Born–Jordan frequency distribution based E-GMANN controller for vibration analysis in SSI based asymmetric floor buildings with ATMD

The Active Tuned Mass Damper (ATMD) system minimizes seismic vibrations in asymmetric floor buildings but faces challenges in precise frequency matching and optimal control under dynamic conditions. To address these issues, a novel Born Frequency Distribution-Ebola Neural Network Controller with Frequency Cyclic Linear Analysis is proposed. The ATMD design struggles with aligning damper and building frequencies during dynamic shifts, leading to mode coupling. So, a Born-Jordan Vibration Dose-based Eigensystem Frequency Distribution is introduced for optimal frequency matching and isolating Sole Causative Factors (SCF), thus enhancing the building's resilience during earthquakes. Existing control algorithms often focus on short-term metrics, ignoring peak amplitude and long-term structural integrity, which leads to sub-optimal control force estimation. To overcome this, an Ebola-optimized Graph Multivariable Adaptive Neural Network (E-GMANN) is utilized to dynamically estimate the damper mass and control force in real-time, and improves KPI tracking and system responsiveness. Moreover, Soil-Structure Interaction (SSI) in asymmetric buildings leads to liquefaction, twisting vibrations, and poor structural resistance. Thus, Frequency Domain-Cyclic Equivalent Linear Analysis (TD-CELA) is employed, which addresses non-linear soil behavior, detects moisture-induced variations, thereby enhancing the deformation prediction. As a result, the suggested model outperforms existing methods, significantly reducing vibration, acceleration, and displacement.