Playbook: Cooling Tower Performance, Control, and Freeze Protection

Library-authored playbook. Mechanisms and safety constraints are grounded in DOE FEMP/PNNL tower O&M guidance, BEE chiller guidance, NREL tower-control discussion, and SPX/EVAPCO manufacturer material. Numeric rule thresholds remain commissioning or site/OEM values as stated on each card.

Freeze/electrical safety: follow the site and OEM freeze plan before field inspection. Do not bypass low-water cutoff, thermostat, over-temperature, vibration, fire, or other interlocks. Do not force a contactor or manually energize an unverified immersion heater. Confirm water level and electrical isolation using qualified personnel and the site’s lockout/tagout procedure.

Step 1 — Verify identity, applicability, and mode

1. Confirm point topology first: tower leaving water is the cold stream going to the chiller condenser; tower entering water is the warm chiller return.
2. Confirm every fan command, proof, speed, and outlet temperature belongs to the same tower object/cell. Do not pair a fleet OR or common header with one arbitrarily selected fan.
3. Confirm normal automatic ownership. Exclude maintenance, hand/local, tests, free cooling/waterside economizer, storage charging, emergency heat rejection, drain-down, deicing, or another approved low-water mode.
4. For TOWER-0006, establish that this is a wet tower with water intentionally present, the monitored heater/equivalent is part of the active freeze plan, basin level and low-water cutoff are healthy, and the configured basin limit and response time come from that site’s OEM plan. Otherwise report NO_EVAL.
5. Check freshness, time alignment, scaling, and sensor calibration. A command echo is not run proof; a heater contactor proves less than measured current, and current proves less than delivered basin heat.

Step 2 — Select the diagnostic branch

A. No fan proof or unexpected fan operation — TOWER-0004

1. Read yFailToStart versus yUnexpectedRun; the repairs differ.
2. Failure to start: check VFD/starter faults, disconnect/overload, vibration or OEM interlock, motor/belt/gearbox, final cell-stage command, and proof source.
3. Unexpected run: check local/manual mode, service override, second controller, stuck output/contactor, VFD internal command, and normal coast-down timing.
4. Compare with TOWER-0003. Cycling plus intermittent proof often points to a drive/overload/control issue before it points to tower thermal performance.

B. Overcooling while fan is loaded — TOWER-0005

1. Confirm the active setpoint is the final target for the same cold outlet.
2. Read yOvercooled and yFanLoaded separately. Cold water with the fan off is normal free convection and must not be dispatched as waste.
3. Check stale/local setpoints, VFD minimum speed, cell-stage deadbands, fan hand/override, sensor bias, and isolation/bypass valve sequence.
4. Review setpoint, mode, fan-state, or cell-count changes before the alarm; add a site holdoff longer than the normal response if required.
5. Do not assume colder condenser water always wastes whole-plant energy. Tune tower fan versus chiller lift from measured plant performance.

C. Basin freeze protection — TOWER-0006

1. Treat yHeaterFailToRun as electrical/command-proof evidence and yLowBasinTemp as thermal evidence. A proven heater cannot mask cold water.
2. From a safe state, verify basin water level, low-water cutoff, local thermostat/controller state, disconnect/breaker/fuses, contactor, current or power proof, and representative submerged temperature.
3. Confirm the sensor is away from the heater plume and not in a dry or stagnant pocket. Confirm local OAT represents the tower exposure.
4. Remember that a basin heater protects the basin/discharge area only; it does not by itself protect exposed piping, pumps, or heat exchangers.
5. Command false with status true is outside the current graph but can indicate an uncontrolled or dry heater hazard. Follow the OEM/site safety workflow immediately; do not treat it as a harmless energy-only condition.

D. Performance degradation — TOWER-0001, TOWER-0002, CHW-0005

1. For approach high, confirm full fan capacity and compare matched-load, matched-wet-bulb history against the commissioned design approach.
2. For range collapse, check condenser-water flow and actual heat rejection first; excess flow or an unloaded loop mimics tower failure.
3. For chiller condenser approach high with normal tower approach, investigate condenser tubes, flow, refrigerant-side non-condensables, and saturation- temperature derivation before cleaning tower fill.
4. If tower approach is high at capacity, inspect fill, spray/nozzle distribution, louvers, drift eliminators, air recirculation, fan delivery, and water-treatment records.

E. Fan short-cycling — TOWER-0003

1. Confirm independent motor starts and a complete warmed-up trailing hour.
2. Review leaving-water deadband, cell-stage sequence, and minimum on/off times.
3. If sequence corrections do not resolve it, inspect VFD/starter, overload, belt/gearbox, vibration switches, and proof chatter.

Step 3 — Correct remotely where authorized

1. Remove only approved stale overrides and restore normal automatic ownership.
2. Correct wrong point/setpoint mapping before changing physical controls.
3. Tune fan minimum, cell staging, deadband, and minimum on/off timers against measured response and manufacturer limits.
4. Never remotely defeat a protective interlock or raise/lower a freeze limit merely to clear an alarm.

Step 4 — Confirm resolution

1. Fan command and independent proof should agree after their configured direction-specific delays; cycling should remain below its warmed-up limit.
2. During normal mechanical heat rejection, leaving water should recover inside its active setpoint allowance without unnecessary loaded-fan persistence.
3. For degradation, compare approach/range and chiller kW/ton at matched load, flow, and weather over the next suitable operating period.
4. For freeze protection, follow the OEM/site commissioning procedure and verify water level, all safeties, independent electrical proof, and representative basin temperature. Do not close the finding solely because the graph cleared.