Discharge Temperature: the 135°C Line, and Why Some Machines Get No Number

135 °C (275 °F) is the discharge temperature at which a process gas compressor duty stops being a routine selection and goes to an engineering review. It is not a screening figure invented for convenience: it is the hydrogen-rich limit in API 618, which caps the predicted discharge temperature at 135 °C (275 °F) for services of molar mass less than or equal to 12, and the Chinese petrochemical engineering specification SH/T 3143-2012 carries the same figure for the same service. The general limit in the same standards is higher — and that difference is the part most often missed.

The reason to know where it comes from is that the line does two jobs at once: it is a thermal safety boundary, and it is the number that decides how many stages a duty needs. The same clause that fixes the temperature also fixes the alarm and trip set points above it, which is what makes the line observable in service rather than merely designed. This article covers both jobs, and then the three machine families for which the question does not arise in the same form.

Where 135 °C actually comes from

The figure appears in two standards, and they are not independent — the Chinese petrochemical engineering specification for reciprocating machines follows the API document closely on this point.

Three things follow that are easy to miss.

Hydrogen is why the general number is not the number you use. API 618's own note gives the reason: non-lubricated hydrogen services generally run at higher discharge temperatures than lubricated ones, because of internal leakage and because hydrogen can release heat when it expands — a departure from the behaviour of the heavier gases in this category. A duty in hydrogen-rich gas therefore sits on a limit 15 K below the general one, and in the sixth edition 15 K below even that.

The limit is a reliability instrument, not a thermal ceiling. What it protects is the non-metallic part set — rings, packings and the diaphragm — whose life falls as the temperature rises. That is also why the same standard's speed clause runs in the same direction: the note to §6.4 says that, in general, rotating speed and piston speed in non-lubricated services should be less than in equivalent lubricated services. Going oil-free costs speed, and speed is what the non-metallic parts trade against.

The alarm structure is part of the limit. 135 °C on its own is a design number. The +10 °C alarm and +20 °C trip in SH/T 3143-2012, or the +20 K and +30 K recommendation in API 618, are what make it visible in service — and both documents require the measurement to be taken per cylinder, at the discharge outlet, close to the valve.

Used as a screening line rather than as a service limit, 135 °C is the conservative reading. For a gas that is neither hydrogen-rich nor in non-lubricated service, the applicable limit in the fifth edition is 150 °C, so a duty screened at 135 °C carries about 15 K of margin that a later, more careful pass can spend.

One gas overrides the line in the other direction. Vinyl chloride monomer is limited to 100 °C, because above that the gas starts to polymerise and the deposit goes onto the surfaces the machine needs (2).

Pressure ratio decides the temperature; suction temperature decides the pressure ratio

The working estimate is:

T2 = T1 × r^((k−1)/k)

where T1 is the suction temperature in kelvin, r is the absolute pressure ratio (discharge ÷ suction) and k is the gas's isentropic exponent (2). Two things follow that are easy to miss.

The exponent is not the lever; it does not vary enough. For most process gases k sits between 1.28 and 1.41 — carbon dioxide at the low end, nitrogen and hydrogen at the high end. Suction temperature and pressure ratio move the answer far more than the gas does, within that band. Monatomic gases are the exception: helium and argon at k ≈ 1.66 heat up much faster per unit of pressure ratio.

Suction temperature sits in front of the whole expression. Every kelvin of suction temperature is a kelvin of discharge temperature before compression starts, and it also reduces the pressure ratio that can be carried at each stage. This is why a warm suction is not a small correction to a compressor duty.

How many stages, and where the line falls

Take a common suction condition of 20 °C (293 K) and solve for the pressure ratio at which the discharge reaches 135 °C (408 K). The answer is the maximum ratio one stage can carry with that gas, assuming each intercooler returns the gas to the same temperature and k stays constant — both approximations, and both worth stating in an enquiry.

Two readings of that table matter more than the numbers.

The light, monatomic gases run out of stages first. Helium reaches the 135 °C line at about 2.3:1 per stage, which turns a 40 bar duty into a five-stage problem — beyond the four stages a single crankshaft carries in the reference data (1). At that point the arithmetic runs out before any family does, and the duty goes to engineering review rather than to a machine. Smaller helium duties still favour the diaphragm family: helium passes every sliding seal, and the diaphragm head is the construction with none.

