A chiller rarely fails without warning. Condenser tubes foul, refrigerant charge drifts, oil degrades, and bearings wear, and each change shows up in the data weeks before a trip or a shutdown. A predictive maintenance programme catches those signals, but most rollouts stall because sensors are bought before goals, thresholds, and response paths are defined. This 25-point checklist sequences the setup into five phases so monitoring is live in about 30 days, with alerts that feed straight into work orders in Oxmaint maintenance management software.
Chiller Condition Monitoring Rollout
Chiller PdM Checklist: 25-Point Monitoring Program Setup
A phased setup plan for commercial chiller plants covering sensors, thresholds, water treatment, oil sampling, refrigerant monitoring, and CMMS integration.
Days 1 to 5FoundationPoints 1 to 5
Days 6 to 15Sensors and dataPoints 6 to 12
Days 12 to 20Thresholds and alarmsPoints 13 to 17
Days 15 to 25Water, oil, refrigerantPoints 18 to 21
Days 22 to 30CMMS and reviewPoints 22 to 25
What a Chiller Plant Tells You Before It Fails
Every common chiller fault leaves a measurable trail. The matrix below links symptoms to likely causes, and it explains why the checklist starts with the right measurements rather than the most expensive hardware.
| Early signal | Likely cause | Measure with |
|---|---|---|
| Condenser approach rising | Tube fouling, scale, or non-condensables | Condenser water and saturation temperatures |
| Evaporator approach rising | Evaporator fouling, low charge, or oil in the evaporator | Chilled water and suction saturation temperatures |
| kW per ton worsening at similar load | Efficiency loss from fouling, charge, or controls drift | Power, flow, and delta-T |
| Vibration trend climbing | Bearing wear, misalignment, or imbalance | Accelerometers on compressor and motor |
| Oil contamination or acid number rising | Moisture ingress, overheating, or wear | Laboratory oil sample |
| Purge run-time or make-up refrigerant increasing | Air ingress or refrigerant leak | Purge counters and refrigerant log |
Phase 1: Foundation (Days 1 to 5)
Define what you are protecting and why before choosing instruments. Skipped foundations are the main reason pilots produce data nobody uses.
1
Build the chiller asset register. Record manufacturer, model, compressor type, capacity, refrigerant, charge size, year, and control platform for every machine.Done when: every chiller has a unique asset record.
2
Rank chillers by criticality. Score impact on tenants, process loads, and redundancy so the first sensors go on machines where failure costs most.Done when: a ranked list is approved by operations.
3
Capture baseline performance. Gather commissioning reports, AHRI rating data, design conditions, and recent log sheets to define normal for each chiller.Done when: a baseline file exists per machine.
4
Review failure modes by chiller type. Centrifugal, screw, and scroll machines fail differently, so list the faults you need to catch for each type.Done when: failure modes map to a measurement.
5
Set programme goals and KPIs. Choose targets such as fewer unplanned trips, tracked efficiency, and fewer emergency call-outs.Done when: three to five KPIs have owners.
Phase 2: Sensors and Data Capture (Days 6 to 15)
Many plants already hold much of this data in the building management system. Audit what exists, then add only what closes gaps.
6
Water temperatures. Confirm entering and leaving temperatures on both the evaporator and condenser, with sensors verified against a reference instrument.Done when: four temperatures log reliably.
7
Refrigerant pressures and saturation temperatures. Capture suction and discharge readings so approach temperatures and lift can be calculated.Done when: pressures trend without gaps.
8
Electrical data. Log motor current, power, and drive status where fitted, because power is the basis of every efficiency calculation.Done when: kW is recorded at one-minute or finer intervals.
9
Vibration monitoring. Fit accelerometers or route handheld readings on compressor and motor bearings, following OEM guidance for mounting points.Done when: a baseline vibration spectrum is saved.
10
Oil pressure, temperature, and level. Record lubrication system values, since oil faults develop quickly and damage compressors fast.Done when: oil values appear on the monitoring screen.
11
Flow measurement. Verify chilled and condenser water flow, because delta-T without flow cannot give true cooling load.Done when: flow and load calculate per chiller.
12
Data pipeline and time sync. Pull points through the BMS using open protocols such as BACnet or Modbus, with consistent timestamps and sampling rates.Done when: data arrives complete for seven days.
Turn Chiller Data Into Scheduled Work
Oxmaint links asset records, inspection readings, and work orders so a rising trend becomes an assigned task with a due date.
