RADIASOFT LLC — Department of Energy SBIR Phase I: 23d

Description

Modeling beam loading in RF cavities is currently a difficult supercomputing problem. By developing new hybrid algorithms, we can reduce the beam loading problem to a desktop problem, enabling design of advanced accelerator concepts such as energy recovery linacs and enhanced beam dumps.Using a pseudo-spectral Green’s function type approach, the longest length scales in simulating beam loading in an RF cavity can be removed, a million-fold reduction in the size of the problem. This makes large parameter scans and design efforts reasonably accessible, allowing both studies of beam break-up instability and a design study for an enhanced resistive wall beam pipe for minimizing radioactivation of the beam dump. A pseudo-spectral algorithm will be derived suitable for simulating beam loading using novel variational methods. This algorithm will be benchmarked for accuracy against existing methods, and implemented in a development library. That work will then be used to set up simulations of beam break-up instability in energy recovery linacs and to make preliminary designs for an enhanced resistive wall beam pipe. Designing cavities for radio frequency or microwave devices can require a supercomputer because of the large range of lengths involved. By intelligently removing the largest lengths, these problems can be moved to a desktop, saving considerable costs for research groups and businesses that lack access to supercomputer resources. Commercial Applications and Other Benefits: Developing an open source library for fast, accurate beam loading and cavity modeling will save considerable resources by moving a problem from a supercomputer to a desktop. Additionally, the design of an enhanced resistive wall beam pipe for reducing the power reaching the beam dump will reduce radioactivation. As private companies are averse to generating too much radiation, such a device is critical for industrial applications of high average power accelerators.