Flange Type Thermal Gas Mass Flow Meter

A thermal gas mass flow meter reads mass directly. There is no pressure transmitter, no temperature transmitter, and no density calculation bolted on afterwards. Two resistance sensors sit in the stream, one heated and one reference, and the heat carried away by the gas is proportional to the mass flowing past. That single fact removes most of the error sources that plague volumetric meters on gas.

We build this one as a flange type body with the sensor head inserted through the wall, which is why it fits ducts and large pipes without a full-bore spool piece. Sizes run from DN15 to DN2000. The same head works on compressed air at 8 bar and on flue gas at 400 C, and it needs 1 to 2 diameters of straight run rather than the 15 that a vortex meter wants.

What the meter actually measures

Mass flow, in kilograms per hour or normal cubic metres per hour. Not volume at line conditions. This distinction matters because gas is compressible, and a volumetric meter on a compressed air line reports a number that changes when the compressor setpoint changes even though the same mass of air is moving.

Energy managers care about mass because leaks are priced in mass. A 3 mm hole in a 7 bar air line passes roughly 1.2 kW of compressor power, and that cost does not depend on whether the header is at 6.5 or 7.5 bar. A thermal gas mass flow meter reports that loss directly. A volumetric meter reports a number you have to correct before you can argue about it.

Emissions work is the second big use. Flue gas mass flow combined with an oxygen reading gives you the mass of a pollutant per hour, which is what a permit is written against. Volumetric flow plus a stack temperature and pressure gives you the same thing only after three corrections that each carry their own error.

How the sensing element works

Two platinum resistance elements are wound on the same probe. One is heated above gas temperature. The other measures the gas temperature and acts as the reference. The controller holds a constant temperature difference between them, and the power needed to hold that difference is the measurement signal.

More mass flowing past the heated element carries more heat away, so more power is needed. The relationship between power and mass velocity is non-linear but very repeatable, and it is linearised in firmware with a twelve-point correction curve. That curve is what you are really buying when you buy a good meter.

Constant temperature difference, not constant current

Older designs held a constant current through the heater and measured the temperature rise. They drifted with gas composition and ambient temperature. The constant temperature difference approach is self-compensating for ambient swings, which is why a modern thermal gas mass flow meter holds its zero through a winter night on an outdoor skid.

The trade-off is response time. Holding a constant difference means the control loop has to work, and a fast loop can be noisy. Ours settles in about one second, which is fast enough for leak surveys and slow enough to average out compressor cycling.

Specifications

Parameter	Range
Nominal diameter	DN15 to DN2000 (insertion type above DN300)
Measured quantity	Mass flow, standard volume flow, total
Accuracy	plus or minus 1.0 percent of reading, plus or minus 1.5 percent at the low end
Repeatability	0.2 percent of reading
Turndown	1:100 standard, 1:1000 on request
Velocity range	0.1 to 120 Nm/s
Medium temperature	-40 C to +200 C standard, up to +510 C with the high temperature probe
Nominal pressure	1.6 MPa standard, 4.0 MPa optional
Straight run	1 to 2 DN upstream, 1 DN downstream
Output	4-20 mA, RS485 Modbus, HART optional
Power supply	24 VDC or 220 VAC
Protection	IP65 standard, IP67 optional
Response time	about 1 second

The turndown figure is the one that changes projects. A 1:100 turndown on a header that swings from 2 to 150 Nm/s means one meter covers the whole operating envelope, where an orifice plate would need two runs and a bypass. On compressed air systems with strong shift patterns, that alone often pays for the difference in price.

Why no compensation is needed

The sensor measures the heat capacity flow, and heat capacity per unit mass is a property of the gas. So the output is already mass. Change the pressure and the mass reading does not move. Change the temperature and the reading does not move, as long as the gas stays within the calibrated range.

Compare that with a volumetric meter. On a compressed air header at 7 bar, a 10 C temperature rise changes the density by about 3 percent, so an uncompensated volumetric meter reads 3 percent wrong. Someone has to notice, correct, and document it. The thermal gas mass flow meter just reads the mass.

There is one caveat worth stating plainly. The calibration is done on a specific gas, usually air. If you use thermal gas mass flow meter on a gas with a very different specific heat, such as hydrogen or helium, the reading shifts and you need a correction factor. Natural gas and flue gas are close enough to air that the factory factor is within a percent or two, and we supply the factor for anything further away.

Accuracy in a real duct

The published accuracy assumes a stable velocity profile at the sensor. In a real installation the profile is the dominant error source, and it is worth understanding before you pick a location.

Velocity profile

An insertion probe measures at one point, or at a few points if you traverse it. In a fully developed turbulent profile the centreline velocity runs about 1.2 times the average. thermal gas mass flow meter applies a profile factor to convert point velocity to average velocity, and that factor is only right if the profile is the one it assumes.