Phase 3: Thresholds and Alarm Logic (Days 12 to 20)
Fixed limits copied from a brochure create false alarms. Build thresholds from each machine's own baseline at comparable load.
13
Approach temperature limits. Set evaporator and condenser approach triggers as deviation from baseline, not a single universal number.Done when: limits exist per machine.
14
Efficiency trending. Track kW per ton normalised for load and condenser water temperature so seasonal change is not mistaken for degradation.Done when: a normalised trend chart is live.
15
Vibration alert levels. Use OEM limits and ISO 20816 guidance, then refine them after observing normal operation.Done when: alert and alarm levels are documented.
16
Three-tier alarm structure. Define advisory, warning, and critical levels, each with an owner and an expected response time.Done when: every alarm has a named responder.
17
Nuisance alarm control. Add start-up delays, persistence timers, and sensor-fault checks so alarms mean something.Done when: test alarms cause no false escalations.
Alarm Response Ladder
Use this ladder as a starting design and adjust response times to your plant's redundancy.
AdvisoryTrend moving away from baseline. Create a planned work order and review at the next scheduled visit.
WarningThreshold crossed or rate of change high. Technician inspects within a defined short window and logs findings.
CriticalProtection limit near or safety risk. Notify engineer on duty, prepare to switch to standby, and open an urgent work order.
Phase 4: Water Treatment, Oil, and Refrigerant (Days 15 to 25)
These measurements catch slow-developing faults that electrical and temperature data alone can miss.
18
Condenser water treatment programme. Log conductivity, pH, inhibitor, and biocide results, and align the plan with ASHRAE Standard 188 and Guideline 12 on Legionella risk.Done when: water results link to the chiller record.
19
Tube condition and cleaning schedule. Use approach trends to time tube cleaning, and schedule eddy current inspection according to OEM advice and risk.Done when: cleaning and inspection tasks are scheduled.
20
Oil sampling programme. Set sample frequency, lab, and test suite covering viscosity, acid number, moisture, and wear metals.Done when: first samples are sent and results filed.
21
Refrigerant monitoring and leak records. Maintain machinery-room detection under ASHRAE Standard 15, log refrigerant additions, and review purge run-time on low-pressure machines. Check current EPA Section 608 and AIM Act requirements for your refrigerant and charge size.Done when: every top-up is recorded against the asset.
Phase 5: CMMS Integration and Review (Days 22 to 30)
Monitoring only pays when an alert reliably becomes a job that somebody completes and closes.
22
Map alerts to work orders. Define which alarm creates which work order type, priority, and assignee inside the CMMS.Done when: a test alarm produces a correct work order.
23
Blend condition-based and time-based tasks. Keep legally required and OEM-mandated preventive maintenance, and let condition data adjust the remaining intervals.Done when: each chiller has a combined task plan.
24
Stock critical spares. Identify long-lead items such as sensors, seals, contactors, and filters, and set reorder points in inventory.Done when: critical spares have minimum levels.
25
Start the review cycle. Hold a review at 30 and 90 days to tune thresholds, retire noisy alarms, and report KPIs to management.Done when: review dates are in the calendar.
Monitoring Priorities by Chiller Type
The checklist applies to all machines, but the emphasis changes with compressor design. Adjust points 9, 10, and 21 accordingly.
| Chiller type | Watch closely | Typical extra check |
|---|---|---|
| Centrifugal | Bearing vibration, oil system, approach temperatures, and surge behaviour | Purge run-time on low-pressure machines |
| Screw | Oil separation and filtration, slide valve position, and compressor vibration | Oil analysis at shorter intervals |
| Scroll and reciprocating | Motor current, discharge temperature, and short-cycling | Start counts and run-hours per circuit |
| Air-cooled | Condenser coil condition, fan operation, and ambient temperature | Coil cleaning trigger from discharge pressure |
Permanent Sensors or Handheld Rounds?
You do not need full online monitoring on day one. Match the method to the criticality ranking from point 2.
Manual roundsTechnicians log readings on mobile inspection forms. Low cost, good for non-critical machines, but coverage is periodic.
BMS trendingExisting points are logged continuously. Best value where the control system already measures most variables.
Dedicated sensorsAdded vibration, oil, or refrigerant instruments on critical machines where failure is costly and early warning matters most.
Who Does What During the 30 Days
Clear ownership keeps the rollout moving. Assign named people rather than departments.
