13.1 Floating Head Pressure
Key Takeaways
- Floating head lets condensing pressure fall with wet bulb (evaporative) or dry bulb (air-cooled) instead of a year-round fixed high setpoint, which cuts compressor BHP/ton as pressure ratio falls.
- The weather sets a thermodynamic floor at ambient plus condenser approach; the plant sets a higher floor from oil-separator velocity, thermosiphon or high-side liquid feed, expansion-valve ΔP, hot-gas defrost, and liquid-injection oil cooling.
- Do not float so low that hot-gas defrost loses driving pressure or liquid-injection oil cooling cannot meter liquid; clamp at the most restrictive of those process limits.
- High head with condenser fans already at 100% and a low wet bulb is an approach, fouling, or non-condensable problem—not a reason to abandon floating.
- Holding extra head to save condenser-fan kW usually costs more compressor shaft power than the fans return.
13.1 Floating Head Pressure
Industrial ammonia plants used to hold a fixed high condensing pressure all year—often near 185 psig, about 96°F saturated condensing temperature (SCT)—so expansion valves, hot-gas defrost, and high-side liquid feeds always had surplus driving force. That setpoint is easy to live with in August. In January, when the evaporative-condenser wet bulb is 40°F, it is an energy tax. Floating head pressure is the operator and control strategy that lets condensing pressure fall with wet bulb (evaporative condensers) or dry bulb (air-cooled condensers) instead of fighting the weather with a high clamp.
The payoff is BHP/ton. Compressor shaft work tracks pressure ratio. When discharge pressure drops, the screw does less work per pound of ammonia, volumetric efficiency improves, and oil-cooler load usually falls with discharge temperature. Plants that actually float see the kW/ton KPI drop on cool, dry days without anyone changing suction setpoint. That is why CIRO treats floating head as a monitoring-and-performance topic, not a luxury control trick.
What floating looks like on a CIRO screen
A typical package screen lists a 300 HP screw, 480 VAC, power factor around 0.86, motor efficiency near 93%, thermosiphon oil cooling, and an evaporative condenser. You do not need any vendor's sample numbers memorized as trivia. You need to read the relationship:
- Wet-bulb (or dry-bulb) versus SCT / condensing pressure
- Condenser fan (and spray-pump) percent
- Discharge versus receiver pressure
- kW, BHP, and $/hr on the same timestamp
If wet bulb is 58°F, fans are 70–100%, and SCT sits near 70–75°F (ammonia in the neighborhood of 114–130 psig), the plant is floating. If wet bulb is 58°F, fans are 100%, and SCT is still 95°F (~181 psig), the head is not following ambient. That is condenser approach, fouling, or non-condensables—diagnose those in condenser approach. Do not raise a winter fixed setpoint and call it safe.
Condenser approach is the gap between SCT and the ambient that can actually reject heat. On an evaporative condenser, approach is SCT minus wet bulb. Floating does not repeal approach; it lets SCT ride wet bulb plus a realistic approach, instead of wet bulb plus 40°F of wasted lift.
Why BHP/ton falls
At a fixed suction, lowering condensing pressure cuts compression ratio. The P-h diagram shows a shorter compression lift and a larger refrigerating effect per pound because liquid leaving the condenser is cooler (more subcooling opportunity, less flash after the expansion device). Both effects improve BHP/ton: the numerator falls, and the denominator often rises. Fan energy is real, but condenser-fan kW is small next to a 300 HP screw. Turning fans down to save the condenser while holding 185 psig in 40°F weather is usually a net loss.
A directional plant trend—not a manufacturer map and not a RETA-published constant—is that compressor input often drops on the order of about 1.5–2.5% per 1°F of SCT in the industrial band operators actually float through. Treat that as a reason to watch kW/ton as head moves, not as a number to recite as code. Plot kW against SCT on your own screens; CIRO items will ask whether the trend is consistent with floating or with a stuck high setpoint.