Two diameters downstream of a long radius elbow, the profile is skewed and the factor is wrong by several percent. That is why the 1 to 2 diameter straight run claim holds only for the flange type design with its flow conditioning, and why we still ask for a drawing before signing off on an insertion installation in a short duct.

Low flow and zero drift

At very low velocity the heat loss by convection gets small compared with conduction along the probe, and the signal flattens. thermal gas mass flow meter still reports a number, but that number is mostly the zero. A thermal gas mass flow meter with a 0.1 Nm/s floor is honest about where the useful range ends.

Zero drift is the maintenance item. Clean the probe annually on dirty gas and every two years on clean gas. A coated probe insulates the sensor and reads low, and the error grows slowly enough that nobody notices until a comparison audit.

Insertion or inline

Above DN300 the insertion version is almost always the right call. The cost difference is large, the pressure loss is negligible, and thermal gas mass flow meter can be removed through a ball valve without shutting the line.

When insertion wins

Large ducts, stacks, and headers where the pipe cost dominates. Retrofits where cutting a spool piece means a shutdown. Anywhere the line cannot be stopped for maintenance, because the insertion probe can be pulled and refitted hot.

When inline wins

Small lines where the probe would occupy too much of the cross section and disturb the very profile it is measuring. High accuracy custody points where a full-bore calibration is worth the money. Lines under DN80, where an insertion probe is simply too intrusive.

The practical break point is DN300. Below it, inline. Above it, insertion, unless the accuracy requirement is tighter than one percent.

Sizing and installation

Size on the maximum velocity, not the maximum flow. A thermal gas mass flow meter cares about velocity because the sensor measures heat transfer, and heat transfer depends on velocity rather than on total flow.

• Keep the maximum velocity under 120 Nm/s. Above that the probe starts to vibrate and the reading becomes noisy.
• Keep the minimum above 0.1 Nm/s. Below that the signal is mostly zero drift.
• Fit the probe at least 1 to 2 diameters downstream of any disturbance and 1 diameter upstream of the next one.
• Insert the probe to the correct depth. Mark the insertion depth on the probe before you start, because a probe set at the wrong depth reads the wrong part of the profile.
• Align the sensor port with the flow direction. The arrow is on the head, not on the body.
• On condensing or wet gas, install the probe at a slight upward angle so liquid drains away from the sensor.
• Leave room to withdraw the probe. A probe that cannot come out is a probe that never gets cleaned.

On compressed air, fit thermal gas mass flow meter downstream of the dryer and filter. Saturated air carries oil aerosol that coats the sensor within months. If thermal gas mass flow meter has to go upstream of the dryer, plan on cleaning it twice a year.

Two more rules apply to every thermal gas mass flow meter installation. Size the probe for the duct rather than for the flange you happen to have. A thermal gas mass flow meter in a DN500 duct needs a probe long enough to reach the centreline, because a short probe reads the boundary layer and reports a number several percent low. Then check the turndown against the quietest hour rather than the design minimum. A thermal gas mass flow meter sitting at 0.05 Nm/s overnight is reporting zero drift, not flow.

Compressed air: the biggest application

Compressed air is the most common use of a thermal gas mass flow meter, and for good reason. Air is clean after the dryer, it is expensive, and leaks are invisible. A plant that spends 200,000 a year on compressor power typically loses 20 to 30 percent of it to leaks, and the only way to find them is to measure mass flow at each branch.

The pattern we see on nearly every survey: a baseline flow at night when production is stopped. If that baseline is 25 percent of daytime flow, a quarter of the compressor bill is going through holes in the pipework. Fix the biggest leaks, watch the baseline fall, and repeat quarterly.

Permanent meters on the main header and on the two or three largest branches give you that baseline continuously. Portable insertion meters give you the detail, branch by branch, in a two-day survey. Most plants end up with both.

Flue gas and high temperature

Stack flow is harder than air because the gas is hot, dirty, and often wet. The high temperature probe handles the heat. The dirt is the real problem, because fly ash deposits on the sensor and insulates it.

For clean combustion gas from a gas fired boiler, the standard probe is fine. For coal or biomass, specify the high temperature version with a longer probe and plan on a retraction mechanism so the sensor can be cleaned without a shutdown. Soot blowers help, but they do not reach every probe location.

Measure at a point where the flow is stable. Stack flow is notoriously non-uniform, and a single point probe in a large stack can be several percent off. If the number is going into a permit report, use a multipoint or a traverse and average.

