This article was printed in CONTROL's October 2009 edition.
"Ask the Experts" is moderated by Béla Lipták, process control consultant and editor of the Instrument Engineer's Handbook (IEH). The 4th edition of Volume 3, Process Software and Networks, is in progress. If you are qualified to contribute to this volume, or if you are qualified to answer questions in this column or want to ask a question, write to [email protected].
Q: Please explain the role of the damping parameter in a differential pressure transmitter. What happens if I increase the damping value from 0.4 sec to 1.2 seconds? We observed a lot of oscillations between 0% to 100% even though the actual level of liquid hydrogen (LH2) is not varying in a cooler. We tried increasing the damping parameter from 0.4 sec to 1.2 seconds and the oscillation was reduced.
A: I have dealt a lot with liquid hydrogen because of my invention, the RFC. (www.controlglobal.com/Media/0807/CG0807_LLfig1.jpg).
I suspect that your d/p transmitter is just fine, but your installation is not. I assume your cooler is a double-walled high-vacuum insulated tank (Figure 1). In these types of level measurements, we do not try to keep the liquid in the lower sensing line in the liquid state, because it is difficult to do so and requires perfect insulation. Instead, we let the liquid hydrogen boil in the lower sensing line. This way, as the sensing line approaches the outside, the temperature will rise, and the liquid hydrogen will vaporize. This will create a liquid-vapor interface in the sensing line. Beyond that interface the line will be filled with vapor.
In order to obtain a stable and noise-free level signal, you should make sure that this interface develops at a point in the sensing line that is well-defined and enlarged (1 in. or 25 mm). At this point, the sensing line should have a slight upward slope as it approaches the outer wall of the vessel. The sensing line should be enlarged because during level changes, the interface between the liquid and vapor can be turbulent.
As shown in Figure 1, the approach to the wall should be such that under all process and ambient conditions, the hydrogen temperature reaches vaporization temperature. To account for transient conditions that temporarily move the level of the interface, a "dry" loop is often provided in the warmer portion of the sensing line, so as to catch and quickly evaporate any temporary liquid carryover.
Figure 1.The correct installation of a d/p cell type level sensor on a tank containing liquid hydrogen.
A: The damping parameter in a DP transmitter is the time constant of a filter applied to the analog signal before the signal is transmitted. The value (in seconds) determines how long the output signal will take to reach 63% of the final value if a step change were applied to the input. Most people set the value to the minimum and apply the filter in the control system, so that it's easily changed and monitored, but there are times when setting the value in the transmitter is preferred.
It's likely that your oscillation problem has little to do with this parameter. You are just masking your problem by damping out the oscillations with the filter and making the signal look better. You can set the damping value so high that the signal will have almost no oscillation at all, but the transmitter would take forever to respond to level changes. Based on your service, you are dealing with very low temperatures and rather low DP calibration. If that is true you should check the following things:
- If you're using diaphragm seals, are they rated for the temperatures you're applying to them?
- If you're using straight tubing, is the transmitter rated for those temperatures?
- Is there a chance that the insulation on the sense lines (or vessel flanges if you are using diaphragm pads) is insufficient? It sounds like you are boiling the liquid hydrogen in the sense line, and thus, the transmitter is seeing the blast of pressure as the liquid boils off and the collapse as the pressure releases and fresh hydrogen liquid rushes in.
A: Damping is a first-order filter. High frequencies (short periods) are attenuated, but low frequencies will not be reduced. For example, a damping of 1s will reduce amplitude of oscillation by:
- ~1.4 if oscillation period is 6s (corner period is 2*PI)
- ~10 times if oscillation has a 0.6s period
- ~100 times if oscillation is 0.06s period
- ~no attenuation if period is longer than ~10s.
Usually in a tank, oscillations from waves are in the range of 5s-20s; to reduce amplitude, damping should be ~10s; this at the same time would increase apparent dead time and destabilize the control loop. In your case, increasing damping from 0.4s to 1.2 s will improve only marginally.
Most of the time, oscillation periods in tanks and reservoirs are in minutes, and are caused by bad tuning of the level controller. If so, damping will not help.
For level control, usually, proportional gain should be exceed 1.0 (PB,100%), and integral time should be longer than 5 minutes/repeat.
Michel Ruel PE
Q: Is a 24-VDC power supply used for control and PLC inputs less susceptible to static as compared to an un-grounded power supply?
A: All power supplies should have a safety ground connected to the power supply's shield to protect against electrical shock, as well as protection against EMI (electromagnetic interference). All signal processing systems (analog or digital) should have a single-point signal ground to prevent ground loops. The signal ground may also be used for shielding signal wires to provide EMI immunity, but should not be connected to more than one safety ground point.
A: Even though this question needs more definition on what is meant by "static" and how the supply is intended to be connected, I will say the following:
- If what is meant by static is over-voltage surges (equipment damage), the same power supply connected as a floating, or ungrounded, power source would normally be less susceptible than when grounded. Its performance would also probably be less affected. I have "floated" supplies to minimize lightning damage in SCADA/telemetry equipment many times.
- If what is meant by "static" is power-line noise/interference, susceptibility will depend mostly on the internal design of the supply and system interconnections. I recall a flowmeter design where an optional internal supply ground could be used to trade off noise performance against UL certification.
- The choice of whether to use a grounded or a floating source generally will have more to do with electrical codes and user safety than system noise interference or damage from over-voltage. Still, using a floating power source will make it easier to avoid system ground loops.
Al Pawlowski, PE
Q: I want to know the effect of not following the straight pipe/meter run requirements for the following flowmeters: orifice, magnetic, vortex, turbine, ultrasonic and PD meters. Is there any general rule for meter run requirements? Whose flowmeter has the smallest requirement?
A: Refer to the flowmeter selection table (Table 2.1b) in Vol. 1, 4th edition, of my book Process Measurement and Analysis. There you will find a complete list of performance data for all flowmeters.