MRI Cooling System Failure: The Chiller Never Alarmed

A regional outpatient imaging center operated a high-volume 1.5-tesla MRI system six days a week.

The dedicated process chiller received routine preventive maintenance. Service records showed no unresolved refrigeration faults, and the chiller controller consistently displayed an acceptable leaving-water temperature.

Then the MRI recorded two intermittent primary-water-flow warnings within several weeks.

Both warnings cleared before a service technician arrived. The chiller remained online, and the MRI returned to normal operation after a reset.

The initial conclusion was that the scanner had experienced a nuisance alarm or a momentary sensor issue. That appeared reasonable because nothing in the mechanical room indicated an active MRI chiller failure.

It was also the wrong conclusion.

The third warning occurred near the end of a long, thermally demanding imaging sequence.

The scan was completed, but the MRI cooling-water temperature remained elevated longer than expected. The equipment-room temperature also increased slightly. Once again, the process chiller had not alarmed.

The warnings were not random. They occurred during longer studies that placed a sustained thermal load on the MRI cooling system. Shorter scans did not create enough heat to expose the problem.

The MRI was supported by a dedicated air-cooled process chiller with approximately 45 kilowatts of nominal cooling capacity.

The facility could confirm that the chiller was running. It could not confirm that the MRI was receiving the cooling flow it required.

The investigation evaluated the complete cooling system rather than treating the chiller as an isolated piece of equipment. The evidence included MRI alarm and event records, chiller-controller history, preventive-maintenance reports, pump amperage readings, supply- and return-water temperatures, pressure measurements at the chiller and MRI, strainer differential pressure, bypass-valve position, and the imaging schedule associated with each warning.

A clear pattern emerged: the MRI warnings appeared when sustained imaging activity placed the greatest demand on the cooling system.

Two Reasonable Assumptions Proved Incorrect

During ordinary imaging activity, the reduced flow was just sufficient to carry the load. During longer sequences, MRI return-water temperature climbed from approximately 63°F toward 69°F. The scanner recognized the loss of cooling margin and generated a primary-water-flow warning.

Pressure measurements identified an abnormal pressure loss across the field-installed Y-strainer.

Differential pressure across the strainer had increased to approximately 9 psi. After the strainer was isolated and cleaned, the pressure drop fell to approximately 2 psi.

The debris had not stopped flow completely. Instead, it gradually reduced MRI cooling-water flow without producing an obvious mechanical failure.

At the same time, the minimum-flow bypass had been left farther open than necessary following an earlier service event.

How the Failure Path Developed

1. The loaded strainer restricted water flow toward the MRI.
2. The open bypass provided an easier return path near the chiller.
3. The pump continued operating and maintained apparently reasonable pressure near the mechanical equipment.
4. The chiller maintained its leaving-water temperature setpoint.
5. Actual water flow through the MRI continued to decline.
6. The MRI became the first component capable of recognizing the problem.

The MRI warning was not the underlying problem. It was the final visible indication of a deteriorating cooling path.

Had the restriction continued to increase, the facility faced several credible consequences:

• Interrupted or aborted patient scans
• Protective shutdown of MRI subsystems
• Extended recovery and functional testing
• Emergency manufacturer service
• Cancelled or rescheduled procedures
• Increased thermal stress on water-cooled components
• Reduced tolerance for another equipment failure

Related reading: When the Chiller Trips, the MRI Clock Starts Ticking.

Why This Matters to Imaging Centers and Health Systems

Medical imaging cooling sits at the intersection of clinical operations, facility infrastructure, equipment service, and patient scheduling. A problem that appears small in a mechanical room can create a disproportionate operational effect when it threatens a scanner booked throughout the day.

For imaging administrators, the important issue is not merely whether the chiller has been maintained. It is whether the complete cooling path has enough verified margin to support the actual scanning schedule, including long and thermally demanding studies.

For facility and reliability teams, chiller status alone is an incomplete performance indicator. Supply temperature, return temperature, flow, differential pressure, strainer condition, bypass position, pump availability, and recovery behavior must be interpreted as one system.

None of those checks independently proves that adequate cooling reaches the MRI under sustained load. The same principle applies wherever cooling failure can interrupt clinical work: MRI, computed tomography, positron emission tomography, linear accelerators, and other high-value medical systems. Equipment-specific requirements differ, but cooling must be verified at the protected load—not inferred from the condition of one upstream component.

Performance Verification

Cleaning the strainer and rebalancing the bypass restored water flow from approximately 14 gallons per minute to approximately 22 gallons per minute. The cooling system was then observed through several sustained, high-demand imaging sequences.

Supply-water temperature remained stable, MRI return-water temperature peaked near 63°F, cooling-water flow remained consistent, the standby pump was functionally tested, both refrigeration circuits staged normally, and no primary-water-flow warning occurred. The system also recovered predictably after the load decreased.

This provided more meaningful evidence than simply restarting the equipment and confirming that the chiller ran. It demonstrated that the complete cooling path could support the MRI under the conditions that had previously exposed the problem.

Condition-Based Maintenance and Early Warning

The original monitoring strategy could identify a complete chiller failure. It could not identify a gradual loss of cooling delivery.

The improved MRI cooling system reliability plan included MRI supply-water temperature, return-water temperature, actual cooling-water flow, differential pressure at the MRI and across the strainer, pump operating status, and standby-pump availability.

Alarm thresholds were structured to provide warning before the MRI reached its internal protective limits. Strainer maintenance was also changed from an arbitrary calendar schedule to a condition-based approach using differential pressure.

A useful reliability review should follow the cooling path from the source to the protected equipment. The purpose is not to impose one universal set of values. MRI manufacturer requirements, project design documents, and the engineer of record remain controlling. The purpose is to verify that the installed and operating system can meet those requirements under realistic conditions.

The result should be a defensible picture of the complete system—not a collection of separate service reports that leave the interfaces unexamined.