Electrical &

Power
FA L L E D I T I O N
Contents
3 — What do you need to know about electrical, power systems
in data centers?

7 — Cummins Power Generation Interruptible Facility

8 — Wastewater Treatment Plant Reduces Emissions with

Cummins Gas Solution

12 — Learn the pros and cons of centralized and decentralized

generator systems

23 — Discover Galaxy VS 3-Phase UPS | Schneider Electric

24 — Driving Climate-Positive Innovation in Data Centers and

Industries: A Sustainable Approach

37 — Your questions answered: What to consider when

designing a paralleling power system

44 — Five more reasons generators fail when you need them

the most

53 — Electrical acceptance testing collaboration leads to

improved project delivery

2
What do you need to know
about electrical, power systems  Back to TOC

in data centers?
Will data centers get larger? More efficient? Use less energy? Learn about
the trends here

Bill Kosik, PE, CEM, BEMP, Senior Energy Engineer, DNV, Oak Park, Illinois – Brian Rener,
PE, LEED AP, Principal, Mission Critical Leader, SmithGroup, Chicago, Illinois – Ameya
Soparkar, Market Leader, Mission Critical, Affiliated Engineers Inc., Rockville, Maryland –
Robert Sty, PE, LEED AP, Vice President, HDR Inc., Phoenix, Arizona
3
What do you need to know about electrical, power systems in data centers?

What are some key differences in electrical, lighting and power systems  Back to TOC
you might incorporate in this kind of facility, compared to other projects?

Brian Rener: On higher density computation facilities or hyperscale facilities there is a

need for extensive medium-voltage distribution systems not typically seen except on
large campuses or health care sites.

Robert Sty: The amount of power density compared to other project types, the com-
plexity of mission critical, uninterruptible power supply (UPS), standby and associated
codes and standards.

How does your team work with the architect, owner’s rep and other
project team members so the electrical/power systems are flexible and
sustainable?

Robert Sty: Some of the most challenging problems are adding on future phases
while the first phase of the data center is operational. If the original design did not
incorporate methods for installing, testing and commissioning major electrical gear
during the basis of design, then there is a risk of interruption of services during the
construction of the latter phases. This takes coordination and should include all project
stakeholders. Using BIM and virtual reality technologies can assist in the planning of
adding major electrical equipment during operation.

Resilient or resiliency is a buzzword when discussing data centers.

What are owners requesting to make the building meet resiliency
goals? How are you designing data centers to be more resilient?
4
 What do you need to know about electrical, power systems in data centers?

Bill Kosik: Depends on the  Back to TOC

type of data center: com-
pute load, storage, on-site,
mirroring. Determines time
duration from beginning
of utility failure to start of
on-site power generation.
Or maybe no on-site gener-
ation, just need time to shut
down. Also, use of energy
storage for cooling. The
resiliency of the compute,
storage and networking
systems will inform the level
of reliability for the systems.
Defining resiliency is specific
to the function of the data
center and understanding
the impact of different levels

CONFIDENCE ON
of reliability. This applies to
multiple aspects of build-
THEY’RE NOT JUST POWER EXPERTS. ing structure, maintenance,
THEY’RE DATA CENTER POWER EXPERTS.
As a top global provider of backup power for the data center industry, Cummins has
geo-redundant.
built the largest dedicated support network. We train teams from around the world as
data center specialists, experts who know how to fine-tune every component of your
Cummins generator sets to your data center’s exact power demands. Cummins data
center specialists work where your data lives, making sure your confidence is always on. Brian Rener: We are look-
TURN YOUR CONFIDENCE ON.

12270_DataCenter_7X24.indd 1 8/12/20 4:31 PM

ing carefully at climate
5
What do you need to know about electrical, power systems in data centers?

change models for  Back to TOC

the next 10-20 years.
This includes design
day temperatures
which may be above
current ASHRAE
standards and for
additional electrical
reliability even if it’s
not officially Uptime
Institute Tier rated.
Also wind and flood-
ing levels. Clients are asking how to make A thermosyphon uses refrigerant and condenser
fans to provide heat rejection for liquid-cooled
the facilities resilient now and in the future
high-performance computing or water-cooled
for climate change. chillers. Its efficiency is like that of cooling towers
with outdoor temperatures generally at about 50°F
or less, but uses no water. Courtesy: SmithGroup
Robert Sty: Balancing uptime with capi-
tal expenditures. The industry has matured
enough to recognize some tried and true electrical architectures, each with own pros
and cons. On-site power generation and sustainability at the grid scale are the next
challenges we expect to face.

Consulting-Specifying Engineer

6
Cummins Power Generation Interruptible Facility

 Back to TOC


Cummins Power Generation
Interruptible Facility
The Cummins Power Generation plant in Fridley, Minnesota
has relied on this integrated power design for decades.

7
 Wastewater
Treatment
Plant Reduces
Emissions with
Cummins Gas
Solution
CUMMINS PROVIDES COGENERATION
SOLUTION AND SOLVES EXCESS GAS
FLARE AT SEWAGE FACILITY
WHERE: North Davis Sewer District (NDSD) collects and treats
Syracuse, Utah (USA)
wastewater from around 80 square miles in North Utah,
serving approximately 200,000 people. The district owns
SUPPLY: and operates around 100 miles of sewer collection lines
2 x C1100N6C, QSK60G gas generator sets which deliver wastewater to the Syracuse, Utah, treatment
facility, located near the shoreline of the Great Salt Lake.
PURPOSE: The facility has the daily capacity to treat around 34 mil-
Upgrade old generators and provide
lion gallons of wastewater, therefore it requires a reliable
cogeneration solution, meeting the customer’s
detailed specification. power solution.
Wastewater Treatment Plant Reduces Emissions with Cummins Gas Solution

The two previous generators  Back to TOC

had been running for a com-
bined average of 157,000 hours,
producing high emissions read-
ings of 19g/hp-hr NOx and is-
sues with flaring excess digester
gas. The crucial requirements of
this upgrade were to tackle the
emissions and excess gas flare
problems.

Upon the completion of the

biding process, Cummins Power Generation was selected to complete the project’s in-
stallation due to the unit’s ability to meet all desired requirements; a max rpm of 1200,
no excess gas flaring issues, and a turndown ratio option. In addition, the generators
utilize lean-burn gas technology
which provide low exhaust emis-
sions. This meant the pre-exist-
ing high NOx emissions would
be replaced with just 0.5gr/hp-hr.
Cummins’ ability to provide local
support throughout the installa-
tion, as well as NDSD’s proxim-
ity to a branch for future spare
parts, cemented the selection
process.
9
Wastewater Treatment Plant Reduces Emissions with Cummins Gas Solution

The QSK60G gas generator range  Back to TOC

offers a fully integrated power
generation system which utilizes
state-of-the-art technology to re-
sult in the optimum performance
and efficient use of fuel needed
for the facility’s cogeneration
needs. The treatment facility,
as a by-product of its digestion
process, produces approximately
300,000 cubic feet of methane
gas per day. This excess gas is
utilized to power the generators which, on average, supplies 50-60% of the plant’s
electrical needs. The excess heat is also harnessed to support the plant’s daily activi-
ties, providing heat to the primary digesters as well as supplying the hydronic heaters
used for all on-site buildings. The
cogeneration solution creates
considerable financial savings for
NDSD as the digesters must be
kept at a constant temperature of
95-100°F in order to maintain the
biological digestion process.

“NDSD chose the two Cummins

generator sets due to their abil-
ity to run at 1200 RPM and have
10
Wastewater Treatment Plant Reduces Emissions with Cummins Gas Solution

the largest turndown  Back to TOC

ratio, while main-
taining emissions. By
converting from the
older naturally aspi-
rated gas engines,
the District lowered
its NOx emissions
from 19grams/BHP
to 0.5grams/BHP,
allowing the facil-
ity to drop below
the threshold to be
considered a Title
5 Plant by Utah’s Air Quality Requirements. Utah has struggled to meet Air Quality
Requirements and the District feels it’s doing its part with the conversion to these
Cummins generators. ”
- Myron Bachman
NDSD Plant Superintendent

11
Learn the pros and cons of
centralized and decentralized  Back to TOC

generator systems
Centralized, remote and distributed standby generator designs are all
options for power systems. Which one is best?

