17.2 Overfeed Ratios and Recirculation Rate

17.2 Overfeed Ratios and Recirculation Rate

Liquid overfeed (also called liquid recirculation) is flooded-coil operation with a pump — mechanical or gas-powered — that delivers more liquid mass than the coil evaporates. The extra liquid keeps every circuit wetted. Two-phase fluid leaves in a wet suction and the separator (surge drum / LPR / recirculator) splits vapor to the compressors from liquid back to the pumps.

This is the industrial answer to ammonia's small evaporated mass per ton. You do not try to meter exactly the boiled pounds the way a TXV does. You overfeed so quality in the tube stays well below 1.0 and the wall stays wet.

The ratio, in one line

Recirculation rate and overfeed ratio are the same idea on CIRO:

Overfeed ratio = mass circulated ÷ mass evaporated

A 4:1 overfeed means four pounds of liquid are pumped to the evaporators for every one pound that boils. Three pounds return as liquid. One pound returns as vapor (plus the vapor that formed in the coil). Typical industrial ammonia overfeed is 3:1 to 4:1. That band is common industrial practice. If a stem or a manufacturer nameplate gives a different ratio for a specific coil, use the stem.

Say the numbers so you do not reverse them. Circulated mass is the numerator. Evaporated mass is the denominator. A ratio of 4 is not four tons of refrigeration per pound. It is a mass ratio.

If the feed entering the coil is essentially saturated liquid at evaporator pressure, outlet quality (vapor mass fraction) is about 1 / ratio. At 4:1, roughly 25% of the leaving mass is vapor and 75% is still liquid. At 3:1, quality is about 0.33. At 1:1 you have evaporated everything — DX thinking — and dry-out is one maldistributed circuit away. That quality picture is why wet suction must be sized and pitched as two-phase, not as DX dry vapor.

Worked mass-flow clip

1 ton of refrigeration = 200 Btu/min. Ammonia's latent heat is large; look up h_fg on the on-screen P-h or saturation chart at the evaporating temperature, not at room temperature.

Example class (freezer, saturation near −20°F). Suppose the chart shows h_fg ≈ 580 Btu/lb (order of magnitude; use the chart on exam day):

• Load on one penthouse: 40 tons
• Evaporated mass ṁ_evap = (40 × 200) / 580 ≈ 13.8 lb/min
• At 4:1 overfeed, ṁ_circ = 4 × 13.8 ≈ 55 lb/min
• Liquid returned (not evaporated) ≈ 41 lb/min

At 3:1, ṁ_circ ≈ 41 lb/min. The coil still sees far more liquid than a DX coil that would feed only the 13.8 lb/min evaporated.

If you accidentally use volume (gpm) without converting density, you will mis-size the pump and lie to yourself about ratio. Overfeed is mass. Convert pump gpm to lb/min before you claim a ratio. Liquid ammonia specific gravity is often around 0.6–0.7 in the usual low-side range — look it up on the chart for the actual temperature. A teaching conversion is lb/min ≈ gpm × 8.33 × SG. A pump clocking 20 gpm is not automatically 4:1 on a 40-ton coil until you finish that conversion and divide by evaporated mass.

Field checks that actually support a ratio claim: a liquid flow meter on the pump discharge (best), a pump curve plus differential head (honest if the curve is for that ammonia and those orifices), or a crude amp draw trend that only tells you the pump is or is not on its usual operating point. Amps without a curve do not equal 3:1.

Why 3:1 to 4:1, not 1:1

At 1:1 you are trying to feed exactly the evaporated mass — that is DX thinking. Any maldistribution, any oil film, any frost-blocked circuit, and that circuit dries out. Quality reaches 1.0 before the outlet. Inside HTC falls. That circuit's TD rises: the space or product cannot get the capacity; frost dies out before the suction header; other circuits may still be wet. The room sees lost capacity and a high evaporator TD.

