Data Center

Basic Power Infrastructure

Compiled and Written by: Qasem Aljaber

March , 2025
 “The charity of knowledge is to spreading it”

Disclaimer:
This document is based on publicly available information from multiple sources and is intended for
educational purposes. While every effort has been made to ensure accuracy, errors are possible.
Readers are encouraged to verify details with official standards and guidelines. The purpose of this
document is to share knowledge and promote a better understanding of the subject matter.

I|Page
Contents
Data Center Power Infrastructure ...................................................................................................................... 1
1.0 Electrical Power System Overview ............................................................................................................ 1
1.1 From Generation to Data Center .......................................................................................................... 1
1.2 Increasing Energy Demands of Data Centers ......................................................................................... 2
2.0 Data Center Rating Levels Overview ......................................................................................................... 3
2.1 Definitions of ‘N’ in Data Center ........................................................................................................... 3
2.2 Rating Levels Overview ......................................................................................................................... 3
3.0 Power Redundancy Strategies for Data Centers ........................................................................................ 4
3.1 Rated-1 Data Center: General Power System Designs ........................................................................... 4
3.2 Rated-2 Data Center: General Power System Designs ........................................................................... 5
3.3 Rated-3 Data Center: General Power System Designs ........................................................................... 6
3.4 Rated-4 Data Center: General Power System Designs ........................................................................... 7
3.5 Substations and Power Feed Requirements .......................................................................................... 8
3.6 Choosing the Appropriate Rating Level ................................................................................................. 8
3.7 New Approaches to Redundancy and Rating ........................................................................................ 9
3.8 Redundancy Level Design and Risk Management ................................................................................. 9
4.0 Backup Power System: UPS .................................................................................................................... 10
4.1 UPS Types: Dynamic UPS vs. Static UPS .............................................................................................. 10
4.2 Types of Static UPS System ................................................................................................................. 12
4.3 UPS Modes VFI ................................................................................................................................... 14
4.4 UPS Parallel Configurations ................................................................................................................ 15
4.5 UPS Energy-Efficiency & Green Options .............................................................................................. 18
5.0 Backup Power System: Batteries ............................................................................................................. 19
5.1 Battery Types for UPS ......................................................................................................................... 19
5.2 Battery Design Criteria........................................................................................................................ 20
5.3 Battery Bank Calculation..................................................................................................................... 22
5.4 Battery Charging Methods and Management ..................................................................................... 24
5.5 Battery Configuration and Paralleling ................................................................................................. 25
5.6 Battery Maintenance and Monitoring ................................................................................................. 26
5.7 Battery Safety and Compliance ........................................................................................................... 26
6.0 Backup Power System: Generator........................................................................................................... 27

II | P a g e
 6.1 Importance of Generators in Data Centers.......................................................................................... 27
6.2 Types of Generators Used in Data Centers .......................................................................................... 27
6.3 Generator Principles and Components ............................................................................................... 28
6.4 Generator Redundancy ...................................................................................................................... 29
6.5 Generator Paralleling and Synchronization ......................................................................................... 29
6.6 Generator Room Requirements .......................................................................................................... 30
6.7 Generator Sizing ................................................................................................................................. 31
6.8 Generator Principles and Compliance ................................................................................................. 31
6.9 Generator Principles and Practices ..................................................................................................... 31
6.10 Generator Monitoring and IoT Integration ........................................................................................ 31
6.11 Automatic Transfer Switch (ATS) ....................................................................................................... 32
6.12 Static Transfer Switch (STS) ............................................................................................................... 33
7.0 Power Redundancy Strategies for the Rack............................................................................................. 34
7.1 Common Techniques for Power Availability ........................................................................................ 34
8.0 Power Distribution Within the Data Center ............................................................................................ 36
8.1 Standard Cabling vs. Busbar Trunking (BBT) ........................................................................................ 36
8.2 Three-Phase / Single-Phase Usage ...................................................................................................... 38
8.3 Cable Distribution Methods ................................................................................................................ 38
8.4 Grounding in the Data Center ............................................................................................................. 39
9.0 Power Distribution: PDU & DB ................................................................................................................ 40
9.1 Isolation Transformer ......................................................................................................................... 41
9.2 Bonding of Natural & Ground ............................................................................................................. 41
9.3 Data Center Power Distribution .......................................................................................................... 42
9.4 Form Factors ...................................................................................................................................... 43
10.0 Power Supply Quality ........................................................................................................................... 44
Power Quality Standards (EN 50160) ........................................................................................................ 44
Common Mode Noise (CMN) ................................................................................................................... 44
10.1 Harmonics ........................................................................................................................................ 45
10.2 CBEMA / ITI© Curves ....................................................................................................................... 46
11.0 Protection Devices ............................................................................................................................... 47
11.1 Basic Single-Line Diagram ................................................................................................................. 47
11.2 Protection Devices in Power Systems................................................................................................ 47

III | P a g e
 11.3 Over-current Protection Devices ....................................................................................................... 48
11.4 Residual Current/Earth Leakage Protection ...................................................................................... 51
11.5 Sizing of Breaking Components ........................................................................................................ 52
11.6 Surge Protection ............................................................................................................................... 53
12.0 Electromagnetic Fields (EMF) ............................................................................................................... 55
12.1 Common Sources of EMF in Data Centers ......................................................................................... 55
12.2 EMF Effects on Data Center Components ......................................................................................... 55
12.3 Best Practices for EMF Mitigation in Data Centers ............................................................................ 56
12.4 EMF Compliance and Measurement Standards ................................................................................ 56
12.5 EMF Shielding Performance & Attenuation ....................................................................................... 57
13.0 General Power Knowledge in Data Centers ........................................................................................... 58
13.1 Real Power vs. Apparent Power ........................................................................................................ 58
13.2 Labeled vs. Actual Power Consumption ............................................................................................ 58
13.3 Data Center Power Density Guidelines ............................................................................................. 58
13.4 Light ................................................................................................................................................. 59
13.5 Thermographic scanning .................................................................................................................. 62
14.0 Data Center Power Efficiency ................................................................................................................ 63
14.1 Power Usage Effectiveness (PUE) ...................................................................................................... 63
14.2 Factors Influencing PUE .................................................................................................................... 63
14.3 Understanding PUE Values ............................................................................................................... 64
14.4 Issues with PUE ................................................................................................................................ 64
14.5 PUE Measurement Classes by The Green Grid .................................................................................. 64
14.6 Additional PUE Considerations ......................................................................................................... 65
14.7 Mixed-Source Environments ............................................................................................................. 65
14.8 Additional Metrics for Measuring Sustainability ............................................................................... 65

IV | P a g e
"Mastering data center power infrastructure is not a destination but a journey —
the more you learn, the stronger your foundation becomes."

V|Page
Data Center Power Infrastructure
1.0 Electrical Power System Overview
1.1 From Generation to Data Center
Ensuring a reliable and high-quality power supply is critical for data center operations. The power journey
begins at the generation source and follows a structured path through multiple stages before reaching IT
equipment. Below is a breakdown of this journey:

1. Power Generation
Electricity is generated from various sources, including:

• Traditional Power Plants (coal, natural gas, nuclear)

• Renewable Energy Sources (solar, wind, hydroelectric, geothermal)
• Step-Up Transformers at power plants increasing voltage for transmission
2. High-Voltage Transmission
The generated electricity is transmitted through high-voltage transmission lines (110kV–765kV) to minimize
energy loss over long distances. This process involves:

• Transmission Grid delivering electricity to various locations

• Substations regulating and routing power to different regions
3. Medium-Voltage Distribution

When power gets closer to the destination, it passes through:

• Step-Down Transformers at substations reducing voltage to distribution levels (13.8kV–33kV)

• Distribution Lines carrying power to cities, industrial areas, and critical facilities like data centers
4. Data Center Power Entry & Protection
Once the power reaches the data center, it goes through several steps to ensure it’s safe and reliable:

• Primary Power Feed: The data center receives power from the utility grid at medium voltage.
• Step-Down Transformers: Reduce medium voltage to low voltage levels required for IT and facility
loads (typically 400V Three Phase / 240V Single Phase).
5. Backup Power System & Redundancy

• Uninterruptible Power Supply (UPS): Battery-based systems that provide temporary power during
short outages or switching delays.

• Generators: Provide long-term backup power if the utility grid fails.

Additionally, several pieces of equipment, such as ATS, STS, and batteries, support power backup and
redundancy. We will explore each of these in detail in the following chapters.

1|Page
1.2 Increasing Energy Demands of Data Centers
Data center power consumption has been rising significantly, driven by increasingly power-intensive IT
equipment and the rapid expansion of artificial intelligence (AI) technologies. This surge in energy demand has
led to overloading in many facilities, accompanied by increased heat loads that pose substantial cooling
challenges.

Key factors contributing to this trend include:

• Higher transistor density and faster clock speeds increase power consumption, leading to greater heat
dissipation.
• Power consumption in data centers is projected to double approximately every four years.
• Higher equipment density due to more compact hardware designs results in increased power demand
per square meter.
• Rising heat loads and greater Delta-T intensify cooling challenges, requiring more efficient thermal
management solutions.

2|Page
2.0 Data Center Rating Levels Overview
The Data Center Rating Levels, defined by the TIA-942 standard, classify data centers based on their
infrastructure reliability. These levels range from basic systems with single paths to fault-tolerant setups with
full redundancy, ensuring minimal downtime. This chapter will explain each rating level in detail.

2.1 Definitions of ‘N’ in Data Center

N refers to a system that meets the minimum required capacity (base need) with no additional redundancy. If
any component fails, operations will be disrupted.

N+1 Redundancy means there is one extra unit, module, path, or system beyond the minimum required for
normal operation. This ensures that if any single component fails or needs maintenance, the system can
continue running without disruption.

2N Redundancy means that the system has two fully independent units, modules, paths, or systems for every
essential component of the base system. This ensures that if one unit fails or needs maintenance, the other
can take over without disrupting operations.

2(N+1) Redundancy means that there are two fully independent units, modules, paths, or systems, each with
an extra unit (the "+1"). Even if one unit fails or needs maintenance, the additional capacity ensures that
redundancy is maintained, and operations continue without disruption.

2.2 Rating Levels Overview

• Tier I or Rated-1 Data Center: Basic
o Single path for power and cooling distribution (N), with no redundancy.
o Multiple single points of failure, which could lead to downtime.
• Tier II or Rated-2 Data Center: Redundant Components
o Single paths for power and cooling distribution, but with redundant components (N+1).
o N+1 at the component level still has SPoF at various levels as the distribution path remains single.
• Tier III or Rated-3 Data Center: Concurrent Maintainable
o Dual power (N+1) and cooling distribution (N+1) paths, but only one path is active at a time.
▪ N+1 as path implementation leads to higher availability than N+1 at the component level
o Redundant components with the ability for concurrent maintenance, preferably isolated.
o One active path and one passive path: It should be possible to switch from the active path to the
passive path for scheduled maintenance without interrupting the ICT load.
o Switching between paths is automatic, though not required by standard.
• Tier IV or Rated-4 Data Center: Fault Tolerant
o Dual active power (2N) and cooling distribution (2N) paths, redundant components, and fault
tolerance.
o All paths are active and utilized, ensuring full redundancy.
o Switching between paths is automatic, with the ability to handle one failure at any given time.
o 2(N+1) redundancy provide two complete units, modules, paths, or systems with “+1” additional
capacity.

3|Page
3.0 Power Redundancy Strategies for Data Centers
3.1 Rated-1 Data Center: General Power System Designs
A Rated-1 data center features a single power distribution path with no redundancy, making it vulnerable to
multiple points of failure and human errors. Power is supplied from the national grid through a transformer. To
ensure continuity during a grid failure, a standby generator is in place, remaining off under normal conditions.

If the grid fails, a sensing circuit detects the outage and automatically starts the generator to restore power.
Both power sources—grid and generator—are routed through an Automatic Transfer Switch (ATS), which
seamlessly switches between them to maintain power availability. However, since there is only one UPS
serving the IT load, any maintenance or failure requires a complete shutdown.

Key Characteristics:

• No redundancy: Any failure or maintenance results in an IT load shutdown.

• Power switching: Grid and generator operate through an ATS for automatic transition.

• Availability: 99.67% uptime, allowing for 29 hours of downtime per year.

• Maintenance schedule: Two planned shutdowns annually, each lasting 12 hours.

• Use case: Suitable for small enterprises with internal server rooms that have a low criticality index.

4|Page
3.2 Rated-2 Data Center: General Power System Designs
A Rated-2 data center improves power uptime by introducing redundancy for critical components like UPS and
generators. However, it still has multiple points of failure and remains susceptible to human errors.

