Home › Mechanical Engineering Services › Data Center Design Services
Home › Mechanical Engineering Services › Data Center Design Services
For operators and cooling OEMs: CDUs, liquid cooling, and CRAH equipment
Find hot spots before the racks go in. Free assessment and roadmap.
The challenge
Data center hot spots form when uneven airflow leaves some racks too hot while cooling runs hard elsewhere.
Review: Commissioning
→
What changes
Our data center design services model where the air goes, so cooling reaches every rack.
The challenge
Cooling plants sized with big margins waste power every hour the hall runs.
Review: Energy target
→
What changes
Heat transfer analysis sizes cooling to the real load, with margin only where it is needed.
The challenge
Denser racks push heat loads past what the original design ever planned for.
Review: Capacity plan
→
What changes
CFD tests future layouts and densities, so you know what the hall can take next.
The challenge
Retrofits in a running data center leave no room for trial and error or downtime.
→
What changes
Each change is tested on a model before it is made, so the live hall never becomes the test.
Cooling and power OEMs, modular builders, and startups: we act as your engineering team on CDUs and modules.
Your extra design team.
Get Your Free RoadmapOperators and developers: we define and design the cooling equipment your halls need, then work with your OEM.
Your halls, our design.
Get Your Free RoadmapOperator and cooling OEM in one design loop, so airflow, heat rejection, and rack density are solved together.
One loop, both sides.
Get Your Free RoadmapBring your own manufacturer, or work with our partner network. We take the designs we engineer as far as you need. We don't manufacture, so we never compete with the OEMs we partner with.
Facts about how we work, not project results. One team engineers your product from analysis to delivery, and calculations carry a PE stamp where required.
01
A cooling equipment design with the airflow and heat data behind it, ready for your OEM to build.
02
CFD and heat transfer results you review and approve at each gate, before anything is built or changed.
03
You own the IP by default. When an OEM or partner manufacturer is involved, the contract spells out who owns what.
04
If we missed it in our QC, we fix it. If your supplier can't make the part within three years, or supply dries up, we fix that too.
You send
What we need to start
Layouts: halls, racks, and containment
Loads: IT power, density, and growth plans
Cooling equipment and its data sheets
Temperature data or known hot spots
Targets: capacity, energy, or schedule
→
You get
What you hold at the end
CFD results for airflow and temperature you approved
Cooling and airflow design changes
Future layouts tested for growth
Drawings and a package ready to quote
Support through commissioning and changes
We price to the value we deliver, not the hours we log. If the value isn't there for you, we'll say so in the assessment.
Model airflow, containment, and cooling units in data halls, so every rack gets the cool air it needs.
Cool air to every rack.
Used For
Airflow
Containment
CRAH Units
Size chillers, heat exchangers, and cooling loops to the real heat load, with margin only where needed.
Cooling sized right.
Used For
Chillers
Exchangers
Loops
Build a model of your hall that tracks real data, so changes are tested on the model, not the live hall.
Test it on the model.
Used For
Live Data
Layouts
Changes
Take a new CDU or cooling module from concept to a tested prototype, with the build run through partners.
Concept to tested unit.
Used For
CDUs
Modules
Testing
Public benchmark
Airflow retrofit in a national-lab data center: +21% cooling capacity, 8% less fan energy (DOE, 2010)
What it means for you
The gains came from finding where the air went wrong and fixing its path, not from adding cooling. We model your cooling equipment and airflow the same way before anything is built or changed.
The public record shows what the method can do. What it does for your equipment depends on your constraints, and the free assessment tells you.
Operator, OEM, or both: we set the model, the builder, and the hand-off.
Scope, layouts to test, dates, and approval gates.
CFD and heat transfer models, checked in our QC.
Design changes, drawings, and a package to quote.
Commissioning support, new layouts, and changes.
You approve the design at every gate, so nothing that affects function surprises you. If we missed something in our QC, we fix it. If a supplier can’t make a part within three years, we fix that too.
New hall or rising density? Send the layout. Free roadmap.
Turbines, heat exchangers, and plant equipment studied for heat, flow, and fatigue before an outage.
Fewer forced outages.
Applications
Turbines
Exchangers
Piping
Scrubbers, ducts, and fans designed to your permit limits, without oversizing the system or the fans.
Right-size your controls.
Applications
Scrubbers
Ducts
Fans
Aerospace to data centers: see how one engineering team solves hard constraints across ten industries.
Find the one like yours.
Industries
Aerospace
Mining
Data Centers
Fixed scope for one project. A lower-risk place to start when you want to see our work on your data.
Start with one project.
Fits When
One Project
Fixed Scope
From $35K
Scheduled monthly engineering for recurring needs: drawing updates, ASME calcs, or FEA for each sale.
Work on a set schedule.
Fits When
Monthly
Recurring
From $15K/mo
A dedicated PPS team working your program daily, built for long, multi-year development and testing.
Your program, every day.
Fits When
Daily Work
Multi-Year
From $800K/yr
Gold · One project: a hall with hot spots, or a cooling upgrade that must work.
Platinum · Ongoing work: modeling support as racks, loads, and layouts change.
Diamond · A long program: a dedicated team across a campus or new build.
FEA, CFD, CAD, DFMA, and testing under one roof. Engineering is a team sport, and you get the whole team.
One team, every skill.
Under One Roof
FEA
CFD
CAD
DFMA
Testing
Veteran-owned SDVOSB status gives defense and federal buyers a set-aside path to proven engineering.
Verify our SDVOSB status.
Credentials
SDVOSB
VOSB
CAGE 8ZK73
NAICS 541330
Missed something in our QC? We fix it. Supplier can't make the part within three years? We fix that too.
Support for the long run.
