Natural Gas Flow Measurement

Put a pressure gauge on it and adjust the density accordingly? I thought heating value was per weight rather than per volume.

When I used to work on coke ovens, we used to install something called, iirc, a Wobbe meter. I think it measures the heating value of the gas. It may be rather expensive for your operation (since I don't know how big your operation is), but perhaps it could be useful.

Here you are. Learn about the Wobbe Index.

Pressure and temperature are already compensated for in the Multivariable transmitter. It is the gas density that changes over time that we are unable to account for right now.

Other than pressure and temperature, what else would change the density nearly as much? Is the composition (e.g., what fractions of various hydrocarbons) really all over the map? Or lots of water vapor or other contamination?

Density varies as much as 20% with the change in gas constituency. Gas composition varies from 95% methane and negligible ethane/propane/pentane/butane to 74% methane and 15% ethane/4% propane. When trying to run at 1.2% excess O2 in flue gas, these swings can really mess with combustion control.

Bot, that is a wide variation. Is this typical of gas supplier QC?

Cosa supplies a variety of calorimeters to determine the BTU content

http://www.cosa-instrument.com/products/calorimeters.html

First, that Rosemount of yours can be integrated into a HART loop to make program changes more accessible. Emerson has those guys now if you want to read up on it or look at some of their other products.

Second, there are other flow measurement options to consider. Here's one from Sick that is primarily for custody transfer but it's a good place for you to start. Hang on though, they're proud of their stuff. From what you've said, you'll probably have to go ultrasonic to get the composition information you're looking for. Good luck!

Thanx for the comments so far guys. We have a direct reading of HV telemetered to our site via the gas supplier, so we have the information on HV (not absolute constituency), just not a clean way to get it into the gas table built into the Rosemount transmitters. We will entertain changing the gas meters on all of our boilers and dryers to Coriolis meters, but in the interim, I was hoping for a technically feasible and safe method of utilizing existing transmitters until we can arrange delivery of new metering and outage time to install (24/7/350 operation). We want to avoid adding a function within the DCS as the burner control programming is "locked down" by the combustion train provider and they are reluctant to stray from their template.

You have not told us what air/gas ratio equipment you have, but this sounds like it's crying out for Oxygen trim ratio control.

Regards,

Gasman

O2 trim is implemented on the boilers as well as CO feedback. As mentioned, we target 1.2% O2 in the flue, very near stoichiometric. If we see this swing in heat value, it can occur in a matter of 10 seconds sending O2 down to 0.8% and CO climbing. This also causes HP steam header to spike suddenly (more energy in). We can live with the process upsets and are working with the fuel supplier to iron out these swings, but we can't deal with the offsets and errors in our metering as people are making decisions based on bad information.

Excellent. The Xentaur option appears a bit clumsy, but the Elster option shows definite promise if we can convince the supplier to install and maintain to give us the 10 seconds of resonance time required for realtime display. Thanx much Gasman.

You can hook up a gas chromatograph to the flow computer. The GC can analyze the gas and dump components, gravity and heating value into the flow computer. Most gas measurement service companies can help you set this up. Emerson will have contacts for contractors who can do this.

Sounds like you need data right away to adjust plant processes. If not, another less expensive alternative is to procure periodic samples of the gas and have them analyzed. Manually adjust the constituent mol percents in the flow computer as you get the results back from the analysis. You can develop more accurate average readings of these mol percents with an automatic sampling system which accumulates small volumes of a sample into the same cylinder based on time or volume.

Keith Fry

Senior Product Specialist

Quorum Business Solutions