SLAC, Menlo Park, California, USA
SLAC, Menlo Park, California, USA
Simulation of thermal effects in accelerator cavity is normally performed assuming steady state solution where a static thermal solver suffices to evaluate temperature gradients and impacts on mechanical design. However, during the rf pulse ramp up or the machine system cool-down process, when the field in the cavity changes rapidly, transient effects need to be taken into account. A parallel time domain thermal solver has been developed in the finite element multi-physics code suite ACE3P with integrated electromagnetic, thermal and mechanical modeling capabilities. The implementation takes advantage of the parallel computation infrastructure of ACE3P and shares most of the ingredients in mesh generation, matrix assembly, time advancement scheme and postprocessing. In this paper, we will outline the finite element formulation of the transient thermal problem and verify the implementation against analytical solutions and existing numerical results. The thermal solver has also been coupled to ACE3P mechanical solver, allowing stress and strain analysis during the transient stage. Application of the transient thermal solver to realistic accelerator cavities will be presented.
LBNL, Berkeley, California, USA
SLAC, Menlo Park, California, USA
Photo injector design is an important consideration in the construction of next-generation accelerators. In current injector optimization, components (e.g. RF cavities) are individually shape-optimized for performance subject to requirements such as peak surface field, shunt impedance, and resonant frequency. Once these component shapes are determined, beam dynamics simulations optimize the injector lattice by adjusting parameters such as the amplitude and phase of the driving fields. However, this form of beam dynamics optimization is restricted by the fixed geometry and field profile of the components. To optimize accelerator design more generally, a coupled optimization of the cavity shape and beam parameters is required. For this coupled optimization problem, we have created an integrated ACE3P-IMPACT workflow. Within this workflow, component geometries are adjusted, field modes are computed with Omega3P (a module in the ACE3P suite), and beam dynamics are simulated with IMPACT-T. This workflow is encapsulated into a multi-objective optimization algorithm using the DEAP* and libEnsemble** Python libraries to yield a Pareto-optimal set of solutions for a simple injector model.
* F.-A. Fortin et al, DEAP: Evolutionary Algorithms Made Easy, J Mach Learn Res, 13, 2171-2175, July 2012
** S. Hudson et al, libEnsemble User Manual, Argonne National Laboratory, Rev 0.7.1, 2020
