GAVIN PIKES; Monte Carlo Simulations in the Modelling and Optimisation of Linac Bunker Shielding

Описание: GAVIN PIKES
Monte Carlo Simulations in the Modelling and Optimisation of Linac Bunker Shielding

Supervisors:
Dr. David Pfefferlé
Dr. Pejman Rowshanfarzad

Abstract
Introduction: Monte Carlo modelling allows for an efficient method of numerical integration through the use of random number sampling. As a result, it becomes ideal for the determination of linear accelerator bunker shielding capabilities where non-standard geometry and beams are in use, meaning the traditional manual calculations provided by the National Council on Radiation Protection and Measurements become less useful. An accurate knowledge of the bunker shielding is essential for ensuring that the dose rate outside the bunker remains at an acceptable level for the safety of both radiation workers and general staff/patients/public.
Methods: The aim of the project is to develop a realistic model of the Marri linac head and surrounding bunker at Sir Charles Gairdner Hospital in GATE, a Monte Carlo simulation toolkit developed for medical purposes in Geant4. The data produced will was tested against measured results to verify the model’s accuracy, before modifications were made to features of the bunker (including wall thickness, maze geometry, type of shielding concrete and addition of new shielding layers) to improve potential weak spots or make recommendations for a bunker design of increased cost/space efficiency.
Results: Results show Monte Carlo simulated data of tenth value layers and percentage depth doses within the primary barrier and water phantom respectively to be within 2% of the measured/documented values. Additionally, a comparison between the dose rates present within the bunker due to scattered radiation provides a high level of agreeance between the simulated and measured values, further validating the model. Overshielding was shown to be present in the primary and secondary barriers by 1 HVL, allowing for a possible cost reduction of $295,000 AUD for future bunker designs. Changes to the maze geometry were shown to decrease the dose to the maze entrance, while higher density bunker materials allowed for increased photon attenuation in the walls. Additional shielding layers were shown to allow for substantial reductions in required wall thickness for steel and concrete/lead layering, while lead layering also reduced the dose to the maze entrance by up to 88%.
Conclusion: Monte Carlo methods can provide a high level of accuracy in determining the shielding capabilities and energy distributions within linac bunkers and may act as a tool for improving the efficiency of their design. This project has established some areas for potential improvement, although further research will need to be conducted for a properly optimised bunker design.