FOCUS on POWER

POWER INFRASTRUCTURE

Getting electrical systems right

Cummins’ Richard Hallahan says there are three things you must keep in mind when designing power distribution infrastructure

When determining what power distribution infrastructure will look like, writes Cummins’ senior manager of strategic accounts Richard Hallahan, you have to keep in mind the size of the infrastructure, reliability of the architecture and operational complexity. Hallahan takes a deep dive on each of these factors in his white paper titled Data Center Design Decisions and their Impact on Power System Infrastructure.

SIZE MATTERS Definition of the power infrastructure for the entire facility is closely related to heat generated by the IT equipment. This is because IT and cooling infrastructure are the biggest energy consumers in the data center and their peak consumption defines the facility’s power needs.

Hallahan recommends either of two methods for calculating electrical consumption of IT equipment. The first is multiplying the data center floor’s heat load by its area. Heat load in this equation is expressed in watts per square foot. Thus, a 48,000 sq ft data center floor with heat density of 150 watts per sq ft will consume 7,200,000 watts, or 7.2MW.

The second method is multiplying the heat load of an IT rack by the total number of racks on the floor at full capacity. Using this equation, IT equipment in a 600-rack data center, with capacity at 12kW per rack, will also consume 7.2MW.

Hallahan suggests a simple rule of thumb for using the IT power number to calculate the approximate power the entire data center will consume. Since mechanical systems of

Electrical consumption of IT equipment will influence power infrastructure decisions

a facility often consume about the same amount of power as the IT gear on the raised floor, total data center power can be arrived at by simply multiplying IT power by two.

RELIABILITY The second key factor – reliability architecture – considers which tier of the Uptime Institute’s Tier Classification system for infrastructure reliability the facility will be designed to. The decision to build to one of the four tiers will, to a large extent, dictate the power infrastructure’s design. At Tier II, for example, some components of the power system will be redundant. The reliability architecture further defines power system infrastructure. The same Tier II data center will most likely have paralleling switchgear to integrate both generators into the infrastructure instead of using a transfer switch. The higher the chosen reliability tier, the more expensive the system will be.

Another aspect that influences power system architecture and the facility’s initial cost is the

choice of operating voltage. Since there is no rule of thumb for determining what cost a certain operating voltage will translate to, Hallahan recommends comparing the cost of a number of designs with different operating voltages to determine the optimal one.

THE COMPLEXITY FACTOR Hallahan’s third key factor is the system’s operational complexity. This is closely tied to the second factor of reliability level. A higher tier, for example, will create the necessity to install paralleling switchgear and other automatically controlled devices.

More equipment means a more complex operational sequence for reacting to changes in normal operating conditions. When utility power is lost, the sequence is designed to transfer the data center load to redundant power-path interconnections, which keep the facility running during an outage. The amount of power paths in different systems can be two or more, depending on design complexity.