Case Studies › Commercial & Institutional
Case StudyChiller DiagnosticsCommercial HVACIndependent · No Chemistry Sold

An energy consultant said replace the chiller. The panel data said clean it.

A commercial office tower ran two 500-ton water-cooled centrifugal chillers. An energy consultant flagged one machine’s specific power well off design and recommended replacing it, and a capital project was already moving. The water treater was asked to explain the chemistry. A structured diagnostic built only on data the chiller already reports found the real problem: biofilm on the condenser tubes, insulating at roughly eight times the fouling the machine was rated to tolerate. The cause was a pump energy retrofit, not the chemistry program. The machine was cleaned, the cause was corrected, and the replacement was cancelled.

Asset
Two 500-ton centrifugal chillers
On The Table
Replace the chiller
Hidden Finding
Biofilm at ~8× rated fouling
Condenser Approach
4.6 → 1.9 °F
Outcome
~$450K cost avoided
Challenge — What Was Happening

A cooling tower can be inspected. A chiller cannot. The condenser and evaporator bundles sit inside a pressure vessel with refrigerant on the other side of the tube wall, so their condition is invisible until someone pulls a waterbox.

An energy consultant flagged specific power on one machine well off design and recommended replacing it. The owner was being asked to choose between two guesses: the mechanical side said “old machine,” and the usual chemistry answer, “the index was high, so it must be scale,” would have been a guess too.

Nobody had trended approach temperature against the machine’s own baseline, so a problem that had been building for fourteen months had no visible history.

Intervention — What We Did

Applied the Chillers in the Dark diagnostic method: one stabilized data set from the control panel, compared to the machine’s own commissioning baseline at the same load.

Closed the heat balance first to prove the data could be trusted, then worked the decision trees to separate waterside from refrigerant-side and flow causes before anyone opened the machine.

Quantified the fouling against the AHRI-rated allowance, wrote down the ranked hypothesis, and confirmed it with the cheapest tests that could change the answer, at an outage the site was already having. No lab budget and no teardown to reach the diagnosis.

What The Panel Data Showed
  • Heat balance closed within 0.5% at a duplicated 91% load, so the data was trustworthy
  • Condenser approach 1.8 °F at commissioning, 4.6 °F now; evaporator approach barely moved
  • Condenser waterside ΔP unchanged from baseline, which points to a surface film, not a restriction
  • Specific power up from 0.639 to 0.672 kW/ton, about 5%, consistent with the added lift
What Was Ruled Out, And What Wasn’t
  • Flow at the pump confirmed by ultrasonic reading, but the pump had a VFD added 14 months earlier
  • At part load, tube velocity sat below the OEM minimum for most of each day
  • Saturation index history mild and cycles steady all season, so there was no scaling driver
  • Basin dip slide counts up an order of magnitude year over year. Branch taken: biofilm
Root Cause Finding

Physical cause: biofilm fouling with early under-deposit attack. With flow verified and bundle area from the OEM submittal, LMTD rose from 5.26 to 8.55 °F and the overall heat transfer coefficient fell from about 312 to 190 Btu/h·ft²·°F, a 39% loss. That is a fouling factor near 0.0021, roughly eight times the 0.00025 AHRI-rated condenser allowance. At the outage, the tube ends and lower pass carried soft gray-brown slime and no hard crust. Dilute acid produced no fizz, loss on ignition returned a high organic fraction, and shallow pits with black paste under a thin crust showed sulfate-reducing bacteria starting to attack.

Latent root cause: a mechanical decision, not a chemistry failure. Condenser flow had been cut to save pump energy with no review of tube velocity, biocide program, or filtration, and no approach trending that would have caught it in month three instead of month fourteen. The same gate also caught the second machine: its computed flow came back at 969 gpm against a nameplate 1,100. Using nameplate would have inflated its fouling factor by about a third and produced a cleaning recommendation on a machine that only needed its flow checked.

  • Mechanical cleaning with the correct brushes for enhanced tubes
  • Raised minimum pump speed to keep tube velocity inside the OEM window at every load
  • Continuous oxidant control on a site-calibrated setpoint
  • Added side-stream filtration
  • Monthly logging of both approaches and waterside ΔP at a duplicated load, against the documented baseline
  • Written commitment: condenser approach back within 0.5 °F of baseline within two weeks. It reached 2.1 °F at 14 days and 1.9 °F at 45
~$450K
In capital cost avoided. The chiller replacement was cancelled, and the machine was returned to within 0.1 °F of its commissioning approach.
Cost Avoided
~$450K
Condenser Approach
4.6 → 1.9 °F
Fouling Found
~8× rated
Specific Power
~5% recovered
Results Delivered
  • Cancelled — a chiller replacement that would not have fixed the problem
  • Restored — condenser approach to 1.9 °F against a 1.8 °F baseline
  • Stopped — early under-deposit attack by sulfate-reducing bacteria before it pitted through
  • Prevented — an unneeded cleaning on the second machine, caught by the heat-balance gate
Economics
  • Cost avoided: ~$450,000, the typical installed cost to replace a 500-ton water-cooled centrifugal (equipment ~$285K, removal, rigging and installation ~$110K, engineering, permitting and commissioning ~$55K)
  • Energy: about 5% specific power recovered, roughly $4,900/yr on one machine at the site’s blended rate
  • Diagnosis cost: panel data, a spreadsheet, and tests already on the service truck
  • Monthly approach trending now catches the next problem in months, not a year
The chiller is sealed, but it is not silent. It reports its own condition every minute in four temperatures, two pressures, two differential pressures, and a power reading. Read correctly against its own baseline, that data settles the argument before anyone spends money on the wrong fix. “The chiller lost efficiency and we don’t know why” is usually a documentation problem, not a technical one. Try the method on your own machine with the free guided chiller diagnostic.

Free PDF Download

Get the full case study

Complete data, methodology, and results — one page.

Download PDF

Being told to replace a chiller that might just need cleaning?

Independent chiller and cooling-water diagnostics for commercial, institutional, and industrial systems. We read the data your machine already reports and tell you which problem you actually have before you spend money on it. No chemistry sold.