4 replies to this topic

[vinnay1999]

Hello 

 

We have a 21 metre high tank which has an  incoming pipeline connection that is connected on the roof top with a 1000 dia HDPE pipe and continues to go down into the tank for 20 m.The top of the pipe is @ 31.3 m elevation (tank bottom was @ 8 m elevation) .We have two pumps @ 5500 m3/hr each that are supposed to send water into this tank after going through ultra filtration system. We get a flow of around 8000 m3/hr (with both pumps operating) . I have attached a representative graphic of the system.

 

We expected that we would get a siphon with this arrangement and thereby helps the pump with the flow rate. I have calculated the pressure drop using network analysis and I am attaching the results graphic for reference.

 

However , we find that while the pumps are started the water is falling freely into the tank creating an air pocket on the top.We have a vent connection at the top of the tank , but when it is being opened to vent off the air , we actually find vacuum that is withdrawing more air into the pipe. We tried to suck air with a vacuum pump but without much improvement. Is there a way to remove the air from this pipe. Now we think to throttle the pump discharge to generate more velocity at the top so that we can have full diameter flow. Am I thinking in the right direction. Please throw some light 

 

below are the links for the drawings 

 

https://docs.google....dit?usp=sharing

https://docs.google....dit?usp=sharing

 

 

[latexman]

V = 8000/(3.14*0.5²)/3600 = 2.83 m/s

 

N_Fr = 2.83/(9.81m/s²*1m)^0.5 = 0.9

 

I calculate your Froude Number is 0.9.

 

In an open system, i.e. the end of a vertical line is not liquid sealed or the line is vented at the top, Froude numbers between 0.31 and 1.0 can result in pulsating vapor/liquid flow and significant vibration. This region is to be avoided for open systems unless the piping is specifically designed for the transient loads.

 

IF the end of the vertical line is liquid sealed AND the line is NOT vented at the top, all the air will eventually flow down the pipe with the liquid, since your N_Fr > 0.31.  HOWEVER, with atmospheric pressure at the 22 m elevation, there will be a strong vacuum (with possible boiling, cavitation, vibration, noise, etc.) at the 31.3 m elevation.

 

I'm afraid both options, open or closed system, could be in a "no man's land" (1.0 > N_Fr > 0.31 or cavitation) from a design standpoint.

 

Can the vent be opened and modifications made such that N_Fr < 0.31 (restrict flow) or N_Fr > 1.0 (increase flow) AND process objectives be met?

[katmar]

I agree with the excellent summary put forward by Latexman.  Relying on the pressure recovery that is theoretically available in a siphon system like this tends to be risky.  There are also problems at start-up which can prevent the system from ever achieving the expected conditions (more on this below).

 

In the network diagram you show the outlet pressure at the bottom of the siphon as 1.3702 barg.  Taking the siphon pressure recovery and the friction losses in this final down flow section gives the pressure at the top of the tank as -0.8859 barg. This pressure corresponds to a water boiling temperature of about 50 °C.  This is all heavily dependent on the tank level remaining perfectly constant.  If the tank level drops a metre or two the pressure at the top of the leg can drop and cause the water to boil - as discussed by Latexman.  On the other hand, if the tank level rises you get less recovery and the pumps see an increased head and the flow rate can drop off.  You therefore have to hold the tank level perfectly constant in order to keep a constant flow rate and if the tank is intended as a buffer of sorts it may not achieve its goal.

 

The start-up problems can be serious.  The pressure at the top of the final down leg will be close to atmospheric at start-up.  This means that the pump delivery head is 0.8859 bar higher than shown on your flow diagram.  Depending on the shape of your pump curve this will cause the flow rate to decrease or even stop completely if this is above the shutoff head.  This decreased flow may not be sufficient to fill the final down leg and initiate the siphon and then you can never get to the desired flow rate.  Fortunately, it seems that you are able to achieve sufficient flow to get the siphon working to some extent - as evidenced by the vent valve drawing air into the system. But wIth your Froude number of 0.9 it may take a while to flush all the air out. Throttling your pumps will decrease the flow - you need to somehow increase the flow to flush the air out.

 

In order to achieve a stable system the general practice is to have a large vent at the top of the final down leg so that the pressure there is always atmospheric.  It does cost more to run the pumps at a higher delivery pressure but the benefit is a system that works predictably under a wide range of flow rates.

[vinnay1999]

Thank you people for your valuable time and invaluable insights.

 

 I think we were unsuccessful with the vacuum pump because of the vapor formation.

 

To increase the flow we can  use the  standby pumps(2 nos.) available with us and eventually increase the froude number.But should we keep maintaining the froude number above 1 or is it an initial requirement till the siphon forms ?

 

On the other hand, what modifications should be put into place to decrease the froude number below 0.3. We can shut down a pump and decrease the flow. Then we have a flow dominated more by gravity  and how can this help?

 

As for katmar's viewpoint, In the network analysis ,  the pressure of 1.37 bar g,  shown was at the bottom of the tank (8 m elevation) considering the 14 m level of liquid column . At the 14 (22 m elevation) m level the pressure was shown as 0 bar g. We do have a outlet from the tank that goes into another system (reverse Osmosis) and this can help to withdraw liquid from the tank for maintaining level.

 

Sometimes I do hear ( a closed roof tank with vents) a sudden fall of liquid into the tank (like a water fall)  and then it remains silent.

 

I'll try to run my stand by pumps and increase the flow and see.

[katmar]

Vinnay, sometimes I jump straight to the details of the technology and underlying mechanism and I forget what the original problem actually was. I seem to have the same mistake here. What exactly is your problem?  Are you getting insufficient flow, or experiencing vibration, or some other problem?  Or are you just wanting to understand what is going on?  We could all talk in endless circles about vacuum levels and Froude Numbers if we don't focus on the true goal of the exercise.