Problem with venturi flow meter 1

[tgmcg]

thread698-332862

I am analyzing the performance of a centrifugal gas compressor equipped with a venturi flow element in the compressor discharge line. Gas MW = 11 (H2 + CO2). Flowing pressure = 650 psig. Flowing T = 340ºF. The venturi is about 30 years old. The compressor is connected in series with another compressor also equipped with venturi flowmeter. The flow through the second compressor is about 140 mmscfd, which places the operating point right on the compressor curve. However, the first compressor is only indicating 110 mmscfd which places the operating point well below the curve. The second compressor receives all its flow from the first compressor and its recycle valve is fully closed. Simple mass conservation suggests theres something wrong with the flow reading for the first compressor. I've connected a freshly calibrated FT in parallel with the plant FT to measure raw DP across the venturis on both compressors. My FT's are seeing close to the same DP as the dual redundant plant FT's. The beta ratio of the troublesome FE is 0.70, with a Reynolds number (Rd) = 6.7e6. The DP should be approx 62 inH2O, but is only reading about 35 inH2O.

I'm looking for explanations/experiences as to how a venturi flowmeter might be reading significantly low and any suggestions as to how to inspect for defects without shutting down the plant.

The flow reading is very erratic, with the DP constantly and randomly varying +/-5%. The connected process feed should be fairly constant but the fluctuation persists even with the recycle valve stationary in full manual control.

Flow separation??? Vortex shedding??

 

[danw2]

Not likely a DP measurement issue with two transmitters showing the same value, but I hafta ask . .

Is the hard piped impulse tubing or manifold valve's equalizing valve leaking?

If both DP transmitters use a different equalizing valve (or none), then rule it out.

 

[tgmcg]

Each of the three FT's is equipped with its own 3-valve manifold.

I too thought that maybe one of the FT's might be leaking and pulling down the DP for the other FTs. But that would take a large leak, unless one of the root valves at the FE is barely open.

 

[Engdoitbetter]

Hi tcmcg,

It seems that your Reynolds is much greater than the maximum allowed by current standards... Was the Venturi calibrated for the service?

Could you also please advise regarding ID and type of construction (welded, machined ...)?

Regards,

Stefano

 

[roydm]

Is it possible you have an error in the reading device e.g. doing the square root extraction in the transmitter and also in the reciever, that will give you an error of about 10%

 

[tgmcg]

rodym,

I'm measuring raw DP and then calculating the flow. There's no square root extractor.

 

[tgmcg]

Engdoibetter,

It's a 12" Fluidic Techniques HHR-F venturi tube, with beta=0.70.

 

[Engdoitbetter]

I suppose that Venturi was calibrated for this particular service, because Reynolds value is beyond ISO 5167 standard. In addition, beta ratio is close to the usual upper limit (0.7-0.75).

As far as the primary element, i.e. the tube, is concerned, I think there are two reasons for this low reading:

- an increased discharge coefficient, i.e. the coefficient is not the same as the one established by calibration
- particular flow patterns near or inside the pressure taps.

I would rule out the first one since the coefficient should increase very much beyond usual values.
With regard to the second, how many pressure taps are available on the device? Is there any piezometric ring?

Regards,

Stefano

 

[hacksaw]

running the meter calculation with 21.5% CO2 /78.5% H2 for a MW=11.04 at your conditions
gives a flow of 122 mmSCFD / 32 Deg F std conditions or 129 mmSCFD w/60 Deg F std conditions for the 35" W.C. reading.

The erratic d/p readings are a concern, but could make the peak flow closer to 132 mmSCFD than the average reading. What sort of upstream piping layout do you have?

 

[hacksaw]

in dealing with the flow changes, you might check to see where your surge line is located, your compressor mfg can usually advise if you don't have the details

 

[ccfowler]

Venturi meters are usually quite trouble free, so I would pay close attention to upstream and downstream piping configuration details. It is not unusual for space constraints to result in dreadful compromises to metering accuracy and stability.

Any possibility of accumulation of debris, burrs or partial clogging of sensing ports,...?

This may seem like a really silly question, but has the Venturi meter been inspected internally recently including verifying meter dimensions? I've encountered some very strange errors including stamped dimensions on meters being seriously wrong.

Valuable advice from a professor many years ago: First, design for graceful failure. Everything we build will eventually fail, so we must strive to avoid injuries or secondary damage when that failure occurs. Only then can practicality and economics be properly considered.

 

[hacksaw]

no doubt about venturi's being good meters

 

[ccfowler]

hacksaw, I agree!! I would glady substitute a decent Venturi meter for any other type under nearly any circumstance. If you can't get both great accuracy and great turndown performance from a well installed and instrumented Venturi meter, you are in big trouble!

Valuable advice from a professor many years ago: First, design for graceful failure. Everything we build will eventually fail, so we must strive to avoid injuries or secondary damage when that failure occurs. Only then can practicality and economics be properly considered.

 

[hacksaw]

flow meter engineering, an art, not just a science

 

[ccfowler]

hacksaw, that is possibly the truest statement ever made!

Valuable advice from a professor many years ago: First, design for graceful failure. Everything we build will eventually fail, so we must strive to avoid injuries or secondary damage when that failure occurs. Only then can practicality and economics be properly considered.