17.3 Surge Drums, Separators, and Wet Suction Returns

17.3 Surge Drums, Separators, and Wet Suction Returns

The surge drum, low-pressure receiver (LPR), recirculator, and suction accumulator/separator are the same family of vessel on an industrial ammonia overfeed plant: a low-side vessel that separates vapor going to the compressors from liquid going to the pumps, and that holds a surge volume for load and defrost swings. Names vary by manufacturer and by whether the vessel is mounted on a single evaporator or serves the whole engine room. CIRO cares about the jobs, not the paint stencil.

Do not confuse this vessel with the high-pressure receiver after the condenser. Different pressure, different liquid, different cutouts.

What the vessel must do

1. Disengage vapor from the wet return so compressors take dry suction (or as dry as internals and velocity allow).
2. Store liquid for the recirculation pumps with enough submergence for NPSH (Chapter 18 owns the pump math).
3. Surge: accept a slug from a defrosting coil, a pull-down, or dumped evaporators without immediately filling the compressor suction.
4. Protect with a high-level cutout that stops compressors (and often pumps) before liquid carries over.
5. Give oil a place to settle so it can be drained to an oil pot — without confusing a high oil pad with a high ammonia level.

If the vessel fails any one of those jobs, the plant still has coils and compressors, but it no longer has a safe overfeed architecture.

Wet suction returns

Overfeed evaporators do not send superheated vapor home. They send a two-phase mixture in wet suction (overfeed return) mains. Those lines must be pitched and sized for two-phase flow, not for DX dry-suction velocity. Liquid is expected. Traps, pockets, and leftover two-inch because it was cheaper are how you log liquid and oil in the returns, starve the separator of returning mass, and then wonder why the LPR is low while coils are holding a flood.

Wet returns typically enter the vessel above the liquid level, often through a drop pipe, baffle, or inlet that aims the mixture so liquid falls and vapor turns toward the outlet. A return dumped below the liquid level can agitate the pool, generate foam, and carry over. A return aimed at the compressor outlet nozzle is a slug machine.

Dry suction leaves the top of the vessel to the compressors. That line should be vapor. If you see frost or a cold pipe all the way to the compressor with zero superheat and a climbing LPR level, you are looking at carryover, not at a healthy flooded coil. A small amount of superheat at the compressor suction after a well-designed separator is normal machine protection; it is created in the dry suction, not by drying out the evaporator.

Vertical shells throw droplets downward against gravity and are common as recirculators. Horizontal shells give more surface area for a given height of level change — useful surge — but vapor must still turn and slow before the outlet. Either shape still needs inlet baffling and an outlet that is not a vacuum cleaner over the liquid.

Internals and residence time (conceptual)

CIRO will not hand you a proprietary sizing spreadsheet. RETA does not publish a mandatory residence-time minute count in the exam facts you are using. Think in time and velocity:

• Disengagement space (vapor volume above the liquid) must be large enough, and vapor velocity low enough, that droplets fall back instead of riding to the outlet. Mesh pads, vanes, or baffles help, but they are not a license to undersize the shell. A collapsed mesh pad after a slug is a real failure mode: the vessel looks unchanged from the outside while carryover starts.
• Residence time in the liquid is how many minutes a slug of returning liquid can sit while pumps continue to feed coils and compressors continue to pull vapor. Too small, and a single hot-gas defrost dump drives high level and a trip — or worse, carryover before the trip. Too large, and ammonia inventory soars (PSM/RMP pounds) and oil has more volume to hide in, delaying oil return and creating a false I-never-need-to-drain-the-pot culture.
• Surge volume between normal working level and the high-level cutout is the operational cushion. If operators run the vessel at the high alarm because the pumps like it, there is no surge left. The next defrost dump is a trip or a slug.

A teaching picture: if wet returns can dump several cubic feet of liquid in a short defrost equalization, the distance from normal level to trip must absorb that dump at the vessel's surface area. You do not need the exact cubic feet on the exam. You need to know why a small drum on a large overfeed plant will high-level during defrost, and why staggering defrost is a separator issue as much as a frost issue.

