A new paper co-authored by Ercan Atam presents an intelligent seismic protection approach for multi-story buildings, a high-impact topic for Türkiye as an earthquake-prone country.
The study uses deep reinforcement learning (PPO) to adaptively control semi-active friction dampers, and benchmarks against constant-force and rule-based strategies. In simulations on a three-story building subjected to 114 test earthquakes (PGA > 0.15 g), the PPO-based controller achieves strong average reductions—43% lower cumulative drift and 60% lower cumulative acceleration, alongside improvements in peak displacement/drift/acceleration. For details, please refer to: https://ieeexplore.ieee.org/document/11370190
