As the data center market continues to grow, the trend to improve efficiency, increase reliability, decrease the footprint for critical infrastructure, and reduce capital costs continues to be a focus for data center engineers. In years past, many of these goals were compromised to achieve others. As reliability increased, so did infrastructure costs. Increased efficiency resulted in decreased reliability, as the higher-efficiency UPS systems typically operated in a line-interactive mode (economy mode). Using today’s technologies, many of these objectives can be met without compromise by applying proper design engineering and following a few key rules: keep it simple, work as a team, conduct factory testing and consolidate the individual pieces into integrated sections.
KISS: Keep It Simple, Stupid
We have all heard this saying many times, but it especially rings true for data center infrastructure designs. As technology improves, many more-complicated options using power tie systems and system synchronization become viable. Unfortunately, when the critical power infrastructure becomes complicated, it can also become difficult to manage. Keeping designs simple reduces operator error and adds reliability. For example, some data centers use closed transition (Diagram 1) transfers between the generator and the utility for times of planned switches. This approach adds a layer of difficulty and introduces a point of failure to the system. The system’s UPS is designed to bridge the gap during this type of transfer. Creating an open transition (Diagram 2) simplifies the process and allows the UPS to make the transition. This is one of many examples where the design becomes more complicated and ultimately reduces reliability.
Simple solutions include isolated A/B distribution systems, each operating in an N+1 configuration. The more isolated each distribution system, the more reliable the solution will become. Since most loads can operate with redundant power supplies, this solution is simple yet reliable for most data centers. For those critical loads that have a single power supply, however, the A and B systems will need to merge at some point to offer redundancy. This merging is typically accomplished through static transfer switches at some point after the UPS.
To read this article in its entirety please click here