Many data center plants leave “free cooling” value on the table in shoulder seasons. The reason is rarely the headline idea (use ambient conditions to reject heat). It’s usually the details: approach temperatures, reset logic, and part-load behavior that keep the chiller running when it could be trimming—or bypassed.
This guide walks through a practical workflow to (1) pick a waterside economizer architecture, (2) budget approach temperatures so you understand economizer hours, (3) implement chilled-water and heat-rejection resets that actually reduce lift, and (4) verify savings with trending.
Key Takeaway: A waterside economizer improves PUE only when the whole plant (HX + tower/dry cooler + pumps + controls) is designed and reset for low lift at part load—not just sized for a few design points.
Step 0 — Define constraints and inputs (before you pick equipment)
Define reliability and risk constraints
Availability: N+1/2N expectations, maintenance modes, and failure scenarios (power loss, valve fail positions, winter operation).
Water risk: corporate WUE targets, water pricing/availability, and whether you have a mature program for evaporative systems (treatment, inspection, documentation).
Compliance and health programs: if you’re considering a cooling tower, include Legionella risk management requirements as a first-class design constraint (not an O&M afterthought).
Define your operating envelope
If you’re using a cooling tower, your “economizer window” is governed primarily by outdoor wet-bulb temperature.
If you’re using a dry cooler, the limiting variable is outdoor dry-bulb temperature.
The difference matters because it changes both the number of economizer hours and the fan/pump energy you’ll carry in partial-free-cooling mode.
Establish a verification plan (so savings are defensible)
Decide up front what you will trend at 1–5 minute intervals:
Chiller kW and, if available, chiller kW/ton or kW/kW of cooling
Chilled-water supply/return temperatures and flows
Heat-rejection loop supply/return temperatures and flows
Pump kW (CHW and CW/heat-rejection)
Tower or dry cooler fan kW
If you want a clean method for converting plant energy changes into an annualized PUE impact (with assumptions explicit), this internal reference shows a neutral modeling approach: Precision cooling shifts 5-year PUE cost.
Step 1 — Pick your waterside economizer heat-exchanger approach
A waterside economizer is not a single component. It’s a set of loops and controls that allow ambient heat rejection to cool the chilled-water loop directly or indirectly.
According to ENERGY STAR’s guidance on water-side economizers, these systems can reduce mechanical cooling hours by using tower or other water-side heat rejection when conditions are favorable.
Choose the economizer configuration you’re actually building
Typical configurations include:
Plate-and-frame HX with cooling tower water: the classic waterside economizer (tower loop ↔ HX ↔ chilled water loop).
Dry cooler loop + HX: a closed heat-rejection loop rejects heat to ambient air; HX isolates the facility chilled-water loop.
Integrated vs non-integrated economizer control:
Non-integrated: economizer replaces the chiller only when it can carry the full load.
Integrated: economizer pre-cools and the chiller trims the remainder, capturing more “partial free cooling” hours.
Pro Tip: When you model savings, treat economizer hours as a spectrum (integrated trim → full economizer), not a binary on/off season.
Budget approach temperatures (this drives economizer hours)
The limiting factor is usually the total approach budget between ambient conditions and the chilled-water supply temperature your data hall needs.
In practice, “approach” shows up in multiple places:
Tower approach (wet-bulb → leaving tower water)
HX approach (tower/dry cooler loop → chilled water loop)
Terminal/coil approach (chilled water → air or secondary loop)
A useful way to think about it: if the sum of your approaches is too large, the economizer doesn’t fail at design day—it fails by shrinking the number of shoulder-season hours where it can carry meaningful load.
Carrier’s modeling note, How to model a waterside economizer application, is a practical reference on why combined approaches matter when determining whether you can meet return/supply temperatures with free cooling.
Design for fouling, filtration, and maintenance
Waterside economizers often depend on low approach temperatures; fouling and rising pressure drop can quietly erase economizer hours.
