Density compensation for steam drum levels
How can level in a steam drum be measured and reported accurately? This measurement has several challenges:
- The fluids are at high temperature and pressure.
- Any tubes colder than the steam and containing air (or any other gas) will collect condensed steam.
- The steam in the drum is dense enough that its weight is not negligible.
- The water in the drum is significantly less dense than water at room temperature.
Wet leg level measurement
The standard solution to these problems is to use "wet leg" level measurement, where both tubes running from the drum to the level transmitter are filled with water ("water" in this article implies liquid water). Unlike "dry leg" level measurement – where a tube from the top of the vessel to the transmitter is filled with air or another gas – "wet leg"measurement means that the high pressure port on the DP (differential pressure) transmitter should be connected to the top of the drum and the low pressure port connected to the bottom.
The measurement is from the top down – if the vessel fills with cold water up to the top tap, the ports of the DP cell will have the same water head above them so the DP will be zero. Falling level will raise the measured DP – the pressure on the low tap will fall relative to that of the high tap still full of liquid.
A wide pipe (perhaps 1" or more in diameter) should connect the top tap of the steam drum to a condensing pot. The pipe will allow steam to flow to the pot where it condenses to water because the exposed surface of the pot radiates heat away. This water keeps the tube running down to the DP cell full, and any excess flows back down the pipe into the drum. While the wet leg will naturally fill with water once the water in the drum boils, it is best to pre-fill the tube with water so that the level can be measured before the water boils. Otherwise it will read highly negative!
DO NOT INSULATE THE CONDENSING POT! Insulated condensing pots do not condense! If water boils out of the wet leg, the level measurement can be thrown far off!
Other than the pipe to the pot, which should slope slightly up to the pot, using horizontal tube runs should be avoided. If possible, the tubes should come down into the transmitter rather than curving back up under it. This will allow gas bubbles to rise to the drum rather than collecting where there should be water.
The transmitter can be scaled either 4 mA at zero DP (drum full) or maximum DP (drum empty). But the CCS (combustion control system) should scale the raw input so that it reads zero when the drum has cold water at the design operating level. Levels below that operating level should read negative, and levels above should read positive.
The example here is for a main steam drum, but these methods can also apply to deaerator storage and other similar vessels. This wet leg scheme solves the problems of high temperature (the liquid at the transmitter is at ambient temperature) and tube condensation. However, the transmitter must be able to withstand pressures higher than could ever exist in the drum.
Why compensate for density?
The wet leg scheme will work very well when the drum contains cold water. However, as the water heats up and pressure rises, the density of the water falls, and the density of the steam rises. Meanwhile the temperature of the water in the wet leg stays at the ambient temperature of the boiler house. Water boiling at 1310 PSIG is 580°F and is only 73% as dense as 100°F water. Saturated (not superheated) steam at 1310 PSIG is 5% the density of 100°F water.
For example, if a drum were full of 580°F water up to the top tap 50" above a bottom tap, it would register over 13" of DP, instead of the 0" DP it would read if the densities were the same. If the water is at the operating level 18" above the lower tap, the weight of water on the low pressure tap would be only 13.2", and the weight of the 32" of steam above it would add 1.6", for a total of 14.8" above the lower tap, reading -3.2" relative to operating level when it is really at operating level.
Modern CCS’s can compensate for this density error, but the math behind this compensation is hairy:
How to compensate for differences in density
Below is the formula for "true" (compensated) level, and is based on the following parameters:
- Lt: True Level
- Lr: Raw Level from DP cell
- Lb: Level of the bottom tap (in this example, a constant set to -18)
- Lc: Level of the condensing pot / top tap (in this example, a constant set to +32)
- Ds: Density of Steam under current conditions, looked up from Steam Table based on drum pressure
- Dw: Density of liquid Water in drum under current conditions, looked up from Steam Table based on drum pressure
- Dr: Reference Density of liquid Water in the wet legs at ambient temperature (in this example, a constant set to 61.96 pounds/cubic foot)
We know how to figure out what the raw level from the transmitter will read based on the true level:
But we want to know the true level based on the raw level … solving that for Lt:
Note that there is a potential division by zero error at the end. This makes physical sense because above supercritical pressures and temperatures (3189 PSIG and 705°F) the supercritical fluid in the drum would not have a surface level to measure.
Difference from Eye-Hye measurement
Drum level is frequently also measured by Eye-Hye – a device like a sight glass with relays to indicate whether each point (separated by an inch or two) has water or not. Because the temperature of the water in the Eye-Hye is colder and more dense than the water in the drum, the water in the Eye-Hye tube will be several inches lower than the level in the drum when at high temperatures. Operators should be trained to expect this difference and understand why the compensated level reported by the CCS is more accurate. If necessary, the CCS can also de-compensate the true level to report what the expected Eye-Hye level should be.
The level of vessels containing water and saturated steam is best measured by a DP cell with wet legs running to both ports. The DP will be near zero for a full vessel and increase as the level drops. Level should be compensated for density for any vessel requiring accurate level measurement over a range of operating pressures. If you have any questions or comments, I’d love to hear them!
Chris Hardy is an electrical engineer out of Georgia Tech. At Cross Company Integrated Systems Group since 1994, Chris has process control experience with boilers, alternative energy, water/waste water, chemicals, pharmaceuticals, security, textiles and automotive. Chris also programs controllers/HMIs and writes custom windows applications for communication, data collection, display, trending and reporting.
Cross Company Integrated Systems Group is a CSIA member as of 3/5/2015