This study investigates the dynamic behavior of a Vacuum Packed Particles Torsional Damper (VPPTD) filled with a 1:3 plastic (ABS - Acrylonitrile Butadiene Styrene) and rubber (NBR - Nitrile Butadiene Rubber) granulate mixture. Experimental results revealed that the system exhibits a symmetric hysteresis loop, with maximum torque increasing systematically with applied underpressure. The response was characterized by viscous damping, linear and nonlinear stiffness, and friction-related torque. The main novelty of the paper is the implementation of the Physics-Informed Neural Network (PINN) to accurately identify model parameters based on experimental data. In addition, to final parameter identification, a detailed analysis of the training process was conducted, revealing how the model progressively converged toward the experimental hysteresis loop. The identified model showed good agreement with measured data, and theoretical model. Parametric trends revealed a near-linear dependence of all model parameters on underpressure. These findings suggest that the model can be generalized by expressing parameters as functions of underpressure, paving the way for adaptive, pressure-aware control strategies.