Hot water level measurement using pressure sensor in closed tanks

[elda5ly]

I want to measure the level of water inside a closed st.st tank with an over flow opening in the top of the tank

at the begining of filling the tank with fresh water the level measurement process being so good.

But after filling the tank, the water needs to be heated to 90 degrees, during the heating process iam noticing that the level of the water is increasing due to the vapour pressure which indicates to a fake level reading.

Is there any tips to overcome this problem?

 

[Bit_Bucket_07]

I assume that you are inferring tank level via a pressure measurement?

If vapor pressure is causing a higher than normal pressure reading, then could it be that the tank needs ventilation? Does the overflow opening you mentioned not provide ventilation?

 

[Steve Bailey]

Use a differential pressure sensor. Or else add a separate pressure sensor at the top of the tank and subtract its reading above atmospheric from the depth sensor.

 

[Bit_Bucket_07]

use a differential pressure sensor. Or else add a separate pressure sensor at the top of the tank and subtract its reading above atmospheric from the depth sensor.

+1

 

[danw]

A pressure sensor at the top of the tank to measure the vapor pressure of heated fumes would be a relatively expensive transmitter capable of reading in the tens of inches of water column.

That's usually a DP transmitter with the low side open to atmosphere, providing a gauge pressure measurement.



Installing a DP with an impulse tube so the low side reads the vapor area pressure and the high side reads the liquid head down at the bottom of the tank would provide a conventional liquid level measurement that is always corrected for vapor pressure.

Level measurement derived from pressure is always subject to an error due to liquid density changes, because pressure/level measurements assume a constant density. If the liquid density changes over the range of the temperature changes, then the level reading will have error in a direct proportion to the specific gravity/density change. If the specific gravity drops 1.5% during the temperature change, the level reading has 1.5% error.

 

[elda5ly]

I assume that you are inferring tank level via a pressure measurement?

If vapor pressure is causing a higher than normal pressure reading, then could it be that the tank needs ventilation? Does the overflow opening you mentioned not provide ventilation?

The tank has an over flow pipe, but it's a tight tube so the vapour from heating still making a a pressure leading to a fake height measurment

 

[tarik1978]

The "guided radar" measurement method
permits accurate and reliable point level
detection in the steam boiler since it is not
affected by changes in density.

 

[elda5ly]

A pressure sensor at the top of the tank to measure the vapor pressure of heated fumes would be a relatively expensive transmitter capable of reading in the tens of inches of water column.

That's usually a DP transmitter with the low side open to atmosphere, providing a gauge pressure measurement.



Installing a DP with an impulse tube so the low side reads the vapor area pressure and the high side reads the liquid head down at the bottom of the tank would provide a conventional liquid level measurement that is always corrected for vapor pressure.

Level measurement derived from pressure is always subject to an error due to liquid density changes, because pressure/level measurements assume a constant density. If the liquid density changes over the range of the temperature changes, then the level reading will have error in a direct proportion to the specific gravity/density change. If the specific gravity drops 1.5% during the temperature change, the level reading has 1.5% error.

Thanks for your concern and replying.

What about inputting the water density to the plc with each degree increase.

Let us assume the temp is 40 degrees so i will input 0.9925 to the water density parameter, and if the temp increases to 41 i will input 0.99186 and so on so on.

Will this affect the height reading ??

Attached the file of the water density relative to each degree C

 

[elda5ly]

A pressure sensor at the top of the tank to measure the vapor pressure of heated fumes would be a relatively expensive transmitter capable of reading in the tens of inches of water column.

That's usually a DP transmitter with the low side open to atmosphere, providing a gauge pressure measurement.



Installing a DP with an impulse tube so the low side reads the vapor area pressure and the high side reads the liquid head down at the bottom of the tank would provide a conventional liquid level measurement that is always corrected for vapor pressure.

Level measurement derived from pressure is always subject to an error due to liquid density changes, because pressure/level measurements assume a constant density. If the liquid density changes over the range of the temperature changes, then the level reading will have error in a direct proportion to the specific gravity/density change. If the specific gravity drops 1.5% during the temperature change, the level reading has 1.5% error.

The "guided radar" measurement method
permits accurate and reliable point level
detection in the steam boiler since it is not
affected by changes in density.

Of Course the guided radar is a solution to this problem, but the guided radar is too much expensive if it's compared to the pressure transmitter

 

[elda5ly]

Use a differential pressure sensor. Or else add a separate pressure sensor at the top of the tank and subtract its reading above atmospheric from the depth sensor.

I will try this solution if the solution of inputting the density of water to the plc with each degree increase doesn't solve my problem.