To enhance reliability, two UPS systems are deployed in a parallel redundant mode, operating concurrently,
and sharing 50% of the load. If one UPS system fails or is taken out of service for maintenance, the other
seamlessly takes over 100% of the load, ensuring continuous power protection. Similarly, generators are
connected to the same bus, each sharing 50% of the load and providing backup power in case of a grid failure.

Key Characteristics:

• Partial redundancy: UPS and generators operate in parallel, allowing for seamless load transfer in case
of failure.
• Availability: 99.75% uptime, allowing for 22 hours of downtime per year.
• Maintenance schedule: Three planned shutdowns every three years, each lasting 12 hours.
• Use case: Suitable for small to medium-sized enterprises requiring improved reliability.

5|Page
3.3 Rated-3 Data Center: General Power System Designs
A Rated-3 data center provides full redundancy for all power components, including UPS, generators, and
distribution paths. In case of a failure in the main distribution path, an alternate path ensures uninterrupted
power to the IT load, minimizing single points of failure (SPoF).

To further enhance resilience, a Tier III data center can incorporate two independent power feeds from the
national grid. However, not all IT systems are designed with dual power supplies. For such systems, a Static
Transfer Switch (STS) can be deployed, automatically switching between the primary and backup power
sources when needed, similar to an Automatic Transfer Switch (ATS).

Key Characteristics:

• Full redundancy: UPS, generators, and distribution paths ensure high availability.

• Concurrent maintenance: Two independent distribution paths allow maintenance without downtime.

• Availability: 99.98% uptime, with a maximum of 96 minutes of downtime per year.

• Power distribution:

o Major mechanical loads are powered by the grid or redundant generators.

o Critical mechanical loads (e.g., drain pumps, fans) are backed by UPS for continuous operation.

• Static Transfer Switch (STS): Ensures seamless switching for IT systems with a single power supply,
preventing disruptions.

6|Page
3.4 Rated-4 Data Center: General Power System Designs
A Rated-4 data center features the most robust power distribution design, ensuring the highest uptime of
99.99%, with a maximum of 48 minutes of downtime per year. It is designed for full fault tolerance, eliminating
single points of failure (SPoF) by incorporating complete redundancy across generators, grids, and distribution
paths.

In this setup, two independent power feeds (Feed A and Feed B) remain active at all times, supplying power
simultaneously. Under normal conditions, IT equipment in the computer room draws power from both feeds
concurrently, ensuring continuous power availability. This architecture enhances resilience, making the data
center capable of withstanding faults, human errors, and fire incidents without impacting operations.

Key Characteristics:

• Full fault tolerance: Generators, grids, and distribution paths are fully redundant, eliminating SPoF.
• Simultaneous power feeds: IT equipment receives power from both Feed A and Feed B under normal
operation.
• Concurrent maintainability: Supports maintenance and repairs without service disruption.
• Availability: 99.99% uptime, with a maximum of 48 minutes of downtime per year.
• High cost: The extensive redundancy increases construction and operational expenses

7|Page
3.5 Substations and Power Feed Requirements
1. Power Feed Redundancy Ratings

• Rated-3: Allows a single substation to supply both the active and passive power paths to the data center.
• Rated-4: Requires two separate active power feeds, each coming from different substations.
o Note: Onsite generation can be considered as a replacement for one of the substations in Rated-4
configurations.

2. Feed Routing Requirements

• Separate Routing: Power feeds must follow distinct paths (e.g., using separate trenches for each feed is
essential). A single trench for both is not acceptable.
• Building Entry: Feeds should enter different parts of the data center building to further ensure
redundancy.

3. Power Source Considerations

• No Requirement for Multiple Power Companies: ANSI/TIA-942 does not mandate using multiple power
companies for data center power feeds.

3.6 Choosing the Appropriate Rating Level

Primary Selection Criteria

The choice of rating level should be based on several business factors:

• Cost of Downtime: Consider the revenue lost per hour and the annual operational hours (8760 hours in a
year).
• Risk Tolerance: Assess the acceptable level of risk for business continuity.
• Budget: The available budget will influence the level of redundancy and rating that can be realistically
achieved.
o Some organizations may not be able to fully meet the desired rating level due to budget
constraints. In such cases, compromises may be necessary.

Targeting Different Rating Levels for Categories

• For TIA-942 compliance, it's possible to target different rating levels for different categories. For
example, Rating-2 for architectural elements and Rating-3 for other infrastructure components.
However, the overall data center rating will be based on the lowest rating level across all categories.

8|Page
3.7 New Approaches to Redundancy and Rating
ICT Layer Redundancy

• Modern Technologies: Advances in ICT (Information and Communication Technology) have made it
possible to achieve high redundancy at the ICT layer rather than at the physical facilities layer.

o Example: Mirroring systems, applications, databases, and storage between two Rated-2 data
centers can provide a Rated-4 level of availability from the end-user perspective.

Containerized and Virtualized Solutions

• Lower Physical Layer Ratings: Some containerized and virtualized solutions may have lower ratings at
the physical (facilities) layer but can achieve high redundancy and availability at the ICT layer, meeting
or exceeding higher rating standards from a functional perspective.

3.8 Redundancy Level Design and Risk Management

Understanding Redundancy Levels

• Rating Levels and Availability: While rating levels are important, they do not guarantee high
availability on their own. The performance and uptime of a data center depend significantly on how
it is maintained and operated.
• Operational Processes: Well-defined operational processes, continuous skills development, and
regular maintenance are essential for ensuring high availability in the data center.

Risk Reduction Strategies

To minimize the risk of IT system failure and breakdown, focus on the following key areas:

• Avoiding Human Error: Implement systems and processes that reduce the likelihood of human mistakes.
• Maximizing Maintenance: Invest in proactive and regular maintenance to keep systems in optimal
working condition and prevent unexpected failures.

9|Page
4.0 Backup Power System: UPS
Uninterruptible Power Supplies (UPS) are crucial in data centers for ensuring continuous, stable power. UPS systems
protect against power outages, surges, and other disruptions that could cause downtime and loss of data.

4.1 UPS Types: Dynamic UPS vs. Static UPS

There are two main types of UPS systems used in data centers:

Dynamic UPS Systems

• Composed of mechanical components such as diesel engines, power generators, and flywheels.
• Primarily available in higher kVA ratings (200 kVA – 3000 kVA).
• Heavy, with a 600 kVA system weighing approximately 8 tons.
• Noisier than other UPS systems.
• Simple design setup.
• No need for large battery banks; only a few batteries are required for starting the generator section.
• Does not require a controlled environment.
• Long lifespan and good efficiency.

Operating Dynamic UPS System

Normal Utility Mode:

• Power flows through the UPS, and the alternator maintains stable frequency and voltage at the output.

Switch to Diesel Mode (in case of power disturbance):

• The input breaker opens if there is a power disturbance.

• The flywheel provides ride-through energy, ensuring a smooth transition.

• The diesel engine activates, driving the alternator to generate power.

• The diesel engine continuously drives the alternator to supply consistent power to the load.

• The diesel engine is monitored and digitally controlled to maintain stable output voltage and frequency.

Return to Normal Mode:

• After a few minutes of operation, the diesel engine cools down and then return to standby mode.

10 | P a g e
Static UPS Systems
• Composed of electronic components such as rectifiers, inverters, control logic, and batteries.

• One of the oldest UPS technologies used in data center environments.

• Available in ratings from 500 VA to 1.6 MVA in single systems.

Key Features:

• Relatively complex power and control electronics.

• Requires a controlled environment for optimal performance.

• Relies on batteries that need to be replaced:

o Typically, every 3-4 years for a 5-year design life.

o Every 7-8 years for a 10-year design life.

o Batteries contribute to chemical waste and have an unpredictable cost factor.

Static UPS Technologies

• Offline UPS

• Line Interactive UPS

• True Online Double Conversion UPS (ideal for data centers)

• Other types include Delta Conversion, Hybrids, High-frequency transformer-less, Eco/smart mode UPS, and
Compressed Air systems.

European Norm for Classification (EN50091-3 / IEC 62040-3):

• VFD-Class (Voltage and Frequency Dependent) – Offline / Standby UPS

• VI-Class (Voltage Independent) – Line Interactive UPS

• VFI-Class (Voltage & Frequency Independent) – True Online Double Conversion UPS

Note:

• Only VFI-class UPS systems are ideal for mission-critical data centers.

• "V" stands for Voltage, "F" stands for Frequency, "D" means Dependent, and "I" means Independent.

11 | P a g e
4.2 Types of Static UPS System
1. Offline / Standby UPS (VFD) Desktop Level

Overview:

• Primarily used for non-critical applications, such as desktop PCs.

• Offers minimal or no power regulation. The incoming utility power flows directly to the load, and if the main
supply fails, the inverter starts, using the battery as a backup power source.
• Not suitable for devices that require stable and regulated power, especially if the UPS does not provide a true
sine wave output, as Power Factor Correcting (PFC) power supplies may fail.

Pros: Very cheap

Cons:

• Load not isolated from the utility

• Large transfer time (up to 12-14 ms)
• Large ripple voltage on battery
• Modified sine wave / true sine wave exceptional
• No frequency regulation
• No true battery test capability
• No bypass
• Only able to run for a short time (risk of overheating)
• Generally smaller kVA output (less than 2kVA)

2. Line Interactive UPS (VI) Office Level

Overview:

• Primarily used for non-critical workstation applications.

• Incoming power passes through the unit and is sent directly to the load.
• The unit provides voltage regulation but does not regulate frequency.

Pros: Cheap

Cons:

• 2-4 ms transfer times (some cases 4-8 ms)

• Load not isolated from the utility
• Ripple voltage on battery
• Some level of surge protection
• Some level of voltage regulation
• No frequency regulation
• ‘Modified sine wave’ (PWM) or true sine wave
• Generally smaller kVA output (less than 5 kVA)

12 | P a g e
3. True Online Double Conversion (VFI) Data Center

Overview:

• The only suitable option for mission-critical facilities, such as data centers.
• Provides continuous power without interruption, even when internal issues like high temperature or overload
occur.
• In case of a problem inside the UPS, the microprocessor activates the static bypass to protect the UPS
electronics, ensuring no disruption to the load.
• When in bypass mode, raw utility power is supplied to the load, which is not ideal but ensures availability until
the issue is resolved.
• The maintenance bypass, activated by service personnel, allows power to bypass the UPS for service and repair.

Pros:

• Isolates Load from Input AC: Completely isolates the load from incoming utility power.
• Rectifies Input AC: Converts incoming AC power to DC, providing better regulation.
• True Sine Wave Output: Generates a true sine wave output via the inverter, ideal for sensitive equipment.
• Always Full Load on Inverter: Designed to maintain a full load on the inverter, ensuring efficient operation.
• Long Runtime Capability: Designed to support long runtimes, making it reliable during extended power outages.
• Regulated & Isolated Voltage & Frequency: Provides regulated and isolated power to the load, ensuring stability.
• Synchronizes to Input Phase: The system synchronizes with the input phase when online, ensuring smooth
operation.
• Bypass Functionality: Includes both static and maintenance bypass options for flexibility in operation.
• Redundant Power Input: In most cases, the bypass/reserve power line input is connected to the same source as
the main input. (Rated 3-4 units may require a separate input for redundancy).

Cons: More Expensive: Higher cost due to the complexity and reliability required for mission-critical applications.

Quick Comparison Table

VFD VI VFI
Grid problem
Off-line UPS Line Interactive UPS On-line UPS
Surge Indefensibility Limited defense Full Defense
High voltage pulse Indefensibility Limited defense Full Defense
Transient over voltage Indefensibility Limited defense Full Defense
Voltage falls Limited defense Limited defense Full Defense
Noise Limited defense Limited defense Full Defense
Deviation of frequency Indefensibility Indefensibility Full Defense
Continuous low voltage Full Defense Full Defense Full Defense
Interruption of the utility Full Defense Full Defense Full Defense

13 | P a g e
4.3 UPS Modes VFI
UPS in Normal Mode

• When the UPS is running in ‘normal mode’ the input

power goes to the rectifier, and the rectifier converts
the AC power to DC power which normally charges
the battery and powers the invertor.
• The static and maintenance bypass are at this stage
not providing any power to the output so there is full
conditioned power going to the IT load

UPS Running on Batteries

• Energy is drawn from the battery bank, which supplies

power to the inverter.
• This transition is seamless, with no transfer time,
ensuring uninterrupted operation of IT equipment.
• When input power to the UPS is restored (Normal
Mode):
o The rectifier powers the inverter.
o The batteries are simultaneously recharged.

UPS Running on the Static Bypass

• The static bypass is regulated by the UPS control logic.