W hen designing or assessing campus-style distribution systems, the topic of cen-

tralized versus decentralized standby power systems is a necessary focus. This
article will explain the various attributes associated with a variety of standby genera-
tor systems. It is intentionally written in a format that starts by making a case for each
arrangement by addressing the positive and beneficial aspects then finishes with the
negative compromises and limitations.

Decentralized standby generators

It is very common for a campus to grow in a manner that results in multiple distributed
standby generator systems spread across the campus both electrically and geograph-
ically. The typical arrangement is an essential electrical system (EES) consisting of one
or more generator sets, serving a defined “building.” Other buildings, usually separat-
ed by age, have a separate EES.

The primary reason for distributed systems is due to the cost and complexity of central-
ized systems and the lack of capacity in the older generator systems to accommodate new
building additions. Many projects lack the budget to overhaul and upgrade the existing
EES. The primary benefit of decentralized arrangements, other than initial cost to each
building, is both electrical and geographic diversity to prevent single points of failure.
12
Learn the pros and cons of centralized and decentralized generator systems

 Back to TOC

If a natural disaster damages the east side of a building Figure 1: Two 600 kW 4,160 V
standby generators were relocated
or campus, the utility and standby power systems may be
to a remodeled space that was
damaged beyond immediate repair. While the portions sized for a third generator and
of the building or campus served will be without power, larger generators. This system is
planned for expansion to operate
the other areas of the building may still sustain power,
as a centralized location in a remote
allowing for continued operation. This approach is pop- utility plant. Courtesy: IMEG Corp.
ular in mission critical and health care campuses where
evacuation and shutdowns are not acceptable.

For campuses that experience frequent growth and renovation, the distributed con-
figuration can be less expensive due to localized and right-sized electrical distribution 13
Learn the pros and cons of centralized and decentralized generator systems

providing a lower cost to feeders and electrical gear. Because the needs of each build-  Back to TOC
ing on a campus may vary, additional savings may be had for systems that do not need
to meet more stringent requirements of Level 1 systems, as defined in NFPA 110: Stan-
dard for Emergency and Standby Power Systems.

For example, a health care campus will likely have Level 1 EES to serve the areas that
are within the International Building Code (IBC) occupancy classification of I-2 (institu-
tional), and may also have IBC occupancy classifications of B (business), A (assembly), S
(storage), etc.

The other occupancy types may be able to have a Level 2 EES as defined by NFPA 110
or may only require emergency loads as defined by NFPA 70: National Electrical Code
(NEC) Article 700 and would not fall under Article 517 or NFPA 99: Health Care Facili-
ties Code, which both govern health care designs.

Providing different levels for EES serving different buildings on a campus may also
allow for different fuel types and storage amounts. In this example, the Level 1 EES
would likely consist of diesel standby generator sets with on-site storage sufficient for
96 hours of operation (with or without refueling during the 96 hours).

A Level 2 EES for a smaller medical office building may be a single natural gas standby gen-
erator with a local temporary connection cabinet. Buildings that do not need an EES at all
could get by with battery backup of emergency and egress lighting and fire alarm systems.

When consulting with a campus, one simple question to ask the client is if the building
can be evacuated at any time, without hesitation. If the answer is yes, then it’s likely the
14
Learn the pros and cons of centralized and decentralized generator systems

 Back to TOC

backup power requirements can be less stringent when Figure 2: Four 600 kW 480/277
V diesel generators located in a
compared to areas that cannot be evacuated.
central utility plant serving 4,000
A paralleling gear in an adjacent
There are, however, negative aspects of a distributed or room. Courtesy: IMEG Corp.

decentralized standby power system and the most glar-

ing one is the lack of shared capacity. For example, if a campus has 4 megawatts (MW)
of total standby power through four 1-MW generator sets and if none of those sets can
be connected in parallel, the available capacity at any point in the system is limited to
1 MW. This can have the effect of increasing costs for future projects, additions or ren-
ovations if the closest and most convenient generator system lacks capacity.
15
Learn the pros and cons of centralized and decentralized generator systems

Additionally, if one generator is taken out of operation for maintenance, the other  Back to TOC
generator sets on campus are not able to back feed, requiring temporary connections.
NEC Article 700.3(F) recognizes the vulnerability of a single generator and requires
installed, temporary connection capability that can be avoided with paralleled genera-
tors of adequate size.

Decentralized generator systems also spread out the equipment causing additional
maintenance hassles and increased footprints. Fuel storage is required for each, fuel
treatment is required for each, battery systems are multiplied and testing is multiplied.
Each generator location needs to be coordinated with outside air intakes for the heat-
ing, ventilation and air conditioning (HVAC) system and noise concerns within the cam-
pus and adjacent properties.

Centralized standby generators

A central utility plant, or CUP, can be located adjacent to the buildings served or can
be remote from the buildings served. The CUP can contain chilled water systems, heat-
ing water or steam systems and electrical distribution systems for both normal (utility)
and standby (emergency) sources.

Centralized electrical distribution can be independent of mechanical systems and lo-

cating them inside of a CUP is just one option. It is common for normal utility sources
to be centralized, typically at the 15 kilovolt (kV) distribution voltage class. This may or
may not be equipment owned by the client, though it is common for the facility to own
and operate the medium-voltage distribution for the normal service. Likewise, standby
generator sets can be centralized and distributed at the same voltage as the normal
service.
16
Learn the pros and cons of centralized and decentralized generator systems

 Back to TOC
Large buildings or campus arrangements are subject to voltage drop concerns when
distributed across the space, therefore, one of the first decisions the team should
consider is the distribution voltage for both the normal and standby systems. In the
U.S., typical building service voltages are 208/120 V for smaller buildings and 480/277
V for larger buildings (with stepdown transformers for 208/120 V loads). The centralized
generator system can distribute at the building service voltage or at higher voltages to
reduce wire size, which will aid with costing and voltage drop.

Alternatively, centralized systems can take advantage of 5- or 15-kV voltage ranges to

greatly reduce the cost to distribute over longer distances, even several miles away.
When using medium-voltage distribution, the buildings served will need to plan for
large transformers to step-down to the building service voltage. Step-down trans-
formers can be part of single-ended or double-ended unit substations. The type and
location is based on budget, space and desired redundancy. Additionally, 15 kV fused
switches are provided at the transformer primary and can include radial or loop out-
puts for future expansion capability.

A centralized and remote location can improve the exterior aesthetics of the various
buildings served by removing exterior enclosures from the areas around each building.
Freeing up space around buildings allows for more flexibility for future expansions and
alleviates concerns with noise and exhaust emissions near buildings.

The ability to parallel standby generator sets together is the strongest argument for cen-
tralized generator systems. The size and quantity of standby generator sets required to
achieve N+1 sizing can be less than decentralized. Let’s work through an example:
17
Learn the pros and cons of centralized and decentralized generator systems

• Building A has 550 kilowatts (kW) of emergency load.  Back to TOC

• Building B has 375 kW of emergency load.

• Building C has 900 kW of emergency load.

Total emergency load is 1,825 kW. Assume the generator sizing calculations (for volt-
age dip and frequency dip) require a generator sized for 2,400 kW, minimum. A central-
ized system can achieve N+1 with two 2,500-kW or three 1,250-kW standby generators.
For a decentralized generator system to achieve N+1, each building would need a pair
of generators, bringing the campus total to at least six.

When determining the number of generators to parallel to serve the total campus
load, the designer then needs to evaluate several factors, such as type of building,
loads served, how they interact with the generator that creates voltage dip and how
long they can sustain being without power before the EES is online.

For instance, in a Level 1 system that feeds a health care occupancy, the engineer must
determine the loads required to connect within 10 seconds or the sum of the critical
and life safety branches. This load is what drives the smallest engine size allowed. Let’s
revisit the example, this time with the load broken out by essential branches, as de-
fined by NFPA 99 Article 517.

