The availability of a fusion power plant is dependent on the frequency and
duration of both planned and unplanned outages. Planned outages result from
a maintenance plan for major component replacement, and are entirely predi
ctable once divertor and blanket lifetimes are established. Unplanned outag
es result from failures of one or more of many diverse components, and can
be estimated only by a probabilistic approach taking into account distribut
ions of failure frequencies and repair times. A Monte Carlo model, PAMPAS,
has been developed which will enable studies of new conceptual power plant
designs once sufficient design detail and component reliability data are av
ailable. It can already be used for more generic studies and results are sh
own which give insight into the magnitude of the problem of achieving adequ
ate plant availability. A simplified model of plant availability has been i
ncorporated into the PROCESS systems code, in order to assess the economic
impact of plant unavailability Initial results show a strong influence on o
ptimum plant parameters dependent on the speed of maintenance operations. T
his work is continuing as part of preparations for a European fusion power
plant conceptual study. (C) 2000 Elsevier Science B.V. All rights reserved.