Suction temperature moves the whole table. Re-run nitrogen with 40 °C (313 K) suction instead of 20 °C and the maximum ratio per stage falls from about 3.2 to about 2.5. Do the same for helium and it falls from about 2.3 to about 1.95. A 40 bar duty that fitted four stages at 20 °C suction now sits right on the four-stage limit — and a little more of either pressure ratio or suction temperature pushes it past.

That is the whole reason the discharge temperature line is treated as a specification input rather than a machine characteristic. It is a property of the duty.

Why three families do not get a number

Liquid ring machines. The gas is in contact with the working liquid through the whole compression, so the discharge temperature stays close to the liquid's own temperature rather than tracking the pressure ratio. The variable that a project controls is the working liquid — water, or a liquid chosen to suit the gas — and the family's published pressure ceiling is what limits the duty instead: to 13 bara for the dedicated compressor line (1). Asking a liquid ring machine for its discharge temperature at a given pressure ratio is asking the wrong question.

Roots machines. Compression takes place inside a pressure differential band — 1.0 bar for a single stage, 2.0 bar for two stages — so the temperature rise is bounded by the construction rather than by the compression ratio. The binding constraint is the differential pressure, and above 2.0 bar the machine is not a candidate at all (1).

Oil-injected screw machines. The injected oil removes most of the heat of compression as it is produced, so the discharge temperature is set by the oil flow and the oil's own temperature limit rather than by the pressure ratio alone. The figure a project watches here is the oil system's, not the gas side's.

The two families that do get a gas-side number every time are the diaphragm and the dry screw. For the diaphragm, the binding number is the standard limit itself — 135 °C for a hydrogen-rich duty, and 120 °C in the sixth edition of API 618 for hydrogen-rich or non-lubricated service — so there is little thermal headroom to trade on that family. For the dry screw, no lubricant sits in the compression chamber, so the family's practical band sits wider — and the constraint that binds instead is the 15 bar differential pressure limit published for oil-free machines (1).

The second temperature limit nobody puts on the sheet

Two gases carry a temperature limit that is not about the machine's materials at all.

Vinyl chloride monomer, at 100 °C (2). Polymerisation rate climbs with temperature, and polymer deposits on valve plates, clearances and seals. The limit protects the machine by protecting the gas.

Oxygen. The temperature that matters here is not a machine limit but a materials-prohibition boundary: oil in contact with oxygen is a combustion hazard, which is why oxygen duties are oil-free without exception and copper alloys are excluded at high pressure (2). No discharge-temperature figure makes an oily machine safe on oxygen.

What to put on the enquiry

Three lines make the temperature question answerable, and all three are routinely left off:

1. Suction temperature, in °C, as a normal operating value and as a design maximum.
2. The pressure basis of both pressures — bar(a) or bar(g).
3. The gas composition, or at least the majority component and the contaminants, so that k can be taken from data rather than assumed.

With those three, a stage count can be bounded without any machine being sized, and the enquiry can be routed to the right family. Without them, the discharge temperature is not estimable and the stage count is a guess.

Where this goes to a human

An estimated discharge temperature above 135 °C moves the duty to an engineering review, as does a vinyl chloride duty above 100 °C, a stage count above four, or any request to design the intercooling (2). Those cases need the gas composition, the cooling medium and the intercooler outlet temperature — the machine cannot be selected around a temperature that nobody has fixed.

What this article does not cover

Note that this article does not calculate stage count, intercooler duty, power or efficiency, and it does not design a cooling arrangement. It sets out where the 135 °C line comes from, how it bounds the pressure ratio per stage, and why three machine families are constrained by something else instead.

Related on this site

• Hydrogen Compression Reliability: Where the Downtime Actually Comes From
• Diaphragm vs Piston vs Screw Compressors: A Comparison Table
• Reciprocating or Diaphragm for High-Pressure Hydrogen?
• Run a duty point: gascompressorselect.com

Sources