Condensate Flow Meter Selection for Steam Systems

By Wu Peng, Senior Process Instrumentation Engineer · Last reviewed August 16, 2026

A condensate flow meter measures the hot water that steam traps return to the boiler feed tank. The measurement is harder than the fluid suggests. Condensate leaving a trap at 10 barg into an atmospheric return line flashes: 16.1 percent of the mass turns back into steam, and that steam then occupies 99.7 percent of the pipe volume. A meter that reads volume in that line is reading the vapor, not the water.

Two things follow from that number. Where you put the meter matters more than which meter you buy. The technology then has to be picked against three condensate properties at once: temperature, phase and electrical conductivity.

This page gives the flash arithmetic and a technology table with real temperature and conductivity limits. It then covers sizing with a worked velocity check, installation rules, and the energy accounting that justifies the meter. One note on the word first: in oil and gas, condensate means light hydrocarbon liquid recovered from wet gas. Everything below is steam condensate, the returning water of a steam loop.

Contents

• How condensate differs
• What flashing does
• Gravity or pumped
• Technology comparison
• Sizing and velocity
• Installation rules
• Energy accounting
• FAQ

How condensate differs

Condensate is water, and that is exactly what makes it deceptive on a datasheet. Three properties separate it from the cold water a meter is usually specified for, and each one removes a different set of instruments from the shortlist.

Property	What it is in a return line	What it removes from the shortlist
Temperature	At or just below saturation for the return line pressure: 100.0 °C at 0 barg, 120.2 °C at 1 barg, 133.5 °C at 2 barg	Rubber and polyurethane magnetic meter liners (about 80 °C), standard clamp-on transducers (90 °C), standard oval gear builds (80 °C)
Phase	Two-phase downstream of any trap unless the condensate is sub-cooled or the line is held above saturation pressure	Every volumetric technology, at any accuracy class, until the flash is dealt with
Conductivity	Clean condensate is commonly under 10 µS/cm; clean-steam condensate held to the pharmacopoeia purified-water limit is under 2.1 µS/cm	Magnetic flow meters, which need at least 5 µS/cm to produce a signal

Saturation temperatures from published steam tables (IAPWS-IF97). Liner, transducer and conductivity limits are the ratings of the instruments listed further down this page.

The conductivity row is the one that catches people out, because it contradicts what most vendor pages say. A magnetic meter is the obvious choice for a hot, clean, unobstructed liquid line. It is recommended for condensate across the web without any mention of a conductivity floor.

On a well-run steam loop the returning water can sit below that floor, and the meter shows noise instead of flow. Blended make-up water raises conductivity; a polished clean-steam system does not.

What flashing does

When condensate drops from trap pressure to return line pressure, part of it boils. The mass fraction that flashes comes straight from an enthalpy balance:

x = (h_f1 − h_f2) / h_fg2

Trap inlet at 10 barg, return line at atmospheric pressure: (781.3 − 419.1) / 2256.5 = 0.161, or 16.1 percent by mass

Sixteen percent sounds survivable. The volume tells the real story.

At atmospheric pressure saturated steam occupies 1.673 m³/kg and saturated water 0.001044 m³/kg. The flash from one kilogram of condensate therefore takes 0.269 m³, while the remaining water takes 0.000876 m³. That is 307 volumes of steam for every volume of water, and it means the vapor holds 99.7 percent of the pipe.

Trap inlet	Saturation	Flash by mass	Vapor by volume	Steam:water
1 barg	120.2 °C	3.8 percent	98.4 percent	63 to 1
2 barg	133.5 °C	6.3 percent	99.1 percent	108 to 1
3 barg	143.6 °C	8.2 percent	99.3 percent	144 to 1
5 barg	158.8 °C	11.1 percent	99.5 percent	201 to 1
7 barg	170.4 °C	13.4 percent	99.6 percent	248 to 1
10 barg	184.1 °C	16.1 percent	99.7 percent	307 to 1

Trap discharging to an atmospheric return line. Computed from IAPWS-IF97 saturation enthalpies and specific volumes.

The table also explains a reading that looks like an instrument fault and is not. Even half a percent of flash by mass fills 89.0 percent of the pipe with vapor and multiplies the volume passing the meter by 9.0.

A turbine spins far above its rated range. A vortex meter reads a mixture whose density is not known. A totalizer built from that volume runs high all shift. The instrument is working; the line is not liquid.

There are three ways to give a meter a liquid to measure. Sub-cool the condensate below the saturation temperature of the return line. Hold back pressure on the line so that saturation sits above the condensate temperature. Or send the discharge through a flash vessel and meter the liquid leg leaving it.

The first two are the sizing constraint below.