• Building A has 20 kW of life safety branch load, 175 kW of critical branch load and
355 kW of equipment branch load.

18
Learn the pros and cons of centralized and decentralized generator systems

• Building B has 10 kW of life safety branch load, 75 kW of critical branch load and  Back to TOC
290 kW of equipment branch load.

• Building C has 35 kW of life safety branch load, 350 kW of critical branch load and
515 kW of equipment branch load

Now we can see that we have 665 kW of load that needs to be online within 10 sec-
onds and after running generator sizing calculations, we may find that the minimum
generator size needed is 900 kW, making 1,000 kW (or 1,250 kW in our example) the
preferred choice for each generator. Next, the designer needs to determine the avail-
able growth that this minimum size allows. If the campus is expected to double in size
over the next 20 years, then the recommended minimum generator size needs to in-
crease accordingly.

The team needs to determine how the centralized generators and associated equip-
ment will be protected. Keeping the paralleling controls and distribution switchgear
inside conditioned (and access controlled) buildings is preferred. While generators
may be in exterior enclosures and generators can feature onboard paralleling controls
to reduce the footprint of the distribution, electronics always prefer a conditioned en-
vironment with low dust, humidity and a moderate temperature range.

Remote buildings, such as a CUP that house the generator sets, fuel polishing systems,
day tanks, engine controllers and associated electrical distribution should take future
expansion into consideration. The layout of the generators, for example, should al-
low for additional units or accommodate replacement with larger units as the campus
grows. This requires a larger footprint or at least the planning for building expansion,
19
Learn the pros and cons of centralized and decentralized generator systems

which drives up the  Back to TOC

costs associated with
centralized systems.

Centralized systems
are not without their
challenges. Initial
costs can be higher
when a greenfield
campus is initially con-
structed because the
central system needs
to be sized (or scalable) for all projected campus Figure 3: Two 2 MW 13.2 kV diesel
generators in outdoor walk-in enclosures
growth. It is also common for centralized systems to provide campus backup of utility feeds
be a retrofit solution for an existing campus looking serving a central utility plant. The
to expand further. generators serve 15 kV-class metal-clad
switchgear with operable breakers for
automatic operation with loss of utility.
In addition, feeder routes must be carefully coordi- Courtesy: IMEG Corp.
nated with other utilities and future expansion plans.
The routing could be underground duct banks, routing through utility tunnels or a
combination thereof, but the pathway should be separated from normal utility pathway
to prevent a single excavator or event from damaging both normal utility and emer-
gency services.

Underground duct banks are also subject to ground water ingress and the need for
periodic maintenance holes (due to wire spool length limitations) means that designers
20
Learn the pros and cons of centralized and decentralized generator systems

should consider methods of providing adequate drainage of conduit and maintenance  Back to TOC
holes. Drainage systems may include connection to stormwater drainage or pumping
solutions.

Centralized standby generators for emergency and utility

backup
Some campuses want a full backup of the normal utility through standby generators
that operate at the service entrance voltage, which is usually 5- or 15-kV class. Solu-
tions that offer this level of backup are often subject to more rigorous requirements,
such as U.S. Environmental Protection Agency (EPA) Tier emission levels.

An advantage to having a lower emission EPA Tier standard allows for load curtail-
ment, which can provide financial savings through agreements with the local utility.
These generators are typically independent of emergency standby generators; how-
ever, we can consider an arrangement in which large, standby generators operating at
the utility service voltage can provide both essential standby and utility backup power.

When using campus backup generators for emergency backup, there are several as-
pects to be aware of when doing this. There are several reasons why it is not accept-
able to simply rely on the campus backup generator to provide the essential electrical
branch for each building.

• Division of branches is not possible with utility backup.

• Two sources of power within a patient care area, where required by NFPA 99, is not
possible.
21
Learn the pros and cons of centralized and decentralized generator systems

• Transfer switches are still required in each building, for each required branch.  Back to TOC

• Branch segregation required within NEC Article 700, 701 and 702 (or NEC 517) are
still required at the branch circuit level.

• Each building will need transformers if the generator voltage is different from the
building voltage. These transformers are separate from the incoming service trans-
formers. This is commonly accomplished using unit substations when the generator
voltage is 5 or 15 kV, but can occur with traditional exterior or interior step-down
transformers.

• The building transfer switches need additional programming to inhibit retransfer to

normal when the campus backup is online.

Richard Vedvik
Richard Vedvik is a senior electrical engineer and acoustics engineer at IMEG Corp.
Vedvik has experience in the health care, education, commercial and government sec-
tors. He is a member of the Consulting-Specifying Engineer editorial advisory board.

22
Discover Galaxy VS 3-Phase UPS | Schneider Electric

 Back to TOC


Discover Galaxy VS 3-Phase UPS
Thanks to patented technologies, this UPS delivers up to
97% efficiency in double conversion mode and up to 99% in
ECOnversion mode, equivalent to 66% higher energy savings. It’s
compatible with Galaxy Lithium-ion battery cabinets which have
a longer battery lifetime and higher temperature tolerance than
classic battery solutions. With its robust modular design, superior
performance, scalability, and Live Swap options, Galaxy VS is the
ideal backbone for your critical infrastructure.

23
Driving Climate- A s the digital landscape continues to expand and tech-
nologies like cloud computing, artificial intelligence
(AI), and the Internet of Things (IoT) become more preva-

Positive lent, the demand for data is expected to surge further. Over
the next five years, the data center market is expected to

Innovation in have an annual growth of 4.66%, resulting in a market

volume of US $410.40 billion by 2027. Additionally, industrial

Data Centers
segments such as manufacturing, oil & gas, and semicon-
ductor are undergoing a digital transformation that requires
data continuity. However, the energy-intensive infrastructure

and Industries: needed to operate and sustain the equipment needed to

facilitate the rapidly increasing demand we’re currently wit-

A Sustainable
nessing poses a significant challenge.

Uninterruptible Power Supplies (UPSs) are crucial in ensur-

Approach ing the continuous and reliable operation of data centers

and industrial facilities. However, UPS systems themselves
consume electricity even when operating under normal
conditions. This additional energy usage adds to the over-
all power consumption of data centers, which can have a
negative impact on their overall environmental impact if
these UPS systems are not energy efficient.

This staggering energy consumption has prompted a

growing emphasis on developing sustainable practices,
hardware, and software technologies to minimize the en-
Driving Climate-Positive Innovation in Data Centers and Industries

vironmental impact and enhance the efficiency of data centers and industries glob-  Back to TOC
ally. An April 2022 Forrester Consulting study surveyed 1,033 global sustainability
decision-makers at data center colocation providers to explore sustainability efforts
in the industry. Colocation providers are in a unique position to improve sustain-
ability at their sites. Their scale allows them to implement sustainable technologies
and practices more efficiently, spreading the cost and effort across a larger custom-
er base. However, the study found that most of these organizations are still early in
their sustainability journey, as many have a goal to achieve sustainable operations,
but they still need a concrete plan to reach it. In fact, despite the generally positive
attitudes toward adopting sustainability practices, only 33% of organizations have
indicated that they possess a strategic sustainability plan in place. The same goes for
commercial and industrial segments. Take the oil & gas segment, for example. Op-
erations decarbonization is a must. A study by IHS Markit estimated that oil and gas
operations achieved energy efficiency improvements of only 10% from 2015 to 2020.

Advancements in energy storage technologies and battery monitoring systems have

emerged to help improve efficiency and sustainability. By understanding the strengths
and considerations of each technology, industry leaders can make informed decisions
that align with their power needs and sustainability goals.

Making Uninterruptible Power Supplies More Eco-Friendly

To ensure an always-on, resilient power supply, UPSs require a reliable energy storage
system that can supply power to the inverter in the event of a utility power failure or
when the power quality falls outside acceptable limits. It is crucial to consider factors
such as power requirements, backup time, reliability, IP rating, and cost-effectiveness
to ensure uninterrupted operations during power outages and fluctuations.
25
Driving Climate-Positive Innovation in Data Centers and Industries

The energy stored must possess the following key characteristics:  Back to TOC

• Immediate Availability: The stored electricity should be readily accessible

to handle micro-breaks, short voltage drops, and utility outages without any
delay.

