On many gas-handling facilities, compression is among the largest equipment investments and power consumers. It is also central to getting gas to its destination: each required pressure increase comes with a power and operating cost. Compression decisions are therefore economic decisions as well as mechanical ones.

Two ways to squeeze a gas

The industry sorts compressors into two families. Positive displacement machines trap a fixed volume and shrink it. The reciprocating compressor is a piston in a cylinder, a syringe in steel, at home at high ratios and modest flows, delivering its gas in pulses. Its most elegant detail is also its weakest point: the suction and discharge valves are thin spring-loaded plates with no timing gear at all, opening purely on pressure difference, millions of times a week.

Dynamic machines trap nothing. A centrifugal compressor's impeller, spinning at thousands of rpm, flings gas outward and the diffuser trades that velocity for pressure. Per wheel the ratio is modest, so wheels stack in series, and in exchange you get enormous, smooth flow from a continuously rotating machine. Choosing between the families is mostly two questions: how much gas, through what ratio. Large-flow export duties often favour centrifugal machines, while reciprocating machines are well suited to many lower-flow, higher-ratio duties. Gas lift and injection can use either family, depending on the application.

The wall is made of heat

Compression adds energy to a gas and raises its temperature. The discharge temperature depends on the gas properties, inlet temperature, pressure ratio, efficiency, and machine design. Manufacturer limits vary by service; neither 150 °C nor a pressure ratio of three or four is universal. Multistage compression with intercooling helps manage temperature and power. The number of stages is selected against the full operating envelope, mechanical limits, and economics, rather than a single rule of thumb.

And cooling a rich gas condenses liquid. So after every cooler stands another scrubber, and the repeating pattern on every compression skid (scrubber, compressor, cooler, again) stops being decoration and becomes visible logic.

The enemy in the title

A compressor cylinder squeezes gas into a small clearance space. Liquid is far less compressible than gas. When a slug arrives, the piston drives it into a space it cannot fit, and something structural must yield. Damage can include failed valves, bent rods, and structural damage. The failure sequence is not predictable or benign. Machines have been wrecked in a single stroke. Even fine mist is a slow poison, washing the lubricant film off cylinder walls and eroding impellers. The cost arrives later, as an overhaul that comes years early.

This is why suction separation and liquid-level protection matter. High liquid level is commonly linked to protective shutdown logic, with the sensing arrangement and response specified for the machine and facility. A scrubber is part of the protection system, not a guarantee that liquid can never reach the compressor.

Surge: the moment the machine fights back

A centrifugal compressor lives on a curve: at a given speed, more head means less flow. Walk up that curve far enough and you reach its end: the machine can no longer hold back the system, and flow reverses. High-pressure gas slams backward through the compressor, pressure collapses, flow recovers, and the cycle repeats in violent oscillations. That is surge, and repeated cycles hammer thrust bearings on a rotor that runs with clearances measured in fractions of a millimetre.

The protection is disarmingly simple and deliberately wasteful: before the operating point reaches the surge line, a recycle valve routes discharge gas back to suction. The compressor sees plenty of flow and stays stable, while the facility burns power compressing the same gas twice. That wasted energy buys the machine's life. So a running centrifugal lives between two invisible fences: surge on the low-flow side, choke on the high-flow side, with compressor control as the quiet full-time job of staying in the corridor.

Machines drift before they fail

Large compressor trains are closely monitored: vibration, temperatures, pressures, lubrication, and seal-system conditions can all reveal developing trouble. Many failures are preceded by useful trends; some are sudden. Monitoring helps teams act earlier, but it complements rather than replaces inspection, maintenance, and protective systems.

I’m a practising oil & gas facilities engineer. If you have run one of these machines, or been woken up by one, the YouTube comments are exactly where that story belongs.

Further reading

• Ariel: Application Manual