Data Center
Modeling
July 2025
Unrestricted | © Siemens 2025 | Grid Software
This presentation is intended solely for informational
Disclaimer purposes only and does not constitute legal or regulatory
advice. Attendees should refer to specific standards,
requirements, or other obligations and should consult
qualified professionals for guidance specific to their
organization or situation. The views expressed herein are
those of the presenters and do not necessarily reflect the
views of any regulatory authority or other organization.
Page 2 Unrestricted | © Siemens 2025 | Grid Software
How to use your
1 Submit questions and comments via the
23 Webinar Questions panel at anytime during the
webinar. Questions will be answered
2 at the end of the session.
Panel: Download the handout from today’s webinar:
Including the presentation slides
Page 3 Unrestricted | © Siemens 2025 | Grid Software
Speaker Introductions
Amar Kyle Nick
Patel Thomas Giffin
Product Manager, VP Eng. & Compliance, Lead Engineer,
Siemens, PSS®E Elevate Energy Consulting Elevate Energy Consulting
Dr. Jay Dr. Feng Dr. Krishnat
Senthil Dong Patil
Senior Engineer, Senior Engineer, Senior Engineer,
Siemens, PSS®E Siemens, PSS®E Siemens, PSS®E
Page 4 Unrestricted | © Siemens 2025 | Grid Software
Poll Questions #1 and #2
01 02
What sparked your curiosity Which best describes your usage
about today’s topic? of PSS®E?
Page 5 Unrestricted | © Siemens 2025 | Grid Software
Agenda
1 Background
2 Typical Structure of Large Data Centers
3 Reliability Concerns with Large Data Centers
4 Load Modeling Challenges and Considerations
5 Summary and Recommendations
Page 6 Unrestricted | © Siemens 2025 | Grid Software
Background
Page 7 Unrestricted | © Siemens 2025 | Grid Software
Typical Structure
and Components
of Large Data
Center Loads
Data center loads are mostly
power electronic in nature,
with unique characteristics of the
protection and control systems
surrounding the computational
components of the load.
Source: Elevate Energy Consulting, GridLab
Page 8 Unrestricted | © Siemens 2025 | Grid Software
Typical Load Composition of Large Data Centers
Category Load % Description
Servers (30 – 40%)
Rack-mounted and blade servers for compute workloads driven by CPUs,
GPUs, and TPUs for AI, machine learning, and cloud
Power
40 – 60% Storage Systems (5 – 10%)
Electronics Hard disk drive (HDD) and solid-state drive (SSD)/flash memory
Networking Equipment (5 – 10%)
Switches, routers, firewalls, and other communication equipment
Computer Room A/C or Handlers (CRAC/CRAH Units) (10 – 20%)
Traditional air-cooled chillers, condensers, and compressors to maintain server room temperatures
Cooling Immersion and Direct-to-Chip Cooling (5 – 10%)
20 – 40% Emerging for high-density AI workloads, which reduces air cooling demand
Systems
Cooling Tower Pumps & Fans (5 – 10%)
Includes water-cooled systems and evaporative cooling units
Uninterruptible Power Supplies (UPS) (5 – 10%)
Power UPS systems convert AC electricity to AC and then to AC, providing fast transition to backup energy
10 – 15%
Infrastructure Power Distribution Units (PDU) (2 – 3%)
PDUs convert high-voltage power to IT-friendly levels
Page 9 Unrestricted | © Siemens 2025 | Grid Software
Reliability Concerns with Large Data Centers
Ride-Through
Characteristics Load Cycling Ramp Rates Power Quality
Impacts from UPS and Quasi-periodic Fast ups and downs Harmonics, voltage
BTM generation oscillations fluctuations/flicker
Subsynchronous
Transient Stability Oscillations Small Signal Stability Other
Voltage recovery, Exciting torsional modes Forced oscillations at Software process
frequency, damping of nearby units, control low frequencies loads, etc.?
instability, etc.
