UPS Availability and Resilience

How enhanced module availability and resilient system design combine to maximize perceived UPS power availability

UPSs exist to deliver continuous, uninterrupted clean power to their critical load. An ideal UPS system is one that delivers this power for the duration of a load's demand, with zero time lost for maintenance or failures. From the user's point of view, this is 100% power system availability. No real UPS system can achieve 100%, but the advanced high availability design of the PowerWAVE9000DPA comes very close, with a figure of 99.9999% or better. This is achieved using two strategies - optimizing availability at UPS module level, and designing resilience to failure into the system.

This can be explained by considering two PowerWAVE9000DPA configuration examples. In the first, a single 50kVA module is used to supply a 40kVA load. To maximize availability of clean power to the load, the only option is to maximize the availability of the UPS module. At the module design level, a more rigorous definition of availability is used, where availability is calculated as the relationship between Mean Time Between Failures (MTBF) and Mean Time To Repair (MTTR). For example the PowerWAVE9000DPA modules achieve a dramatic increase in availability from their 'hotswap' feature. A module can be exchanged for repair or maintenance in less than half an hour, compared with the 6 hours MTTR typically required by a traditional standalone system. A simple calculation shows that this alone can yield an availability of 99.9999% (6 nines), compared with a traditional UPS value of 99.99952% (5 nines). This is the availability that the load enjoys, subject to that of the mains power supply and, if applicable, the standby generator.

In the second configuration example, redundant capacity is added for enhanced protection of a highly critical load. The load is 140kVA and it is supplied by a PowerWAVE9000DPA rack containing four 50kVA modules. This is known as an n+1 redundant configuration, where n=3 in this case. Because any three of the four modules provide more than enough capacity for the load, uninterrupted power to the critical load will survive the failure of any one module.

From the user's point of view, and intuitively, this increases the system's availability. Its uptime becomes a greater proportion of its operational time. However if a module does fail then the system has suffered a failure requiring attention even though the load enjoys uninterrupted power. The addition of redundant modules does not of itself reduce MTTR or increase MTBF. In fact MTBF increases slightly due to the additional number of components in the redundant configuration. Instead, the addition of redundant modules is said to provide the UPS with resilience, because it allows the UPS to continue delivering power to the load even in the event of a module failure. This resilience, by shielding the user from the effects of a module failure, allows him to perceive an availability of uninterrupted power which is higher than - but still based on - the calculated availability of the UPS modules.

To summarise, the above shows that an n+1 redundant PowerWAVE9000DPA system achieves a power availability to the user of better than 99.9999% primarily through the high availability of its hot swap modules but also through the resilience to failure of its redundant design. However it also shows that reduced MTTR is the biggest factor in maximizing availability at both module and system level. That is why hot swap modules with their huge reduction in MTTR are so important in maximizing the availability ultimately enjoyed by the UPS user.