# Experts weigh in on distillation plant measurement problems

March 3, 2016
Bela Liptak and his crew tell how to resolve difficulties with measuring column differential pressure and flows of flare and natural gas.

This column is moderated by Béla Lipták, automation and safety consultant and editor of the Instrument and Automation Engineers’ Handbook (IAEH). If you have an automation-related question for this column, write to [email protected].

## Question:

I always use your valuable books in my field as a control & instrumentation engineer. I have three questions, and if you could help me it will be so appreciated:

1) Installation differential pressure transmitters on tall towers (inlet and outlet) for vapor service has been a big problem for me.I have mounted a differential pressure (DP) transmitter above the outlet line of the tower with its low-pressure side connected by a sloping tube to the outlet pipe, and high-pressure side connected to the tower inlet pipe with a 14-meter-long tube. We have two pressure gauges on the inlet and outlet pipes showing 3 PSID while the DP cell reads 600 mbar (9 PSID). What is the problem, condensation? What is the remedy, increasing slope or insulation? A similar case has not been resolved, and I had to use two pressure transmitters (PT) to get software differential value. When I touch the low tapping, its temperature is colder than the other leg. Butene is the measured vapor.

2) In another process we have a flare flow measured by a pitot tube. The flowrate is between zero and 24,500 kg/h. We need a rangability of 20:1, which cannot be measured by the DP-type sensor, because it shows 0 kg/h at all times. Should we select and substitute a thermal mass or  ultrasonic flowmeter? We want to see from the lowest possible flowrate (zero if possible) to 24,500 kg/h. Inlet pressure is less than 0.5 bar, fluid temperature -20/250 C. What is your suggestion?

3) I want to select and size a flowmeter for natural gas service with maximum 6,000 kg/h flowrate in 4-in. pipe that has one PT and temperature transmitter (TT). This flowmeter is for custody transfer of natural gas at  6 bar pressure and  normal flowrate of 1,500 kg/h, operating at ambient temperature.

Rahim Romel
[email protected]

On question 1, the left side of Figure 1 shows my understanding of what you have now and on the right I am showing what I would do to fix your problem. In general, I do not like to use conventional DP cells to measure the pressure difference between points that are at different elevations, because the lead lines can introduce errors due to condensation, or due to errors caused by differences in temperature or density on the two sides of the cell. These problems can be reduced by using capillaries or pressure repeaters, but why bother? The best solution in my experience is to use two good pressure transmitters and measure the difference between their outputs, as shown on the right of Figure 1.

On question 2, you are right, pitot tubes can not be used for rangeabilities above 3:1. I assume that you do not need to measure the flare flow very accurately, so a wet and cooled, time of signal passage-type, multipath and intelligent ultrasonic flowmeter, with compensation for composition, pressure and temperature (density) variations, would be an option. If you want direct mass flow measurement, use a thermal (lower cost and accuracy) or Coriolis meter (higher cost and accuracy).

On question 3, for high-accuracy mass flow measurement when the composition and density of the gas varies, use a Coriolis flowmeter.

Béla Lipták
[email protected]

Figure 1: Tower pressure drop measurement

Flowmeter engineering is intertwined with deep understanding of process, process behavior, operational ranges, process thermodynamics, and flow system engineering as to what is the objective, for example, custody transfer, process control or flow indication as the questions indicate. Instrumentation engineering is applied science,  requiring more knowledge than just transmitters and orifice plates.

On question 1, your transmitter leg on the high side is very long compared to the low side, and because of this the effect of the fluid thermal expansion/contraction in the impulse line has a potential of giving you this error. As you know, the coefficient of thermal expansion varies by the cube of the expansion coefficient. Insulation of the impulse tubes may help—my suggestion would be to use capillary filled tubes of equal length on both sides, low and high side. Make sure the capillary lengths of both legs are the same, and are exposed to the same ambient condition. One leg should not be in shade and the other in sun, because we want to negate the thermal effects. The capillary should eliminate the effect of phase changes etc. so it would give a more accurate measurement.

On question 2, here, using a Pitot tube is wrong, because of its limitations on rangability or flow turndown. In most flare header lines, because of the EPA regulations, an ultrasonic meter is used, however, you have to compensate it  with composition analyzer, pressure and temperature to get accurate measurement of flare gases. You mention a temperature range of -20 to 250 C. The high temperature is not suitable for ultrasonic sensors. On the low side, you may want to contact Flexim, which supplies cryogenic ultrasonic flowmeters and may be able to add cooling extension plates between the sensor and pipe to make it also suitable for your high temperature. Multiple thermal mass flowmeters for varying temperature conditions may be another option, but accuracy would suffer. I am not sure what can handle the high side of your temperature range.

On question 3, variation in composition and range from (pilot flow to max flow) is a daunting task for any flowmeter. Most people settle for ultrasonic flow, however, for results to be meaningful you would need a composition measurement, which could be expensive. Orifice is a wrong for this application if you need high accuracy and wide turndown (orificce plate is 3:1). Your best bet is an ultrasonic flwometer with pressure and temperature compensation.

Romel S. Bhullar, senior technical director, Fluor Corp.
[email protected]

On question 1, since the low pressure side of the dp cell is connected to the pressure tap at the top of the column, it is recommended that the cell itself be located above that point so that this low side will be self draining back into the top outlet pipe. With this configuration, the high pressure side of the dp cell will have a very lead line connecting it to the pressure tap at the bottom of the column inlet pipe. Depending on the thermal insulation used on the lead lines and on the temperatures of the process streams at the in and outlets of the column, there might be condensation in these tubes, which might introduce some errors. Yet, this configuration is still better than locating the dp cell down at the level of the inlet pipe, because then the lead line to the low side will not be self draining, but can fill and if it does, that side will become the high pressure side. In any case you may need to recalibrate the transmitter.

On question 2, the biggest issue here is that generally flare gases have a mixture of products so the density variations as well as moisture content can affect the measurement. I would recommend a Coriolis meter in case you want precise measurements (see Micro Motion ELITE Coriolis Flow and Density Meters). A Pitot tube will not give you better than 3:1 rangeability and even than, density compensation will be needed. That can be obtained by periodic lab testing.

Alex (Alejandro) Varga, project & construction management, Devco
[email protected]