Page 10 Unrestricted | © Siemens 2025 | Grid Software
Load Modeling
Model Types and Modeling Approaches
Page 11 Unrestricted | © Siemens 2025 | Grid Software
Types of Loads
Static Dynamic
Loads Loads
Page 12 Unrestricted | © Siemens 2025 | Grid Software
Load Modeling Approaches
Component-Based
Performance-Based
Page 13 Unrestricted | © Siemens 2025 | Grid Software
Review of ®
PSS E
Load-Related Models
Load Modeling Challenges and Considerations
Page 14 Unrestricted | © Siemens 2025 | Grid Software
Static Load Models in PSS®E
In Power Flow In Dynamics
Constant MVA
Constant current
Constant admittance
Not very accurate
for dynamics
Page 15 Unrestricted | © Siemens 2025 | Grid Software
Dynamic Load Models in PSS®E
Model Description
ACMTxxU1 User-written performance-based model of single-phase air conditioner motor
CIM5xx Induction motor model
CIM6xx Induction motor model
CIMWxx Induction motor model (WECC)
CLODxx Complex load model
CMLDxxU2 Composite load model
CMLDxxDGU2 Composite load model with DER included
EXTLxx Extended-term load reset model
IEELxx IEEE load model
LDFRxx Load frequency model
Page 16 Unrestricted | © Siemens 2025 | Grid Software
Composite Load Models
(CMLDxxU2 or CMLDxxDGU2)
The main components of the composite
load model are the following:
• Substation transformer with load tap changer
(LTC) control
• Substation shunt
• Distribution feeder equivalent The electronic load model varies the load
• Feeder compensation as a function of voltage
• Motor loads (up to four) • Constant P, Q down to V = Vd1
• Up to three different types of three-phase • P, Q reduces to zero linearly between Vd1 and Vd2
induction motors with built-in protection
• Single-phase air conditioner compressor motor
with built-in protection
• Electronic loads
• Static load
• Load shedding
Page 17 Unrestricted | © Siemens 2025 | Grid Software
Composite Load Models
(CMLDxxU2 or CMLDxxDGU2)
Motor A
Three-phase motors driving constant torque Load Bus
loads, such as commercial/industrial air Substation 3-ph Motor A
conditioner compressors and refrigerators
High-Voltage Low
3-ph Motor B
System Bus Side Bus
Motor B LTC
Three-phase motors driving torque Distribution Feeder Equivalent
proportional to speed-squared loads 3-ph Motor C
with high inertia, such as fans
1-ph AC Motor D
Motor C Bf1 Bf2
Three-phase motors driving torque
Undervoltage Bss Feeder Electronic Load
proportional to speed-squared loads Load Shedding Compensation
with low inertia, such as pumps Underfrequency
Load Shedding Substation
Static Load
Motor D Shunt
DG
Single-phase compressors, often used
in residential air conditioning
Page 18 Unrestricted | © Siemens 2025 | Grid Software
PSS®E Load Component Models
Starting in PSS®E version 35.6 a load can be broken into different components.
(i.e., modular form of load models)
L3PMxxU
Load Main Load Distribution L1PMxxU Load Individual
MAINxx10U Component DISTxxU Component LSTCxxU Component
Models Models LELCxxU Models
DERAxxU
PL, QL
Distribution Model
Load
Component models
L System (far end)
.. Bus Low Side Bus Bus
. Substation
Transformer Rfdr Xfdr
L
Distribution Load
System
System Components
Model Xxf Distribution
1:T Feeder
Load bus S
.. Bss Bfar
System bus .
