Reliability and availability considerations in the RF systems of ATW-class accelerators

In an RF-driven, ion accelerator for waste transmutation or nuclear material production, the overall availability is perhaps the most important specification. The synchronism requirements in an ion accelerator, as contrasted to an electron accelerator, cause a failure of an RF source to have a greater consequence. These large machines also are major capital investments, so the availability determines the return on this capital. RF system design methods to insure a high availability without paying a serious cost penalty are the subject of this paper. The overall availability goal in our present designs is 75% for the entire ATW complex, and from 25 to 35% of the unavailability is allocated to the RF system, since it is one of the most complicated subsystems in the complex. The allowed down time for the RF system (including the linac and all other subsystems) is then only 7 to 9% of the operating time per year, or as little as 613 hours per year, for continuous operation. Since large accelerators consume large amounts of electrical power, excellent efficiency is also required with the excellent availability. The availability also influences the sizes of the RF components: smaller components may fail and yet the accelerator may still meet all specifications. Larger components are also attractive, since the cost of an RF system usually increases as the square root of the number of RF systems utilized. In some cases, there is a reliability penalty that accompanies the cost savings from using larger components. We discuss these factors, and present an availability model that allows one to examine these trade offs, and make rational choices in the RF and accelerator system designs