• Adequate Power Level: The energy storage system should be capable of deliv-
ering power at a level equivalent to that of the UPS itself, ensuring that the entire
load is supplied with sufficient power.

• Backup Time: Typically, UPSs need to provide backup power for approximately
ten minutes. This duration should be tailored to meet the specific requirements of
the supported loads, considering any alternative power sources, such as an engine
generator set for longer backup durations.

Several energy storage technologies are currently available for UPS applications, each
with its own strengths and considerations. These technologies include:

• Batteries: Different types of batteries can be employed for energy storage, such
as lithium-ion, sealed lead-acid, vented lead-acid, and nickel-cadmium. Each
battery type offers unique characteristics and trade-offs that should be carefully
evaluated based on the specific needs and constraints of the UPS system.

• Ultracapacitors: These devices provide rapid energy discharge and recharge ca-
pabilities, making them suitable for quick power delivery applications. Ultracapaci-
tors offer advantages in terms of high power density and long operational lifespan.
26
Driving Climate-Positive Innovation in Data Centers and Industries

 Back to TOC
• Flywheels: Flywheel-based energy storage solutions can be used in UPS systems.
Traditional flywheels operate at low speeds (1500 rpm) and are often combined
with engine generator sets for extended backup durations. Alternatively, medi-
um-speed (7000 rpm) or high-speed (30 to 100,000 rpm) flywheels can be utilized to
meet specific UPS requirements.

Figure 1 shows the fields of application for the

different technologies.”

27
Driving Climate-Positive Innovation in Data Centers and Industries

 Back to TOC

Choosing the Right Battery for The table below compares the different
solutions in terms of their capacity to meet the
UPS Applications energy-storage requirements of static UPSs.
When selecting the appropriate battery for
Uninterruptible Power Supplies (UPS), several options are available. Sealed lead-acid
batteries have been widely used in UPS systems. They offer various advantages, such
as maintenance-free operation, easy implementation, and compatibility with different
types of rooms, including computer rooms and technical spaces not specifically de-
signed for batteries.

Vented batteries, whether lead-acid or nickel-cadmium, have their own advantages.

They provide extended service life, longer backup times, and higher power ratings.
However, their installation requires compliance with specific regulations, including
dedicated rooms and proper maintenance practices.
28
Driving Climate-Positive Innovation in Data Centers and Industries

Nickel-cadmium batteries have also been employed in UPS applications. However, their  Back to TOC
usage has become less common compared to other battery types due to their higher
cost and the emergence of more advanced alternatives like lithium-ion batteries.

Lithium-ion batteries have gained popularity as a viable alternative to lead-acid batter-

ies. They are based on mature and proven technology, offering significant advantages.
Lithium-ion batteries excel in terms of the number of cycles they can endure, operat-
ing temperature range, and reliable monitoring capabilities, including State of Health
(SoH) and State of Charge (SoC). They also have embedded and dedicated monitoring
systems, enhancing their overall performance.

When choosing a battery, several factors need consideration:

1) Operating conditions and requirements: Consider factors such as the availabili-

ty of special rooms, battery cabinets, or racks.

2) Required backup time: Determine the necessary backup time based on the
application’s specific needs, considering factors like system shutdown procedures
and the duration of power-system failures.

3) Cost considerations: Evaluate the cost-effectiveness of different battery options

in relation to the required performance.

The selection of the most suitable battery type depends on the unique requirements of
each application. It’s essential to consider the specific requirements of different appli-
cations when determining the appropriate backup time:
29
Driving Climate-Positive Innovation in Data Centers and Industries

• For computer systems: The backup time should allow for necessary file-saving  Back to TOC
and system-shutdown procedures to ensure a controlled computer system shut-
down. The computer department usually determines the required backup time
based on their needs.

• For industrial processes: The backup-time calculation should consider the econom-
ic cost of process interruption and the time required for restarting the operations.

• For applications requiring long backup times: Engine generator sets can be
used in conjunction with batteries for backup times exceeding 30 minutes to one
hour. This combination should be carefully evaluated to optimize the generator
rating and ensure proper operation.

By carefully considering the advantages and limitations of different battery types and
tailoring the backup time to specific requirements, data centers can select the most
suitable UPS solution that ensures reliable power backup during critical situations.

Increasing the performance and reliability of battery monitoring

The Galaxy Li-ion battery solution has been specifically designed to deliver exception-
al performance and reliability for critical applications. Compliant with UL9540A stan-
dards, this advanced battery solution offers numerous benefits, including a significant
reduction in battery footprint and weight of up to 70%. By optimizing space utilization,
businesses can make more effective use of their infrastructure – which is especial-
ly beneficial as the data center industry is witnessing a shortage of available land to
build. With its Green PremiumTM certification, businesses can trust that its eco-friendly
features align with their sustainability goals.
30
Driving Climate-Positive Innovation in Data Centers and Industries

One of the standout advantages of the Galaxy Li-ion battery solution is its ability to  Back to TOC
double battery life. By operating at higher temperatures, these lithium-ion batteries
significantly reduce cooling requirements, resulting in a lower total cost of ownership
while simultaneously reducing the data center’s overall carbon footprint. With extend-
ed battery life and reduced maintenance, businesses can experience more significant
cost savings and enhanced operational efficiency.

The included battery management system provides a new level of visibility, predictabili-
ty, and manageability for the battery system. Embedded monitoring at the cell, module,
and cabinet-level ensures comprehensive insight into battery health, enabling predictive
maintenance and minimizing the risk of unexpected failures. The modular and touch-safe
design of the Galaxy Li-ion battery solution simplifies maintenance procedures, promot-
ing increased safety during servicing. With built-in fuse protection at the battery cell and
cabinet level, businesses can confidently carry out maintenance tasks.

By proactively identifying potential failure points and maintenance requirements in

various components such as UPS power supplies, battery backups, and more, opera-
tors can prevent issues before they arise. This proactive approach helps reduce upfront
costs and the need for costly investments to replace failed assets. Utilizing analytics
decreases the occurrence of failures and the need for interventions, resulting in im-
proved uptime rates and longer intervals between failures.

Furthermore, the Galaxy Li-ion battery solution seamlessly integrates with EcoStrux-
ure Power Monitoring Expert (PME) for Backup Power Monitoring. This integration
offers real-time visibility into energy storage availability, allowing businesses to analyze
historical data, monitor alarms, and create proactive maintenance plans.
31
Driving Climate-Positive Innovation in Data Centers and Industries

Case study for looking to a low-carbon horizon  Back to TOC

with EcoDataCenter
The Swedish company EcoDataCenter exemplifies the effective integration of en-
ergy-efficient UPS systems within a sustainable data center, showcasing the practical
implementation of eco-friendly practices and technologies in line with the broader sus-
tainability goals of the industry. EcoDataCenter was founded in 2014 with the goal of
enabling the green transition by being a supplier of sustainable digital infrastructure.

EcoDataCenter takes pride in its social and environmental responsibility when estab-
lishing new data centers by offering both colocation services and high-performance
computing. EcoDataCenter is working to change how data centers are conceived and
built, pushing highly sustainable and efficient designs.

It started with recognizing the need for an ultra-low carbon footprint data center to pow-
er its HPC colocation services in Falun, Sweden. Leveraging the Nordics’ cooler climate
and renewable energy sources such as solar, wind, and hydropower, EcoDataCenter aims
to become a leading data center operator in the region within the next five years.

To make that happen, EcoDataCenter partnered with Schneider Electric to use its
EcoStruxure architecture to create its greenest and most forward-thinking data center
yet. EcoDataCenter’s use of EcoStruxure has led it to have four Galaxy VX UPSs with a
capacity of 1250kW each, providing reliable backup power at an impressive efficiency
rate of 99% when running in ECOnversion mode.

