No abstract

No abstract

3 through existing network analysis codes, and the simulation of the actual mechanical motion requirements is being developed through computer-sided design (CAD) system pro grams. These are discussed in Sect. 3.2. Finally, the estimation of future failure and repair rates through formal interrogation of experts is discussed in Sect. 3.3.The status and prospects for obtaining data from existing or near-term fusion experi mental facilities were reviewed primarily through discussions with the operators of those devices. Section 4 reports the results of these discussions, including the procedures used and recommended to collect data and the availability of data. Existing fusion programs relevant to the availability program and general attitudes toward the formation of a comprehensive 3ata collection system are abo discussed. CONCLUSIONS AND RECOMMENDATIONSThe review of existing availability calculation models shows that they are sufficient to meet the needs of the fusion program. The models are not used to their potential because the quality of the input data for fusion systems does not warrant sophisticated availability analyses. In the near term, the best source of reliability input data is expected to be expert opinion, and the best source of repair data is expected to be maintenance simulation. In the long term, the development of a comprehensive and validated reliability, availability, and maintainability (RAM) data base will be essential to the fusion program, but the value of data from existing experimental devices is arguable. It is recommended that selec tive collection of experimental project data be initiated as a test of the timeliness and current value of a comprehensive data base. Detailed conclusions and recommendations follow. Availability ModelsExisting availability calculation models are sufficient to meet the needs of the fusion program. The deterministic mode! developed by the Electric Power Research Institute (EPRI) for advanced power system applications, UNIRAM, is recommended for rapid assessment of the effect of changes in design and input data on the overall system availa bility. Two stochastic codes developed specifically for fusion systems, the University of Wisconsin (UW) and the Fusion Engineering Design Center (FEDC) models, are recom mended for more detailed systems analysis. These stochastic codes can assess more com plex operating and maintenance scenarios, and they provide a more realistic model of the plant than the deterministic code does. Both the UW and FEDC codes are in development, and neither has reached the limits of the stochastic model capability. It is recommended that use and development of both continue so that the codes' progress can be reviewed by sample case comparisons. Data RequirementsAvailability models are not routinely used in design studies, and they are not used to their full potential. The reason is that the models are already well advanced compared with the pvailability and quality of the input data. Effective use of the models requires detailed design defin...

No abstract