Plan for:
Strainers/filtration suitable to the chosen loop (open vs closed)
Isolation valves and a bypass around the HX (maintenance, degraded performance, commissioning)
Trending points that let you detect approach degradation early (temperatures, flows, and HX ΔP)
Step 2 — Choose dry cooler vs cooling tower (tradeoffs that change PUE and risk)
Cooling tower: when it’s the better answer
A tower-based economizer is generally compelling when:
Wet-bulb conditions are favorable for a meaningful part of the year
You need colder leaving water temperatures to expand economizer hours
Water cost and water-risk programs are manageable
This is the thermodynamic advantage of evaporative heat rejection: it can cool closer to wet-bulb than a dry cooler can cool to dry-bulb.
Cooling tower: what you must be prepared to operate
Be explicit about the obligations:
Water use and WUE impact: blowdown, drift, and treatment are ongoing costs.
Program maturity: water chemistry, inspection cadence, documentation, and response procedures.
Health risk management: cooling towers can be associated with Legionella risk if not properly managed, so policies and controls can be as important as equipment.
Dry cooler: when it’s the better answer
A dry cooler-based economizer is often favored when:
Water availability, water pricing, or corporate WUE targets are dominant constraints
The site prefers a closed-loop heat rejection strategy with simpler water-side management
Local constraints make evaporative systems undesirable
Dry cooler: the energy and performance limitations
The tradeoff is typically:
Higher approach vs ambient (dry-bulb limited)
Higher fan energy in warmer conditions
Freeze protection requirements in cold climates (often glycol), which can add pumping penalty and reduce heat transfer
Step 3 — Implement reset strategies that reduce lift without breaking reliability
The goal of reset is straightforward: run the plant at the lowest lift compatible with your load and risk constraints.
Alfa Laval’s overview, Energy-saving waterside economizers for data centers, summarizes the role of economizers in reducing reliance on mechanical cooling. The next step is making that real through control logic.
Chilled-water supply reset (CHWS)
Principle: run the warmest CHWS temperature you can, because warmer supply typically reduces compressor work and increases economizer hours.
Practical implementation patterns:
Reset CHWS upward when terminal valves/coil authority indicate headroom.
Constrain resets with clear “do not exceed” limits based on data hall requirements.
If you’re writing specifications, use the phrase your stakeholders will search for in RFPs: a chilled water reset strategy tied to measurable signals (valve position, supply-air limit, or IT inlet constraint), with documented bounds.
Heat-rejection reset (tower leaving / condenser water / dry cooler leaving)
Principle: reset heat-rejection leaving temperature to the minimum necessary, not the minimum achievable.
For towers: reset against wet-bulb plus approach, while respecting chiller head-pressure constraints.
For dry coolers: reset against dry-bulb plus approach, while ensuring freeze protection and stable control.
Differential pressure (DP) reset and VFD sequencing
If you hold fixed DP everywhere, you often pay pump energy you don’t need—especially at part load.
Typical best practice:
DP reset to the “critical” branch or most remote/most-open valve group.
Sequence VFDs so you don’t overpump while simultaneously overcooling.
Changeover, deadbands, and protection logic
Mode hunting is a common real-world failure mode (especially in shoulder seasons).
Include:
Deadbands and minimum run/off timers for economizer enable/disable
Freeze protection limits (tower basins, coils, outdoor piping)
Clear alarming and fallback modes
Step 4 — Use part-load curves to predict (and prevent) disappointment
Most plants don’t operate at design load for most of the year. That’s why part-load curves matter.
What to look at (beyond full-load efficiency)
Evaluate the combined curve:
Chiller kW vs lift (and how economizer trim shifts the operating point)
Pump kW vs flow (cube-law effects under VFD control)
Tower/dry cooler fan kW vs leaving temperature setpoint
A frequent “gotcha” is that as lift drops, compressor energy falls—but fans and pumps can become the dominant term if setpoints and DP aren’t reset.
Common part-load failure modes
Free cooling isn’t free: parasitic fan/pump kW can dominate in marginal economizer conditions.