EcoStruxure serves as Schneider Electric’s system architecture and platform, designed

to be open, interoperable, and IoT-enabled. With a vast deployment of over 480,000
32
Driving Climate-Positive Innovation in Data Centers and Industries

sites and the support of 20,000+ system integrators and developers, EcoStruxure con-  Back to TOC
nects and manages more than 1.6 million assets.

By embracing Schneider Electric’s EcoStruxure for Cloud & Service Providers solution,
EcoDataCenter seamlessly integrates hardware, software, and monitoring systems. This
open architecture platform provides comprehensive analytics and services, facilitat-
ing informed decision-making and optimal operational efficiency. EcoDataCenter now
benefits from connected sensors and meter data, which generate detailed reports on
operational efficiency and sustainability performance.

The resulting data center boasts an impressively low seasonal design power usage
effectiveness (PUE) value of 1.15. In 2022, data center managers reported an average
annual PUE ratio of 1.55. PUE is a metric used to assess the energy efficiency of a
data center. It represents the ratio of the total amount of power consumed by the data
center (including IT equipment, cooling systems, lighting, etc.) to the power consumed
by the IT equipment alone.

EcoDataCenter’s PUE of 1.15 indicates a high level of energy efficiency; for every unit
of power consumed by the IT equipment, only an additional 0.15 units of power are
used for cooling, lighting, and other non-IT functions. This is achieved through re-
dundancy measures and using Schneider Electric’s efficient cooling systems, such as
CRAH units operating at an energy efficiency rating (EER) of 132 in normal operation
mode.

Surplus energy is redirected to clients, further optimizing resource allocation. Notably,

EcoStruxure also plays a vital role in recycling the facility’s heat waste, allowing EcoDa-
33
Driving Climate-Positive Innovation in Data Centers and Industries

taCenter to contribute energy back to the local grid and solidify its position as one of  Back to TOC
the most sustainable data centers in the Nordics.

As part of a unique energy ecosystem, the data center collaborates with a local bio-
fuel facility and pellet mill, repurposing surplus low-grade waste heat. By storing and
utilizing this heat during colder months, the facility reduces its reliance on fossil fuels
for heating pellet production. EcoDataCenter actively offsets the energy system’s
need for oil and gas by capturing and redirecting heat from server rooms to Falu En-
ergy and Water.

Schneider Electric’s support in assembling the colocation facility has contributed to

EcoDataCenter’s competitive pricing, driven by significantly lower energy costs. The
integration of various technologies, including medium-voltage switchgear, Galaxy
VX UPS units, Masterpact MTZ breakers, scroll compressors, CRAH units, lithium-ion
batteries, and Uniflair chillers, all labeled with Schneider’s “Green Premium,” ensures
environmental transparency and sustainability at every step.

Transparency is paramount in environmental sustainability, especially in light of new

public and regulatory pressures to disclose ESG (Environmental, Social, and Gover-
nance) and emissions data. Schneider’s “Green Premium” products serve as a mecha-
nism to ensure environmental transparency and sustainability at every step by providing
detailed information on regulatory compliance, environmental impact, and circularity
attributes. With increased public awareness and concern about sustainability, green-
washing can lead to reputational damage and loss of trust. Transparent disclosure helps
companies avoid accusations of greenwashing by providing concrete evidence of their
environmental efforts, backed by reliable data and measurable outcomes.
34
Driving Climate-Positive Innovation in Data Centers and Industries

EcoDataCenter’s mission revolves around climate-positive solutions that benefit their  Back to TOC
clients, the surrounding communities, and the environment. Beyond efficiency and reli-
ability, EcoDataCenter’s partnership with Schneider Electric extends to making sustain-
able contributions to the community. The data center actively gives back and reinforc-
es its commitment to climate positivity by providing job opportunities and supporting
the local energy circle.

Through innovative design, well-tested technical systems, and the power of EcoStruxu-
re, EcoDataCenter and Schneider Electric have successfully created an environmentally
conscious data center that sets new standards for efficiency and sustainability while
benefiting both customers and the local community.

The path forward

Data centers play a vital role in today’s digital age by storing, processing, and trans-
mitting vast amounts of information. Driving climate-positive innovation is crucial for
mitigating their significant and notorious energy consumption and environmental im-
pact. As the demand for IT infrastructure continues to grow, the industry must prioritize
sustainable practices and technologies to enhance efficiency and reduce carbon foot-
prints. For example, the selection of appropriate energy storage technologies, such as
lithium-ion batteries, can further improve the reliability, performance, and longevity of
UPS systems.

EcoDataCenter, in partnership with Schneider Electric, has emerged as a leader in

sustainability, designing and building data centers that are highly eco-friendly and
energy-efficient. By leveraging innovative solutions like EcoStruxure, EcoDataCenter
has set new standards for efficiency and sustainability in the industry. EcoDataCenter
35
Driving Climate-Positive Innovation in Data Centers and Industries

employs sophisticated monitoring systems that track various parameters, including en-  Back to TOC
ergy consumption, temperature, and airflow. This data-driven approach enables them
to identify potential inefficiencies and implement targeted improvements, leading to
significant energy savings and reduced environmental impact.

Addressing the environmental impact of data centers and industrial facilities requires a
holistic approach encompassing energy storage technologies, battery monitoring sys-
tems, and sustainable operational practices. With continuous advancements and adop-
tion of these solutions, these industries can make significant strides toward a more
sustainable and resilient future.

36
Your questions answered: What
to consider when designing a  Back to TOC

paralleling power system

When designing a paralleling power system, electrical engineers need to
consider many aspects including transfer switches, grounding and protection

T his course on Sept. 15, 2022, titled “What to consider when designing a paralleling
power system” provided electrical engineers with an overview of power systems
and covers the elements to consider for designing generator paralleling systems, auto-
matic transfer switches, withstand/closing ratings, grounding and protection.

Several questions were left open, and they’re answered here by Hassan Obeid, Senior
Global Technical Advisor – Energy Management Solutions, Cummins Power Generation

Do you have an example of generator application in a microgrid with

solar plus battery energy storage?

Yes, we have a few examples where we have implemented generator sets, battery en-
ergy storage system and solar.

Can generators of different kW ratings be paralleled?

Yes, they can. All the generators will take their share of the load based on percentages.
The larger the generator, the more load it takes. The generators onboard controller
does that on a millisecond timescale.
37
Your questions answered: What to consider when designing a paralleling

Can you parallel generators from different manufacture via autonomous  Back to TOC
paralleling or do you have to use min controller?

Yes, that is possible. It is also possible to add the Cummins paralleling controller on
top of other controls to minimize the number of controllers. It is also important to con-
sider the winding pitch of the alternators.

With Onboard Paralleling with generator set mounted breakers using a common col-
lector bus, what is the best practice for generator maintenance? Should a disconnect
be provided to isolate the generator from the collector bus? Should both gensets be
shut down if work is being performed on the circuit breaker?

Yes, providing a disconnect switch to isolate the generator set for maintenance is one
of the items to consider. Please refer to this white paper: “Design considerations for
generator set mounted paralleling breakers.”

Can generators parallel with BESS (battery energy storage system)?

Yes, Generator sets can be paralleled with BESS.

Quickly touch on lean burn NG stabilization: How do parallel lean burn

gensets react to changing loads/minimum load requirement when oper-
ating in an island configuration?

As a rule of thumb, lean burn generator sets should be loaded about 60%. Lean burn
generator sets react slower than diesel when to comes to load removal for example.
38
Your questions answered: What to consider when designing a paralleling

Having said that, the controls are  Back to TOC

tuned precisely to accommodate
for that. Therefore, that is not an
issue. The most important thing
to consider is if a lean burn is
paralleled with a diesel generator,
then special adjustments must be
made in the generator and engine
controls to account for that.

