Search NASA⌕ Search

Engineering topics

Ostiguy, J -F

Publications and source records attributed to Ostiguy, J -F.

Study of Effects of Failure of Beamline Elements and its Compensation in CW Superconducting Linac

Project-X is a proposed high intensity proton facility to be built at Fermilab in United States. The first stage consists of a superconducting linac (SC) operating in continuous wave (CW) mode to accelerate a H{sup -} beam from 2.1 MeV to 3 GeV. Failure of any beamline element during operations induces a downstream mismatch of the beam which is especially severe when the failure occurs at low energy. A large mismatch causes emittance growth and ultimately results in beam losses. In a worst case scenario, the operability of the machine may be affected and long downtime may be needed to replace the failed element. To minimize possible downtime, the optics can be designed in a way that allows local retuning to make the machine operable. This paper presents studies performed to investigate retuning scenarios after failure of an accelerating cavity or a focusing magnet at critical locations in the Project-X CW superconducting linac.

43 PARTICLE ACCELERATORS↗

PXIE Optics and Layout

The Project X Injector Experiment (PXIE) will serve as a prototype for the Project X front end. The aim is to validate the Project-X design and to decrease technical risks mainly related to the front end. The paper discusses the main requirements and constraints motivating the facility layout and optics. Final adjustments to the Project X front end design, if needed, will be based on operational experience gained with PXIE.

43 PARTICLE ACCELERATORS↗

Analysis of beam loss mechanism in the Project X linac

Minimization of the beam losses in a multi-MW H{sup -} linac such as ProjectX to a level below 1 W/m is a challenging task. The impact of different mechanism of beam stripping, including stripping in electric and magnetic fields, residual gas, blackbody radiation and intra-beam stripping, is analyzed. Other sources of beam losses are misalignements of beamline elements and errors in RF fields and phases. We present in this paper requirements for dynamic errors and correction schemes to keep beam losses under control.

43 PARTICLE ACCELERATORS↗

Stability of Flat Bunches in the Recycler Barrier Bucket

We examine the stability of intense flat bunches in barrier buckets used in the Fermilab Recycler. We consider some common stationary distributions and show that they would be unstable against rigid dipole oscillations. We discuss the measurements which identify stable distributions. We also report on experimental studies on the impact of creating a local extremum of the incoherent frequency within the rf bucket. We considered two typical stationary distributions and found they were not adequate descriptions of the Recycler bunches. From the measured line density distribution we find (a) the tanh function is a good fit to the line density, and (b) the coherent frequency of the rigid dipole mode for this distribution is within the incoherent spread at nominal intensities. Stability diagrams when the beam couples to space charge and external impedances will be discussed elsewhere. Our initial experimental investigations indicate that longitudinal stability in the Recycler is, consistent with expectations, influenced by the ratio T{sub 2}/(4T{sub 1}) which determines the location of the extremum of the incoherent tune. The coherent tune depends strongly on the distribution in the bunch tails which is difficult to measure. Numerical studies using both a conventional tracking code and a Vlasov solver are in progress and should provide more insight into conditions that may lead to unstable behavior.

43 PARTICLE ACCELERATORS↗

Recent Improvements to CHEF, a Framework for Accelerator Computations

CHEF is body of software dedicated to accelerator related computations. It consists of a hierarchical set of libraries and a stand-alone application based on the latter. The implementation language is C++; the code makes extensive use of templates and modern idioms such as iterators, smart pointers and generalized function objects. CHEF has been described in a few contributions at previous conferences. In this paper, we provide an overview and discuss recent improvements. Formally, CHEF refers to two distinct but related things: (1) a set of class libraries; and (2) a stand-alone application based on these libraries. The application makes use of and exposes a subset of the capabilities provided by the libraries. CHEF has its ancestry in efforts started in the early nineties. At that time, A. Dragt, E. Forest [2] and others showed that ring dynamics can be formulated in a way that puts maps rather than Hamiltonians, into a central role. Automatic differentiation (AD) techniques, which were just coming of age, were a natural fit in a context where maps are represented by their Taylor approximations. The initial vision, which CHEF carried over, was to develop a code that (1) concurrently supports conventional tracking, linear and non-linear map-based techniques (2) avoids 'hardwired' approximations that are not under user control (3) provides building blocks for applications. C++ was adopted as the implementation language because of its comprehensive support for operator overloading and the equal status it confers to built-in and user-defined data types. It should be mentioned that acceptance of AD techniques in accelerator science owes much to the pioneering work of Berz [1] who implemented--in fortran--the first production quality AD engine (the foundation for the code COSY). Nowadays other engines are available, but few are native C++ implementations. Although AD engines and map based techniques are making their way into more traditional codes e.g. [5], it is also probably fair to say that map-oriented codes are still perceived as a specialized niche.

43 PARTICLE ACCELERATORS↗