Room: Exhibit Hall | Forum 6
Purpose: To investigate the effect of modeling considerations on accuracy of magnetic fluid hyperthermia simulations and to optimize the predictive power of simulations through a dynamic approach that implements the dependence of volumetric power dissipation on temperature.
Methods: Magnetic fluids of various concentrations of magnetite nanoparticles (10 - 60 mg ml-1) with size 8.6 Â±2.9 nm, were prepared and exposed to an external alternating magnetic field with frequency 280.5 kHz and amplitude 19.4 kA m-1. From the first temperature derivative of each experimental heating curve, the respective volumetric power dissipation was estimated using the initial slope criterion. A numerical model was developed in COMSOL Multiphysics 5.2a, employing the â€œnon-isothermal-flowâ€? module that couples the temperature elevation and the stirring of the magnetic fluid to simulate the heating process. Time-dependent studies were carried out, exploiting parametric scanning to evaluate and refine the experimentally derived values of the volumetric power dissipation. Subsequently, an optimization study incorporating Gaussian filtering was leveraged to define the volumetric power dissipation as a function of temperature. The optimization solver utilized the SNOPT algorithm to determine the best fitting of the simulated heating curves to the experimental data, with respect to the least-squares method.
Results: Our simulations demonstrate that the initial slope method underestimated the heating efficiency of the examined samples. Depending on the sample concentration the temperature relative error ranged from 9% to 25%. The suggested optimization method reduced the relative error significantly, ranging from 2% to 7%.
Conclusion: This study indicates the opportunity in optimizing magnetic fluid hyperthermia simulations, through numerical models that consider the dependence of heating efficiency on the prevailing thermal energy.