Achieving a realistic design for a superconducting gantry with large momentum acceptance for proton therapy

Abstract

Purpose:A systematic design procedure for a superconducting gantry is presented. To increase the momentum acceptance of the gantry beamline, alternating-gradient combined-function magnets were used to suppress dispersion. Methods:The main difficulties and challenges occurring in the design procedure, such as the difference between low- and high-order optics and the difference in beam behavior between simplified optics models and realistic magnetic fields, were analyzed. On the basic of these analyses, a step-by-step iterative design approach was proposed. In these steps, a genetic algorithm (GA) was used for multi-objective optimization, and graphics processing unit (GPU) acceleration was adopted for realistic model simulation. Results:The results of the design demonstrated that momentum acceptance of −7%∼5% was achieved for particle tracking in realistic magnetic fields. Conclusions:Using a GA, multi-objective optics fitting up to the fifth order became feasible. The optimization procedure was also more intelligent compared with the traditional one, which depends on experience and manual intervention. The use of GPU acceleration in magnetic field calculation and particle tracking dramatically reduced the overall optimization time. Finally, design optimization for a large momentum acceptance gantry beamline with realistic magnetic fields was achieved.