• If the microprocessor detects an issue (e.g., high
temperature, overload, or component failure), the
UPS activates the static transfer switch (static bypass).
• In this state, there is no power conditioning or battery
backup, making it an undesirable UPS mode.
• Immediate action is required to restore proper
operation.
• Avoid touching anything upstream, as it may cause a voltage drop and disrupt power delivery to the IT load.

UPS Running on Maintenance Bypass

• The maintenance bypass is a manually activated

switch that ensures safe servicing of the internal unit
by turning off power to the electronics.
• After servicing, switching is done in reverse order,
with the inverter running in parallel with the
maintenance power.
• The final step is to turn off the SW-MB switch,
allowing the load to run fully on the inverter.
• This procedure ensures uninterrupted power to the
load during servicing.

14 | P a g e
4.4 UPS Parallel Configurations
Benefits of Paralleling UPS Systems:

• Increased Availability of Conditioned Power

• Increased Power Rating

Types of Parallel Configurations:

• Bypass / Hot-Standby Parallel

• Parallel Redundant

• Isolated Parallel Redundant

Centralized Battery Bank:

• Multiple UPS systems connect to the same battery bank to reduce costs.

• Not Advisable: Sharing a centralized battery bank between multiple UPS systems is generally not recommended
due to reliability concerns.

When to Use Parallel UPS Systems:

• When a high KVA rating is required, paralleling UPS systems becomes necessary, as a stand-alone UPS has a limit
on the KVA it can deliver.

• Ideally, each UPS system should have its own dedicated battery bank for maximum availability and reliability.

Bypass / Hot-standby parallel (Not Recommended)

Overview:

• An outdated concept with higher energy efficiency.

• Considered the least reliable parallel configuration.

How It Works:

• One UPS remains active, while the other stays in standby.

• The redundant UPS is feeding the bypass line of one or multiple UPS

• If the active UPS fails, it switches to bypass, drawing power from the standby UPS.

• Conditioned power is still supplied from the backup unit when the active UPS is in bypass.

Key Concerns:

• Requires the standby UPS to have the same capacity as the entire cluster.

• A failure in the active UPS (e.g., system card or static switch issue) can cause a total cluster failure.

• Overload risk: If the active UPS cannot handle the load and fails over to the standby UPS, the standby unit may
also overload, leading to a complete system collapse.

15 | P a g e
Parallel Redundant UPS (50% Good)

Overview:

• All UPS units in the cluster share the load.

• Potential cost savings.

• Requires all UPS systems to be from the same vendor.

How It Works:

• Each UPS has its own rectifier, inverter, and battery bank.

• Only one single static and manual bypass cabinet is used for maintenance.

• All UPS units must be synchronized to ensure matching output sine wave and frequency.

Key Concerns:

• Single Point of Failure (SPOF): Each of these UPS has control over this cabinet and if errors appear they could
instruct to go to bypass. This particular bypass cabinet is now a SPOF.

• Scalability Limitations: Not flexible for future expansion.

• If errors occur, multiple UPS units may instruct the system to switch to bypass, potentially causing system failure.

Isolated Parallel Redundant (Good)

Overview:

• All UPS units in the cluster share the load.

• Each UPS operates as a fully standalone system.

• Easy to scale.

• Requires all UPS systems to be from the same vendor.

How It Works:

• Every UPS functions independently while working in parallel.

• Similar to a parallel redundant cluster, all UPS systems must be synchronized.

• The output sine wave and frequency must be identical for the configuration to function properly.

16 | P a g e
Requirements for Parallel Configurations

Key Principles:

• No Single Point of Failure (SPOF):

o Avoid centralized control, bypass cabinets, or components that could cause a SPOF, including:

▪ Single static switch cabinets

▪ Single parallel (master) control board

• Fully Automated Master/Slave Principle:

o If the master unit encounters issues, control will automatically transfer to one of the slave units, which
will then become the master, preventing SPOF.

o Any UPS unit can take on the master role at any time.

Load Distribution:

• Load difference between each UPS module should not exceed 5-10%.

• Each UPS in the cluster should share an equal load.

• Cabling Requirements:

o Input cables and output cables must be of the same length for each UPS to avoid balance regulation
issues.

▪ Input VS output cables don’t need to be at the same length

o While input and output cables don’t need to be of the same length, both input cables and output cables
for each UPS must match in length.

Communication Cabling:

• Communication cables must be redundant or configured in a loop with a broken indicator.

Cluster Size & Risk Reduction:

• The total number of units in a cluster should be kept low (fewer than 4 units) in a cluster to reduce risks such as
rotating currents.

Synchronization of Output to Input

• The output sine wave of the UPS inverter is synchronized to the bypass input voltage (not the rectifier).
• This synchronization ensures that in the event of a problem with the UPS (e.g., high temperature on the
inverter), the system can switch to bypass without interrupting the load.
• Since both the bypass and inverter are in sync, a make-before-break switch over can occur seamlessly, preventing
any disruption to the load.

17 | P a g e
4.5 UPS Energy-Efficiency & Green Options
As data centers consume significant energy, optimizing UPS systems for efficiency is critical for both cost savings and
environmental sustainability. Energy-efficient UPS systems are now available, offering improved performance while
minimizing energy waste.

Energy-Friendly Options
Energy-Efficient UPS Features

• Eco-mode: Automatically switches to a more energy-efficient mode when power quality is good, saving energy.

• Smart-mode: Adjusts to optimize efficiency based on the load and input power.

• Enhanced Eco-mode: Allows the UPS to switch from high-efficiency VFI mode to a lower-efficiency VFD mode
when conditions allow, providing additional savings.

However, these modes are less effective in areas with poor power quality, where energy savings may be minimal or
unavailable.

UPS Selection: What to Look For

UPS Specifications to Look out for

• UPS Class (VFI, etc.): Check IEC-62040-3 for the classification (e.g., VFI for mission-critical applications).
• Efficiency & Efficiency Modes: Look for Eco-mode or Smart-mode for energy savings.
• Crest Factor (Peak vs. Nominal): Aim for 3:1 or better for handling high transient loads without overloading.
• Power Factor (Input/Output): Higher power factor = more efficient, ensuring minimal energy loss.
• Input Total Harmonic Distortion (THDi): Lower THDi = better performance and less stress on the system by
minimizing harmonics.
• Output Total Harmonic Distortion (THDv): Lower THDv ensures clean output, especially for nonlinear loads.
• Rectifier Technology: Look for 6/12 pulse rectifiers and IGBT or SCR/Thyristor technology, which ensure efficient
conversion of AC to DC power.
• Inverter Bridge Technology: IGBT technology is ideal for efficient DC to AC-conversion, and isolation transformers
improve protection and noise reduction.
• Fault Logging & Measurements: The UPS should be capable of logging faults and providing measurements to
help with monitoring and troubleshooting.
• Isolation Transformer: Adding a K-13 or higher isolation transformer to the bypass line or output improves
power isolation and enhances protection.

18 | P a g e
5.0 Backup Power System: Batteries
Batteries are a crucial component of Uninterruptible Power Supply (UPS) systems in data centers, providing
backup power during outages and ensuring continuous operation of critical equipment. They bridge the gap
between utility power failure and backup generator activation. Without reliable batteries, UPS systems cannot
perform their primary function of maintaining power during short disruptions. As data centers grow in size and
complexity, the need for efficient, durable, and easily maintainable battery systems has become even more
important to ensure minimal downtime and maintain uptime reliability. This section will outline the role and
importance of batteries in supporting UPS systems for data center operations.

5.1 Battery Types for UPS

There are four main battery technologies used in UPS systems:

1. Flooded Cells (Old Technology)

Also Known As Wet or Vented Cells, typically used in UPS systems of 500 KVA and above.

• Pros:
- Long lifespan (15-20 years)
- Can operate at high currents
• Cons:
- Must be stored in vented rooms to safely expel hydrogen gas, which is highly explosive when mixed
with air.
- Needs maintenance, including regular water replacement
- Electrolyte handling is hazardous (can cause burns)
- Must be used upright for safety and performance
2. Sealed Lead Acid (SLA) / Valve Regulated Lead Acid (VRLA)
(Suitable for Small-Medium Data Centers)

• Design Life: 5 or 10 years (Note: Design life does not equate to service life)
• Service Life: 5 years design: Approx. 3-4 years, 10 years design: Approx. 7-8 years
• Service life depends on factors like discharge frequency, UPS ripple factor, and ambient temperature.
• Pros:
- Cost-effective
- No memory effect
- Low self-discharge rate
• Cons:
- Slow recharging time (+8 hours)
- Requires constant charge to prevent sulphate-destroying plates
- Can’t stand high temperature

19 | P a g e
3. Lithium-ion (Li-ion)
• Pros:
o Long lifespan (>10 years)
o Smaller footprint and lighter weight
o 10x more discharge cycles than SLA
o Faster charging capability
• Cons:
o Expensive (approximately 3 times more than SLA)
o Requires a battery management system (BMS) to ensure safe charging/discharging
o Stricter transportation regulations due to safety concerns

4. Nickel-Cadmium (Ni-Cd)
• Pros:
o Long lifespan (20 years)
o Can operate in high ambient temperatures
o Reliable and predictable performance
o Fast charging capabilities
• Cons:
o Requires special chargers in UPS systems
o Susceptible to memory effect (needs to be fully discharged before fully charging)
o Expensive (4-6 times more than SLA)
o Disposal is costly due to high chemical content

5.2 Battery Design Criteria

When designing a battery bank, several critical factors must be considered to ensure reliable performance, longevity, and
cost-effectiveness.

Key Design Criteria

1. Capacity (Autonomy Time): Defines how long the UPS must support the load during a power failure.

2. DC Bus Voltage: Determined by the UPS manufacturer and influences the number of battery blocks in a string.

3. Redundancy: Ensures backup power reliability (e.g., N+1 configurations).

4. Service Capability: Batteries must meet operational needs, including charge/discharge cycles.

5. Recharge Time: The faster the battery recharges, the quicker the UPS can restore full protection.

6. Battery Technology: Choice of Lead-Acid (Flooded/VRLA), Lithium-Ion, or Nickel-Cadmium, depending on

performance requirements.

7. Ambient Temperature: Affects battery performance and lifespan; higher temperatures reduce battery life.

8. Expected Lifespan: Batteries typically come with 5-year or 10-year design life options.

9. Fireproofing: Standard ABS casings are not fire-rated; fire-resistant enclosures may be needed.

10. Weight Load: Autonomy time and battery arrangement (floor-spread or rack-stacked) affect structural planning.

11. Cost: All the above factors influence overall system cost and budget considerations.

20 | P a g e
Battery Bank Terminology
• A battery bank consists of all the batteries connected to a UPS system. It provides the necessary backup power
and is designed to meet capacity and redundancy requirements.

• Battery String: A battery bank contains at least one string, but most UPS systems have multiple strings connected
in parallel to enhance capacity or redundancy.

• Battery Block: Each string is made up of individual battery blocks connected in series.

• Battery Cell: Each battery block contains at least one cell. For example, a 12V SLA/VRLA battery consists of six
cells (each cell providing approximately 2V).

Designing Battery Banks Tips

When designing a UPS battery bank, it is important to keep the sizing reasonable to avoid operational inefficiencies and
potential system complications.

1. Avoid Excessive Autonomy Times

• UPS systems are primarily used to bridge the gap between power loss and generator startup.

• Longer autonomy times are usually unnecessary and can introduce issues such as:

o Increased recharging requirements.

o Delayed recharging capability, often requiring additional chargers.

o The need for multiple parallel battery strings, which can create balancing challenges.

2. Limit Parallel Battery Strings

• Too many parallel strings lead to unequal charging/discharging, reducing battery efficiency and lifespan.

• To mitigate this, keep parallel battery strings to 2-3, and preferably do not exceed 4.

3. Charging Limitations

• UPS systems have a limited battery charging current capacity.

• Oversized battery banks may result in:

o Extended recharge times beyond acceptable limits.

o Inability to fully charge the batteries.

o The need for additional chargers, which introduces further complexity.

By maintaining a balanced battery bank design, you ensure optimal UPS performance, reliable recharging, and minimal
system risks.

21 | P a g e
5.3 Battery Bank Calculation
To calculate the battery bank, the following factors must be determined:

• Inverter Efficiency

• DC-Bus Floating Voltage of the UPS

o Defines the number of cells in the battery string.

• Minimum Discharge Voltage on the DC-Bus

o Defines the cut-off voltage per cell:

Min VPC = Battery Minimum Discharge Voltage / Number of Cells

Calculation Steps
Step 1: Determine UPS Required Load Capacity

• Calculate UPS Output Power:

WLoad = Load Apparent Power (kVA) × Power Factor (pF)
• Include inverter losses, and determine the energy capacity (Watts) needed:
WBattery = WLoad / Inverter Efficiency

Step 2: Determine the Number of Cells

• Calculate the number of cells required for the floating voltage:

#Cells = Battery Floating Voltage / Cell Floating Voltage
• Note: The number of cells is not necessarily equal to the number of battery blocks.
o Example: A 12V battery block consists of 6 cells (2V each).

