Should You Repair, Overhaul, or Replace a Commercial Chiller?

Diagnose the plant before condemning the machine

Low capacity, high pressure, trips, poor efficiency, and unstable temperatures can originate in the chiller, but also in flow, fouled heat exchangers, pumps, tower performance, valves, sensors, controls, load changes, or poor staging. A proposal should show the measurements and operating condition behind the diagnosis.

Ask which findings are confirmed, which are inferred, and which connected systems were evaluated. A replacement chiller installed into an unresolved plant problem can reproduce the complaint at a higher capital cost.

Understand what the chiller can and cannot solve

A chiller removes heat from chilled water and rejects that heat through an air-cooled condenser or a condenser-water system. The building still depends on pumps, control valves, coils, air handlers, towers or fluid coolers, sensors, and control sequences to move that cooling to the load. A complaint at the occupied space can therefore travel through several systems before it reaches the chiller.

The machine type also changes the decision. Air-cooled and water-cooled plants have different heat-rejection systems and maintenance demands. Reciprocating, scroll, screw, and centrifugal compressors respond differently to low load, lift, fouling, oil condition, surge, staging, and overhaul. The evaluation should identify the actual machine and use its service literature rather than a generic age rule.

Define the operating problem in measurable terms: insufficient tons, unstable leaving-water temperature, high kilowatts per ton, nuisance trips, refrigerant loss, tube leakage, vibration, noise, oil contamination, starter failure, or inability to obtain parts. Each symptom points to a different evidence path.

• Record entering and leaving chilled-water temperatures under a known load.
• Verify evaporator and condenser flow rather than assuming design flow.
• Review approach temperatures, pressure readings, alarms, starts, run hours, and load percentage.
• Inspect heat exchangers for fouling, scaling, corrosion, tube damage, and water-side pressure loss.
• Check tower or air-cooled condenser condition, ambient conditions, and heat-rejection control.
• Compare sensor readings with calibrated field measurements before trusting the trend history.

Use current load and operating data

Existing nameplate tonnage may reflect an old design, safety factor, removed process, changed occupancy, envelope improvement, or expansion that never occurred. Review peak and part-load behavior, operating hours, weather, process demand, chilled-water temperatures, and simultaneous loads before selecting capacity.

Part-load performance matters because many plants spend limited time near peak. The design should also state redundancy, future growth, low-load stability, and how multiple chillers, pumps, and towers will stage.

Build a condition assessment that survives scrutiny

A useful condition assessment separates an immediate fault from the remaining condition of the machine. A failed starter does not establish tube condition. A refrigerant leak does not establish motor condition. Conversely, restoring one failed component does not erase widespread corrosion, repeated tube repairs, unstable controls, or a declining compressor.

For a major decision, request a written record of the refrigerant circuit, compressor or driveline, motor, starter or drive, oil system, heat exchangers, controls, safeties, vibration history, water treatment, leaks, previous repairs, open recommendations, and manufacturer support. Where appropriate, the record may include oil analysis, vibration analysis, eddy-current tube testing, megohm or winding tests, refrigerant analysis, leak-test findings, and teardown measurements.

Not every test is necessary for every machine. The contractor or engineer should explain which tests are relevant, what question each test answers, and what uncertainty remains when a test is omitted.

Define repair, overhaul, retrofit, and replacement as real scopes

A repair addresses a confirmed defect. An overhaul restores defined components and tolerances. A retrofit changes a major subsystem such as controls, drives, starters, tubes, or refrigerant strategy. Replacement introduces a complete project around new equipment. Labels alone are not comparable.

Require each option to state inclusions, exclusions, remaining original components, expected service horizon, warranty, parts support, performance target, downtime, and verification.

Know what each investment leaves behind

A focused repair can be the rational choice when the defect is isolated, the rest of the machine is stable, service support is available, and the repair preserves an acceptable operating horizon. The proposal should identify the failed part, root cause, collateral damage, repair method, warranty, and test that will prove the repair.

An overhaul can make sense when a rebuildable machine has a known deterioration pattern and the restored scope is well defined. Ask which bearings, seals, impellers, rotors, valves, gaskets, tubes, controls, oil-system components, or motor elements will be renewed; which tolerances will be measured; and which original components remain outside the work.

A retrofit may improve a sound machine through controls, a variable-speed drive, starter replacement, heat-exchanger work, compressor modifications, or another manufacturer-supported change. A retrofit should have a documented compatibility and performance basis. It should not be used to hide a deteriorated pressure vessel, unsupported controls architecture, or an unresolved plant problem.

Replacement becomes stronger when several major systems are deteriorated, service and refrigerant support are weakening, capacity no longer matches the building, reliability risk is unacceptable, or the full lifecycle comparison favors a new plant. Even then, the replacement design must correct the connected-system causes that limited the old machine.

• What confirmed problem does this option correct?
• Which original components and risks remain?
• How much outage time and temporary cooling are required?
• What service life and performance are expected, and on what basis?
• What warranty applies to parts, labor, refrigerant, controls, and performance?
• What test establishes successful completion?

Challenge the lifecycle-cost assumptions

DOE procurement guidance emphasizes both full- and part-load efficiency for electric chillers. Energy modeling should use the actual utility tariff, demand structure, operating schedule, weather or process profile, plant sequence, and expected load distribution—not only a catalog efficiency point.

