Playbook: Chilled-Water Plant Diagnosis

Adapted from HVAC FDD Reference v1.0, Remediation Playbooks (pp. 161–163), with library-authored command/proof, tracking, and cycling triage.

Step 1 — Establish commanded versus running machines (CHW-0008)

1. List each chiller separately: final BAS stage command, independent run proof, local/remote mode, active lockout, and timestamp freshness. Plant enable or a fleet-OR status is not enough.
2. For yFailToStart, follow the sequence in order: lead/lag selection, isolation valves, chilled/condenser-water pumps and flow proof, oil system, starter/drive, and chiller safeties. Do not bypass anti-recycle protection.
3. For yUnexpectedRun, check local/manual mode, service overrides, a second controller, welded/stuck outputs, and whether the command was bound upstream of the machine’s real control owner.
4. Verify the configured 300/120-second proof windows exceed this machine’s normal start and stop sequences plus point-delivery latency.

Step 2 — Confirm flow, permissives, and measurement boundaries

1. Confirm evaporator and condenser flow are established for each running machine and minimum-flow interlocks are satisfied.
2. Verify isolation and bypass valve positions, strainers, pump proof, and branch/header topology before diagnosing the refrigerant circuit.
3. Confirm chiller_load, temperatures, and power/tons belong to the intended machine. On parallel plants, do not read a mixed-header value onto each chiller without proving that boundary is the controlled target.
4. Rule out sensor, scaling, timestamp, flow-meter, and power-meter error.

Step 3 — Inspect CHWST tracking direction (CHW-0007)

1. Compare the individual evaporator leaving-water temperature with the final active target delivered to that controller. A common header is acceptable only when the staged machines genuinely regulate that same point.
2. Exclude startup pull-down, reset ramps, staging transfer, ice-making, and current/lift/surge/freeze/demand limiting before interpreting the alarm.
3. yTooWarm: check capacity, flow, fouling, refrigerant, permissives, sensor bias, and whether the setpoint actually reached the local controller.
4. yTooCold: check aggressive staging, local-loop tuning, reset delivery, sensor bias, and a machine controlling a different target than the BAS.
5. A CHW-0008 fail-to-start direction makes tracking non-evaluable; an unexpected-running machine can still have meaningful tracking evidence.

Step 4 — Review starts, timers, and staging (CHW-0009)

1. Trend per-machine proof at a cadence that resolves the shortest OFF/ON dwell. At the defaults, use a fixed 60-second evaluator and count_scale=evaluation_window/tick.
2. Compare each start with load, CHWST/setpoint, plant enable, lead/lag selection, minimum on/off timers, and lockout/safety history.
3. Look for low-load inability to turn down, narrow deadbands, insufficient loop volume/storage, unstable proof, or safety trip/auto-reset cycling.
4. Apply the manufacturer’s starts-per-hour and minimum on/off limits. The library’s three-start threshold is a commissioning placeholder.

Step 5 — Then evaluate efficiency, reset, delta-T, and approach

1. Review CHW-0001 kW/ton against its per-machine fitted baseline and verify the fit period, tons conversion, load domain, and meter boundary.
2. Check CHWST reset (CHW-0002) and loop DP reset (CHW-0003). A setpoint locked too low increases lift; a poor DP sequence can waste pump energy.
3. Review low delta-T (CHW-0004) with coil valves, bypass/decoupler flow, staging, and return-water temperature. Do not attribute a plant/header signature to one chiller without branch evidence.
4. Compare condenser (CHW-0005) and evaporator (CHW-0006) approaches with their design/commissioned values, load floor, refrigerant P-T provenance, water temperatures, and flow.
5. Check tower fan staging, condenser-water reset, and pump operation. Higher condensing lift can explain both approach and kW/ton degradation.

Step 6 — On-site service after controls and sensors are cleared

1. High condenser approach: inspect condenser flow and strainer, clean tubes, verify water treatment, and purge non-condensables as applicable.
2. High evaporator approach: inspect chilled-water flow/strainer and clean evaporator tubes.
3. Normal approaches with high kW/ton: leak-test and verify refrigerant charge, compressor/VFD current, power quality, oil system, and mechanical condition.
4. For persistent proof or cycling faults, inspect starter/drive histories, contacts, compressor protections, run-proof wiring, and local controller event logs before replacing equipment.

Step 7 — Confirm resolution

1. Verify command and independent proof agree through normal start/stop cycles.
2. Observe at least one full CHWST tracking delay after startup and staging; the active machine should remain within its commissioned band.
3. Observe at least one complete cycling window with manufacturer-compliant start count and minimum on/off times.
4. Confirm flow, delta-T, setpoint reset, and both approach temperatures are in their commissioned domains.
5. Monitor kW/ton on a disjoint post-repair period; do not refit the baseline on the faulty interval merely to make CHW-0001 clear.