Low ΔT syndrome: excessive flow and low return temperature reduce economizer effectiveness and waste pumping energy.
Approach drift: fouling/air-side degradation increases approach over time, silently reducing economizer hours.
What to trend to verify improvement
Trend and report results by operating mode:
Economizer-only hours vs integrated trim hours vs mechanical-only hours
Plant kW by component (chiller, pumps, fans)
Approach temperatures and HX ΔP over time
Step 5 — Integration opportunities: heat reuse and higher water temperatures
Heat reuse is site-specific, but it’s worth evaluating early because it can change your “best” setpoints.
Heat reuse becomes more realistic at higher water temperatures
If your system can operate with higher leaving-water temperatures (within your thermal and condensation constraints), it can expand the set of viable reuse sinks.
How economizer choices affect reuse
A tower may deliver lower temperatures more often (good for minimizing lift), but that doesn’t always align with reuse temperature needs.
A dry cooler strategy may align better with “moderate temperature” heat rejection and tighter water governance.
Treat “lowest PUE” and “highest reuse value” as two objectives that may not be perfectly aligned. Controls need a declared priority.
Common mistakes (and how to avoid them)
Assuming aggressive approach temperatures year-round without modeling fouling, part-load, and commissioning realities.
No verification plan: if you can’t trend and normalize, you can’t defend savings to procurement.
Treating tower operation as purely mechanical rather than a program (chemistry, inspection, documentation).
Skipping mode-transition logic (deadbands, minimum run times), leading to hunting and operator overrides.
Quick decision checklist (one page)
Decision | What to verify | Why it matters |
|---|---|---|
Tower vs dry cooler | Wet-bulb vs dry-bulb economizer hours; water constraints | Determines achievable leaving water temps and water risk profile |
HX approach budget | Tower + HX + terminal approaches at part load | Drives integrated/free cooling hours |
Reset strategy | CHWS reset, heat-rejection reset, DP reset logic | Lift reduction only happens if setpoints move |
Part-load curve | Plant kW distribution at 20–80% load | Parasitics can dominate if not reset |
Verification | Trending points + reporting boundaries | Savings must be provable |
Next steps
If you want a practical starting point, request a waterside-economizer controls point list + commissioning checklist (trend points, enable criteria, deadbands, freeze protection, and acceptance tests). You can also request an economizer-hours estimate using your site weather file and target CHWS temperature.
Appendix A — Example points list and acceptance checks (illustrative)
Because no specific manufacturer HX specs/controls were provided for this article, the examples below are typical/illustrative and should be validated against official submittals and the site’s sequence of operations.
Example instrumentation and control points to include in submittals
A practical pattern is to include enough sensing to validate approach and part-load performance—not just enable/disable economizer mode.
Typical point list:
Temperatures: CHWS/CHWR; heat-rejection supply/return; HX approach
Flow: CHW loop flow; economizer/HX loop flow; condenser/heat-rejection loop flow
Pressure: system DP at critical branch; HX ΔP
Actuators: economizer bypass valve; isolation valves; VFD speeds (pumps/fans)
Environmental constraint (as applicable): dew point signal for condensation avoidance logic
Example high-level sequence of operations (conceptual)
Enable integrated economizer trim when ambient conditions provide approach margin.
Transition to economizer-only mode with chiller bypass when the economizer can meet CHWS.
Apply deadbands and minimum run/off timers to avoid hunting.
Constrain resets and enables with freeze protection and safe minimum temperatures.
Example “verify it worked” acceptance checks
Trend points are calibrated and recorded at 1–5 minute intervals.
CHWS and heat-rejection resets execute smoothly without oscillation.
Plant kW (chiller + pumps + fans) improves vs baseline by operating mode.
Appendix B — Further reading on Coolnetpower.com
For additional context on economizer modeling and verification boundaries, see Precision cooling shifts 5-year PUE cost (already linked above). If you’re comparing higher-water-temperature cooling approaches, see Do rear-door heat exchangers reduce server-room PUE? and Precision cooling vs comfort cooling for server rooms.