Step 3: Determine Energy Required Per Cell

• Calculate energy (W) per cell:

WPer Cell = WBattery / #Cells

• Find the minimum voltage per cell:

Min VPC = Battery Minimum Discharge Voltage / #Cells

Step 4: Selecting the Right Battery Type

• Check battery specifications for the required Min VPC.

• Locate the corresponding value in the battery manufacturer’s discharge table.

• Identify the correct battery type based on required autonomy and power per cell.

22 | P a g e
Example: Battery Bank Calculation
Given:

• Load: 60 kVA / 48 kW
• Inverter Efficiency: 93%
• DC-Bus Minimum Discharge Voltage: 336V
• Number of Cells in Series: 192
• Required Autonomy: 30 minutes

Step 1: Calculate Required Battery Power

WBattery = 48 kW / 0.93 = 51.6 kW

Step 2: Determine Energy Per Cell

WPer Cell = 51,600 W / 192 Cells = 269 W/Cell

Step 3: Identify Minimum Voltage Per Cell

Min VPC = 336V / 192 Cells = 1.75V

Step 4: Selecting the Right Battery Type

Locate the 30-minute autonomy column in the battery discharge table.

Find the row corresponding to 269 W/Cell at 1.75V.

The closest match is 303 W/Cell at 1.75V, which corresponds to a 180 Ah battery .

23 | P a g e
5.4 Battery Charging Methods and Management
Calculating the Charger Current
• Charger current determines the speed of charging up to 85%.
• A typical recharge cycle takes up to 8 hours.
• Charging from 85% to 100% can take up to 16 hours, prolonging battery lifespan.
• Large battery banks may exceed the UPS rectifier’s charging capability, requiring additional external chargers.
• Issues with external chargers: Lack of control over external charger cards may cause unreliable battery tests
unless additional measures are taken.

Charging Current Requirement

• Rule of Thumb: Charging current should be ~10% of the battery bank’s total Ah capacity.
• Example: A 300Ah battery bank requires 30A charging current.
• Charging current availability depends on UPS output load.
o Ask the vendor: "What is the maximum available battery charging current at 100% UPS load?"

Charging via Rectifier

• The rectifier’s maximum current limits available power for both:
o Output load
o Battery charging
• Lower UPS load → Higher charging current available
• In VFI UPS systems, charging current depends on:
o Rectifier capacity
o Inverter load consumption

Charging via DC/DC Converter

• Charging current is limited by both:
o The rectifier’s maximum current
o The DC/DC converter’s maximum current
• Even if the rectifier provides sufficient current, the DC/DC converter may be a bottleneck.
• Ask the UPS supplier for the available charging current at 100% load.

24 | P a g e
5.5 Battery Configuration and Paralleling
Handling Insufficient Capacity in a Single Block
• If required W/Cell is not in the discharge table, consider parallel strings:
o Example: If 750 W/Cell is needed:
▪ 2× 230Ah batteries in parallel provide:
2 × 387 W/Cell = 774 W/Cell (sufficient capacity).

Parallel Strings and Redundancy

• Parallel strings can improve maintainability and add redundancy.
• Keep parallel battery strings to 2-3, and preferably not more than 4.
• If more than 4 parallel strings are needed, switch to 2V battery blocks (higher capacity per cell).

Parallel Strings in Battery Banks (Incorrect Setup)

• Common but problematic configuration:
o Strings connected in a daisy-chain configuration.
o Unequal cable lengths from battery bank to UPS.
• Issues:
o Unequal charge/discharge between strings.
o Battery swelling due to imbalanced loading.

Parallel Strings in Battery Banks (Correct Setup)

• Best practices:
o All interconnecting cables must be the same length and diameter.
o Connections to the UPS must also have equal-length cables.
o This ensures equal impedance, leading to uniform charging and discharging across strings.

Guidelines for Parallel Battery Banks

• Minimize the number of parallel strings.
• Equalize cable lengths for:
o Inter-string connections
o UPS-to-battery bank connections
• Each string should have its own fuse/breaker.
• Use DC-rated fuses/breakers (AC types require careful review).
• Shorter cables = Lower voltage drop.
o DC cables have greater voltage loss than AC.

25 | P a g e
5.6 Battery Maintenance and Monitoring
• UPS battery monitoring is basic, often based on string monitoring instead of block-based monitoring:
o Voltage-based monitoring provides limited insights (UPS reading).
o Impedance/resistance-based monitoring offers better early failure warnings (manually handheld tester
every 3-6 months at block level).
• Test under different conditions:
o Floating: Normal UPS mode (measure voltage before)
o Under Load: Input power off, UPS running on batteries (measure voltage after)
• Testing Approaches: Spot checks and Continuous monitoring

Internal Resistance & Capacity

• Ask the battery supplier for the nominal ohmic value.
• Battery capacity based on resistance increase:
o 50% increase (i.e. 150% ohmic value, 1 milli Ohms-> 1.5 milli Ohm) → 80% capacity remaining,
o 100% increase (i.e. 200% ohmic value, 1 milli Ohms-> 2 milli Ohm) → 50% capacity remaining
o 150% increase (i.e. 250% ohmic value, 1 milli Ohms-> 2.5 milli Ohm) → 25% capacity remaining
o >150% increase → Consider replacing the battery

Battery Bank Configuration & Replacement

• All battery blocks in a bank must match in:
o Capacity (Ah rating)
o Age
• Replacing individual batteries:
o Allowed only if the new battery’s internal resistance is within 5-10% of existing blocks in the string.
o Entire strings must be replaced together.
• Unequal resistance in either blocks or string can lead to:
o Overcharging
o Battery damage

5.7 Battery Safety and Compliance

• Standard casing: ABS (Acrylonitrile Butadiene Styrene)
• Flame retardant casing options (UL-94 rated):
o V2: Burning stops within 30s, allows flaming drips.
o V1: Burning stops within 30s, no drips allowed.
o V0: Burning stops within 10s, no drips allowed.
• Fire Risk Considerations:
o Standard ABS casing may continue burning if ignited.
o Flame-retardant cases are recommended when no gas-based fire suppression is available.

26 | P a g e
6.0 Backup Power System: Generator
6.1 Importance of Generators in Data Centers
Generators provide backup power to maintain the seamless operation of data centers during utility power outages.
Without generators, the cooling systems in data centers would cease to function, leading to overheating and potential
hardware failures. To prevent such risks, data centers employ redundant generator sets, ensuring that power remains
available even if one generator fails.

6.2 Types of Generators Used in Data Centers

Data centers use different types of generators based on their power requirements, fuel availability, and operational
demands. The three main types include:

1. Standby Generators

o Most commonly used in data centers.

o Designed for use when utility power is generally reliable but may experience occasional outages.

o Typically rated for a maximum of 80% average load factor and around 200 hours of operation per year.

2. Prime Generators

o Provide variable load power for an unlimited number of hours per year.

o Should not exceed 70% of the prime power rating over any 250-hour operational period.

o If the engine is running at 100% prime power, yearly hours should not exceed 500 hours.

o Have a 10% overload capability available for 1 hour within a 12-hour cycle.

3. Continuous Generators

o Designed to supply power at a constant 100% load for unlimited hours.

o Used where grid power is unavailable, making them suitable for off-grid data centers.

Additionally, high-power generators with high-voltage ratings and lower RPMs (1,500 RPM for 50Hz, 1,800 RPM for 60Hz)
offer cost savings and durability. These generators experience less wear and tear, resulting in a longer lifespan and lower
maintenance costs.

Furthermore, environmental factors impact generator performance and must be considered during selection. Extreme
weather, altitude, air quality, and natural disasters all play a role. Cold climates cause fuel to gel, while hot climates risk
overheating; altitude reduces oxygen, requiring engine adjustments. Pollutants damage components, and natural
disasters demand disaster-resistant generator rooms for reliable operation.

27 | P a g e
6.3 Generator Principles and Components
Generators function based on the principle of electromagnetic induction, where a rotating coil within a magnetized
environment produces electrical energy.

The key components of a standby generator include:

1. Combustion Engine: The prime mover that drives the alternator to generate electricity, using diesel or gas as
fuel. While diesel is the standard in data centers, natural gas is gaining traction for its cleaner emissions.

2. Alternator: The alternator is responsible for converting mechanical energy into electrical energy.

o Proper Sizing – Prevents glazing and operational inefficiencies from low-load operation.

o Harmonics – High harmonics require an oversized alternator to ensure stability.

o Future Growth – Allows for increased load capacity without performance issues.

3. Governor: Regulates the fuel supply and maintains generator frequency.

o Mechanical governors: ±5% speed variance.

o Electronic governors: ±0.25% speed variance, enabling better efficiency and synchronization in parallel
configurations.

4. Fuel Storage:

o Fuel is stored in separate day tanks and storage tanks for redundancy (Tier3/4).

o Redundant fuel pumps and crossover circuits enhance reliability.

o Fuel refuel points must be secure, above flooding area, and easily accessible.

5. Voltage Regulator: Ensures stable voltage output.

6. Starter motor and Battery: cranks the engine using battery power to initiate operation until it runs
independently.

7. Control & Monitoring System: Oversees generator performance and automates operations.

8. Air Intake, and Exhaust System: The air intake supplies oxygen for combustion, while the exhaust system
removes combustion gases for safe operation.

Effective load management, variable speed drives (VSDs), exhaust heat recovery, and fuel efficiency all contribute to
optimizing generator performance. Automatic load-sharing ensures generators stay within optimal load ranges for better
efficiency and longevity. VSDs adjust generator speed based on demand, improving fuel efficiency and reducing wear.
Exhaust heat recovery repurposes waste heat for building or water heating, enhancing energy efficiency. Using high-
efficiency engines and clean-burning fuels optimizes consumption and minimizes emissions.

28 | P a g e
6.4 Generator Redundancy
Redundancy ensures power continuity during generator failures, with varying reliability levels.

• N+1 Redundancy: One extra generator than needed. If one fails, the spare ensures sufficient power. Balances
cost and reliability.

• 2N Redundancy: Enough generators to fully backup the load if any generator fails. Ideal for mission-critical data
centers with no tolerance for downtime.

• 2N+1 Redundancy: Extra backup units for maximum reliability. Used in high-uptime data centers, though more
costly.

6.5 Generator Paralleling and Synchronization

1. Automatic Transfer Switch (ATS)

• Automatic Detection: Detects power loss from the main supply and activates the backup generator
automatically, ensuring seamless switching between primary and backup power sources

• No Downtime: Transfers power without interruption, ensuring smooth operation.

• Fast Response: Switches power in milliseconds to prevent disruptions in critical systems.

• Prevents Downtime: Essential in mission-critical environments like data centers and industrial facilities.

2. Synchronization Basics

• Purpose: Ensures multiple generators work together, they must be synchronized to avoid damage and ensure
smooth operation.
• Synchronization Parameters:

• Voltage: Must be identical across all generators to prevent circulating currents.

• Frequency: Must match to maintain stable power output.

• Phase Angle: Must align to prevent electrical faults and imbalances.

• Automatic Paralleling Switch Gear (APSG):

• Starts multiple generators and connects them to a power bus.

• The first generator reaching the correct voltage and frequency becomes the lead unit.

• Other generators synchronize to match the lead generator.

• Advanced APSG Features:

• Monitors load and redundancy levels.

• Activates only the necessary generators based on real-time demand.

• Uses rotating schedules to evenly distribute generator usage and minimize wear.

29 | P a g e
3. Load Distribution and Balancing

• Objective: Evenly distribute the electrical load among multiple generators to prevent overloading.
• Key Mechanisms:

• Shared Power Bus: Connects all generators, allowing load sharing instead of being tied to just one.

• Automatic Load Sharing: Monitors generator load and redistributes power to prevent overloading.

• Dynamic Load Adjustment: Automatically adjusts load distribution in real-time based on demand.

• Essential Equipment:

• Paralleling Switchgear (PSG): Ensures generators operate together and share load efficiently.

• Automatic Load Sharing Controllers: Monitor and adjust load distribution dynamically.

• Current Transformers (CTs): Measure generator output to maintain balance.

• Voltage Regulators: Keep voltage steady for smooth power sharing.

• Benefits:

• Prevents Overloading: Reduces the risk of generator failure.

• Improves Efficiency: Ensures optimal capacity utilization and fuel efficiency.

• Reduces Wear and Tear: Distributes load evenly, extending generator lifespan.