Thanks too much for your concern and replying

 

[Bit_Bucket_07]

I will try this solution if the solution of inputting the density of water to the plc with each degree increase doesn't solve my problem.

Thanks too much for your concern and replying

There are two different variables affecting your estimate of tank level, although both are attributable to heat gain.

Water becomes less dense as it is heated. That will introduce an error into your level calculation. A look-up table for water density would address that.

The head pressure increase due to vapor pressure is a force acting in the opposite direction, putting additional pressure on the water column as it is heated and compressing it. The DP sensor arrangement that danw mentioned would address that.

You need to account for both forces for a perfect level inference model, but perfection may not really be required in your application?

Are you more interested in the volume of water or the mass? Volume, I would assume, since it's a level measurement.

 

[drbitboy]

Water coming out of the overflow pipe indicates a minimum value for a level. Would it be possible to put a temperature sensor on the outlet of that overflow pipe, and when the temperature is high simply use the height of the overflow as the level, perhaps with an alarm?


Pressure differential gives weight of fluid per unit cross-sectional area of the container vessel. If thermal expansion of water is to be part of the calibration, then what about the thermal expansion of the vessel walls changing the area of the vessel?

 

[parky]

Does it really matter, what is the water in the tank used for?, first of all, unless you are relying on a specific level in the tank of less than +-10% for some reason then why does it matter, the increase in level in some respects may be counteracted by the expansion of the tank, I have never known a hot water system to matter in the moderate change in level, is the water dosed out as a complete batch or is it maintained when it fills below a certain level, I can only think of a couple of reasons why you my want to know the true level, for example a batch process where you rely on a specific level (never a good idea as heating will expand it), using the level as some form of metering (again not a good idea). I had a similar discussion with a process engineer on a hot water flash system, there were two debates with him regarding the system.
1. The level in the tank did change (the water was heated to 150.0 Deg. C (under pressure), this was then flashed into a pressure cooker.
2. The temperature would never reach 150.0 as the steam was 150.0 so we only achieved about 148.2 -149.4, my arguments were obvious, the cooker it was fed into had a flow meter & load cells, this controlled the amount of super heated water added, did it really mater that the temperature (which could not be met with the current site steam temperature).
In the end I won, the guy was a P***,

 

[elda5ly]

There are two different variables affecting your estimate of tank level, although both are attributable to heat gain.

Water becomes less dense as it is heated. That will introduce an error into your level calculation. A look-up table for water density would address that.

The head pressure increase due to vapor pressure is a force acting in the opposite direction, putting additional pressure on the water column as it is heated and compressing it. The DP sensor arrangement that danw mentioned would address that.

You need to account for both forces for a perfect level inference model, but perfection may not really be required in your application?

Are you more interested in the volume of water or the mass? Volume, I would assume, since it's a level measurement.

Did you mean that i need to use input the density of water relative to each temperature degree in addition to the top of tank pressure transmitter??

So i must do the two tips, because only one tip will still lead to an error in the reading?

Did i get your point correctly?

 

[elda5ly]

Does it really matter, what is the water in the tank used for?, first of all, unless you are relying on a specific level in the tank of less than +-10% for some reason then why does it matter, the increase in level in some respects may be counteracted by the expansion of the tank, I have never known a hot water system to matter in the moderate change in level, is the water dosed out as a complete batch or is it maintained when it fills below a certain level, I can only think of a couple of reasons why you my want to know the true level, for example a batch process where you rely on a specific level (never a good idea as heating will expand it), using the level as some form of metering (again not a good idea). I had a similar discussion with a process engineer on a hot water flash system, there were two debates with him regarding the system.
1. The level in the tank did change (the water was heated to 150.0 Deg. C (under pressure), this was then flashed into a pressure cooker.
2. The temperature would never reach 150.0 as the steam was 150.0 so we only achieved about 148.2 -149.4, my arguments were obvious, the cooker it was fed into had a flow meter & load cells, this controlled the amount of super heated water added, did it really mater that the temperature (which could not be met with the current site steam temperature).
In the end I won, the guy was a P***,

The system is a CIP system composed of three tanks.
Acid tank, soda tank, hot water tank.

At soda tank for example, the process starts with filling the tank with fresh water with normal room temperature up to 2000L for example.
then we start to heat them up to 80 degrees for example,

the first issue is during the heating process the volume of the water starts to increase and this problem appears on the system HMI.

The second issue is when we start to pump this heated water from the tank, the pump may still operating while the tank is empty because of the fake volume resulting from the vapour pressure.

I know that you may say, it's easy you can installing a flow switch or a level switch to protect the pump from dry run operation, i know, but iam asking of how can we overcome the level and volume errors on the hmi or the scada system.