IIAR 2 governs machinery-room and vessel design for ammonia; it does not replace operator judgment about running level. Do not invent a code-required diameter if the stem does not give one.

Level: operating band, alarm, cutout

A high-level cutout that is bypassed because it nuisance-trips on defrost is how screws ingest liquid. The correct response is vessel volume, defrost stagger, level setpoint, and return piping — not a jumper. Ladder-diagram interlocks (Chapter 7) should show this cutout in the compressor start circuit.

Dual level devices are common: a continuous transmitter plus independent float switches for trip. Chapter 18.3 owns instrument types. Here, remember redundancy on the trip that protects the compressor. A single DP transmitter that was never re-zeroed after an oil drain is not a safety system.

Oil logging versus liquid refrigerant

Ammonia liquid and immiscible oil both collect in the bottom of a cold vessel. They are not the same high-level problem.

• Ammonia liquid high: the float and the transmitter both see a rising liquid height of refrigerant. Compressors are in danger of slugging liquid ammonia. Pump NPSH may actually look better just before the trip. The wet suction is bringing refrigerant.
• Oil logging: oil pads on the bottom. Pumps may lose ammonia NPSH even though a sight glass looks up, or a DP transmitter calibrated on ammonia density reads wrong because the bottom is oil. Oil can block a drop leg. Compressors may still get dry vapor while pumps cavitate. Draining the oil pot (Chapter 18) restores ammonia liquid at the pump.

Sight glasses can lie: oil stratifies; a glass in the vapor space frosts differently than a glass in liquid. A high-level trip must be set on a device that sees liquid height that would carry over, not on the oil pot is due. Conversely, do not drain ammonia into a bucket because you thought a high glass was oil.

Foam from a violent inlet or from oil-ammonia agitation can fool some level technologies. Inlet baffling and reasonable vapor velocity are part of making the trip honest.

Single-coil surge drums versus plant recirculators

A surge drum on a flooded chiller or a large flooded coil sits at the evaporator. Level in that drum is the evaporator level. Suction to the compressor (or to a plant wet return) leaves the top. Liquid makeup comes from a level valve.

A plant LPR / recirculator sits in the engine room or penthouse mechanical space and serves many overfeed evaporators. Wet suctions come home; dry suction goes to the machines; liquid pumps go out. High level here protects every compressor on that suction. Low level here protects every recirculation pump on that vessel.

Two-stage plants add another low-side vessel story (intercoolers, Chapter 19). Do not treat an open intercooler as a substitute explanation for a high-stage LPR. Each suction level that takes wet return needs its own separation and its own high-level protection.

What you watch on a CIRO screen

Normal: LPR level in band, pump amps stable, suction at the compressor showing vapor (separator doing its job), coil TD healthy, wet returns cold.

Abnormal patterns:

• Level rising with pumps on and compressors loaded: extra liquid from the high side (makeup valve, hot-gas defrost drain), overfeed too high, or load collapse so evaporated mass fell.
• Level falling: underfeed from the high side, pump-out, leak, or evaporated mass high with makeup stuck.
• Level steady but compressors slugging: internals failed, mesh collapsed, vapor velocity too high (undersized vessel or too many machines on one drum), or a suction takeoff too close to the inlet.
• Pumps cavitating with a good-looking glass: oil, vortex, blocked suction strainer, or false level.

Exam traps

• Wet suction should be wet. Dry suction should be dry. The separator is the boundary.
• High-level cutout protects compressors. Low-level cutout protects pumps.
• Bigger is not automatically safer: inventory and oil logging grow with volume.
• Oil at the bottom is not extra refrigeration charge you can ignore, and it is not the same as a liquid-ammonia high level.
• Do not invent a mandatory residence-time minute count. Describe surge volume, vapor velocity, and defrost dumps.