6.6 Generator Room Requirements

The generator room must be designed to accommodate maintenance activities and ensure safety. Key considerations
include:

• Adequate Airflow: The room must have sufficient and unobstructed airflow to prevent overheating.

• Oil Trays: Oil trays should be installed to prevent oil spills on the floor.

• Temperature Control: Proper temperature control is essential to avoid overheating.

• Noise Levels: Noise levels should comply with local regulations.

• Exhaust Location: The exhaust should be positioned to prevent fumes from entering the data center.

• Security: The generator room should be secure, with access controlled via locks and monitored by CCTV.

30 | P a g e
6.7 Generator Sizing
Proper sizing of generators depends on multiple factors:

• Total kVA/kW requirement of the connected load.

• Expected inrush currents during operation.

• Harmonic levels returning from the connected installation

The total generator capacity can be calculated using the sum of the following:

• UPS load (1.8x or 2.5x multiplier based on harmonics level).

• Cooling and light loads (typically 1.5x multiplier).

6.8 Generator Principles and Compliance

Fuel storage regulations define the amount of fuel required based on data center ratings:

• Rating 1: Minimum onboard fuel tank filled up to 80%.

• Rating 2: 24 hours of fuel storage.

• Rating 3: 72 hours of fuel storage.

• Rating 4: 96 hours of fuel storage.

Data centers must follow local fuel storage regulations and have refueling agreements to stay certified. A smaller fuel
reserve is usually acceptable if a refueling plan is in place. To ensure reliable backup power, generator sizing should be
based on full-load fuel consumption with extra capacity.

6.9 Generator Principles and Practices

Routine testing and maintenance are critical to ensure reliability:

• Free-running tests: At least monthly for 15-20 minutes.

• Full load tests: At least quarterly using load banks or actual data center load.

• Fuel quality checks: Monitoring for contamination, microbial growth, and gelling.

• Emergency lighting: Available at generator locations for safety.

6.10 Generator Monitoring and IoT Integration

• Real-Time Data: IoT sensors track key parameters (fuel consumption, temperature, vibration) and send data to a
centralized system for monitoring.

• Predictive Maintenance: IoT helps predict failures by detecting early signs (e.g., unusual vibrations, temperature
spikes), reducing downtime and repair costs.

• Efficiency Optimization: Historical data helps optimize load distribution, fuel settings, and cooling systems for
peak generator performance.

31 | P a g e
6.11 Automatic Transfer Switch (ATS)
• A device that automatically selects power from one source or an alternate source (e.g., switching
between utility and generator power).
• Used in power redundancy designs to ensure continuous operation.

Operating Principle:
• Typically used when the two sources are not synchronized (e.g., utility power and generator power).

• Mechanically Operated:
o Uses electromechanical contactors or relays to perform switching.
o Switching time is slow (typically >50 ms), which requires UPS support to prevent power
interruptions.

ATS Configurations & Types

• Open Transition ATS (Break Before Make) – Most Common
o Fully disconnects from the primary source before switching to the alternate source.
o Used in standard applications where short power interruptions are acceptable (covered by UPS).
• Closed Transition ATS (Make Before Break)
o Temporarily connects both power sources during transfer to create a seamless transition.
o Used when the two sources are synchronized (e.g., utility-to-utility transfer or generator-to-
generator).

ATS Considerations
• Switching Speed:
o Slower than STS; requires a UPS to cover transfer time.
• Redundancy:
o Used in N+1 or 2N power systems to enhance availability.
• Monitoring & Control:
o Intelligent ATS units can be monitored remotely via BMS or DCIM for fault detection and
automation.

32 | P a g e
6.12 Static Transfer Switch (STS)
Function of the STS Panel
• STS – Static Transfer Switch
o A device that electronically switches power from one source to an alternate source with very
fast response times.
o Used in critical applications where even a slight power interruption is unacceptable.
o Faster than ATS since it uses semiconductor components (SCRs, IGBTs) instead of mechanical
contacts.
• Operating Principle:
o Break Before Make (BBM) Operation:
▪ Similar to ATS, but operates electronically.
▪ Switching time < 4 ms – 20 ms, significantly faster than ATS.
o Seamless Switching:
▪ Can instantly transfer between power sources if they are synchronized (e.g., two UPS
systems).
▪ Typically used inside data centers at rack, row, or room levels.

STS Configurations & Types

• Single-Phase vs. Three-Phase STS
o Single-Phase: Used for smaller loads, often at the rack level.
o Three-Phase: Used at room or facility level for large-scale power distribution.
• Dual-Input vs. Multi-Input STS
o Dual-Input STS: Standard design with two power sources (e.g., UPS A and UPS B).
o Multi-Input STS: Allows more than two sources for increased redundancy.

STS Considerations
• Ultra-Fast Switching:
o Prevents power interruptions and eliminates the need for UPS intervention during transfers.
• Synchronization Requirement:
o STS requires both sources to be in sync to avoid voltage and frequency mismatches.
• Placement in Data Centers:
o Deployed at rack-level or row-level for critical IT loads.
• Efficiency & Redundancy:
o STS helps optimize redundancy (e.g., in a 2N power architecture, STS can alternate between
two UPS feeds).

33 | P a g e
7.0 Power Redundancy Strategies for the Rack
7.1 Common Techniques for Power Availability
To ensure high availability and minimize downtime in data centers, implementing power redundancy and
reliability is essential. The following best practices should be adopted to create a robust power infrastructure:

• Dual Power Sources:

o Route dual power sources to each rack whenever possible to ensure redundancy.
o Ensure all conductors are of the same size
o For high harmonic content or phase imbalance, use double-sized neutral conductors to maintain
electrical balance.
• Use of Dual or Multi-Cord Equipment:
o Equip mission-critical servers and devices with dual-cord or tri-cord power supplies to enhance
redundancy and ensure uninterrupted power even if one power source fails.
• Static Transfer Switch (STS):
o Deploy STS for single-source equipment to provide seamless power switching in case of a fault.
STS operates as a "break-before-make" switch to ensure power continuity during the transition.
• Minimizing Single Points of Failure (SPOF):
o Reduce SPOF wherever possible by ensuring that all critical components have redundancy and
backup systems in place.
• Power Distribution Design:
o Use separate circuits for UPS and non-UPS power to avoid confusion and ensure clear
segregation of critical and non-critical loads.
o Ensure that each power rail or strip in the rack has its own breaker for protection.
• Breaker Placement for Isolation:
o Place circuit breakers as close to the load as possible to isolate faults and prevent disruptions.
o Avoid using shared breakers to reduce risks of power failures affecting multiple loads.
• Spare Breakers:
o Keep an adequate stock of spare breakers to quickly replace any faulty components and
maintain system uptime.
• Regular System Inspections:
o Conduct regular checks for load imbalances, earth leakage, and power quality issues to maintain
the health of the power distribution system.
o Before adding or relocating loads, ensure proper load distribution and balance across all phases
to avoid overloading any single phase.

By following these practices, data centers can achieve a high level of power redundancy and reliability,
ensuring continuous operation even during failures or maintenance activities.

34 | P a g e
Option 1: Low Availability
This setup relies on a single power source, where all equipment is powered
through one UPS and a single PDU. If either the UPS or PDU fails, the entire
power system is disrupted, leading to downtime.

Option 2: Slightly Better Availability

In this setup, two power rails/strips are connected to a single power source.
If one PDU fails, it will shut down the equipment connected to it, while the
other power rail remains operational, keeping the equipment powered from
the alternative side.

Option 3: Improved Availability

This option involves two power feeds (A+B) to the computer room, greatly
enhancing redundancy by eliminating a single point of failure. However, since
the equipment is still single-source, an STS (Static Transfer Switch) is used to
automatically switch between the power feeds, further improving the
system's resilience.

Option 4: High Availability (Example A)

This option enhances redundancy by using an STS. If a power strip fails,
whether from the A or B feed of the PDUs, the STS will seamlessly switch to
the alternate feed, ensuring continuous operation of all single-corded
systems without interruption.

Option 5: High Availability (Example B)

In this configuration, dual-source equipment is directly connected to the A&B
PDUs, while single-source equipment is connected to a third PDU, which in
turn is linked to an STS. However, the STS in this setup represents a single
point of failure (SPOF).

35 | P a g e
8.0 Power Distribution Within the Data Center
Effective power distribution within a data center is crucial to ensure consistent performance and uptime. It
should be designed to be:

• Reliable: To minimize downtime and ensure continuous operations.

• Scalable: To accommodate future growth and changes in power demands.
• Flexible: To allow for easy adjustments or expansion as needed.

Distribution Choices

• Location: The placement of power distribution systems can impact performance, especially in terms of
electromagnetic interference (EMI). To mitigate issues, ensure that power distribution systems are
appropriately spaced:
o Overhead: Distributing power cables above the ceiling, commonly used for large areas and easy access.
o Under Raised Floor: Running power cables below the floor, a common choice for data centers with a
raised floor design, facilitating easy cable management and airflow.
• Distribution Type:
o Standard Cabling: Conventional power cables used for standard power distribution, typically in smaller
setups or less complex environments.
o Busbar Trunking (BBT): A more advanced and flexible power distribution solution that uses a busbar
system to deliver power to various areas of the data center. It offers better scalability and reduced
losses compared to traditional cabling.

8.1 Standard Cabling vs. Busbar Trunking (BBT)

When choosing between standard cabling and busbar trunking (BBT) for power distribution in a data center,
each option offers its own set of advantages and trade-offs.

• Standard Cabling
o Less Flexibility: Once installed, it’s harder to reconfigure without disrupting the system.
o Fixed Power Rating: Typically, the power rating is fixed and can't be adjusted without replacing
components.
o Cost-Effective: Generally less expensive than busbar trunking, making it a good choice for
smaller or less complex installations.
• Busbar Trunking (BBT)
o Greater Flexibility:
▪ Supports both 1-phase and 3-phase power distribution.
▪ Adjustable power ratings to accommodate future growth.
▪ Allows for easy installation of breakers at each tap-off box for enhanced control and
maintenance.
o Higher Initial Cost: More expensive upfront compared to standard cabling.
o Reduced Electromagnetic Interference (EMI): BBT systems generally produce less EMI,
improving overall system reliability and minimizing potential interference with sensitive
equipment.
36 | P a g e
Busbar Trunking (BBT)
• Fully Enclosed Power Bars: Busbar trunking systems feature fully enclosed power bars, ensuring safe
and secure power distribution.
• Tap-Off Box: A tap-off box is connected to the busbar system, providing easy access for power
extraction.
o The tap-off box determines which phase(s) are supplied to the specific equipment.
o Circuit Breakers: Breakers are installed at each tap-off point to provide protection. If an
overload or short circuit occurs, the breaker will trip inside the tap-off box to prevent further
damage.
• Fixed Spacing Configurations: Most busbar systems come with fixed spacing between busbars, which
may not always align with rack spacing, especially in environments with mixed rack widths.
• Flexible Spacing Options: Some busbar systems offer non-fixed spacing, providing greater flexibility to
match specific rack or power distribution needs.

Power Cabling
Power cabling in a data center should be carefully selected based on factors such as material type (XLPE, PVC,
FR), power ratings (voltage and amperes), ambient temperature, physical protection, fire ratings, regulatory
requirements (e.g., fire rating, color coding), and ease of installation. Proper consideration should be given to
voltage drop, as cables have inherent resistance, leading to a drop in voltage over longer distances, which may
require compensation.

Installation Considerations

To ensure safe and efficient power distribution:

• Neutral Sizing: Use a neutral conductor appropriate for the system's needs:

o Half size (cost-saving but less recommended)

o Same as live conductor (recommended)

o Double size (for high harmonics/imbalance, typically x1.7 live wire, to prevent overheating)

• Ground Wire: Should match the neutral wire to enhance safety and minimize electrical noise.

• Cable Routing: Use generous bend radii for power cables and ensure secure fixing and protection via
ladders, trays, or trunking.

• Flexibility & Safety: Allow for movement in seismic zones and ensure proper labeling for easy
identification.

37 | P a g e
8.2 Three-Phase / Single-Phase Usage
Three-phase power distribution within the building consists of five wires: three active (phases), one neutral,
and one ground, with the conversion to single-phase power typically carried out in the distribution boards
(PDU/DB/PPC).

Most data center UPS systems operate on three-phase power, while most ICT equipment is designed for single-
phase power.

Significant imbalances between phases can lead to power quality issues and may affect UPS performance.

Key Considerations:

• Loads should be balanced as evenly as possible across all three phases.

• Ideally, load imbalance should not exceed 10-15% among the 3 phases.
o Imbalances beyond this can result in higher harmonic currents, leading to power problems that
impact the reliability of IT equipment.
• It’s recommended to connect Rack1-Red, Rack2-Yellow, Rack3-Blue, Rack4-Red and so on for both legs
to enhance the load balance.
• Parallel UPS Configuration:
• When running two UPS systems in parallel, balance the load across all phases.
• If one UPS fails or undergoes maintenance, the other will remain unaffected, as long as load balancing
is maintained.