Extra liquid is a wetting insurance policy. It also helps oil move along the bottom of circuits toward the wet return instead of baking onto the tube wall. Ammonia overfeed coils are circuit-balanced with hand expansion / balancing valves on the liquid drops, not with TXVs. Those valves are opened to get design flow (and sometimes a minimum flow for oil), not to hold superheat. Closing them to save ammonia is how you create dry-out.

Too low

Symptoms of insufficient overfeed (ratio sagging toward 1–2, a failing pump, a throttled hand valve, a plugged strainer, a circuit valve cranked shut, or a gas pump that is not transferring):

• Dry-out at the top or outlet of some circuits
• High evaporator TD for the load
• Uneven frost: heavy at the inlet, bald toward the suction header
• Capacity missing while suction pressure is not correspondingly low (the coil is not using the surface)
• Warm leaving air or product with a coil that should be big enough

Do not first blame the compressor slide valve. Ask whether liquid is actually recirculating at the design multiple. A strainer in the pump suction or in a liquid drop will fake a low ratio with a pump that still sounds busy.

Too high

Overfeed is not more-is-always-better. Excessive recirculation (pump wide open, second pump started for no reason, ratio drifting toward 6:1 or 8:1 on a coil designed for 4:1):

• Pump power rises. Recirculation pumps are often modest horsepower individually, but a plant with several pumps running off their curves is real kW. On a CIRO operating-cost screen, unexplained low-side pump kW is a recirculation story.
• Wet suction carries more liquid. Return mains and the separator see a higher liquid fraction. Two-phase pressure drop goes up. The separator can overload: vapor velocity through the disengagement space rises with compressor load at the same time liquid rain increases, so droplets carry over.
• High level in the LPR becomes more likely during defrost dumps and load swings because more liquid is in transit in the returns.
• Coil internal pressure drop can rise, which hurts TD from the other direction.

The separator (Section 17.3) is sized for a design recirculation rate, not for an arbitrary flood.

Mechanical pump vs gas pump

Mechanical liquid pumps (typically canned-motor or sealless ammonia pumps) take liquid from the LPR and push it to the evaporator liquid header. Gas-pumped systems use high-pressure gas (often a transfer-vessel / CPR-style arrangement) to push liquid without a rotating pump. Chapter 18 owns NPSH, cavitation, and oil-pot details. Here you only need: both methods exist to produce overfeed; the ratio definition does not change. If the exam shows a gas-pumped plant, you still compute circulated mass over evaporated mass.

Recirculation language you will see

CIRO-style trap: treating 4:1 as circulate 4 lb and evaporate 4 lb. Wrong. Evaporate 1, circulate 4.

Load changes evaporated mass, not pumped mass (unless you VFD the pump)

A fixed-speed pump delivers roughly a fixed gpm (curve versus head). Evaporated mass follows the load. At light load, ṁ_evap falls, so the ratio rises even if the pump does nothing. Light-load overfeed can become very wet returns — good for wetting, harder on the separator. At heavy load, ṁ_evap rises, ratio falls, and you approach dry-out if the pump cannot keep up. That is why pumps are selected for design load at 3:1 to 4:1, with some margin, and why a plugged liquid line at peak pull-down is a dry-out event.

A VFD on a recirculation pump can track load and hold a more constant ratio. If the drive is in hand at minimum speed during a blast freeze, you have invented low overfeed. If it is locked at 60 Hz on a Sunday night with one room calling, you have invented extra pump kW and a wet separator.

Defrost makes the ratio dynamic. A coil off on hot-gas defrost evaporates little (it is condensing inside). Neighbors may still be at full load. Header flow can shift. After defrost, a cold coil can take a gulp of liquid. The surge drum exists so those gulps do not go to the compressor.

Exam traps

• Ratio is mass, not tons, not gpm until converted.
• Typical industrial ammonia: 3:1 to 4:1.
• Too low → dry-out, high TD. Too high → pump kW, wet suction, separator overload.
• Overfeed is flooded-coil operation with extra liquid; it is not a DX TXV set a little rich.
• Do not invent a RETA-official pump kW formula. Direction of the effect is what is tested.