Pulsating flow - Measurement, crosscheck and uncertainty 1

[hollerg]

The plant I support has two existing speed controlled simplex diaphragm pump installations (manual stroke adjustment) w/ Coriolis mass meter configured so Flow adjusts stroke speed. Flow has a pulsing frequency of ~ 1sec. Because these are simplex pumps, pulsing affects flow more than 5%, in spite of the pulsation dampener. The second flow is ratio controlled from the first flow.

I would like to take advantage of the constant volume characteristic of the pump to compare the flow generated by the pump to the flow measured by the cmm, as a plausibility check of the health of the cmm and pump. In a sense an on-line prover.

This seems straight forward: stroke speed*volume*cmm density where volume is the calibrated volume @ a stroke setting. However the computed value is in effect a smoothed value while the meter is instantaneous. I imagine I need an RMS value from the Micro Motion.

First Any advice on how to extract or generate the value?

Second, how would one compute the uncertainty of the computed value and the flow meter RMS value?

I am trying to determine whether ratio control with the computed values or the RMS values will reduce downstream variability induced by the feed control & pump system. (Best would be to replace the two pumps with a duplex or multiplex proportioning pump so that the flows are in sync and the mass ratio controlled by stroke length, but I do not have that option near term.)

 

[moltenmetal]

A few things to consider: why is it that a real positive displacement pump decreases in flow as the discharge pressure increases? It's due to check valve efficiency, which means that there's a period of REVERSE flow into the pump during the suction stroke. Presuming your pulsation dampener is upstream of the MicroMotion, some of its contents will be discharging forward through the MicroMotion and some backward through the leaking discharge check valve of the pump, so fortunately the Micromotion shouldn't need to be concerned about reversing flow- provided your pulsation dampener is sized properly.

Getting an RMS value from the MicroMotion is something to talk to MicroMotion themselves about. The (modern) units have a digital signal processor in them, so maybe it's possible to get a proper average out of them without making the unit totally custom. A simple moving average "damping" of the output signal may not be accurate enough for your purposes- or it may be- it depends on how big your puslation dampener is, and how accurate you need this to be.

You can model the pump's net forward discharge flow versus stroke speed, stroke length and discharge pressure, but as the check valves wear you may find that this calibration data changes- hence the need for the coriolis meter in the first place, relative to simply using a calibration column on the suction and treating the pump as "ideal" PD.

The other obvious thing to do is to minimize your stroke length and maximize your stroke speed for the range of operation you need. This will effectively maximize the benefit of your pulse dampener.

 

[jmw]

Talk to them too about the effects of pressure pulsations on measurement. In some bent tube mass meters the effect of pressure is to stiffen the tubes which can give rise, or used to give rise to, problems with measurement.

JMW

http://www.ViscoAnalyser.com