Pressure is one quantity. Engineers measure it in at least four units in everyday work — pounds per square inch (psi), bar, kilopascal (kPa), and megapascal (MPa) — and which one ends up on a drawing has almost nothing to do with physics. It is decided by the country the equipment will run in, the discipline writing the spec, the customer’s house style, and the code section the part has to satisfy. Mix them up and the worst case is not a missed conversion factor: it is a hydraulic cylinder ordered for 250 bar that ships rated for 250 psi, an order of magnitude low, with a ten-week lead time.
This guide is about the practical choice — when each of the four common units is the “right” one, the conversion factors worth memorising, the gauge-versus-absolute conventions that trip up cross-market work, and the small drafting habits that keep a multinational design package internally consistent.
Why four units for the same quantity?
The pascal (Pa) — one newton per square metre — is the only SI unit of pressure. The BIPM SI Brochure lists it as a derived unit with a special name, and ISO 80000-4 (Mechanics) treats it as the reference for every pressure quantity in engineering. Yet a pascal is a tiny pressure: standard atmospheric pressure is roughly 101 325 Pa. So in practice nobody uses raw pascals on a drawing. They use one of three SI-derived prefixed forms (kPa, MPa, GPa), or one of two non-SI but widely tolerated units (bar, psi) that survived because they were already entrenched in industry when SI was standardised.
NIST Special Publication 811 is explicit that the bar is “outside the SI” but accepted for use with SI; the psi has no such acceptance and is treated purely as a US customary unit. That status difference — official SI, accepted non-SI, US customary — is the cleanest way to remember which unit belongs where.
PSI: the US default that refuses to die
Pounds-force per square inch is dominant in the United States across nearly every pressure-bearing application that predates the metric switch. Consumer-facing pressures — passenger-car tire pressure, garden hose, bicycle tubes, scuba tanks sold in the US — are quoted in psi by reflex. Industrial pneumatics in the US still defaults to 90 psi shop air; OSHA 29 CFR 1910 limits compressed-air gun pressure to 30 psi for cleaning. ASME’s pressure-vessel codes, when written for US customary practice, use psi for design pressure and stress allowables.
Where psi causes friction is the moment a US-built skid leaves the country. A spec that reads “design pressure 150 psig, hydrotest 225 psig” is unambiguous in Houston and meaningless in Frankfurt without a conversion. The reverse also bites: an inspector in Singapore reading a European drawing in bar and assuming psi has rejected gauges that were perfectly correct for the system.
Bar: the European industrial workhorse
The bar is exactly 100 000 Pa = 100 kPa = 0.1 MPa, and equals about 14.504 psi. It maps almost one-to-one onto standard atmospheric pressure (1 atm = 1.01325 bar), which is the reason it survived: a meteorologist or process engineer can substitute “bar” for “atmospheres” with under 2% error and rarely cares.
In European, Japanese, Korean, and Chinese industrial equipment catalogues, bar is the default pressure unit for pneumatics, hydraulics under about 700 bar, refrigeration, and most general process work. A Festo, SMC, or Bosch Rexroth datasheet will almost always list operating pressure in bar; the psi conversion appears in parentheses for the US market. Bar is not part of the SI but appears in the BIPM SI Brochure’s table of non-SI units accepted for use, alongside the litre and the hour.
A practical reason to choose bar over kPa on a hydraulic schematic: the numbers stay short. A 200 bar system reads more cleanly than 20 000 kPa, and a working pressure of 6 bar for shop air is easier to call out than 600 kPa. Where the numbers get large — high-pressure hydraulics, injection-moulding clamps, pressure-vessel design — engineers usually switch to MPa anyway.
kPa: the SI default for everyday and HVAC pressures
Kilopascal is what you get when you stay strictly inside SI and round to a sensible prefix. It is the standard unit on weather reports outside the US, on European and Asian passenger-car tire placards (a typical sedan reads 220 kPa front, 240 kPa rear), and on HVAC duct static-pressure and refrigerant low-side gauges in metric jurisdictions. Most national building and gas codes outside the US specify pipe and appliance pressures in kPa.
