IPAC2012 - List of Authors (Mukherjee, A.)

Paper	Title	Page
WEPPP081	Fast Beam Tuning for Accelerator Driven Systems	2897
	S. Bhattacharyya, R.K. Yedavalli Ohio State University, USA A. Mukherjee Fermilab, Batavia, USA
	The biggest challenge for Accelerator Driven Systems (ADS) is the stringent availability requirement of >99% compared to ~80% achieved by a typical accelerators. In addition to overall availability, due to thermal stress problems, ADS is also sensitive to the length of each downtime. A significant source of downtime is re-adjustment – “tuning” – of the system to account for drift in component behavior, or substitution of a backup device for one that failed. Tuning at present is done “by hand,” i.e. with human observation, interpretation, and decision, a process which takes hours; whereas ADS requires recovery in minutes. In this research, we apply intelligent controls in a (simulated) proton linac to automate fine-tuning. Beam monitor data is fed into a controller which adjusts magnet currents and RF power to minimize beam loss. We consider fluctuations in ion source characteristics; drift in magnet behavior (mechanical motion, or change in calibration); and failure of an accelerating cavity.

THPPR031	Reliability Modeling Method for Proton Accelerator	4035
	S. Bhattacharyya, R.K. Yedavalli Ohio State University, USA J.S. Kerby, A. Mukherjee Fermilab, Batavia, USA
	Reliability Analysis is an essential part of designing any complex system in order to predict performance and understand availability. However modeling complex systems has been a challenging task due to the large number of components and inter-dependencies. The options have been custom written simulation packages, requiring large investment of programming and debugging time; or standard commercial software running for many days. In our research we developed a hierarchical method to represent the reliability model of “Project X,”* a proposed linear accelerator at Fermi National Accelerator Laboratory. The system is first divided into subsystems small enough to readily simulate. Each subsystem is then separately simulated and parameterized so they can be represented as simple blocks in the top level system diagram. This allows standard, commercial software to model systems with many tens of thousands of components without requiring many days of computer time. Simulation were run and compared with data gathered from existing accelerators. * S.D. Holmes, "Project X: A Multi-MW Proton Source at Fermilab," Proc. of IPAC’10, TUYRA01, p. 1299 (2010).

Paper

Title

Page

Fast Beam Tuning for Accelerator Driven Systems

2897

S. Bhattacharyya, R.K. Yedavalli
Ohio State University, USA
A. Mukherjee
Fermilab, Batavia, USA

The biggest challenge for Accelerator Driven Systems (ADS) is the stringent availability requirement of >99% compared to ~80% achieved by a typical accelerators. In addition to overall availability, due to thermal stress problems, ADS is also sensitive to the length of each downtime. A significant source of downtime is re-adjustment – “tuning” – of the system to account for drift in component behavior, or substitution of a backup device for one that failed. Tuning at present is done “by hand,” i.e. with human observation, interpretation, and decision, a process which takes hours; whereas ADS requires recovery in minutes. In this research, we apply intelligent controls in a (simulated) proton linac to automate fine-tuning. Beam monitor data is fed into a controller which adjusts magnet currents and RF power to minimize beam loss. We consider fluctuations in ion source characteristics; drift in magnet behavior (mechanical motion, or change in calibration); and failure of an accelerating cavity.

THPPR031

Reliability Modeling Method for Proton Accelerator

4035

S. Bhattacharyya, R.K. Yedavalli
Ohio State University, USA
J.S. Kerby, A. Mukherjee
Fermilab, Batavia, USA

Reliability Analysis is an essential part of designing any complex system in order to predict performance and understand availability. However modeling complex systems has been a challenging task due to the large number of components and inter-dependencies. The options have been custom written simulation packages, requiring large investment of programming and debugging time; or standard commercial software running for many days. In our research we developed a hierarchical method to represent the reliability model of “Project X,”* a proposed linear accelerator at Fermi National Accelerator Laboratory. The system is first divided into subsystems small enough to readily simulate. Each subsystem is then separately simulated and parameterized so they can be represented as simple blocks in the top level system diagram. This allows standard, commercial software to model systems with many tens of thousands of components without requiring many days of computer time. Simulation were run and compared with data gathered from existing accelerators.
* S.D. Holmes, "Project X: A Multi-MW Proton Source at Fermilab," Proc. of IPAC’10, TUYRA01, p. 1299 (2010).