivering the best performance in our tests.
Read more about our findings here: https://lnkd.in/eEnu8sEW
Publication: Digital guidework for augmented thin-tile vaulting construction
Masonry vaults are among the most elegant and efficient structural forms, yet their construction has long been slowed down by costly falsework and guidework.
Our team explored a new path: augmented reality as digital guidework. Instead of rigid scaffolding, masons can now follow holographic projections that provide just the right amount of visual support—keeping builders in control of their analog craft.
In field tests, this approach improved productivity by ~30% while achieving remarkable accuracy (within 1% of the vault span). Looking ahead, interactive mixed-reality could further boost precision, speed, and even training opportunities.
This work shows how centuries-old craftsmanship and cutting-edge technology can merge to keep masonry vaulting not just viable, but visionary. Read more about our findings here https://lnkd.in/d4-KcCxS