S
VL
S: Source type component | L: Load type component (the percentages) | PL: Sum of active power System Bus: Bus at which load is attached in power flow record | Load (far end) Bus: internal bus at
of load components | QL: Sum of reactive power of load components | VL: Load bus voltage which components are connected | System bus will be same as load bus if there is no distribution model
Page 19 Unrestricted | © Siemens 2025 | Grid Software
Ride-through and Protection
UPS ensures safety and Utility
resiliency of IT equipment Local
generation
Main
G
switchover
Entire load may transfer for V
terminal voltage deviations
UPS Possible P recovery
Challenging to parameterize
MW
UPS switchover Protection
IT Cooling Main
54% to 87% 46% to 13%
switch mode variable speed
power supply drives
Page 20 Unrestricted | © Siemens 2025 | Grid Software
Standard Library PSS®E Load Relay Models
Model Description
LDSHxx Rate of frequency load shedding mode
LDS3xx Underfrequency load shedding model with transfer trip
LDSHxx Underfrequency load shedding model
LDSTxx Time underfrequency load shedding model
LVS3xx Undervoltage load shedding model with transfer trip
LVSHxx Undervoltage load shedding model
UVUFxxU1 User written undervoltage and underfrequency load
shedding model
Page 21 Unrestricted | © Siemens 2025 | Grid Software
Undervoltage and Underfrequency Load Shedding Model
(UVUFxxU1)
I, ‘USRLOD’, LID, ‘UVUFxxU1’, 13, IT, 1,
20, 0, 7, 18, JBUS, CON(J) to CON(J+19)
Page 22 Unrestricted | © Siemens 2025 | Grid Software
NERC Incident Review
Widespread tripping of voltage-sensitive
protections at many data center loads
Source: NERC
Page 23 Unrestricted | © Siemens 2025 | Grid Software
Limitations of Electronic Portion of CMLD Model
Load disconnect and reconnect logics only Both disconnect and reconnect logics are
consider voltage sags, and not voltage swells merely based on voltage measurements
(i.e., no frequency protections)
Load disconnect logic of electronic portion No logic to simulate the impacts
does not have a time delay of automatic reclosing
Load reconnection logic of power electronic
portion does not have a time delay
Page 24 Unrestricted | © Siemens 2025 | Grid Software
Data Center Performance
AI Load Dynamics
Source: EdgeTunePower
Page 25 Unrestricted | © Siemens 2025 | Grid Software
Benefits of PSS®E User-Written Models
Flexibility and Accuracy
custom models to accurately represent specific PSS®E user-written models
equipment, protection or control systems (user-defined models, UDMs)
not available in the standard library.
enable more accurate and realistic
simulations, while maintaining
Confidentiality confidential information.
black-box models that can incorporate
proprietary, vendor-specific details.
Page 26 Unrestricted | © Siemens 2025 | Grid Software
Large Load Studies
Considerations for Data Center
Interconnection Studies
Page 27 Unrestricted | © Siemens 2025 | Grid Software
Steady-State Analysis
Routine modeling of demand
Consider load variation, especially with network upgrades
Contingency analysis
Able to represent load loss
Review remedial action schemes (RAS),
system operator limits (SOLs), and interconnection
reliability operating limits (IROLs)
Page 28 Unrestricted | © Siemens 2025 | Grid Software
Dynamic Analysis
Voltage stability impacts of interconnected
load and behavior
Frequency stability for load disconnects
Transient and angular stability of affected generation
Impacts on RAS, SOLs, and IROLs
Additional considerations
Source: EPRI (Presentation to NERC LMWG Jan 2025)
Page 29 Unrestricted | © Siemens 2025 | Grid Software
Summary and Recommendations
Data center loads are becoming a major Accurate modeling of data center loads
concern in power systems due to their is essential to ensure validity of simulation
peculiar characteristics and size. results, make sound engineering decisions,
and ensure reliability of the bulk power system.
PSS®E user-written models allow engineers Comprehensive studies of large data center
to build, customize, and refine models tailored loads are essential to identify potential
to new technologies that are not yet available issues and evaluate mitigation solutions.
in standard PSS®E libraries.
PSS®E can be used to assess some
of the challenges associated with large
data center loads presently.
Page 30 Unrestricted | © Siemens 2025 | Grid Software
Poll Questions #3 and #4
03 04
Would you be interested in a As a utility or grid planner, which
follow-up to learn more about one of the following types of
PSS®E? information is most critical for
you to receive from a data center
when processing an
interconnection request?
Page 31 Unrestricted | © Siemens 2025 | Grid Software
Upcoming Webinar – Don’t Miss It!
Webinar 3: Fundamentals and Validation Process for
DER and Dynamic Load Modeling
Date: August 12, 2025
Time: 11:00 AM ET
Location: Online
Join experts from Siemens PSS®E and Elevate Energy
Consulting as they explore:
• Core concepts of Distributed Energy Resource
(DER) and dynamic load modeling
• Insights into NERC guidelines and FERC Order 901
• The latest in model validation efforts and future initiatives
Page 32 Unrestricted | © Siemens 2025 | Grid Software
Questions
Kyle Thomas, PE
VP of Engineering and Compliance Services
Elevate Energy Consulting
[Link]@[Link]
[Link]@[Link]
Nick Giffin, PE
&
Lead Engineer, Studies and Modeling
Elevate Energy Consulting
[Link]@[Link]
[Link]@[Link]
Answers
Amar Patel
PSS®E Product Manager
amar-patel@[Link]
amar-patel@[Link]
Page 33 Unrestricted | © Siemens 2025 | Grid Software