8.3 Cable Distribution Methods

• Trunking: Used to route cables from the PDU to various equipment racks in the data center.
• Trays: Typically used for power and network cables that are routed under the raised floor.
• Ladders: Used for cabling between different floors through cable shafts.
• Baskets: Used for network cabling under the raised floor or hanging from the ceiling.

38 | P a g e
8.4 Grounding in the Data Center
• Grounding/Earthing: The terms "grounding" and "earthing" are interchangeable, both referring to the
process of connecting electrical systems to the earth.
• Ground Wire Functions:
o Voltage Reference: Establishes a reference point for voltage levels.
o Fault Clearing: Facilitates the clearing of electrical faults.
o RF/ESD Discharge: Provides a path for radio frequency (RF) and electrostatic discharge (ESD),
often through a Single Reference Ground (SRG).
o Lightning Protection: Carries lightning currents safely to the ground.
• Grounding Requirements for Data Center Equipment:
o Low Impedance Earth Point: The ground should be at a low impedance, minimizing noise or
electrical interference that could affect the equipment.
o Bonding at Building Entry: Grounding should be bonded as the power enters the building to
establish a solid connection.
o No Unauthorized Taps: Ensure the ground point is dedicated and not tapped into by other
systems.
• Grounding Infrastructure:
o A separate ground bar is created within the data center, and all metal components (e.g., PDUs,
racks, raised floors) are connected to it.
o A separate grounding cable runs from this bar through the building and connects to the earth’s
ground system.
• Ground Resistance:
o The ideal ground resistance should be less than 1 Ohm, with a maximum allowable resistance
of 5 Ohms.
o Current Distribution: Electrical current will flow through the path with the lowest resistance, so
maintaining low resistance is essential for effective grounding.
• Bonding of Metal Objects: All metal parts in the data center, including:
o Cabinets
o PDUs
o Air Conditioning Units
o UPS Systems
o Raised Floors

Should be properly bonded to the ground to prevent electrical hazards and ensure system integrity.

39 | P a g e
9.0 Power Distribution: PDU & DB
Power Distribution Units (PDUs) and Distribution Boards (DBs) are critical components in the electrical system
of a data center, responsible for managing and distributing electrical power to various loads.

(EN60439-1)

• Ambient operating -5 up to +40 °C

• Average temperature over 24 hours should not exceed +35 °C
• Indoor installations require clean air
• Relative humidity not exceeding 50% at a maximum temperature of +40°C
• Buy local and install local (Compliance regulations)
Key Requirements:

• Isolators for maintenance and safety: Essential to disconnect parts of the system for servicing without
disrupting the entire infrastructure.
• Over-current protection devices: Safeguard against overloads and short circuits to prevent equipment
damage.
• Residual current leakage protection: Protects from electrical shocks and ensures personnel safety.
• Surge protection: Safeguards equipment from voltage spikes, protecting sensitive IT equipment from
damage.
• Compliance with form factors and standards: Ensures adherence to safety, operational, and reliability
standards.
• Internal isolation transformers: Minimize common mode noise (CMN) and harmonics, improving the
power quality delivered to sensitive equipment.
• Metering for monitoring power usage: Enables tracking power consumption for efficiency and load
management.
• Indicator lights and alarm interfaces for EMS/BMS integration: Allow real-time monitoring and
control, ensuring smooth operation.
• Spare breakers or expandable breaker slots for live additions: Ensure scalability of the power
distribution system as the data center grows.

This structured approach ensures reliable, scalable, and safe power cabling within the data center.

40 | P a g e
9.1 Isolation Transformer
Isolation transformers are crucial for enhancing the quality of power in a data center and protecting
equipment from electrical disturbances.

Key Functions:

• Galvanic isolation: Provides electrical isolation between input and output to prevent ground loops and
reduce electrical noise.
o Filtering of power disturbances: Helps reduce common power issues like surges, transients, and
harmonics, improving overall power quality.
o Reduction of Common Mode Noise (CMN): Helps in reducing noise that can affect sensitive
equipment.
• K-Factor Transformers:
o These transformers are specifically designed to handle high harmonic loads.
o K-13 transformers are the most common in data centers, as they can tolerate higher harmonic
content without overheating.
o Consider doubling the size of neutral conductors when there is a high harmonic load to prevent
overheating and improve safety.

When integrated with PDUs or at strategic points in the power distribution network, isolation transformers
provide the best performance in reducing electrical noise and maintaining system stability.

9.2 Bonding of Natural & Ground

Bonding refers to the connection of the neutral and ground systems to ensure safety and reliability of the
power system.

• Best Practices:

o Bonding should be done as close as possible to the ICT equipment to ensure the effective
operation of safety devices and minimize noise.

o Isolation transformers are based on delta-wye play a role in managing CMN and ensuring
effective grounding by separating neutral and ground at strategic points.

o Proper neutral-to-ground bonding enhances system performance and ensures that any stray
currents or noise are minimized, protecting sensitive equipment.

• Guideline: The IEEE recommends that the bonding point for the ground system be within 25 meters of
the ICT load, ensuring minimal impact from electrical noise.

41 | P a g e
9.3 Data Center Power Distribution
Good Performance:

• A centralized isolation transformer offers good performance for power distribution, depending on its
distance from the ICT equipment. The closer the transformer is to the load, the better the performance.

• In many data center setups, isolation transformers are used in conjunction with UPS systems. The
bonding point is typically located at the output of the transformer, with cables running from there to
various distribution panels throughout the data center.

• The UPS room should be positioned near the distribution boards (DB). IEEE standards recommend that
the bonding point should not be more than 25 meters away from the ICT load to ensure effective
power distribution and minimize voltage drop.

Best Performance:

• Power Distribution Units (PDUs) with built-in isolation transformers provide the best performance.

o These PDUs help achieve superior control over Common Mode Noise (CMN), ensuring better
power quality and reliability for sensitive IT equipment.

42 | P a g e
9.4 Form Factors
Form factors are important in ensuring the physical layout and safety of electrical systems, especially within
switchboards.

• Key Forms:

o Form 1 – No Separation: The switchboard is in a secure location. Failure of the switchboard

disrupts other areas fed from the same switchboard.

o Form 2 – Separation of busbars from other functional units: Separation ensures that a fault in
the switchboard does not affect all functional units, improving operational resilience.

o Form 3 – Separation between busbars and all functional units: Provides additional protection
by isolating functional units, ensuring that failure of one unit doesn't cause unacceptable
disruption.

o Form 4 – Complete Separation: Ensures that busbars, functional units, and their terminations
are fully separated, allowing for easy isolation and safe maintenance.

These forms help enhance safety by isolating faults and reducing the risk of widespread disruptions,
contributing to reliable and continuous power supply.

43 | P a g e
10.0 Power Supply Quality
The power supplied to data centers is not always stable, which can harm IT equipment and affect its
performance. Poor power quality can cause problems regularly, especially in areas with unreliable electrical
systems. These issues can lead to:

• System freezes or keyboard lockups that disrupt operations

• Reduced hardware availability due to unexpected shutdowns or failures
• Damage to sensitive electronics that rely on stable power.
• Loss of important data if the power cuts off suddenly.
• Grounding problems, which might still occur even with backup power systems, as proper grounding
depends on the system’s design, not just the UPS.

Power Quality Standards (EN 50160)

Key Parameters:

• Voltage: 230V ±10%

• Frequency: 50Hz ±10%
• Common Mode Noise: <1% of Phase-to-Neutral Voltage
• Total Harmonic Distortion of Voltage (THDv): <8% (for non-linear loads)
• Total Harmonic Distortion of Current (THDi): <12%

Common Mode Noise (CMN)

• Definition: CMN refers to the voltage difference between neutral and ground in an electrical system.
• Values and Recommendations:
o Preferred value: < 1 Volt (according to IEEE 1100 and EN 50160).
o Acceptable limit: up to 1% of the line/phase to neutral voltage.
o Vendor recommendation: typically < 3 Volts.
• Causes:
o Long cable runs, harmonics, and phase imbalance can increase CMN.
• Impact:
o CMN can cause heat generation in power supplies, reducing efficiency and potentially damaging
equipment.
• Grounding: Proper bonding of ground and neutral is done at the building entrance to minimize CMN.

Common Mode Noise (CMN) is a voltage imbalance between neutral and ground, which can lead to heating in
power supplies and system inefficiencies. Keeping CMN low is important for maintaining equipment health,
and proper grounding and cable management are essential in minimizing this issue.

44 | P a g e
10.1 Harmonics
Harmonics are waveform distortions caused by non-linear loads that draw current in pulses rather than a
smooth sinusoidal manner. These distortions create unwanted frequencies that are multiples of the
fundamental frequency (50Hz or 60Hz), affecting power quality.

Total Harmonic Distortion (THD): THD measures harmonic distortion in a power system and is categorized as:

• THDi: Measured on the input side of devices like UPS systems, it represents the extent of harmonic
currents drawn by non-linear loads.

o 6-Pulse Thyristor Rectifier: typically around 30%

o 12-Pulse Thyristor Rectifier: typically around 12%

o IGBT-based Rectifier: typically between 3-5%

• THDv: Measured on the output side of a UPS system, it indicates how much the voltage waveform is
distorted due to harmonic currents in the system.

o With linear loads, typically < 3-5%

o With non-linear loads, typically < 5-8%

High THDi leads to THDv, causing voltage distortions, overheating, and reduced system efficiency. Excessive
harmonics may require oversizing power components such as generators and transformers to handle
additional losses.

Harmonic Effects on Electrical Infrastructure

• Overheating of transformers, cables, and motors due to excessive harmonic currents.

• Voltage distortion, which can cause malfunctioning of sensitive electronic equipment.
• Increased losses in the power system, reducing overall efficiency.
• Poor power factor, leading to higher energy consumption and costs.
• Resonance conditions that may amplify certain harmonic frequencies, worsening power quality.

Harmonic Filtering and Reduction Techniques

Passive Harmonic Filters

Use inductors, capacitors, and resistors to block or absorb harmonics at specific frequencies.

Pros: Cost-effective, low maintenance, and improves power factor.

Cons: Requires periodic component replacement, Less effective under variable load conditions.

Active Harmonic Filters (AHF)

Use power electronics (IGBTs) to detect and cancel harmonics in real-time.

Pros: Adapts to changing harmonic levels and improves power factor.

Cons: Higher cost, requires regular maintenance, and has efficiency losses (~3%).
45 | P a g e
10.2 CBEMA / ITI© Curves
The CBEMA/ITI Curve defines the tolerance of Information Technology Equipment (ITE) to voltage
disturbances, such as sags, swells, and dropouts. It helps understand how equipment reacts to different levels
of voltage.

Key Points:

1. Prohibited Region: Voltage surges or sags in this range can damage IT equipment.

2. No Damage Region: Voltage variations here won't damage the equipment but may cause operational
issues (e.g., resets, malfunctions).

3. No Interruption in Function Region: Voltage changes in this region won't affect operation or cause
damage.

• Voltage Tolerance: IT equipment can typically handle a ±10% variation from the nominal voltage
without issues.

• 20 ms Tolerance: If the AC supply voltage drops to zero, the system can still run for 20 milliseconds,
allowing a smooth transfer between power sources (using a Static Transfer Switch (STS)) without
interrupting equipment operation.

The CBEMA/ITI curve ensures that IT equipment can tolerate minor voltage disturbances without damage, and
it’s used to design UPS systems and transfer switches that provide reliable power to IT systems in data centers.

46 | P a g e
11.0 Protection Devices
11.1 Basic Single-Line Diagram
• Purpose: Represents the power layout and flow within a facility.
• Essential Components:
o Overcurrent Protection Devices: Includes MCB, MCCB, and other circuit protection
mechanisms.
o Isolators (Switches): Used for safe disconnection of circuits (Note: Overcurrent protection
devices should not be used as isolators).
o Residual Current Protection Devices: Provides earth leakage protection to prevent electrical
hazards.
o Wrap-Around Bypass Switches: Enables maintenance and redundancy in critical systems.
o Cable Type and Size Indications: Ensures proper specification and selection for load handling.
o Metering: Monitors power consumption and system performance.
o Indicator Lights: Provides visual status updates on power distribution.
o BMS/EMS Connection Points: Allows integration with Building Management Systems (BMS) and
Energy Management Systems (EMS) for monitoring and control.