The limits: you cannot float to wet bulb
The atmosphere sets a floor. The process sets a higher one. Float until the most restrictive of the following limits is binding, then clamp. Do not keep lowering a PID setpoint because the KPI still looks tempting.
| Limit | What fails if you go too low | Operator clues |
|---|---|---|
| Oil-separator velocity | Coalescer carryover; oil leaves with discharge | Falling sump, oil in evaporators, high separator ΔP or mist after the vessel |
| Thermosiphon / high-side liquid feed | Oil-cooler loop stalls; remote DX or gravity feeds starve | Rising oil temperature on a thermosiphon package; high superheat at remote DX while the receiver still shows liquid |
| Expansion-valve ΔP | TXV/EEV mass flow collapses with the square root of available drop | Coil superheat high, suction falling, evaporator TD large while head looks efficiently low |
| Hot-gas defrost | Not enough pressure to feed the coil and leave through the drain regulator | Long defrosts, residual ice, low drain/relief pressure |
| Liquid-injection oil cooling | Injection valve cannot meter liquid into the compressor | Oil temperature and discharge temperature climb together at light load |
Oil-separator velocity. Discharge-gas density falls as pressure falls. Mass flow through a loaded screw often rises as ratio falls (more capacity). Velocity through the separator is mass flow divided by density and area, so both effects push velocity up. Coalescing elements and drain legs are sized for a velocity band. Too fast, and oil leaves with the discharge; the sump drops, bearings starve later, and you chase oil in evaporators. A minimum head, a fan strategy, or a separator differential alarm is how plants stay inside that band.
Thermosiphon and high-side pressure for liquid feed. Thermosiphon oil coolers circulate by density difference between a high-side liquid column and the warmer cooler return. If receiver/condenser pressure and liquid level are too low, the loop stalls: oil temperature climbs even though the plant is unloaded and the weather is cold. Remote liquid feeds that rely on high-side pressure rather than a mechanical pump—DX stations, some high-side float arrangements, long roof runs—also lose driving ΔP. Pumped overfeed is less sensitive to head for mass delivery, but the high-side still has to feed the pump receiver properly.
Expansion-valve ΔP. A TXV or EEV is an orifice plus a control loop. Mass flow through the valve scales with the square root of the pressure drop available. Float the head until ΔP is below what the coil needs at that load, and superheat rises, the coil starves, and suction pressure falls even though the condenser looks efficient. The exam trap is blaming the TXV when the real problem is a winter head clamp set too low for that evaporator.
Defrost hot-gas pressure. Hot-gas defrost needs discharge/hot-gas pressure high enough to push gas into the coil, condense, and leave through the defrost drain pressure-relief regulator. If floating head sits near or below what that regulator and the coil pressure drop require, defrosts run long, ice remains, evaporator TD collapses, and you add compressors that would not have been needed if the coil were clean. Some plants raise head only during a defrost window; that is still floating, with a temporary floor. Details of the defrost sequence belong with hot-gas defrost; here the point is that floating head is one of the pressures that sequence drinks from.
Liquid-injection oil cooling. Packages that inject high-side liquid into the compressor for oil and discharge-temperature control need liquid pressure and subcooling at the injection point. Float below that, and injection becomes intermittent: oil temperature and discharge temperature climb, viscosity falls, and the slide or VFD cannot save the bearings. Thermosiphon packages can often float lower than liquid-injection packages. Know which oil cooling the screen shows before you copy another plant's minimum psig.
Operator clamp logic
Set the floating-head minimum to the most demanding of separator velocity, thermosiphon/liquid feed, expansion-valve ΔP, defrost, and injection. Document which limit binds in summer versus winter. Do not copy 185 psig from a 1990s sequence because ammonia likes high head. Do not copy a magazine's low clamp if your defrost regulators and liquid-injection valves were never checked at that pressure.
What floating is not
- It is not uncontrolled head. There is still a transmitter, a PID or stepwise fan/pump sequence, and a low clamp.
- It is not an excuse to ignore non-condensables. NC gas raises condensing pressure at any ambient; floating will look like it failed when the purge is the repair.
- It is not flooded-condenser hold-back used as the year-round strategy. Hold-back is a tool to raise head when a limit is binding, not the default.
Read kW/ton, oil temperature, defrost duration, and liquid-feed superheat together. If BHP/ton improved but defrosts doubled and oil is 20°F hotter, you floated through a limit. Raise the clamp, fix the binding constraint (undersized injection, dirty separator, defrost regulator set too high), then float again.
A January night shows evaporative-condenser wet bulb at 42°F. The plant used to hold about 185 psig year-round. Why should a supervisor enable floating head instead of keeping that fixed setpoint?
Floating-head control is pulling receiver pressure down on a thermosiphon-cooled screw that also uses hot-gas defrost. Which observation should stop further lowering of the minimum-head clamp?
Why can oil carryover get worse as a loaded screw is allowed to float to a lower discharge pressure?