The Organization and Collection of System’s Data in Fusion

Abstract

Future fusion power plants and experiments may have low availability due to complexity of equipment employed. This low availability translates into high cost of electricity (COE) for power plants and long mission times for experiments, i.e. low performance. In order to analyze the availability of systems, we have devised a Monte Carlo computer program that simulates plant operations and maintenance.^1-3 This enables us to predict the system’s performance, run parametric and sensitivity studies and make recommendations as to the system design and criteria that must be met in reliability and maintenance downtimes. However, data (especially reliability data) is nonexistent for many subsystems of a fusion plant.⁴ We’ve had to rely on experts’ opinion and estimates published in some reports. With fusion experiments operating around the world, we can obtain some data from real experience. Before collecting this data, we need to know which data we need for our analysis and where that data can be stored, updated and easily accessed. This is accomplished by means of our Fusion Systems’ Data Base (FUSEDATA). To collect raw data in the field, however, it is more convenient to use a component based collecting data base. For example, the CREDO framework, in use in the U.S. advanced reactor facilities has been used to collect data from the Tritium Systems Test Assembly (TSTA) at the Los Alamos National Laboratory.