Return line pressure	Condensate must stay below
0 barg (vented receiver)	100.0 °C
1 barg	120.2 °C
2 barg	133.5 °C
3 barg	143.6 °C

Gravity or pumped

A steam plant has two condensate lines with almost nothing in common, and vendor pages that talk about condensate metering rarely say which one they mean.

Section	What the flow looks like	Metering verdict
Trap discharge to receiver gravity, roughly 0 to 2 barg	Intermittent slugs of water in a pipe that is mostly flash steam, with long idle periods between trap cycles	Do not meter here for accounting. If a reading is needed per trap, meter energy or condition, not volume
Receiver to boiler feed pumped, roughly 3 to 10 barg	Full pipe, single phase, sub-cooled by the receiver vent, steady while the pump runs	This is the metering point. Every technology below is usable here if temperature and conductivity allow

One detail decides whether the pumped side behaves: what kind of pump. An electric transfer pump gives a steady flow while it runs, so the meter sees a clean rate. A mechanical pressure-powered pump discharges in slugs, and a rate reading between slugs means nothing.

Totalize in that case and report volume per shift rather than instantaneous flow. The difference between a rate and a total is worth being precise about, and it is covered in the guide on totalizers against flow meters.

Technology comparison

The table below lists the same technologies every vendor does. It adds the two columns they leave out: the temperature the instrument tolerates, and what disqualifies it on condensate.

Technology	Accuracy	Turndown	Medium temperature	Condensate limit
Magnetic	±0.5% of reading	Wide	Set by liner: 80 °C rubber, 150 °C PFA, 180 °C PTFE	Needs at least 5 µS/cm. Clean condensate often sits below that
Clamp-on ultrasonic	±1% of reading	Wide	90 °C standard transducer, 160 °C high-temperature transducer	Signal fails on entrained vapor. A 100 °C line already needs the high-temperature set
Vortex	±1.0% of rate	20 to 1	Up to 400 °C by build	Liquid velocity floor of 0.15 m/s, and 40D upstream of straight pipe
Turbine	±0.5% of reading	10 to 1	120 °C standard, 180 °C high-temperature build	Bearings and a strainer are consumables in hot condensate
Oval gear	±0.5% of reading	10 to 1	80 °C normal, 150 to 200 °C high-temperature build	Needs no straight run, but needs a strainer and clean water
Averaging pitot	About 1% of rate	10 to 1	Up to 450 °C	Lowest permanent pressure loss of the inline options

Accuracy, turndown and temperature figures are the published ratings of the instruments linked in this section.

An orifice plate belongs on the list for completeness and is usually the wrong answer on a return line. Flow follows the square root of the differential, which holds turndown to about 4 to 1 or 5 to 1. The permanent pressure loss is also the largest of any option here. Where the pumped line already has a differential pressure instrument on the boiler feed, an orifice plate makes sense; where it does not, it adds a loss the transfer pump has to pay for.

That points at a question engineers have been asking for years and rarely get answered. How do you meter condensate without imposing back pressure, when back pressure is what stalls the trap?

There are three answers. A clamp-on ultrasonic meter puts nothing in the pipe at all. That makes it the only option with zero permanent loss, and the natural pick for a retrofit on a line that cannot be cut. A full-bore magnetic flow meter has an unobstructed bore and loses no more than the same length of pipe, provided conductivity clears 5 µS/cm.

An averaging pitot tube takes a small loss and inserts through one boss, which suits a large line that cannot be cut. A turbine meter or an oval gear meter is accurate and cheap on the pumped side, where the pump supplies the head to pay for the loss.

On the supply side of the same loop the choice is different again, because steam is compressible and needs density compensation. That is covered on the steam flow meter page and, for the vortex option specifically, the steam vortex flow meter.

Sizing and velocity

Condensate return lines are sized for two-phase flow, so they are almost always far larger than the liquid alone needs. Drop an inline meter into that pipe without checking velocity and it will spend its life under its low-flow cutoff. Aim for 1 to 3 m/s at normal flow in the metering section, and check the minimum against the meter, not against the pipe.

v = (m / ρ) / (3600 A)

1,500 kg/h of condensate at 95 °C (density 961.6 kg/m³) in a 4 inch line, bore about 100 mm, area 0.00785 m²: (1500 / 961.6) / (3600 × 0.00785) = 0.055 m/s

That is 5.5 percent of the 1 m/s bottom of that band.

Running the same arithmetic backwards, 1 m/s at this flow calls for a bore of 23.5 mm, which is a DN25 metering section in a 4 inch line. The vortex option is ruled out before a model is chosen. Its 0.15 m/s liquid floor in a 4 inch line corresponds to 4,078 kg/h. That is 2.7 times the maximum flow the plant will ever produce.