For atmospheric and weather pressures, the millibar and the hectopascal are numerically identical (1 mbar = 1 hPa) and both equal 0.1 kPa, which is why aviation altimeters set in “hPa” still read the same number a pilot would expect in millibars. ICAO standard atmosphere is 1013.25 hPa = 101.325 kPa.
kPa is also the right unit to reach for whenever the working pressure is within an order of magnitude of atmospheric: low process pressures, ventilation, plumbing supply, low-pressure gas distribution, cleanroom pressurisation. Once you go above a few hundred kPa it usually pays to switch to MPa to keep the numbers small.
MPa: hydraulics, materials, and metric pressure-vessel codes
Megapascal — one million pascals, equal to 10 bar or 145.04 psi — is the unit of choice for three audiences: materials scientists, hydraulics engineers, and anyone working to a metric pressure-vessel code.
In materials data, yield strength, ultimate tensile strength, fatigue endurance limit, and Young’s modulus are quoted in MPa or GPa across virtually every modern reference. ASME BPVC Section II Part D’s metric tables list allowable stresses in MPa; structural steel grade S355 has a 355 MPa minimum yield by definition; a typical aluminium 6061-T6 ultimate is 310 MPa. Quoting these in psi is correct but unwieldy: 355 MPa is 51 488 psi, which nobody reads at a glance.
In hydraulics, MPa replaces bar once pressures climb past about 400 bar. A 700 bar (70 MPa) hydraulic cylinder, a 250 MPa ultra-high-pressure waterjet, or a 140 MPa metal-forming press all read more naturally in MPa than in bar. Metric ASME work and the European Pressure Equipment Directive (PED) also default to MPa for design pressure once the value is over a few bar.
Geotechnical engineers use both kPa and MPa. Soil bearing capacities and pore pressures are usually in kPa; rock unconfined compressive strengths and concrete cube strengths are in MPa. (A concrete grade C30/37 has 30 MPa cylinder, 37 MPa cube characteristic strength.) Knowing which to expect in which subdiscipline is part of being literate in the trade.
Conversion factors worth memorising
A short list will get you through 95% of conversions without a calculator:
- 1 bar = 100 kPa = 0.1 MPa = 14.504 psi
- 1 MPa = 10 bar = 1000 kPa = 145.04 psi
- 1 psi ≈ 6.895 kPa = 0.06895 bar
- 1 atm = 101.325 kPa = 1.01325 bar = 14.696 psi
- 1 ksi (1000 psi) ≈ 6.895 MPa
The two factors that earn their keep are 14.504 (psi per bar) and 145.04 (psi per MPa). With those two and the trivial 1 bar = 100 kPa = 0.1 MPa, you can hop between any pair of the four units in your head.
Cross-reference table
Common values, all four units, rounded to engineering precision:
| psi | bar | kPa | MPa | Typical context |
|---|---|---|---|---|
| 14.7 | 1.013 | 101.3 | 0.101 | Standard atmospheric |
| 32 | 2.21 | 221 | 0.221 | Passenger-car tire (cold) |
| 90 | 6.21 | 620 | 0.620 | US shop compressed air |
| 150 | 10.34 | 1034 | 1.034 | ANSI Class 150 pipe flange max |
| 3000 | 207 | 20 700 | 20.7 | Aircraft hydraulic system |
| 5000 | 345 | 34 500 | 34.5 | Mobile hydraulic, high end |
| 10 000 | 689 | 68 900 | 68.9 | 700 bar industrial hydraulic |
| 51 488 | 3550 | 355 000 | 355 | S355 steel yield strength |
Gauge versus absolute: a regional vocabulary problem
Pressure can be quoted relative to absolute vacuum (absolute pressure) or relative to local atmospheric pressure (gauge pressure). In US practice the suffixes psia and psigmake the distinction explicit on the unit itself. In SI usage there is no analogous “kPag” or “bara” in the standards — ISO 80000-4 deprecates appending letters to the unit symbol — so the distinction has to be carried in the variable, the prose, or a parenthetical “(gauge)” or “(absolute)”.
In the field you will still see “bar(g)”, “barg”, “kPa(a)”, and “MPag” despite the standards. Treat them as informal but unambiguous. What you must never do is quote a pressure without saying which reference it uses on a P&ID, a vessel nameplate, or an instrument calibration sheet. ASME B40.100 calls this out explicitly for gauge selection.
The default convention also varies by industry. Process plant operators almost always speak in gauge. HVAC refrigerant gauges read gauge. Vacuum and barometric work uses absolute. Aerospace propulsion mixes both, with the convention typically called out in the document’s nomenclature section.
