Читать книгу Maintaining Mission Critical Systems in a 24/7 Environment - Peter M. Curtis - Страница 50
3.6 Quantifying Reliability and Availability
ОглавлениеData center reliability ultimately depends on the organization as a whole weighing the dangers of outages against available enhancement measures. Reliability modeling is an essential tool for designing and evaluating mission critical facilities. The conceptual phase, or programming, of the design, should include a full Probabilistic Risk Assessment (PRA) methodology. The design team must quantify performance (reliability and availability) against cost in order to push fundamental design decisions through the approval process.
Reliability predictions are only as good as the ability to model the actual system. In past reliability studies, major insight was gained into various electrical distribution configurations using IEEE Standard 493‐2007 Recommended Practice for the Design of Reliable Industrial and Commercial Power Systems, or, The IEEE Gold Book. It is also the major source of data on failure and repair rates for electrical equipment. There are, however, aspects of the electrical distribution system for a critical facility that differ from other industrial and commercial facilities. Therefore, internal data accumulated from the engineer’s practical experience is needed to complement the Gold Book information.
Reliability analysis with PRA software provides a number of significant improvements over earlier, conventional reliability methods. The software incorporates reliability models to evaluate and calculate reliability, availability, unreliability, and unavailability. The results are compared to a cost analysis to help reach a decision about facility design. The process of evaluating reliability includes:
Analyzing any existing systems and calculating the reliability of the facility as it is currently configured.
Developing solutions that will increase the reliability of the facility.
Calculate the reliability with the solutions applied to the existing systems.
Evaluate the cost of applying the solutions.