Interesting PLC puzzle involving a pump and pressure transducer

There is a hydrostatic pressure transducer threaded onto a pipe. On that same pipe, there is a pump pumping fluid in a very large cell (500 meters by 500 meters, 2 meters deep). When the pump is running, the reading will be effected due to pipe friction. Is it still possible to monitor the hydrostatic level while the pump is running? What is the best way to mathematically correct this?

So far I have a value from when the pump wasn't running being differentiated with current value reading with pump on. The difference calculated every cycle and then running average is applied to filter noise. After 1 minute the running average stops. The last value calculated from running average is subtracted from the level reading while the pump is running until the pump is turned off. When pump is off the transducer reads as normal, no calculations are applied. I haven't tested this. Does anyone have a better method?

Thanks!

So, the PT is meauring level in the cell? If so I would question locating it in the infeed pipe.

You may benefit from explaining the issue a bit more

cjd1965 said:

So, the PT is meauring level in the cell? If so I would question locating it in the infeed pipe.

You may benefit from explaining the issue a bit more

Click to expand...

Agreed. Doesn't seem like a logical place for a LT.

If that PT needs to reside in the pipe, then you may consider a second sensor for LT purposes only.

I think it will depend on how accurate you need to be. Your scheme should give you a rough estimate of the level, but if it needs to be accurate you will need a separate sensor.

Yes, accuracy while the pump is running is not a big deal. It's less than ideal, but that's how the engineer says it should be done, ill respect that. For reasons I don't know why.

The transducer is measuring the level of the cell through the feed of the cell. The cell is actually more like a dug out pool of waste/feces.

The spec sheet does not specifically say that the cell needs to be monitored while the pump is running. It just notes that there will be inaccuracies due to the pump. This obviously wasn't a mistake. It doesn't say the there should be calculations to compensate for it. But I am going to assume that it should be done since it wouldn't make sense to monitor the level of the cell while it doesn't change. Especially when its being trended.

Sounds like maybe you are pumping clarifier waste into a reed bed??
I guess it's probably a good idea to monitor that, but a lot of sites don't do it.

blackbird307 said:

It doesn't say the there should be calculations to compensate for it. But I am going to assume that it should be done since it wouldn't make sense to monitor the level of the cell while it doesn't change. Especially when its being trended.

Click to expand...

That last sentence might be your key. I would trend the signal at various fill levels. Start the pump and trend the signal. It may spike and then settle down to a value somewhere above where it should fall after the pump is stopped. How far "off" it will be is going to depend on several factors like distance from the pump, pipe diameter etc. You might be able to put together a method of offsetting this value if you gather enough information from it based on trials.

This technique works ok if the friction losses are low. We have used this for small reservoir sites with remote bores back before wireless I/O was around.

However a high flow out of your tank (i.e. burst pipe) can result in inaccurate readings.

Those design engineers love saving $1,000 on a level sensor and spending $2000 on programming time to compensate!

It sounds like you did it exactly right. With the pump off, you have a static system and your transducer is measuring the pressure due to fluid height, regardless of the volume of fluid in your cell. All that matters is the height of fluid above your transducer, since there is no flow in the pipe. With the pump running in steady state (assuming you have a centrifugal pump, and not a positive displacement pump) there will be head loss, or pressure drop due to friction. How much head loss depends mostly on the velocity of the fluid, which will be higher if the level is low in the cell, and lower if the level is high in the cell. But with only 2m of depth, the change in head loss will not be that significant. With a little practice, you should be able to guess the level in the cell, with the pump running, or not running, since the pressure differential will be proportional, unless you develop a blockage, or a rupture in the pipe.