How to Read Steam Tables: A Mechanical Engineer's Guide

Open any thermodynamics textbook to the back appendix and you will find what is, by volume, the densest information in the book: the steam tables. A few pages of small print encode the entire equation of state for water across the pressures and temperatures every power-cycle, refrigeration, and HVAC engineer will ever need. Reading them well is not difficult — but reading them sloppily is the single most reliable way to throw a Rankine-cycle homework problem off by a factor of ten, or to misjudge a real turbine's exit state by enough to matter.

This guide walks through the two table families you will actually use — the saturation tables and the superheated vapor tables — explains what each column physically means, demonstrates linear interpolation with a worked example, shows how to extract steam quality from a known property, and lists the unit and sign-convention traps that catch students and practicing engineers alike.

Two table families: saturated vs. superheated

Water in equilibrium has a peculiar property: along the saturation curve, pressure and temperature are not independent. If you know one, you know the other. That is why steam tables come in two main flavors, and why understanding the difference is the first thing to internalize.

Saturation tables describe states where liquid and vapor coexist — anywhere under the saturation dome on a T-s or P-v diagram. They come in two presentations of the same physics:

• Saturated water — temperature table. Indexed by temperature (e.g. 10 °C, 20 °C, ...). Useful when temperature is your input.
• Saturated water — pressure table. Indexed by pressure (e.g. 1 kPa, 5 kPa, 10 kPa, ...). Useful when pressure is known, which is more common in pump and condenser sizing.

Both tables list the same kind of columns: the saturation pressure (or temperature), and properties of the saturated liquid (subscript f, for "fluid"), the saturated vapor (subscript g, for "gas"), and the difference between them (subscript fg).

Superheated vapor tables describe states above the saturation temperature for a given pressure — the region to the right of the dome, where steam behaves as a single-phase gas. Here pressure and temperature are independent again, so the tables are organized as a grid: each block is a fixed pressure, and within the block, temperature varies. You will need to interpolate twice (once in T, once in P) to land on a state that is not exactly on the grid.

A third table — compressed (subcooled) liquid — covers states where the liquid is below saturation temperature at its pressure. For most undergraduate work, this region is approximated using the saturated liquid values at the given temperature, with a small pressure-correction term that is typically negligible below ~5 MPa.

What each column means

Every steam table you will encounter shares the same set of intensive properties. Memorize the symbols once and the rest is bookkeeping.

• T — temperature, in °C (SI) or °F (USCS).
• P — pressure, in kPa or MPa (SI) or psia (USCS). Always absolute pressure unless explicitly noted as gauge.
• v_f — specific volume of saturated liquid, m³/kg. Tiny number (around 1.0 × 10⁻³ m³/kg near room temperature).
• v_g — specific volume of saturated vapor, m³/kg. Often three to six orders of magnitude larger than v_f.
• h_f — specific enthalpy of saturated liquid, kJ/kg.
• h_fg — latent heat of vaporization, kJ/kg. The enthalpy you must add to fully evaporate one kilogram of saturated liquid at constant pressure. Equals h_g − h_f.
• h_g — specific enthalpy of saturated vapor, kJ/kg.
• s_f, s_fg, s_g — the corresponding specific entropies, in kJ/(kg·K). Same algebraic relationship: s_fg = s_g − s_f.
• u_f, u_g — specific internal energies, kJ/kg. Some tables list them; others omit u and let you back it out from h, P, and v via u = h − Pv.

Note the reference state: by convention, h and s of saturated liquid water at the triple point (T = 0.01 °C, P = 0.6117 kPa) are set to zero. That is why you will sometimes see slightly negative h values for deeply subcooled liquids in some tabulations — it is just a reference artifact, not a thermodynamic anomaly.

Worked example: linear interpolation

You will rarely land on a tabulated row exactly. The standard remedy is linear interpolation between adjacent rows. The arithmetic is identical to interpolating any two-column dataset, but the discipline is what counts.

Problem. Find the saturation enthalpy of vapor h_g at P = 175 kPa, given the following two adjacent rows from a saturated water — pressure table:

Setup. Linear interpolation says: between two tabulated points (x₁, y₁) and (x₂, y₂), the value of y at an intermediate x is

y = y₁ + (y₂ − y₁) × (x − x₁) / (x₂ − x₁)

Here x is pressure and y is whatever property you want. The fractional position of 175 kPa between 150 and 200 kPa is

f = (175 − 150) / (200 − 150) = 25 / 50 = 0.50

So you are exactly halfway between the rows. Apply f to h_g:

h_g(175 kPa) ≈ 2693.1 + 0.50 × (2706.3 − 2693.1) = 2693.1 + 6.6 ≈ 2699.7 kJ/kg

For a sanity check against an authoritative source, the NIST WebBook gives the saturation enthalpy of vapor at 175 kPa as approximately 2700.2 kJ/kg. The 0.5 kJ/kg discrepancy comes from the gentle nonlinearity of h_g(P) along the saturation curve — linear interpolation slightly underestimates it. For typical engineering problems, that error is negligible. For thermal-design work near a critical operating point, use NIST directly or apply a finer table.

Two-way interpolation in superheated tables works the same way, just done twice. Suppose you need h at 175 kPa and 250 °C. Steps:

1. Interpolate h vs. T at the lower pressure (150 kPa) between the two temperature rows that bracket 250 °C.
2. Interpolate h vs. T at the higher pressure (200 kPa) between the same two temperature rows.
3. Interpolate the two resulting h values vs. pressure to land at 175 kPa.