Do you use 4 pole or 3 pole

ATSs on 120/208V Y systems

To determine which configuration

(3-pole or 4-pole) to select, atten-
tion must be paid to the system
grounding scheme and ground
fault protection requirements. If
ground fault detection is required
on either source, a 4-pole transfer
switch is necessary in most cases.
Per NFPA 70: National Electrical
Code 230.95 Ground Fault Pro-
tection (GFP) of equipment is
required at the service disconnect
(utility breaker) for systems with all
of the following: 39
Your questions answered: What to consider when designing a paralleling

• Solidly grounded wye electrical service.  Back to TOC

• More than 150 volts to ground (277/480 or 347/600 VAC).

• Over current device rating of 1000 A or more.

• NEC 700.6 (D) and 701.6 (D) requires Ground Fault Indication (GFI) at the emergency
source.

Ground fault sensing depends on being able to sense ground fault current. To accu-
rately sense ground fault current, it must return to its source on a known path relative
to ground fault current transformers (CTs). In basic emergency standby systems there
are two rules to follow to meet these requirements:

• There can only be one neutral/ground connection on any neutral bus at one time.

• Ground fault sensors (the CTs) must be downstream (or on the load side) of the
bonding connection.

To meet both rules when connected to either the normal or the emergency source, the
neutral must be switched using a 4-pole transfer switch

What does phase angle play in paralleling control?

The synchronizer in the paralleling controller must match the phase angle of the two
sources by adjusting the fuel to the engine. Therefore, the onboard paralleling control-
ler matches phase, voltage and frequency. 40
Your questions answered: What to consider when designing a paralleling

With an ATS having a WCR-100,000 A and time duration of 0.05 sec, is  Back to TOC
it possible to adjust the short time protection of the upstream UL 489
breaker to 50,000 A and 0.3 sec time duration since 50,000 A is lower
than the WCR of the ATS?

The short time element S in LSI trip units in UL 489 breakers doesn’t allow you to go to
that time setting of 0.3 sec. The max will not allow you to go beyond the instantaneous.
The instantaneous is always on and cannot be disabled. Only LVPCB (Low Voltage Power
Circuit Breakers) UL 1066 allows you to adjust that to go up to 0.5 sec for the short-time
S, which in that case you need to specify a short-time WCR for the transfer switch.

If the switchboard is solidly grounded, is the neutral at the generator to

be grounded or not?

The neutral from the generator sets should be brought back to the distribution board
and there should be one neutral to ground pond at the gear with a CT for ground fault.

How do you coordinate a UL 1558 gear with an ATS?

Specify a short-time WCR of 0.5 seconds.

What is the key point to make the decision of paralleling?

Reliability and redundancy.

What is the manufacturer definition of manual paralleling control?

41
Your questions answered: What to consider when designing a paralleling

Manually initiated by the operator, automatically executed by the onboard paralleling  Back to TOC
control for the maximum safety and reliability

How do you parallel generators from different manufacturers?

It can be done and it has been. We need to think about the winding pitch of the alter-
nators (2/3rd or 5/6th), managing which generator closes its paralleling breaker to the
bus and then manage load share (real and reactive) across the units.

Why can’t you install a single HRG on the common neutral bus of paral-
leled generators? The question applies to LV (480 V) system.

In LV (480 VAC) applications the loads need that neutral from the generators and there-
fore the neutral is brought back to the distribution board. In the distribution board, the
bond between the neutral and ground is established along with CTs around that bond
that feed into the ground fault relay.

When do you specify 4-pole ATS?

If ground fault protection is needed then specify a 4-pole transfer switch.

What is a basic formula to calculate the available fault current from a

generator?

As a rule of thumb, the fault current is about 8 to 10 times the rated current. This is a
very conservative estimate.
42
Your questions answered: What to consider when designing a paralleling

Generator Set Available Fault Current =  Back to TOC

(kW*1,000)/( √3*V*p.f.*Xd)

Where:

kW: kilowatt output rating of the generator set

p.f.: generator set rated power factor

V: generator set rated output voltage

Xd: alternator per unit subtransient reactance based on the generator set rating

What controls are available for loading and unloading for retransfer to grid?

All we need to control is engine fuel and field excitation. For a single generator/single
utility Cummins paralleling controller can handle all of that. Base load/peak-shave, ex-
tended paralleling and ramping the load on/off the utility. For multiple generator sets
with a utility, Cummins also offers it is synchronizer that precisely matches the genera-
tor set control to achieve all that was described.

Hassan Obeid
Hassan Obeid is a global technical advisor for systems and controls — critical protec-
tion at Cummins Power Generation, focusing on technical vision, business strategy and
solving a wide range of complex problems.
43
Five more reasons generators fail
when you need them the most  Back to TOC

Here are five additional reasons why generators fail

E ngineers tend to focus on prescriptive design solutions for mission critical and life
safety generator applications. Often the guiding concept is that adding capac-
ity and complexity to the design helps ensure reliability. However, the devil is in the
details. Real-world generator system reliability is often dictated by seemingly simple,
mundane items that are often overlooked.

It’s reasonable to expect that a properly installed generator will function perfectly on
Day One, but it should also be noted that its useful service life will often extend well
beyond 25 years. In addition, that generator is only one component of a larger emer-
gency power supply system. Failure of any individual part of that system could compro-
mise the overall performance and reliability of that system. Given that extended ser-
vice life, the logical question for any engineer is, what parts of that system will become
vulnerable as the system ages and how can the associated risks be mitigated?

NFPA 110: Standard for Emergency and Standby Power Systems is the most applicable
standard in this regard. NFPA 110 addresses installation, testing and (most importantly)
ongoing maintenance requirements for the EPSS. The issues that are examined in this
article echo those identified within NFPA 110 and consist mostly of simple items that
have outsized consequences if not properly addressed.

Here are five common reasons why generators fail, with additional information avail-
able in the previous issue:
44
Five more reasons generators fail when you need them the most

1. Generator or related  Back to TOC

systems are in a vulnera-
ble location
Aside from engineers, few peo-
ple appreciate the appearance of
generators. Often, generators are
consigned to out of sight locations
that are less that optimal — on
roofs, in basements, behind shrub-
bery. As multiple natural disasters
have proved, these types of loca-
tions can make a generator just as
Figure 1: Smart battery charger and
vulnerable as the electrical utility services that they
starting batteries within corrosion resistant
are intended to back up. battery boxes next to a generator. Battery
failure is a leading cause of inoperable
generators. Temperature-compensated
NFPA 110 Section 7.2 addresses some generator
battery charging is important in maximizing
location considerations such as avoiding locations battery service life. Courtesy: McGuire
with the potential for flooding. While this is a valid Engineers

concern, flooding is by no means the only risk that

should be evaluated. There may be other pertinent vulnerabilities associated with spe-
cific geographic locations or applications. In addition, other components of the EPSS
may be just as vulnerable as the generator. For example:

• Outdoor generators may be damaged by flying debris/projectiles in coastal areas

where hurricanes are common.

45
Five more reasons generators fail when you need them the most

• Generators at the top of tall buildings, where elevators are the primary means of  Back to TOC
access, can become nearly inaccessible for emergency repairs if the elevators fail.

• Even if the generator is located above grade, if fuel tanks and associated transfer
pumps are in a basement, flooding may still cause an EPSS failure.

• In extremely cold regions, it may not be possible to maintain diesel fuel tempera-
ture above cloud point if stored outdoors. Cloud point is the temperature at which
wax crystals precipitate in the fuel. This wax precipitation can cause fuel filter plug-
ging and generator failure.

Some situations may be unavoidable. However, a proper risk assessment performed

during the design phase of a project can help identify these types of issues ahead of
time and potentially allow specification/design of appropriate measures to help mini-
mize the associated risks.

2. Weak or dead batteries

Just like an automobile, you cannot start a typical standby generator without a func-
tional battery. And like their automotive counterparts, dead batteries are a leading
reason for generators failing to start. If you doubt this, ask any generator vendor what
the most common cause for generator starting issues and the most likely answer that
you’ll get is batteries.

Batteries are electrochemical energy storage devices. Anything that affects the chemi-
cal reactions that happen within that battery will impact affect its functionality. Unfortu-
nately, the adverse conditions under which most operate (vibration, improper charging,
46
Five more reasons generators fail when you need them the most

wildly fluctuating ambient temperatures) can  Back to TOC

do just that and will also dramatically reduce
their useful service life.

