Citation

Sungheetha, Akey and R., Rajesh Sharma and Yogarayan, Sumendra and Kannan, Subarmaniam (2026) Black Hole Magnetospheric Dynamics: Gravitational Wave Detection and Accretion Flow Optimization. Procedia Computer Science, 283. pp. 5233-5244. ISSN 18770509

Text 20.pdf - Published Version Restricted to Repository staff only Download (474kB)

Abstract

Black hole magnetospheric dynamics represent fundamental phenomena governing energy extraction mechanisms and gravitational wave generation in extreme gravitational environments. Current observational frameworks face significant challenges in accurately modeling the coupled magnetohydrodynamic-relativistic dynamics within ergospheric regions where frame-dragging effects dominate electromagnetic field configurations. This research addresses critical limitations in existing gravitational wave detection methodologies by proposing an integrated computational framework combining adaptive deep reinforcement learning with hybrid physics-informed neural networks for real-time accretion flow optimization. The proposed Multi-Scale Magnetospheric Analysis System employs hierarchical feature extraction through convolutional architectures processing synthetic magnetohydrodynamic simulation data derived from General Relativistic Magnetohydrodynamic codes. Experimental validation demonstrates accuracy improvements of 94.7 percent in magnetic field topology prediction, computational efficiency gains of 3.8-fold compared to traditional numerical relativistic methods, and gravitational wave signal extraction precision enhancement of 89.3 percent for spin-orbit coupling parameters. The framework achieves temporal resolution of 0.03 milliseconds for transient electromagnetic phenomena near event horizons with energy extraction efficiency quantification accuracy within 2.1 percent deviation from theoretical Blandford-Znajek predictions. Results indicate magnetic flux threading optimization yields 87.6 percent correlation with observed quasi-periodic oscillation frequencies in accretion disk systems, enabling predictive modeling for electromagnetic counterparts to gravitational wave events with false positive reduction of 91.4 percent.

Item Type:	Article
Uncontrolled Keywords:	Physics-informed neural networks, Deep reinforcement learning
Subjects:	Q Science > Q Science (General) > Q300-390 Cybernetics
Divisions:	Faculty of Information Science and Technology (FIST)
Depositing User:	Ms Rosnani Abd Wahab
Date Deposited:	03 Sep 2026 02:32
Last Modified:	03 Sep 2026 02:32
URII:	http://shdl.mmu.edu.my/id/eprint/16582

Downloads

Edit (login required)