ASHRAE Journal - June 2021 - 38

TECHNICAL FEATURE
but much of it can directly correlate to the calculation
approach. The level of measurement chosen will impact
the accuracy of the calculation of the PUE during the
planning stages.
Where and How to Improve PUE
The book discussed here focuses on measurement.
Standard 90.4 focuses on efficiency and compliance. It
is not the intent of this article to propose an alternate
method to Standard 90.4. Instead, it proposes using
PUE as an approach for comparison of individual components
and systems as a whole. The transition of PUE
from a metric that is centered around measurement and
monitoring to a metric better suited as a design tool is
straightforward.
An additional consideration in this transition is the
approach of using PUE for electrical infrastructure sizing.
This is a common practice among owners and consultants.
If a good understanding of worst-case infrastructure
efficiencies on an ambient extreme design
day‡ exists, the ability to use PUE to appropriately size
the electrical infrastructure associated with critical IT
capacity exists. The metric which would be (and often is)
used for this purpose is peak power usage effectiveness
(or peak PUE).
(IT Power) +
Peak PUE =


Extreme Day Critical
Systems Overhead Power
IT Power


If a facility has a goal for 90 MW of IT (critical power) at
a peak PUE of 1.6 and 1 MW of ancillary house, site and
general power, the information can be communicated to
the power authority that 145 MW of utility power would
be required.
Utility Power = Critical Power Peak PUE + Ancillary Power
×
This holds true at the block level as well. This means of
sizing electrical infrastructure can be applied to transformers
and generators as well. A 1,500 kW IT load at a
peak PUE of 1.6 would require a 2,500 kW transformer
and a 2,500 kW generator.§ This is just one example of
how diverse the PUE metric is for design professionals.
Consultants are able to simulate energy consumption by
categorizing power values into 8,760 hourly bins per year
of ambient conditions. By calculating total system power
consumption within each bin, then multiplying it by the
number of hours within the respective bin that occur in
a typical year, the total energy expended in a given year
is able to be approximated (Power × Time = Energy). By
dividing total energy over time by the product of the critical
IT load and the total hours in that period, one can
calculate the PUE for that time period. For a year, one can
calculate the annualized PUE with the equation:
Annualized PUE =
Total Energy Consumed
Critical IT Power 8× ,,760 hours
This allows us to simulate energy as a product of
power. Using this means of calculating PUE creates two
useful metrics from the same data set, peak PUE# and
annualized PUE. Both metrics developed by this means
are able to be validated once a facility is operational.
Prescribing this path and paths like it is a first step in
developing a unified metric across the industry.
Equipment manufacturers, particularly mechanical
equipment manufacturers, attempt to use partial PUE
(pPUE) to illustrate the efficiency of their mechanical
systems. By a strict definition of pPUE as described by
the book, this is an incorrect use of the term. Partial PUE
is defined as the total energy consumed divided by the
critical energy consumed within a given boundary in a
mixed-use facility.2 This does not allow for pPUE to be
applied to specific systems or components to illustrate
their contribution to the overall PUE.
To take steps to rectify this common misrepresentation
of the pPUE metric by manufacturers and their
representation, a set of standard calculations needs to
be developed by which standard component efficiencies
are defined or ignored in a prescriptive method. If all
system components are defined as a standard efficiency,
they can be uniformly " swapped " out as individuals
‡The extreme ambient design day is typically defined by owner or engineer preference. ASHRAE extreme conditions are used to define
these figures for various locations throughout the world based on historical weather station data. ASHRAE 20-year extreme and
ASHRAE 50-year extreme are the two most commonly used values; more conservative designs are based around the 50-year extreme.
§Generators and transformers are commonly available in 2,250 kW and 2,500 kW sizes. With 1,500 kW required at a peak PUE of 1.6,
2,400 kW of capacity is required. In this scenario, the next frame size up would be required to meet demand.
#Peak PUE must be derived typically using 20-year or 50-year extreme ambient temperatures, which will not exist in the 8,760 hour data set.
38
ASHRAE JOURNAL ashrae.o rg
J U N E 2021
(1)

https://www.ashrae.org/

ASHRAE Journal - June 2021

Table of Contents for the Digital Edition of ASHRAE Journal - June 2021