The most common starting batteries used for

generators are flooded cell lead-acid type.
These types of batteries consist of lead cath-
ode and anode plates separated by a sulfu-
ric acid electrolyte. Every time a battery is
discharged, the lead on the electrode plates
within the battery will react with the sulfate in
the electrolyte and form lead sulfate. When
the lead on the plates can no longer react
with the electrolyte, the battery is considered
completely discharged. For recharging, the
chemical reaction is reversed by applying an
external voltage source, the lead sulfate is Figure 2: All NFPA 110 compliant
converted back to lead that redeposits on the plates and remote generator annunciators
have a “not-in-auto” alarm.
the sulfate returned to the electrolyte liquid.
Remote annunciators should be
placed in regularly supervised
However, the lead sulfate tends to deposit and accu- locations so that these alarms
don’t go unnoticed. Courtesy:
mulate in crystalline form on the electrode plates in the
McGuire Engineers
battery. That lead sulfate has a relatively high resistance.
If the thickness of the deposits on the electrodes are great enough, the charging volt-
age can be insufficient to break them down. Eventually, the battery becomes unusable.
Proper charging with adequate voltage and current is critical for slowing this deposit
formation and ensure the batteries maximum service life. 47
Five more reasons generators fail when you need them the most

Temperature compensated battery charging is critical in maximizing battery life. NFPA  Back to TOC
110 section 5.6.4.7(6) requires this capability for Level 1 (life safety) EPSSs. Batteries op-
erate optimally in ambient temperatures (68°F to 77°F). Deviations from this optimum
temperature directly impact the speed of chemical reactions that take place during
battery charging and discharging.

Accordingly, a temperature compensated battery charger adjusts charging voltage

based on ambient temperature. When the ambient temperature is high, charging
voltage is reduced to avoid boiling off the electrolyte. When the ambient temperature
is low, the charging voltage is increased to decrease the accumulation of lead sulfate
crystals on the electrode plates.

Even with proper charging, batteries have a finite life expectancy. As such, NFPA 110
appendix section A5.6.4.5.1 recommends that batteries:

• Be tested semi-annually.

• If not tested, replaced every 24 to 30 months when exposed to temperatures ex-

ceeding 81°F for significant periods of time.

• Replaced every 36 to 60 months for cooler temperatures.

3. Bad fuel (diesel)

Most emergency generators see relatively little use over their useful service life. It’s not
unusual to come across 25-year-old emergency generators with less than a few hun-
dred hours of total runtime. In areas with reasonably reliable utility sources, a genera-
48
Five more reasons generators fail when you need them the most

tor’s runtime may be limited to only weekly/monthly testing purposes. However, many  Back to TOC
EPSS applications require 24 hours or more of on-site fuel supply. In those situations, it
may take years for a generator system to burn through a full tank of fuel.

The primary problem with this is that diesel fuel is not maintenance-free. As fuel ages,
it will oxidize through mechanisms similar to how animal fats become rancid. Partially
filled tanks can also attract water through condensation and promote microbial growth.
The resulting water, wax, varnish and sludge can clog fuel filters and cause engine
damage.

Modern diesel engines are manufactured to much tighter tolerances and have preci-
sion components such as high-pressure common rail fuel injectors. As such, modern
engines are much more sensitive to fuel contamination. Most manufacturers have min-
imum specifications for fuel cleanliness based on ISO 4406. ISO 4406 defines contam-
ination by a coding system that quantities the number of particles for difference sizes
ranges within a milliliter sample.

The particle sizes referenced in this standard are not visible to the naked eye. If you
can see contaminants or discoloration of the fuel (offroad diesel for generators is typi-
cally translucent red in color), chances are that fuel quality is less than what the manu-
facturer allows. If the fuel is contaminated, the fuel filters are the last line of defense for
the fuel injection system. The filter’s ability to remove contaminants (efficiency, or num-
ber of particles retained by filter versus what is passed through) is in high +98% range.
For larger engines with high fuel consumption, significant amount of contamination
can still pass through the filter and cause damage over time.

49
Five more reasons generators fail when you need them the most

Ultra-low sulfur diesel fuel typically only has a storage life of 6 to 12 months. The re-  Back to TOC
duced sulfur levels, while good for reducing harmful engine emissions, are less effec-
tive at controlling microbial growth. It is recommended where fuel is stored for extend-
ed periods of time (more than 12 months) that the fuel be periodically pumped out and
replaced with fresh fuel.

Current biodiesel blends potentially have an even shorter storage life, though devel-
opment of hydrotreated vegetable oil biodiesel fuels may address that issue. HVO
complying with European Standard EN 15940 is a further refined biodiesel feedstock
that has improved oxidation stability and is less prone to bacteria growth, making it
better suited to long-term storage such as in generator applications. It is expected
that as carbon neutrality requirements increase, use of HVO biodiesel will become
more pervasive.

If the fuel cannot be replaced at regular intervals, there are some options that may ex-
tend fuel life. Fuel polishing systems can be installed on fuel tanks to filter out contam-
inates. Although not officially recognized by generator manufacturers, fuel additives
and biocides can also help. NFPA 110 appendix section A5.5.3 specifically mentions
concerns with oversizing fuel tanks and even goes as far as suggesting that multiple
smaller tanks instead of a single large may be preferable from a fuel management
standpoint. Appendix section A.7.9.1.4 also recommends that tanks be kept cooler and
relatively full to slow fuel degradation.

4. Generator controls not left in auto

If you have ever stood next to a large generator when it started and quickly accelerat-
ed to full speed, one of your first impressions probably was that it’s a surprisingly loud
50
Five more reasons generators fail when you need them the most

and violent event. Given that generators  Back to TOC

are often installed in skin-tight weather-
proof enclosures or similar confined areas,
maintenance work generally requires ex-
tremely close proximity to those engine
components that could cause severe injury
or death if the engine accidentally start-
ed. As such, the first step in any generator
maintenance procedure is to disable auto-
matic starting.

However, a surprisingly common reason for

generators failing to start is that the service
technician forgot to return the generator to
automatic operation upon completion of
maintenance work. While NFPA 110 re-
quires that “not in auto” be included in the
generator’s remote annunciator alarms, an-
Figure 3: A 2-megawatt diesel
nunciators are often hidden in fire command centers or in generator serving a mission
similar out of the way areas that are not subject to proper critical facility. Locating generators
within conditioned, dedicated
supervision. As such, alarm conditions often go unnoticed
generator rooms can significantly
until it’s too late. extend their service life. Courtesy:
McGuire Engineers

Human error is unavoidable. Design solutions should take

this into account and provide the means to make sure that qualified personnel can be
made aware of these types of conditions in a timely manner. A simple solution such as
51
Five more reasons generators fail when you need them the most

adding a second remote annunciator in the engineers’ office or similar supervised loca-  Back to TOC
tion with knowledgeable personnel can go a long way in addressing this. With older gen-
erators, adding an additional annunciator could be a significant undertaking. Often, they
used dedicated conductors for each and every annunciator alarm point which resulted
in a complicated spaghetti mess of wiring. However, most modern generators use RS485
serial cabling, which dramatically simplifies wiring requirements and associated costs.

5. Lack of testing with load

This is a frustratingly simple problem. How do you know if a generator will work if you
don’t test it? As stated before, most generators have very little run time. While NFPA
110 mandates monthly generator testing and an annual two-hour full load bank test, a
surprising number do not comply with the load bank requirement.

Also, many generator owners perform monthly tests consisting of running the gener-
ator under no-load for 30 minutes. Running a diesel generator excessively with little
or no load can result in wet-stacking. Wet-stacking occurs when cylinder pressure and
temperature does not reach a sufficiently high level, resulting in incomplete fuel com-
bustion. As a result, deposits form within the cylinders and a black, oily liquid may
ooze from exhaust pipe joints. This can lead to reduced power output and accelerated
engine wear. Most manufacturers will recommend that diesel generator be run at a
minimum of 30% load for 30 minutes for every four hours of light load operation. This
will burn off deposits and minimize the chance of wet-stacking.