A reduced metering spool is the way through, as is a technology that is insensitive to velocity profile. Either way the reducer belongs in the installation cost, not in a site surprise.

Installation rules

Most condensate metering complaints trace back to the pipe rather than the instrument. Six rules cover the ones worth writing into a specification.

• Keep the meter full. Put it in a rising leg, or below the receiver so static head keeps the bore flooded. A meter in a falling leg runs part empty and reads whatever velocity the water film has.
• Meter downstream of the pump. The pumped section is single phase; the trap discharge is not. Where the plant only has a gravity line, add back pressure or sub-cool before metering.
• Give it the straight run it needs. The numbers vary by more than most people expect, from none for an oval gear meter to 40 diameters upstream for a vortex meter. The per-type figures are collected in the guide on straight run requirements.
• Strain ahead of anything with moving parts. Scale, valve packing and rust travel in return lines and break small turbine rotors quickly.
• Read the reverse total as well as the forward total. The check valves downstream of a receiver leak often enough that backflow through the meter is common. If only the forward total is logged, the same water is counted twice.
• Compensate volumetric readings for temperature. Water at 95 °C is about 4 percent less dense than at 20 °C, so a volume converted with a cold density understates mass by the same margin. The mechanism is set out under pressure and temperature compensation.

Calibration deserves one line of its own. A meter certified on cold water at 20 °C has not been shown to work at 95 °C, and the difference shows up as a bias rather than noise. Where the reading is used for billing or for an energy report, ask for calibration at process conditions and keep the certificate. The wider framework is in the guide on flow meter calibration.

Energy accounting

The usual argument for metering the return side is that a kilogram of steam becomes a kilogram of condensate, so a cheap water meter can stand in for an expensive steam meter. The mass balance is real, and it does not survive the trip back. Three losses sit between the two numbers.

The first is the flash vented at the receiver, which the table above sizes at 3.8 to 16.1 percent depending on trap pressure. The second is steam blowing through failed traps, which arrives as steam and leaves through the vent. The third is condensate that never returns at all, lost to drains, sampling and process contamination.

Any one of them puts a systematic gap between steam sold and condensate counted, and none of them is small enough to ignore in a billing application.

What the return line does justify on its own is the energy recovery figure. Saturated condensate at 100 °C carries 419.1 kJ/kg, which is 15.1 percent of the 2,779.7 kJ/kg in the 10 barg steam that produced it. Measured against 20 °C make-up water at 83.9 kJ/kg, every kilogram returned instead of dumped saves 335.2 kJ, or 12.1 percent of the energy in the steam.

In the return line of the worked example above, 1,500 kg/h at 95 °C against 20 °C make-up is 314.1 kJ/kg, or 130.9 kW of recovered heat flowing past the meter.

Turning that into a billable energy number needs a temperature measurement as well as a flow measurement. It is the same arrangement used for chilled and hot water energy metering. The arithmetic, the paired-sensor requirement and the accuracy consequences are set out in the guide on energy metering with a BTU meter. If the aim is only to compare technologies for the return line, the overview of flow meter types covers the ones this page has held to condensate service.

FAQ

How to measure the flow of steam?

Steam is measured on the supply side with a vortex, orifice or averaging pitot meter plus pressure and temperature compensation, because steam density changes with both. Measuring the condensate instead is cheaper and uses ordinary liquid instruments. The two numbers are not interchangeable: flash losses, failed traps and condensate that never returns sit between them.

Where should a flow meter be installed?

On the pumped section between the receiver and the boiler feed, in a rising leg or below the receiver so the bore stays full, with the straight run the chosen technology needs. Avoid the gravity line between trap and receiver, where the flow is intermittent slugs in a pipe that is mostly flash steam.

What are the three types of flow meters?

Meters are grouped by what they sense: velocity meters such as turbine, vortex and ultrasonic; volume meters such as oval gear; and mass meters such as Coriolis. Differential pressure devices infer velocity from a pressure drop. On condensate the grouping matters because only mass and volume devices are unaffected by the velocity profile a short pipe run creates.

How does a water flow meter work?

A water meter converts a physical effect into a flow rate. That effect is rotor speed in a turbine, induced voltage in a magnetic meter, transit-time difference in an ultrasonic meter, or a pressure drop across a restriction. Condensate is water, so all of these work in principle. Temperature, conductivity and entrained flash steam decide which of them work in practice.

Request a quote

Tell us the return line pressure, the normal and maximum condensate temperature, the flow range and the line size, and we size the metering section rather than assuming the existing pipe. Reach our application engineers or use the form below.

Written and technically reviewed by Wu Peng and the Instranova engineering team.