No abstract

Aurora is a short-pulse high-power krypton-fluoride laser system that serves as an end-to-end technology demonstration prototype for large-scale ultraviolet laser systems of interest for short wavelength inertial confinement fusion (ICF) studies. The system is designed to employ optical angular multiplexing and serial amplification by electron-beam-driven KrF laser amplifiers to deliver 248 nm, 5-ns duration multi-kilojoule laser pulses to ICF targets using a beam train of approximately 1 km in length.In this paper, we will discuss the goals for the system and summarize the design features of the major system components: front-end lasers, amplifier train, optical train, and the alignment and controls systems.

TheRF controlsystemfor a charged-particle acceleratormustmaintainthe correctamplitudeand phaseof the RF field insidethe acceleratorcavity in the presenceof perturbations, noises,and time varying systemcomponents. For an acceleratorwith heavybeam-loading, fluctuationsin thebeamcurrentcauselarge perturbationsto the RF field amplitudeand phase that must be correctedby the RF control system.The ADTI' applications require a high-current, heavily beam-loaded, continuous-wave(CW) accelerator. Additional concernscreated by the CW operation include system start-up, beam interruption, and fault recovery. Also, the RFcontrolsystemfor an ADTI"facilitymustincludesophisticatedautomationto reduce the operator interactionand support.This paperdescribesan RF controlsystemdesign that addressesthese variousissuesbyevaluationa combinationof feedbackand feedforwardcontroltechniques. Experiencefrom the high-currentGround Test Accelerator (GTA) is drawn upon for this RF control system design. Comprehensivecomputermodelingwith the Matrixxsoftwarehas been usedto predictthe performanceof this RF controlsystem.

Recent work has demonstrated the ability of neural-network-based controllers to optimize and control machires with complex, non-linear, relatively unknown control spaces. We present a brief overview of neural networks via a taxonomy illustrating some capabilities of different kinds of neural networks. We present some successful control examples, particularly the optimization and control of a small-angle negative ion source.

No abstract