Compared with vortex and orifice

	Thermal mass	Vortex	Orifice
Direct mass	Yes	No, needs compensation	No, needs compensation
Turndown	1:100	1:10 to 1:15	1:3 to 1:5
Straight run	1 to 2 DN	15 DN up, 5 DN down	10 to 20 DN up
Pressure loss	Negligible	Moderate	High, permanent
Dirty gas	Coating risk	Good	Good
High temperature	To 510 C	To 400 C	To 540 C
Large duct	Insertion, low cost	Limited	Limited

Against the vortex flow meter, the thermal gas mass flow meter wins on turndown and straight run and loses on very dirty gas and on steam, which vortex handles well and thermal does not handle at all. Against a balanced orifice flow meter, it wins on pressure loss and turndown and loses on initial cost at large sizes and on high pressure applications where the orifice plate is simply more robust.

For steam, neither thermal option applies. Go to vortex or orifice. Our flow meter selection guide covers that decision tree if you are starting from the fluid rather than from thermal gas mass flow meter.

Five mistakes we see every month

Installing downstream of a lubricated compressor without a coalescing filter. The sensor coats in weeks and reads low by five percent before anyone notices.

Sizing on normal cubic metres per hour without converting to velocity. A DN200 meter selected on 2,000 Nm3/h looks fine until you realise the line runs at 8 bar and the velocity is 30 Nm/s, which is fine, or at 2 bar and the velocity is 120 Nm/s, which is not.

Putting the probe in a vertical downflow with wet gas. Condensate runs onto the sensor and the reading goes wild.

Ignoring the insertion depth. A probe set 20 mm shallow reads the boundary layer instead of the core, and the error is consistent enough to look like a calibration problem.

Expecting thermal gas mass flow meter to read steam. It will not. Thermal dispersion needs a gas with a stable specific heat and no phase change.

Keep the calibration certificate with thermal gas mass flow meter, not in the project folder. A thermal gas mass flow meter is calibrated on one gas and one probe orientation, and a thermal gas mass flow meter rebuilt with a different probe length is a different instrument. When a thermal gas mass flow meter comes back for service, send the certificate with it and ask for the new curve on the gas you actually run.

Ordering checklist

• Gas type and composition, especially if hydrogen or CO2 is present
• Line size, flange standard and pressure class
• Maximum, normal and minimum flow, and thermal gas mass flow meters you work in
• Line pressure and temperature range
• Insertion or inline, and whether hot tap removal is needed
• Output and protocol
• Power available on site
• Hazardous area classification

Running a leak survey with the meter

A portable thermal gas mass flow meter turns a leak survey from guesswork into arithmetic. The method is simple and it works on any compressed air system, regardless of size or age.

Start with the plant stopped. No production, no tools running, no purge flows. Read the main header. That number is the leak load. Divide it by the daytime average and you have the leak fraction as a percentage of compressor output.

Then work outward. Take a reading on each branch at the same moment, so the numbers add up. The branch with the highest standing flow has the most leak, and it is rarely the branch anyone expected. Old pipework in a corner nobody visits usually wins.

Fix the top five leaks and repeat the night reading. A plant that starts at 28 percent typically gets to 12 percent in one round of work, which on a 200,000 power bill is worth doing before the next budget cycle. A thermal gas mass flow meter installed permanently on the main header keeps that number visible, and a visible number is the only one that stays fixed.

One caution. Ultrasonic leak detectors find leaks by sound, and they miss leaks inside walls, behind cladding and under insulation. thermal gas mass flow meter finds the total but not the location. Use both: thermal gas mass flow meter tells you whether the work paid, the detector tells you where to work.

Questions from the shop floor

Does it need a straight run like an orifice plate?

No. One to two diameters upstream and one downstream is enough, because the sensor measures at a single point and the flange body includes flow conditioning. An orifice plate on the same line would want ten to twenty diameters.

Can it measure biogas or digester gas?

Yes, with two caveats. The gas is wet, so the probe needs a slope and a drain. And the composition varies, so the specific heat correction drifts. On digester gas, expect two to three percent uncertainty rather than one.

What happens if the gas is dirty?

The sensor coats and reads low. The reading does not fail, it drifts, which is worse. Clean it on a schedule rather than waiting for a discrepancy.

Can it be installed with a hot tap?

Yes, with a tapping valve and a retraction mechanism. That is the main reason insertion meters are popular on headers that cannot be shut down.

Is the flow reading affected by pressure?

No, that is the point. The output is mass. Pressure affects density, and density is already accounted for by the measurement principle. This is why thermal gas mass flow meter needs no pressure transmitter.

Related reading

If you are metering fuel gas rather than air, the gas turbine flow meter is the custody transfer option and the gas Roots flow meter is the low pressure option. For non-invasive measurement on existing pipe, look at the clamp on ultrasonic flow meter.

Reference for compressed air system measurement practice: the U.S. Department of Energy compressed air systems programme publishes the leak rate tables most surveys are based on.

Related reading

• BTU Meter Selection for District Heating Substations and Chilled Water Systems