11.2 Protection Devices in Power Systems

• Over-current Protection (Circuit Breaker/Fuses):
o Protects the installation from overloads and short circuits.
o Ensures that wires and cables do not overheat or burn by cutting the circuit when current
exceeds safe levels.
• Isolators (Switches):
o Provides a means of safely disconnecting power from specific sections of the system for
maintenance or fault isolation.
• Earth Leakage Protection:
o Detects leakage of phase current to earth, typically due to faults in the system.
o Ensures the safety of personnel by preventing electric shock hazards.
o Implemented using Residual Current Breakers (RCBs), which disconnect the circuit in case of
leakage.
• Protection Device Types:
o Triple Pole and Neutral (TPN): Used for three-phase systems with neutral protection.
o Single Pole and Neutral (SPN): Used for single-phase systems with neutral protection.

47 | P a g e
11.3 Over-current Protection Devices
Some definitions:

Over-current:

• The level of current at which the protection device will trip, thereby breaking the flow of energy to
prevent damage or overheating.

Maximum Prospective Short Circuit Current:

• The maximum electrical current that can flow through a system under short-circuit conditions.
• Determined by the voltage and impedance of the supply system.
• Typically, in the range of a few thousand amperes.

Breaking Capacity:

• The maximum current that a fuse or circuit breaker can interrupt without being damaged or causing a
harmful electric arc.

Fuses
Fuse (Fusible Link):
• A fuse is a safety device consisting of a metal wire or strip that melts when heated by an excessive
electrical current, breaking the circuit.
Protection of MCB:
• Fuses are sometimes placed in front of a Miniature Circuit Breaker (MCB) to provide additional
protection.
• Fuses react faster than MCBs in certain cases.
Current and Time-Dependent:
• Fuses are classified based on their current rating and time characteristics.
• Three common categories of fuses:
o Ultrafast, Fast Blow, and Slow Blow.
o Indicated by FF, F, and T.
Blow Sequence:
• When using fuses with similar current ratings, the blow sequence can be determined based on the
shortest time blow, typically near a branch in the circuit.

48 | P a g e
MCB/MCCB/VCB/ACB
MCB: Miniature Circuit Breaker

• Typically, up to 100 Amps

• Trip characteristics are normally not adjustable
• Thermal or thermal-magnetic operation

MCCB: Molded Case Circuit Breaker

• Typically rated for 60 Amps and above

• Operates on thermal or thermal-magnetic principles
• Adjustable trip current
• Offers more flexibility in trip settings

VCB: Vacuum Circuit Breaker

• Rated current up to 3,000 Amps

• Uses a vacuum container to extinguish the arc created by breaking current
• Longer life expectancy between overhauls compared to ACBs
• Adjustable trip settings

ACB: Air Circuit Breaker

• Rated current up to 8,500 Amps

• Fully adjustable trip characteristics, including configurable trip thresholds and delays
• Often electronically or microprocessor-controlled
• Commonly used in main power distribution systems
• Typically arranged in draw-out enclosures for ease of maintenance
• Adjustable trip settings for customized protection

Indicators on over-current Devices

Over-current Break/Trip Point:

• Indicates the current at which the device will break or trip, typically specified in amperes (e.g., 16A – 32A).

Breaking Current Capacity:

• Measured in kiloamperes (kA).

• Based on the maximum prospective short circuit current.
• Depicted using a breaking curve to define how the device will react to different fault conditions.

Timings: The duration required for the device to trip under certain conditions.

Critical Settings:

• Settings must be performed by a licensed electrical worker as they are crucial for the safety and proper
functioning of the system.
• Incorrect settings can result in nuisance tripping or, conversely, inadequate protection for the electrical
installation.

49 | P a g e
Breaking Curve
• The breaking curve represents the maximum inrush current an over-current protection device can
handle within a defined time period.
• Inrush current is the initial surge of current when an electrical device starts, often much higher than
the normal operational current.
• Types of Breaking Curves:
o B – Curve: Handles 3 to 5 times the nominal current (for devices with low inrush currents).
o C – Curve: Handles 5 to 10 times the nominal current (commonly used for ICT loads).
o D – Curve: Handles 10 to 20 times the nominal current (typically used for motor loads).
• Important Considerations:
o Check the appropriate breaking curve for the application to prevent nuisance tripping.
o Inrush currents must be evaluated at both the branch and device level to ensure correct
protection.
Fuses vs Breakers
Fuses:
• Cheaper and cost-effective.
• Ideal for situations where no breaks are expected, such as in battery banks.
• Non-adjustable timing, set to a fixed value.
• Often cannot be replaced without power interruption.
• Once blown, fuses need to be replaced, except for PPTC fuses which are resettable.
Breakers:
• More expensive than fuses.
• Can be reset after an overload or fault, offering reusable protection.
• Some breakers offer adjustable timing to fine-tune protection.
• Inspect and replace breakers that have experienced multiple short-circuit events to maintain safety and
performance.

50 | P a g e
11.4 Residual Current/Earth Leakage Protection
Key Points:

• Earth Leakage protection is critical for human safety as current, not voltage, causes harm.
• Human body resistance varies based on conditions:
o Dry skin: Approx. 10,000 ohms (Ω)
o Moist skin: Approx. 1,000 ohms (Ω)
• Threshold limit for safety is generally 100V-AC.

Residual Current Device/Breaker (Earth Leakage Protection Device)

• Multiple names for the same device:

o RCD: Residual Current Device
o RCB: Residual Current Breaker
o ELCB: Earth Leakage Circuit Breaker
o GFCI: Ground Fault Circuit Interrupter (USA)
o ALCI: Appliance Leakage Current Interrupter (Canada)
o RCBO: Residual Current Breaker with Overload (combines ECB+MCB)
• Function:
o Trips if there is a leakage from live to earth.
o Does not protect against overcurrent or short circuits.
o Test button: Used for periodic testing of the device functionality.

Working of RCB and Key Indicators

• Measures the difference between phase current and the return current on the neutral. Normally,
the EMF fields of both wires cancel each other out.
• Trip mechanism: If a leakage occurs (e.g., when someone touches the phase), current will flow
through the human body to the ground, causing an imbalance in the EMF fields. The imbalance
increases the EMF field, and when it exceeds the preset threshold, the device trips to cut the
current.

Key Indicators on RCB:

• Continuous current: Indicated in amperes (A).

• Fault current: Indicated in milliamperes (mA).

51 | P a g e
11.5 Sizing of Breaking Components
Categories of Breaking Components:

• Isolators

• Over-current protection devices (e.g., MCB/MCCB)

• Residual current (earth leakage) protection devices (RCD/RCB)

Downstream & Upstream Components:

• Downstream: Components located closer to the load.

• Upstream: Components located closer to the main supply.

Proper Configuration:

• From upstream to downstream, the ratings of MCB/MCCB and isolators must decrease.

o Carefully review timing settings and breaking speed (Fuses vs. Breakers).

• On a busbar, multiple breakers might exist, and the total theoretical rating could exceed the upstream
breaker rating.

o Ensure the ratio of main versus sub-breakers is realistic and compliant with local regulations.

o During normal operation, the sum of currents drawn by all devices should not exceed the
upstream breaker setting.

o Regular measurements should be conducted, especially after major load changes, to verify
proper operation.

RCB Device Sizing:

• Ensure that Residual Current Breaker (RCB) devices have a rating equal to or higher than the upstream
breaking devices to maintain system protection.

52 | P a g e
11.6 Surge Protection
Surges are brief voltage spikes often induced by external factors such as lightning strikes or internal factors like
load switching and faults in the electrical supply. Effective surge protection is crucial to safeguard equipment
from these voltage increases.

Origin of Surges and Transients:

• External factors: Lightning, power line crossovers, and other natural causes.
• Internal factors: Switching of electrical loads and faults within the electrical supply system.

Norms and Standards:

• Surge immunity levels and testing methods are defined by standards such as ANSI/IEEE-C62.41 and
ANSI/IEEE-C62.45.

Surge Protection Devices (SPD) and TVSS:

• TVSS (Transient Voltage Surge Suppressor) and SPD (Surge Protection Device) are the common terms
used for surge protection.
o TVSS has been replaced in modern standards by the term SPD (Type 1, 2, 3, and 4).
• Installation: Surge protectors should be installed at various levels within the electrical supply chain,
such as at each Distribution Board (PDU/DB).
o Telecommunication cables should also have surge protection if the equipment does not have
built-in protection, particularly for copper cabling.

Technology Used in Surge Protection:

• Main surge protection technologies include:

o Metal Oxide Varistor (MOV): Offers high energy capacity.
o Silicon Avalanche Diode (SAD): Known for fast response time.
o Gas Tubes and LCR Filters.
o Often, a combination of technologies (e.g., SAD/MOV) is used for enhanced protection.
• Surge protection components vary in:
o Energy capability
o Reliability
o Cost
o Effectiveness

53 | P a g e
Types of Protection Provided:

• Surge protectors must provide adequate protection against:

o Phase-Phase
o Phase-Neutral
o Phase-Ground
o Neutral-Ground
• Some surge protectors may fail to provide adequate protection due to poor installation practices or the
use of unsuitable types.

Cascading Surge Protection:

• Cascading surge protection involves placing SPDs at each level of the electrical system to ensure proper
protection from lightning strikes, transients, and switching loads.
o It also ensures better suppression closer to the critical loads.

Proper Installation Practices for SPDs:

• SPD placement: This should be installed as close as possible to the feed and on every distribution
board.
o Long or thin cables can reduce the effectiveness of the SPD, so they should be kept short.
o Integrated SPDs are the most effective, but if external SPDs are used, ensure short cable lengths
to maintain protection quality.

By adhering to proper installation and cascading strategies, surge protection ensures the safety and longevity
of critical equipment in the data center.

54 | P a g e
12.0 Electromagnetic Fields (EMF)
Electromagnetic waves are composed of two related fields:

• Electric Field (E-field): Created by voltage and measured in volts per meter (V/m).

o Exists even when no current is flowing.

o The higher the voltage, the stronger the electric field.

• Magnetic Field (H-field or B-field): Created by electric current flow and measured in Gauss (G),
milligauss (mG), or Tesla (T).

o 10 milligauss = 1 micro-Tesla (uT).

o The greater the current, the stronger the magnetic field.

12.1 Common Sources of EMF in Data Centers

• Electrical Equipment:

o High Tension / High Voltage (HT/HV) power lines

o Transformers, UPS systems, and generators

o Power cables and lightning strikes

• Wireless Communications:

o Wi-Fi, Bluetooth, and canopy networks

o Radar and transmitting towers (TV, radio, GSM)

12.2 EMF Effects on Data Center Components

Impact on Integrated Circuits & Network Cabling

• Microprocessor Chips: EMF interference can lead to erratic behavior and system failures.

• Copper Network Cabling:

o Magnetic flux from nearby power cables may disturb data transmission.

o Shielded Twisted Pair (STP) cables are effective for high-frequency alien crosstalk but not for
low-frequency EMF.

Electromagnetic Pulses (EMP/HEMP)

• High-Altitude Electromagnetic Pulse (HEMP): Can disable entire mission-critical facilities.

• Sources:

o Atomic, neutron, or e-bomb explosions.

o Man-made via pumped flux compression generators and High Power Microwave (HPM) devices.
55 | P a g e
12.3 Best Practices for EMF Mitigation in Data Centers
During Design Phase
• Create distance between power equipment, cabling, ICT equipment, and personnel areas (e.g., NOC,
security rooms).

• Use optimized cable arrangements like trefoil formation or busbar trunking.

• Conduct EMF-level measurements post-construction to align with international standards.

• Use shielding materials certified by international bodies (e.g., TÜV).

EMF Protection Using Shielding

• Material Selection: Should be handled by experts.

• Techniques:

o Combination of absorption and reflection.

o Avoid welding.

o Consider de-rating when using cable trays.

o Use TÜV-approved materials for guaranteed performance.

12.4 EMF Compliance and Measurement Standards

Acceptable EMF Levels
• ICT Equipment Compliance: Based on IEC 61000-4-8 / NCRP.

• Human Health & Safety: 1.0 μT (microtesla) = 10 mG (milligauss).

Major EMF Sources in Data Centers

• Power Infrastructure:

o HT-supply transformers

o High current distribution boards

o UPS & isolation transformers

o Lightning rods and cabling

o High current power cables (single-phase vs. three-phase)

• Single vs. Three-Phase Radiation:

o Single-phase cables emit higher EMF than three-phase.

o Best practice: Use tri-foil arrangements to minimize emissions.

56 | P a g e
EMF Measurement Techniques
• Low-cost, single-axis meters: Measure composite EMF only.

• Mid/high-range, tri-axis meters: Measure composite and individual axes.

• High-end spectrum analyzers: Provide detailed frequency breakdowns.