This is bilinear interpolation. Order does not matter — you can do P first, T second — but you must do all three steps. Skipping the inner T-step and only interpolating P is the most common student error.

Determining quality from a known property

Inside the saturation dome, neither pressure nor temperature alone fixes the state — you also need to know the proportion of vapor to liquid. That proportion is the quality, x, defined as the mass fraction of vapor:

x = m_vapor / (m_vapor + m_liquid), 0 ≤ x ≤ 1

Any specific property y in the two-phase region is the mass-weighted average of its saturated-liquid and saturated-vapor values:

y = (1 − x) y_f + x y_g, equivalently y = y_f + x · y_fg

That second form is the one you will use most. If a turbine exit condition is reported as "h = 2200 kJ/kg at 10 kPa", and the saturated liquid and latent values at 10 kPa are roughly h_f = 191.8 kJ/kg and h_fg = 2392.1 kJ/kg, then

x = (h − h_f) / h_fg = (2200 − 191.8) / 2392.1 ≈ 0.840

That is, 84.0% of the mass leaving the turbine is vapor and 16.0% is liquid droplets — high enough moisture to threaten blade erosion in a real low-pressure turbine, which is why power plants reheat or otherwise keep x above ~0.88 at the exit.

The same trick works on entropy, internal energy, or specific volume. If you know two intensive properties, one of which is sat-pressure or sat-temperature, you can always solve for x — and then anything else.

Units conventions and how to keep them straight

SI units are now standard in nearly all academic thermodynamics instruction worldwide. United States power-industry practice still carries a strong USCS legacy, and many ASME documents tabulate both.

Two conversions you will use constantly: 1 BTU/lbm ≈ 2.326 kJ/kg, and 1 psia ≈ 6.895 kPa. Note that USCS entropy uses Rankine in the denominator, not Fahrenheit — a degree Rankine and a degree Fahrenheit have the same size, but the absolute scale matters. This is the unit equivalent of the trap that catches first-year students.

Common pitfalls

After grading several thousand thermodynamics problems, the same handful of mistakes appear again and again.

• Interpolating across the dome. If your target state straddles the saturation curve — for example, a superheated row at one pressure and a saturated row at the next — linear interpolation is meaningless. Properties have a kink at the dome. Always check whether your two anchor rows are in the same phase region.
• Mixing absolute and gauge pressure. Steam tables tabulate absolute pressure. If your problem statement gives 100 psig (gauge) and you plug it in as 100 psia, you are off by atmospheric pressure (~14.7 psi or ~101 kPa). For low-pressure systems the error is large in percentage terms.
• Forgetting the latent term in two-phase enthalpy.h = h_f alone treats every state as saturated liquid; h = x · h_fg alone treats every state as starting from zero. Always use h = h_f + x · h_fg.
• Using K when °C is expected. Specific entropy uses K in the denominator (since it is a Δ), but the saturation column itself is often °C. A 273.15 unit confusion will make your isentropic processes look insane.
• Trusting a single decimal of interpolation. Linear interpolation is exact only for genuinely linear data. Properties like h_g near the critical point are decidedly not linear. If you are doing high-precision work near 22 MPa or 374 °C, interpolate from a finer table or query NIST directly.
• Using the wrong reference state. Most modern tables (NIST, IAPWS-IF97-derived, Cengel, Moran) reference saturated liquid at the triple point. Some older refrigeration tables for water/steam use 0 °F or different conventions. For Δh and Δs problems the reference cancels — but only if both endpoints come from the same table.

Where the tables come from

The numbers in any modern steam table are not measured directly — they are computed from a fundamental equation of state that has been fitted to a vast body of experimental data over more than a century. The current international standard for industrial use is the IAPWS Industrial Formulation 1997 (IAPWS-IF97), maintained by the International Association for the Properties of Water and Steam. IAPWS-IF97 divides the (P, T) plane into five regions and provides fast, mutually-consistent equations within each region, designed for use in power-plant simulators where evaluating properties thousands of times per second matters.

For research-grade accuracy, IAPWS also publishes the 1995 Scientific Formulation (IAPWS-95), which is slower but more accurate near the critical point. The NIST Chemistry WebBook serves IAPWS-95 free of charge through a web form — useful for sanity-checking tabulated values, generating problem-set entries that do not appear in any printed appendix, or producing custom property tables for a course.

Cengel & Boles, Moran & Shapiro, and the ASME Steam Tables are all derived from these formulations. Numerical differences between editions are typically in the third or fourth significant figure — enough to be visible in a homework answer key but well below the accuracy of any real measurement of an operating power cycle.

Putting it to work

Nine times out of ten, an engineering steam-table problem reduces to the same loop. Identify the region (subcooled, two-phase, superheated) from the inputs you are given. Pick the right table. Locate the two rows that bracket your state. Interpolate cleanly, in the right order. Carry units explicitly — write "kJ/kg" next to every number until you are done. And if you are doing two-phase work, always solve for quality before solving for anything else; quality is the linchpin that connects every other property in the dome.

Once you internalize that loop, steam tables stop being a wall of small print and become what they really are: a remarkably compact, century-tested reference for the most engineered fluid on Earth.