Test sensitivity to hours, rates, load growth, maintenance exposure, financing, incentives, and analysis period. Operational resilience and lost production may be more consequential than modeled energy savings and should be shown separately rather than hidden inside one payback number.

Measure efficiency at the plant boundary

A catalog chiller rating is useful for equipment comparison, but the utility meter sees the entire plant. Pumps, towers, condenser fans, heat-exchanger fouling, chilled-water resets, condenser-water temperature, sequencing, minimum-flow arrangements, bypasses, and simultaneous operation can materially change plant energy use.

Ask whether the analysis reports chiller-only performance or plant performance. State whether auxiliary power is included, how load bins were created, what entering conditions were assumed, and how the proposed sequence controls pumps and heat rejection. A highly efficient chiller can operate poorly when it is oversized, short-cycling, fighting another machine, or held at unnecessarily difficult conditions.

Baseline data should cover enough operating conditions to distinguish a persistent problem from a single hot day. After the work, trend the same variables so the owner can compare expected and actual performance.

• Chiller input power and produced cooling under defined conditions.
• Pump and tower or condenser-fan power included in the plant view.
• Operating hours by load range, not only annual full-load hours.
• Utility energy and demand charges using the applicable tariff.
• Sequencing, temperature resets, minimum flow, and low-load operation.
• Water, treatment, maintenance, and replacement-part effects where relevant.

Treat refrigerant and service support as specific risks

Refrigerant strategy matters, but a generic statement that a refrigerant is obsolete is not enough. Identify the machine's exact refrigerant, charge, leak history, conversion options, regulatory obligations, recovery requirements, service availability, manufacturer position, and expected supply exposure. A refrigerant conversion can affect capacity, efficiency, oil, seals, controls, relief devices, labeling, and certification; it requires an engineered, manufacturer-supported scope where offered.

Supportability also includes compressor parts, tubes, boards, sensors, drives, software, service tools, and technicians familiar with the machine. Ask the manufacturer or authorized service organization which components are available, which are repairable, what lead times apply, and whether a planned support sunset exists.

Use support risk as one factor in the decision, not as a substitute for condition evidence. A well-supported older machine may deserve continued investment; a nominally repairable machine may be operationally unacceptable if a critical part has an uncertain lead time and the plant has no redundancy.

Price the work around the chiller

Replacement may require structural review, rigging and access changes, refrigerant management, electrical and starter work, pumps, piping, isolation, insulation, controls integration, water treatment, permits, sound mitigation, temporary cooling, and restoration. These costs can exceed the apparent equipment difference between proposals.

Build a responsibility matrix for the owner, engineer, mechanical contractor, controls provider, electrical contractor, water-treatment provider, manufacturer, commissioning authority, and facility staff.

Normalize proposals before comparing price

Chiller bids often look comparable while carrying different boundaries. One may include rigging, demolition, electrical upgrades, controls integration, pump changes, insulation, startup, commissioning, and temporary cooling. Another may cover only the equipment and mechanical connection. Build one comparison sheet and force every bidder to answer the same scope.

List allowances, alternates, owner-furnished work, exclusions, taxes, freight, storage, escalation, overtime, permits, hazardous-material assumptions, structural work, roof or wall openings, crane constraints, noise limits, shutdown windows, and restoration. Identify long-lead items and the date at which the schedule becomes committed.

Require a written change-control process. Discovery is common in central-plant work, but undefined scope should not become a blank check.

• Exact equipment, accessories, ratings, and approved selections.
• Demolition, disposal, refrigerant recovery, and environmental documentation.
• Piping, valves, strainers, pumps, insulation, flushing, treatment, and balancing.
• Electrical service, disconnects, starters or drives, controls, networking, and metering.
• Rigging path, structural work, weather protection, permits, and restoration.
• Temporary cooling, outage coordination, contingency, startup, commissioning, training, and closeout.

Commission the complete plant sequence

Factory or manufacturer startup verifies equipment requirements. Commissioning should prove how the chiller operates with pumps, valves, towers or dry coolers, sensors, safeties, resets, staging, alarms, trend data, and building loads.

When full-load weather is unavailable, define deferred or seasonal tests. Closeout should include current sequences, setpoints, trend points, startup reports, flow or balancing records, training, warranties, and baseline operating data.

What the owner should have before approval

A capital decision is ready for approval when operations, finance, engineering, and procurement are working from the same facts. The record should show why the work is needed, what alternatives were considered, how uncertainty was handled, and how the selected result will be accepted.

If the record cannot explain the present defect, current load, complete installed scope, outage exposure, lifecycle assumptions, and verification plan, the project is not yet ready merely because a budget number exists.

• Written diagnosis and machine-and-plant condition assessment.
• Current load profile, operating requirements, redundancy target, and future assumptions.
• Comparable repair, overhaul, retrofit, and replacement definitions.
• Lifecycle model with visible rates, hours, escalation, maintenance, outage, and sensitivity assumptions.
• Complete scope and responsibility matrix, including controls and temporary operation.
• Risk register covering schedule, long-lead parts, outages, rigging, water, refrigerant, and commissioning.
• Acceptance plan with startup, functional testing, seasonal verification, training, and documentation.

The bottom line

The strongest chiller decision does not begin with new equipment. It begins with a defensible diagnosis and a shared operating basis for every option.

Make bidders show what each path solves, what original risk remains, how the full plant will operate, what the facility must do during the work, and how performance will be proven afterward.