Safe Distance Guidelines

• EMF Exposure Reduction:

o Maintain safe distances based on theoretical models.

o Distance minimizes shielding requirements.

o Shielding is best applied close to the protected area for cost-efficiency.

12.5 EMF Shielding Performance & Attenuation

• Attenuation Calculation:

o Formula: A = 20 log (R/M)

▪ A = Attenuation (dB)

▪ R = Real measured value

▪ M = Maximum acceptable value

o Example: If R = 80mG and M = 10mG:

▪ 20 log (80/10) = 18dB attenuation

• Material Permeability & Saturation:

o Different materials absorb EMF at varying rates.

o EMF shielding material performance is non-linear.

o Shielding tests must be conducted under full operational loads.

o Best time to test: During the Integrated Product Verification Test (IPVT) to ensure shielding
effectiveness at peak load.

57 | P a g e
13.0 General Power Knowledge in Data Centers
13.1 Real Power vs. Apparent Power
• Real Power (Watts): Actual power consumed by equipment.
• Apparent Power (VA): Combination of real power and power reflected back to the supply.
• Power Factor (PF): Ratio of real power to apparent power.
o Formula: W = VA × PF
o PF ranges between 0 and 1 (0 = no power, 1 = Real Power = Apparent Power).
o Example: A 100 kVA UPS with an output PF of 0.8 delivers 80 kW of real power.

13.2 Labeled vs. Actual Power Consumption

• Equipment label ratings indicate maximum power capacity, not actual usage.
• Real power usage is 20-40% lower than the label power.
• Some vendors offer web-based tools to estimate actual power consumption based on hardware
configuration, CPU load, etc.

13.3 Data Center Power Density Guidelines

Power Usage per Floor Space

• Low-density: 500 W/m²

• Medium-density: 1000-1500 W/m²
• High-density: 2000-2500 W/m²
• Extreme-density: >3000 W/m²

Power Usage per Rack

• Standard racks: 1-3 kW

• Mixed racks: 5-8 kW
• Blade server racks: >15 kW (up to 28 kW per rack!)

Electrical Basics: Single-Phase vs. Three-Phase Power

Voltage Conversions

• Voltage Line to Neutral (VLN) → Voltage Phase to Phase (VPP)

o Formula: VLN × 1.732 = VPP (e.g., 240V × 1.732 = 415V)
• Voltage Phase to Phase (VPP) → Voltage Line to Neutral (VLN)
o Formula: VPP / 1.732 = VLN (e.g., 380V / 1.732 = 220V)

Power Formulas

• Single-phase power: Voltage × Current = VA

• Three-phase power: Voltage × Current × 1.732 = VA
• Power Factor Impact:
o Real Power (Watts) = Apparent Power (VA) × PF
58 | P a g e
13.4 Light
• Adequate lighting is essential for maintaining a safe working environment.
• Lumen: Refers to the light strength or output.
• Lux: Measures the light intensity in an area, with 1 Lux being equal to 1 Lumen per square meter.
o Example: 1000 Lumens equals 1000 Lux in one square meter, and 1000 Lumens over 10 square
meters equals 100 Lux.
• Lux is the preferred unit of measurement for data centers and is typically measured 1 meter above the
raised floor.

Standards

• Relevant lighting standards include EN 12464-1 and DIN-5035.

• The recommended lighting levels are:
o Rooms: Minimum of 500 Lux, with 600 Lux recommended.
o Corridors: Minimum of 200 Lux, with 300 Lux recommended.
o Service areas: Minimum of 150 Lux, with 200 Lux recommended.

Light Fixtures and Placement

• Various types of light fixtures are available, but protected types are preferred for durability.
• Lights should be placed:
o At regular intervals
o In all aisles
o With a focus on the back of the racks, as this area is typically more important than the front.
• Power Supply Considerations:
o Lights should be connected to the raw power supply with a backup power supply from the
standby generator set.
o Fluorescent lights should never be connected to the same UPS that powers the ICT equipment
to avoid interference.

59 | P a g e
Emergency Light
• Purpose: Emergency lighting is critical to allow personnel to safely evacuate during a power failure.
• The required lighting level ranges from 1 Lux to 15 Lux, which is equivalent to the light provided by a
candle in the dark.
• Regulations often specify that emergency lights should be installed at ceiling level, although this can be
difficult to see during a fire due to smoke.
• It is advisable to exceed local regulations to ensure sufficient lighting for safety:
o Provide adequate lighting in each aisle or at least at the end of each aisle.
o Install emergency lights at lower levels to assist with evacuation during a fire.
• Emergency Lights Should Be Installed At:
o All critical switchboards
o UPS
o Standby generators
o Cooling equipment
• Torch Lights:
o Use quality torch lights with proper light strength.
o Regularly check and replace batteries.
o Glow sticks are a useful addition.

Types of Emergency Light

• Various types of emergency lights are available:

o Incandescent
o Fluorescent
o LED (Light-Emitting Diode)
• Standalone Emergency Lights:
o These are troublesome for maintenance, as both the light and the battery need regular
attention.
• Centralized Emergency Lights:
o EPS (Emergency Power Supply):
▪ Offers a more robust solution and can be located in the UPS room with its own battery
bank.
o EN 50171 (Central Power Supply Systems):
▪ Allows for easy centralized maintenance.

60 | P a g e
Savings on Lighting Infrastructure
1. Power-Saving Lights

• Replace Incandescent Lights with Fluorescent or LED Lights

o Opt for fluorescent light fixtures with high light output (lux) efficiency.
o Consider LED lights for longer lifespan and energy efficiency.
• Enhance Lighting Efficiency
o Use reflectors behind the light tubes to increase light distribution.
o Install grills instead of matt Plexiglas to reduce light diffusion losses.

2. LED Fixtures

LED lighting is highly energy-efficient and offers multiple advantages for data center environments.

• Advantages of LED Fixtures

o New/Retrofit Kits: Available for easy upgrades.
o Longer Lifespan: LEDs last significantly longer than traditional lighting.
o Color Tones: Select the right color temperature for optimal lighting.
o Check Lux Levels: Ensure lighting provides adequate brightness for data center tasks.
o Beam Spread:
▪ For ceiling heights between 2.4–2.7 meters (7.8–8.9 feet), use a beam spread of >60
degrees.
▪ For higher ceilings, choose a beam spread of 50 degrees or lower for proper coverage.

3. Automated Lighting Control

• Lights-Off Strategy:
o Motion Sensors: Automatically turn off/on lights based on movement in the data center.
o Access Control Integration: Lights can be tied to access control systems to turn off when no
personnel are in the area.
• ANSI/TIA-942 Lighting Protocol:
o Level-1: Unoccupied areas should have minimal lighting.
o Level-2: Initial entry areas should have moderate lighting.
o Level-3: Occupied areas should have full lighting levels.
o Optional override function: Allow for manual control when necessary.

4. Security Considerations

• Security Requirements: Ensure that light levels are sufficient for security cameras to function
correctly, even if low-level lighting is used for energy savings.
• Night CCTV: Even minimal lighting improves CCTV image quality, so a balance between energy
savings and security should be maintained.

61 | P a g e
13.5 Thermographic scanning
• Thermographic scanning is a technology used to detect heat by visualizing temperature variations.
• It enables proactive detection of high-resistance connections and high-current conducting wires,
which can develop heat and potentially cause:
o Breakdowns
o Fires
• Standards
• ANSI/NFPA 70B and SS507 recommend conducting a thermographic scan annually.
• Areas to Scan
• Main switchboards and PDUs
• Connections and breakers of high-power equipment, including:
o Isolation transformer
o UPS
o Generator sets (Gen-sets)
• Main power cables (three-phase) and bus bars, ensuring to include the neutral cable.
• Critical rack/equipment/breakers that require close monitoring.

62 | P a g e
14.0 Data Center Power Efficiency
Data center power usage is typically divided into three main sections:

1. Power Distribution for ICT Equipment: This part supplies power to the Information and
Communication Technology (ICT) equipment, often using an Uninterruptible Power Supply (UPS)
system to ensure clean and reliable power.
2. Cooling Infrastructure: A significant portion of power is dedicated to cooling systems, necessary to
maintain the ideal operating environment for the ICT equipment.
3. Auxiliary Systems: This includes power for lighting, monitoring systems, and other non-ICT related
components.

The total power consumed by ICT equipment is referred to as Total IT Power, while the total incoming power
required for all aspects of the data center, including distribution, cooling, and auxiliary systems, is called Total
Facility Power.

14.1 Power Usage Effectiveness (PUE)

The Power Usage Effectiveness (PUE) metric is widely used to measure data center energy efficiency, as
defined by The Green Grid consortium. It compares the total facility power to the power used by IT equipment.

Formula: PUE = Total Facility Power / Total IT Equipment Power

Since facility power is generally higher than IT power, the PUE value will always be a number greater than 1
(e.g., 1.x, 2.x). The closer the PUE is to 1, the better optimized the facility is.

Can PUE reach 1? Theoretically, a PUE of 1 means that all utility power is used exclusively for the IT equipment.
However, achieving a PUE of 1 is virtually impossible, as some energy will always be consumed by cooling,
lighting, and other auxiliary systems.

The location of a data center plays a significant role in its overall efficiency and sustainability. Factors such as
access to free cooling and the type of power generation source (e.g., nuclear, coal, hydro) can greatly influence
a data center's "green" footprint.

14.2 Factors Influencing PUE

PUE is affected by several factors:

• Legacy vs. New Data Centers: New, optimized data centers typically have lower PUE.
• Location and Ambient Conditions: Data centers in cooler climates often perform better due to the
availability of free cooling.
• Redundancy: High levels of redundancy (such as N+1 or 2N configurations) can increase energy
consumption.
• Computer Room Utilization: Better utilization of the computer room leads to more efficient cooling
and energy use.

63 | P a g e
14.3 Understanding PUE Values
PUE values are indicative of how well a data center is optimized. Here’s a breakdown of common PUE values:

• PUE = 1: Theoretically all power is used for IT equipment (not feasible in practice).

• PUE = 1.6: Optimized and achievable with appropriate measures.

• PUE = 2: Efficient, and a reasonable target for most data centers.

• PUE = 2.4: Average, with room for improvement.

• PUE > 3: Inefficient, indicating significant potential for optimization.

Note: These values assume a tropical climate, which significantly impacts cooling needs.

14.4 Issues with PUE

While PUE is a valuable metric, it has limitations:

• PUE Measures Power, Not Energy: Power measured over time is referred to as 'energy', but PUE does
not account for the duration or the energy consumed.

• PUE Alone Does not Reflect Full Sustainability: A data center might reduce its electricity consumption
for cooling but at the cost of increased water usage, making the data center less sustainable despite a
lower PUE.

• Confusion in Measurement: Despite the Green Grid's efforts to standardize PUE, different
interpretations and practices still lead to inconsistencies in measurements.

14.5 PUE Measurement Classes by The Green Grid

The Green Grid defines three classes of PUE measurement:

Measurement
Level 1: Basic Level 2: Intermediate Level 3: Advanced
Class
IT Equipment
UPS PDU Server
Power
Data center input minus
Data center input Data center input minus
Facility Power shared HVAC plus building,
power shared HVAC
lighting, and security
Measurement
Monthly/Weekly Daily Continuous
Interval
The data center input power is measured at the utility hand-off point, and different incoming power capacities
(e.g., voltage levels) may result in varying losses at onsite transformers.

64 | P a g e
14.6 Additional PUE Considerations
1. Carbon Footprint: The actual carbon footprint of a data center depends on the source of its power
generation (e.g., coal, nuclear, hydro).

2. Mixed-Use Buildings: In mixed-use environments, PUE is calculated in the same way as dedicated data
centers, but energy consumption for non-data center operations (e.g., office space, elevators) may be
excluded.

3. Alternative Power/Cooling Sources: PUE based solely on electricity consumption can be misleading for
data centers using alternative sources such as district cooling or renewable energy. Additional metrics
are required to assess the sustainability of these sources.

14.7 Mixed-Source Environments

In data centers using multiple energy sources, it's essential to apply weight factors to account for the different
energy types. The Green Grid defines the following weight factors:

Energy Type Weight Factor

Electricity 1.0
Natural Gas 0.35
Fuel Oil 0.35
District Chilled Water 0.4

14.8 Additional Metrics for Measuring Sustainability

The Green Grid has also introduced several additional metrics for evaluating the "green" performance of data
centers:

1. CUE (Carbon Usage Effectiveness)

2. DCEPT (Data Center Energy Productivity)

3. ERE (Energy Reuse Effectiveness)

4. DCCETM (Data Center Compute Efficiency)

5. WUEM (Water Usage Effectiveness)

These metrics help provide a more comprehensive picture of the environmental impact of a data center,
especially when considering factors like free cooling and power generation sources.

65 | P a g e