Worked example: re-spec’ing a hydraulic power unit for two markets
Suppose a US OEM has designed a hydraulic power unit (HPU) with the following nameplate:
- Working pressure: 3000 psig
- Pressure relief setting: 3300 psig
- Reservoir pressure: atmospheric (vented)
- Fluid: ISO VG 46 hydraulic oil, operating temperature 50 °C
The same skid is to be sold into the European Union under PED rules and into Japan under the High Pressure Gas Safety Act. Both jurisdictions expect SI units on the nameplate. The arithmetic is straightforward:
- 3000 psi × 0.06895 = 206.84 bar = 20.684 MPa
- 3300 psi × 0.06895 = 227.53 bar = 22.753 MPa
Two design choices follow. First, what unit goes on the nameplate? European hydraulics convention is bar for the working pressure, so “Working pressure 207 bar (20.7 MPa)” is the cleanest read. PED documentation, on the other hand, wants design pressure in bar or MPa with the choice consistent across the technical file. Second, what tolerance do you round to? Rounding 206.84 bar down to 200 bar shaves 7 bar off the rated capacity and might invalidate the system curve; rounding 3300 psig up to 230 bar relief raises the burst-disc inventory cost. The right answer is to keep the rated values at the conversion-exact number and let the gauge faces and relief plate match — typically 210 bar working / 230 bar relief with the conversion-exact figures preserved in the data book.
The same example exposes the gauge-vs-absolute trap. A US engineer reading the European data book will read “210 bar” and assume gauge. An auditor in Tokyo may want “210 bar(g) / 211 bar(a)” spelled out, because the High Pressure Gas Safety Act regulates absolute pressures over 1 MPa. Adding the small atmospheric offset costs nothing in the document and removes one mode of failure during inspection.
Common mistakes that cost time and money
- Mixing units in one document.A vessel rated “design pressure 150 psig” with a relief valve set to “10 bar” is internally inconsistent. 10 bar is 145 psig, very close to design — a coincidence that masks the real risk that someone substitutes a 10 bar relief on a 100 psig vessel and over-pressures it.
- Omitting g/a or the equivalent in prose.“Tank pressure 1.5 bar” in a vacuum drier documentation set could be 1.5 bar above atmospheric (almost full atmospheric over) or 1.5 bar absolute (a partial vacuum). The two are off by 1 bar — a factor of nearly 2 in process behaviour.
- Confusing kPa with kgf/cm² on legacy Asian equipment. 1 kgf/cm² = 98.066 5 kPa, not 100. The 2% drift adds up over a multi-stage compressor train.
- Using ksi where MPa is expected.ksi is 1000 psi, not the same as MPa. A material spec that reads “36 ksi yield” is 248 MPa, not 36 MPa — an order-of-magnitude error if blindly substituted.
- Rounding both ways. Converting 100 bar to 14.5 ksi and back gives 1450 psi = 100 bar exactly only if you carry enough digits. Round each step and a 100 bar spec becomes 99.6 bar after a round trip. For tolerance-tight work, carry conversions to one more significant figure than the source.
- Trusting gauge labels at altitude. A gauge reads zero at local atmospheric. Move it from sea level to a mountain plant and the absolute pressure it reports for a sealed vessel changes by the difference in barometric pressure — about 12 kPa per 1000 m. For low-pressure work, absolute readings are safer.
Picking the unit on a new spec
A workable decision rule, in order:
- Ask which jurisdiction’s code the equipment is built to. ASME US-customary work uses psi; PED and most national metric codes use bar or MPa.
- Ask what the customer’s house standard is. Plant standards almost always pick one of the four and stick with it across every drawing.
- Pick the unit that keeps the numerical value between roughly 1 and 1000 without scientific notation. A 7 bar shop air spec is more readable than 700 kPa or 0.7 MPa; a 350 MPa yield strength is more readable than 350 000 kPa.
- Add the conversion in parentheses if the document will cross a market boundary. “Design pressure 25 bar (363 psig)” costs nothing and saves a phone call.
- State gauge-or-absolute exactly once in the nomenclature, then carry the convention. Don’t leave the reader guessing.
Pressure units are one of the few places in engineering where the right answer is almost entirely a matter of audience rather than physics. Picking the unit your reader uses, sticking with it, and being explicit about the gauge-vs-absolute reference will save you more rework than any conversion calculator.