We Fix
Our QC Misses
Supplier Gaps
Supply Chain
We price to the value we deliver, not hours logged. You buy a design that works, not more iterations.
Pay for the result only.
Collections
Gold
Platinum
Diamond
What do your data center design services cover?
Our data center design services cover the equipment that keeps a hall cool: cooling units, CDUs, cooling loops, and the airflow around the racks. We engineer that equipment, not the building. Most operators feel the gaps as hot spots, high cooling bills, or a hall that can't take denser racks. Rack density keeps rising, and cooling sized by rule of thumb either runs hot or wastes power every hour. Lawrence Berkeley National Laboratory runs a Center of Expertise for Data Center Energy to help designers and operators make these choices. A cooling gap found after go-live is slow, risky, and costly to fix.
Data center CFD simulation shows how air moves through the hall, where hot spots form, and how containment and cooling units perform. Heat transfer analysis sizes chillers, exchangers, and loops to the real load. We design CDUs and liquid cooling loops with the OEMs that build them. Future layouts are tested before they go in, not after. Each model is checked against measured data where it exists.
Can you model a hall that is already running?
Yes. Most changes happen in halls that are already running, where downtime is not an option. Our data center design services start with the hall as it is today: layouts, rack loads, cooling units, and any temperature data you have. In a live hall, trial and error is too risky. Under-floor air and open paths send cool air where nobody planned, and hot spots follow. A DOE airflow retrofit case study notes that CFD airflow models help operators find these gaps. Guessing means new racks, new cooling, and new bills, while the real problem stays in the room and keeps costing you.
We set up a CFD model of the hall and check it against your measured temperatures, so it shows what really happens in the room. Then each change is tested on the model first: new racks, higher density, containment, or a new cooling unit. You see the effect on temperature and energy before you approve it. The model stays yours, so your team can test the next change too.
Can PPS work with our OEM or manufacturer?
Yes. Data center design services from PPS fit the partnerships that are normal in data center cooling projects. We work in one of three models. As your OEM partner, we act as your engineering team. Operator-led, we design the equipment and work with the builder you choose.
When design, build, and test sit with different companies, problems pass back and forth. The builder works from an old drawing, and each party defends its own scope. The Department of Defense's IPPD Handbook organizes programs around integrated product teams for this reason. In a joint program, the operator and the OEM work with us in one loop.
We engineer the design and coordinate the builder in one loop. Bring your own manufacturer, or work with our network of fabricators, machine shops, and test labs. You decide how far we take the design, up to the start of production. We never run production, so we never compete with your OEM. NDAs cover every party, and you own the IP by default, set in your contract.
What do you need from us to start?
To start a project with our data center design services, send what you have, in any format. That usually means hall layouts with racks and containment, IT loads and growth plans, and cooling equipment data sheets. Add temperature data or known hot spots, photos, and your targets for capacity, energy, or schedule. Old commissioning reports help too. Missing data is the most common reason engineering projects stall, because a model built on guessed inputs gives a guessed answer. That is why we don't start the work until every data set is in.
Once it is, the first step is a free assessment of your project scope. We confirm what the work can and can't answer, and tell you what is still missing. Then we send a roadmap proposal with scope, targets, and approval gates. Every project is scoped on its own, so the roadmap sets the schedule, not a stock promise. NDAs are available before you share any files. For gaps in material data, NIST material measurement data is a good first stop.
Who owns the design and the models?
You own the design, the drawings, and the models from your data center design services project by default, as set in the contract. With an OEM involved, the contract says who owns what. Some firms keep the models, or license them back to you, so every later change sends you back to them. That locks your product to one vendor and makes design reviews harder, because your team can't open the evidence behind the decisions. You receive the native CAD, the drawings, the simulation model files, and the documented results. Your engineers can then check the work, rerun a load case, or hand the package to a supplier or customer.
We sign mutual NDAs before any of your data is shared, and your files stay in our controlled project folders while we work. If your program has export-control or customer data rules, tell us at the start and we will set up the data handling to match them. For defense programs, the State Department ITAR guidance sets the rules we follow.
What happens if a prototype fails in test?
If a prototype fails in test, we find out why and fix the design. Hard engineering carries unknowns on any project: real loads, material lots, and assembly effects never match a model perfectly. With data center design services from PPS, you approve the data at each gate. A test result is then checked against predictions you have already seen, not a black box. A failure in test is data. It shows which assumption was wrong: a load, a material property, or a boundary condition. That data improves the model and the design.
When testing is in scope, we plan the test to confirm the margins that matter most. Then we correlate the model to the results, following ASME verification and validation practice. If the miss was ours, because something slipped through our QC review, we fix it. If the test uncovers a condition nobody knew about, we scope the change with you and update the roadmap. Either way, you get a clear answer and a path to a design that holds.
How are data center design services priced?
We price data center design services to the value they deliver, not the hours we log. Hourly billing rewards extra iterations, so the cheapest quote often becomes the most expensive project once tooling, rework, and field failures show up. Our incentive is to get the design right the first time. We engage by fit, not as a ladder. Gold (from $35K): you have one project and need help with it. Platinum (from $15K/mo): you have recurring work at a steady rhythm. Diamond (from $800K/yr): you have a long development program, often multi-year.
Gold is never required, and some clients start at Platinum or Diamond. These are starting points, not quotes; your roadmap sets the price. We price to the value we deliver, not the hours we log. If the value isn't there for you, we'll say so in the assessment. Our engineering consultant pricing page explains how each collection works in more detail.
Hot spots or rising rack density? Send the layout. Free roadmap.
Send your layouts, loads, and cooling data. We review your project scope in a free assessment and reply within one business day with next steps.
On-site assessments that require travel are quoted separately.