John Yoon
John Yoon, PE, LEED AP ID+C; is lead electrical engineer at McGuire Engineers, Chi-
cago. He is a member of the Consulting-Specifying Engineer editorial advisory board.
52
Electrical acceptance testing
collaboration leads to improved  Back to TOC

project delivery
Collaboration between the system integrator, contractor and acceptance
testing agency is critical to creating efficiencies while maintaining quality
assurance in a data center’s electrical system

A thorough check of electrical power systems and components before energiza-

tion can uncover and help correct problems that otherwise would lead to project
schedule delays or larger, more costly issues in the future with impacts to data center
downtime and consumer dissatisfaction.

What is acceptance testing and is it necessary?

Acceptance testing is the physical inspection and testing of newly installed electrical
equipment. This involves thorough visual and mechanical inspections using calibrated
test instruments to ensure electrical components and completed systems operate as
designed. It occurs before electrical system commissioning and startup as well as be-
fore the new equipment is put into operation.

Taking this initial step verifies that manufactured devices are free from defects, operat-
ing as designed and intended and installed correctly as specified. It is important that
acceptance testing be performed by a third-party testing firm that is unbiased and
independent in its evaluation and findings.

53
Electrical acceptance testing collaboration leads to improved project delivery

Common issues when acceptance testing  Back to TOC

Several issues that can be costly to correct after startup can be found with acceptance
testing:

• Failure of cabling damaged during installation.

• Incorrect wiring.

• Mechanical operating problems.

• Nuisance tripping or breakers tripping outside of manufacturer’s curves.

• Improper relay settings and programming.

• Compromised insulation dielectric systems.

• Improper grounding.

• Wrong transformer taps and/or improper ratios for proper voltage.

• Instrument transformer and metering circuit ratio and wiring issues.

• Surge protection device defects.

• Switchgear bus and cable connections not properly assembled.

54
Electrical acceptance testing collaboration leads to improved project delivery

• Bus connections improperly torqued.  Back to TOC

Acceptance testing also avoids unnecessary expenses for data centers. Finding system
and component anomalies during acceptance testing — while equipment is still un-
der warranty and in a controlled environment — is critical. Determining and correcting
deficiencies before startup can save an owner capital and maintenance expenses by
preventing costly outages, equipment repairs and potential safety issues.

Hyperscale data center requirements

The hyperscale data center market continues to grow and with the amount of digital
information being generated, there’s no end in sight. With this growth comes a need
for new data centers and/or expansion of existing facilities at a much greater rate to
keep up with digital demand. The increased demand for data storage drives larger and
more complex power system requirements to handle the loading, not only for the serv-
ers storing the data, but also for the thermal systems required to cool the data center
equipment.

Traditionally, acceptance testing of the critical power distribution infrastructure for

these types of facilities would have been performed fully at the data center. The test-
ing agency would work alongside the electrical contractor during or in many cases
after installation to complete all acceptance testing activities.

With the increased demand and the need to bring these data centers online faster that
model has evolved. It is important to note, however that the quality assurance, testing
and commissioning aspects of the project must always remain the goal without com-
promises being made.
55
Electrical acceptance testing collaboration leads to improved project delivery

Off-site acceptance testing opportunities  Back to TOC

A shift in the use of system integrators creates an opportunity for testing to begin off-
site. Acceptance testing at an integrator’s facility or an equipment provider is becom-
ing more common and offers notable efficiencies. Collaboration with a system integra-
tor at its location while the equipment is being built allows the integrator and electrical
contractor to follow parallel paths for all data center assets.

The electrical contractor can be installing system components while the integrator is
assembling the remaining portions of the power system distribution equipment. These
components can be tested as they are connected off-site, significantly reducing time
on-site. This time savings provides an improvement to the overall construction sched-
ule that benefits all parties involved.

It is important to note that acceptance testing will still need to be conducted at the fi-
nal on-site location to ensure no damage occurred during shipping or final installation.
This multistep approach will help save time overall as the testing firm can uncover any
issues and make corrections off-site, leading to fewer issues being found on-site.

Acceptance testing is best performed by the same independent firm working with the
integrator at its location as well as with the electrical contractor during installation on-
site. Keeping a consistent partner makes for smoother transitions, increased efficien-
cies, standard processes and procedures, communication improvements, etc.

While integrators are providing a customized solution involving numerous components

from various original equipment manufacturers, the testing agency conducting the
acceptance testing must be familiar with a wide range of manufacturers’ equipment.
56
Electrical acceptance testing collaboration leads to improved project delivery

Using qualified technicians who can implement industry-recognized testing methods  Back to TOC
and procedures will help ensure the quality and consistency of testing at the data cen-
ter and off-site.

Equipment tested at the system integrator’s site can include skid-mounted systems in
outdoor enclosures. Each skid includes medium- and low-voltage equipment such as
circuit breakers, trip units, protective relays, wiring, metering, instrument transformers,
panelboards, batteries, etc. The integrator also includes their own system for power,
lighting and heating, ventilation and air conditioning controls.

The electrical contractor can be wiring the loads at the data center and be ready and
waiting for the skids of low- and medium-voltage switchboard and/or switchgear as-
semblies to arrive. Testing at the data center includes electrical contractor-installed
equipment fed from the skids, such as busways, circuit breakers, bus plugs, discon-
nects, grounding, etc.

Tests are often conducted concurrently at the site while integrator testing is underway.
Testing on equipment from the skids that cannot be performed at the integrator in-
cludes items such as interconnect wiring, grounding and complete system functional
testing. These are completed at the data center to close out the on-site testing.

Choose the right partner

Due to the importance of conducting proper acceptance testing, selecting a good
acceptance testing firm for which the experience and knowledge can be trusted is cru-
cial. Consider the following criteria when selecting a suitable acceptance testing firm:

57
Electrical acceptance testing collaboration leads to improved project delivery

• Independence from the manufacturer to ensure an unbiased, thorough assessment  Back to TOC
of equipment.

• Ability to accurately interpret test results to determine best course of action for
each unique customer environment.

• Previous work experience on similar projects to determine the value added.

• Technician certification and experience to ensure thorough testing by a trained

professional.

• Size of the firm and ability to staff with qualified professionals to deliver the project
on time and within budget.

• Awareness of all safety standards to ensure safe work practices and require less
supervision.

• Test equipment calibration program to ensure that equipment is in proper working

order to deliver accurate test results.

• Affiliations with reputable organizations such as InterNational Electrical Testing

Association, or NFPA to confirm knowledge of testing standards.

• Ability to perform new and innovative testing services to ensure a complete evalua-
tion and comprehensive recommendations.

58
Electrical acceptance testing collaboration leads to improved project delivery

• Ability for the same company to be used at the system integrator as well as the  Back to TOC
physical test site.

• Strong financial standing and ability to be a long-term provider.

The above criteria are important when identifying an independent testing company
that can become a trusted adviser to help to extend system life, reduce downtime and
improve the safety surrounding equipment.

Electrical systems are among a data center’s most critical assets and they can have a
big impact on the bottom line. In data centers that power the online economy 24/7,
even a brief disruption can cost millions of dollars. Their production and management
costs are high and failures almost always lead to catastrophic losses.

Hyperscale data centers are experiencing substantial growth, requiring collaboration

and an integrated project delivery to improve consistency and shorten the overall con-
struction schedule. Investing in thorough acceptance testing by using a trusted, inde-
pendent testing firm can save a project money during construction and throughout the
equipment life cycle.

Eric Nation
Eric Nation is general manager at High Voltage Maintenance (HVM).

59
 Electrical & Power
Thank you for visiting the Electrical & Power eBook!
Content Archive

2023 Summer Edition If you have any questions or feedback about the contents
in this eBook, please contact CFE Media at
2023 Spring Edition
customerservice@cfemedia.com
2022 Winter Edition

2022 Fall Edition

We would love to hear from you!