Dry running of a Pump

[nicksmail]

Hello,
To protect a centrifugal pump from running dry, a liquiphant (sensor that detects dryness) was installed in the suction side of the pump after the suction strainer (See drawing). If the liquiphant detects dryness, the pump shuts down. The pump sucks water/oil from a vessel and pumps it to a tank. The suction line is positioned approx. 20 cm below the suction nozzle of the pump. To check the function, the suction slide valve was closed. However, the pump continued to run because the liquiphant did not seem to detect dryness. Unfortunately, I cannot explain this. Is it possible that the suction line is not being sucked dry when the suction slide valve is closed and the pump is running? The liquiphant or sensor is located at the top of the pipe and protrudes about 1 cm into the pipe.
Thank you in advance,
Regards Nick

 

[Latexman]

Is it possible that the suction line is not being sucked dry when the suction slide valve is closed and the pump is running?

Yes. If a vacuum is created, the flow may stop. You need a vent valve downstream of the slide gate with a riser that goes to higher elevation than the slide gate. Teing the vent valve riser into the slop tank vent may be the safest way.

Good Luck,
Latexman

 

[nicksmail]

Hello Latexman,
thank you for your reply. Could it be a problem, that the liquiphant is not installed at the same height as the suction nozzle?

 

[LittleInch]

nicksmail

Think about it. You turn off the incoming liquid , but where is the lack of liquid / creation of an air pocket going to come from?

It would only come from air entering the system due to it being below atmospheric pressure or by pulling a vacuum.

Your system by the look of it does neither, hence the pipe remains full of liquid.

If you want to test it I wold do the same test, but then open a vent line somewhere between the suction slide valve and the pump to allow some liquid to be moved through the pump and air to take its place. The strainer probably has a vent on it somewhere. This needs the pump inlet line to be less than atmospheric pressure, but try it and see if air goes in or liquid comes out....

Inlet strainers are usually not a good move as they reduce NPSHA as the filter gets clogged. you always need to have a DP gauge or transmitter to identify when you need to clean it and set this alarm quite low ( < 0.5 bar).



Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[3DDave]

The place a void will be created is in the most dynamic section - the inside of the pump.

You need a mass flow-meter to detect such massive cavitation or a pressure sensor to detect the loss of feed pressure.

 

[LittleInch]

I can only assume that the intent is to drain the tank completely. Hence why this instrument is used. You need to make sure the inlet velocity in the tank nozzle is quite low otherwise you'll suck in air as it approaches the end of the liquid.

A centrifugal pump may not be your best bet here. I would have gone for a progressive cavity pump or a diaphragm pump myself.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[horacio Torres]

It is a complicated device to handle, sophisticated but extrange

I , as a process designer, will prefer to install a low level switch ( 2 independent if I want to get down the probability to 10-6 )   in the  tank, It is more accurate and simple.

Horacio

 

[TiCl4]

For those who may be wondering, “liquiphant” is an Endress & Hauser vibronic tuning fork design level switch. If OP is not seeing the sensor activate, then liquid must still be present.

OP may wish to consider a flow switch rather than a point level device. A level switch seems poorly suited to this application.

 

[don1980]

My reaction is the same as Horati's - this seems unnecessarily complicated. Why not rely on a low-level sensor in the feed tank (simple, cheap, and highly reliable)?

 

[pierreick]

Hi,
To me because of the strainer ,I would consider a flow switch with time delay .
Pierre

 

[Latexman]

nicksmail,

Why must the tank be run totally empty? If not, horacio Torres and don1980 have a very good point.

Good Luck,
Latexman

 

[nicksmail]

Thank you for the numerous responses. Normally, I also install a level switch in the vessel, which switches off the pump at min. level. However, since a suction strainer is installed, which can become clogged, a level switch must be installed between the suction strainer and the pump according to our internal Standard. That is the reason why I did not choose another solution. Unfortunately, it does not work, as already described.

 

[LittleInch]

If you're worried about no flow because the suction strainer is blocked you should be looking at a low pressure transmitter with an alarm or trip.

Your internal standard must be thinking about some other configuration because it would rarely work unless you can pull a vacuum.
Or it was written by someone with no practical grasp on life...

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[nicksmail]

Yes, you are right. We have to discuss the internal Standard in our company. I‘ll let install a ventpipe in the suction line to the Top of the vessel.

 

[dik]

I understand that with some pumps, if they run dry, they can be damaged.

Rather than think climate change and the corona virus as science, think of it as the wrath of God. Feel any better?

-Dik

 

[Compositepro]

The most common method I have seen to prevent dry running of pumps is a pressure switch on the discharge. If it drops below a certain value it will stop the pump. To start the pump a momentary contact switch or timer is used to allow the pump to start and build pressure. If pressure does not build within a set time, the start switch is released, or the timer times out and the pump stops and stays off.