An efficient point kernel method for rapid dose evaluation in dynamic virtual environment.

Area of Science:

• Computational physics
• Radiation dosimetry
• Virtual environment simulation

Background:

• Accurate gamma radiation dose evaluation is vital for radiation safety.
• Point kernel calculations (PKCs) are efficient but slow for complex 3D models and large datasets.
• Existing methods struggle with arbitrary 3D geometries and high computational loads.

Purpose of the Study:

• To develop an efficient computational approach for rapid gamma radiation dose calculation in dynamic virtual environments.
• To create a high-performance point kernel code (PLSPK) capable of handling arbitrary 3D models.
• To accelerate large-scale PKCs using graphics processing unit (GPU) parallel computation.

Main Methods:

• Utilized triangular mesh models for arbitrary shape geometric representation.
• Implemented maximised-parallelism bounding volume hierarchy for optimized ray tracing.
• Leveraged GPU parallel computation for significant acceleration of PKCs.
• Validated accuracy against established point kernel codes and MCNP Monte Carlo simulations.

Main Results:

• PLSPK achieves substantial speedups (approx. 2000× at 1E+06 PKCs) compared to serial implementations.
• The code efficiently handles complex 3D scenes with minimal impact from complexity.
• Demonstrated reliable dose estimations comparable to existing methods.
• Enabled continuous, rapid updates of radiation fields in dynamic virtual environments.

Conclusions:

• PLSPK is a high-performance tool for efficient gamma radiation field calculations in complex virtual environments.
• The developed approach addresses the limitations of traditional PKCs for arbitrary geometries and large-scale computations.
• PLSPK meets the demands for real-time radiation safety assessment in dynamic virtual settings.