- Programme lead: owns the plan, the timeline, and the weekly progress check.
- Controls technician: verifies BMS points, calibrates sensors, and builds trends.
- Chiller technician: confirms baselines, takes oil samples, and responds to alarms.
- Water treatment provider: supplies test results and recommendations in a usable format.
- Maintenance planner: sets up work order triggers, schedules, and spare parts levels.
- Facility manager: approves alarm response times and reviews KPIs.
Timing the Rollout Around the Cooling Season
Start before the main cooling season if you can, because baselines are most useful under real load.
- Capture baselines at several load levels, not only at full load.
- Complete tube cleaning and water treatment checks before peak demand begins.
- Test alarms and standby changeover while the plant has spare capacity.
- Review the first heat-wave data to retune thresholds quickly.
Data Quality Checks Before You Trust a Trend
A bad sensor produces a convincing but false trend. Run these checks during phase 2 and repeat them at each review.
- Compare chilled water temperature sensors against a calibrated reference at stable load.
- Confirm the heat balance across the evaporator is plausible given flow and delta-T.
- Look for flat-lined values, sudden steps, or missing intervals that suggest a failed point.
- Check that engineering units and scaling in the BMS match the instrument range.
- Confirm that trend data is stored long enough to compare year over year.
From Finding to Maintenance Action
Each finding should lead to a defined job type. Agree these mappings once so technicians do not improvise under pressure.
| Finding | Typical work order | Follow-up record |
|---|---|---|
| Condenser approach trending up | Inspect and clean condenser tubes, and verify water treatment | Approach after cleaning compared with baseline |
| Vibration above alert level | Confirm reading, inspect alignment and bearings, plan repair | Spectrum before and after the work |
| Oil moisture or acid number high | Investigate leak path, change oil and filters per OEM | Repeat oil sample result |
| Refrigerant top-ups increasing | Leak search and repair by a certified technician | Refrigerant log and leak test result |
| Alarm repeats with no fault found | Review threshold and sensor health | Updated alarm settings |
Common Rollout Mistakes and Fixes
Most failed programmes share the same few weaknesses, and nearly all can be avoided with planning in the first week. Check your own rollout against each row before go-live, and revisit the table at the 30-day review to confirm none of these problems has crept in.
| Mistake | Consequence | Fix |
|---|---|---|
| Buying sensors before defining goals | Data collected with no decision attached | Complete points 1 to 5 first |
| Using generic thresholds | False alarms and ignored alerts | Baseline each machine at comparable load |
| No owner for alarms | Alerts sit unread | Assign responders per tier |
| Uncalibrated sensors | Wrong approach and efficiency figures | Verify against a reference instrument |
| Monitoring separate from work orders | Findings never turn into repairs | Integrate alerts with the CMMS |
Oxmaint in a Chiller PdM Workflow
Asset recordChiller details, history, and documents
Inspection or readingMobile rounds capture temperatures and notes
TriggerThreshold breach or schedule creates a work order
ExecutionTechnician completes the job with parts and photos
ReportingDashboards show trends, cost, and compliance
KPIs for the First 90 Days
| KPI | Purpose | Review |
|---|---|---|
| Monitored chillers as share of fleet | Programme coverage | Monthly |
| Alarm to work-order time | Speed of response | Weekly |
| False alarm rate | Quality of thresholds | Monthly |
| Unplanned chiller downtime | Reliability outcome | Quarterly |
| Normalised kW per ton | Efficiency trend | Monthly |
| Oil and water result compliance | Condition of consumables | Per sample |
Frequently Asked Questions
How quickly can a chiller PdM programme go live?
Around 30 days when the BMS already holds most points. Retrofit sensors can extend this.
Do we need new sensors on every chiller?
Not always. Audit existing BMS points first, then add only what is missing. Talk through your plant with a specialist.
How often should chiller oil be sampled?
Follow the OEM schedule as a minimum, then adjust based on trends and machine criticality.
Can PdM replace preventive maintenance?
No. It refines intervals, but regulatory and OEM tasks remain. Manage both in one system.
What standards should the programme reference?
Commonly ASHRAE Standards 15, 147, and 188, AHRI performance standards, and ISO 20816 for vibration.
Get Your Chiller Programme Live in 30 Days
Bring the checklist, your chiller list, and your BMS points. Oxmaint helps you connect monitoring to work orders and records.
