19 replies to this topic

[Sherif Morsi]

Hi,

 

During a HAZOP session, a concern was raised about the possibility to siphon the H2S scavenger tank due to the difference in elevation between the two tanks in the attached sketch.

 

A diaphragm pump injects H2S Scavenger into the water tank which sits on a 25 m lower level than the H2S Scavenger tank. Both tanks are atmospheric tanks. The water tank is equipped with a pressure/vacuum relief valve and a 20" overflow line.

 

What I mentioned is that we have a diaphragm pump installed and the fluid won't flow if the pumps is not operating given the presence of the check valves on the pump. 

 

1. Can I take credit for the diaphragm pump to prevent siphoning the tank?

2. What other measures/controls are required to prevent siphoning?

 

Thanks,

Sherif

Attached Files

•  img-505052412-0001.pdf   491.87KB   76 downloads

[breizh]

Sherif,

a- Can you install a dip pipe inside the Scavenger tank connected to the section of the pump located above the tank? On the upper part of the pipe drill a hole to prevent siphoning .

 

b- On the proposed design you may consider to actuate the valve at the suction of the pump ( solenoid valve) to prevent flow .

 

Hope this helping you .

Breizh

Edited by breizh, 05 May 2017 - 09:34 PM.

[Bobby Strain]

What you describe is not a siphon, but a drain. I see no feature that would prevent flow from the scavenger tank to the water tank when the pump is off. In fact, I wonder why you even have a pump. So, you need to reconfigure the piping from the pump to the water tank so as to prevent flow when the pump is down. The configuration you need should be obvious.

 

Bobby

[Saml]

A diaphragm  pump do have check valves, but they open in the flow direction, so if you have more pressure in the suction side than at the discharge, both check valves (suction and discharge) will open, and there will be flow.

 

So, in addition to the piping option Bobby mentions above (probably the simplest solution) some metering pump manufacturers recommend a back pressure valve or backpressure regulator. Other offer a "munltifunction valve" optional accesory, that I guess is a backpressure regulator.

 

Look on google about "dowhill metering pump".

[Sherif Morsi]

Bobby,

 

How do I reconfigure the piping downstream the pump? Do we route the line vertically?

 

Sherif

[Sherif Morsi]

Hi Breizh,

 

Can you provide me with a sketch for option A?

 

Regards,

Sherif

[Sherif Morsi]

A bit of history on the matter,

 

This tank/pump arrangement had another application and under the proposed arrangement, it will inject regulated doses of H2S scavengers into the water tank. The line runs for about 120 meters; it's a 3/4" SS 316 tubing. The pump is a duplex diaphragm pump with manual adjustable strokes.

 

Downstream from the pump, the line is equipped with check and ball valves.

 

Please let me know if you need more details

[Sherif Morsi]

Just to clarify, I couldn't reply earlier than yesterday cause the site went down just after I posted my query then it was the weekend and my laptop is down so I couldn't reply till yesterday

[Saml]

I don't know the density of your scavenge, but let's say is 1100 Kg/m3 and negligible vapor pressure.

You know that after the backpressure valve, you cannot have a pressure below absolute vacuum.

Then you set a backpressure valve that cracks at 25 psi differential (172.3 kPa)

When you substract the atmospheric pressure (assuming you are at sea level) of 101.3 kPa you are left with 71 kPa

 

Maximum Height of liquid column on the high pressure side of the valve:

 

71000 [Pa] / ( 1100 [Kg/m3] x 9.81 [m/s2] ) = 6.6 m

 

That means. You install a backpressure valve with 25 psi cracking pressure like a 3/4 Swagelock C series check (*) at the discharge of the pump. The valve should be located not lower than 6.6 meters below the maximum level of the scavenger tank. You have to be conservative, so probably that distance has to be shorter, or your cracking pressure a little bit higher.

 

(*) I have no affiliation with Swagelock, nor I endorse this manufacturer, it is just an example for illustration purposes.

Edited by Saml, 09 May 2017 - 08:30 AM.

[Sherif Morsi]

Saml:

 

Thanks for the feedback.  When you say a back pressure valve, you mean a vacuum breaker?

 

Could you elaborate more on this calculation?

The density is 1,065 kg/m³ and the difference in level between the 2 tanks is 28 meters.

[Sherif Morsi]

And what's that 25 psi differential? Could you provide a simple sketch as well?

 

Thanks

[Bobby Strain]

I would relocate the scavenger tank and pump to grade. Solve the problem forever; no band-aid.

 

Bobby

[sgkim]

Sherif, 

 

I would like to recommend a simple additional (or revised) piping route as marked with red pencil in the following sketch:  

 

 Dosing Pump Schematic Piping Plan.png   412.19KB   5 downloads

 

~Stefano

Edited by sgkim, 10 May 2017 - 10:50 PM.

[Sherif Morsi]

Stefano,

 

Thanks for your reply. If I understand it correctly, this change in piping route would make us take the line up for 26m to counter the difference in level, is that correct?

[Sherif Morsi]

I would relocate the scavenger tank and pump to grade. Solve the problem forever; no band-aid.

 

Bobby

 

Bobby,

 

Ideally yes but given that we are offshore and there are a lot of layout and weight limitations for such solution which would drive our cost beyond the current budget.

 

Sherif

[Art Montemayor]

Sherif:

 

A siphon is a tube or conduit bent into legs of unequal length, for use in drawing a liquid from one container into another located on a lower level by placing the shorter leg into the source container and the longer leg into the one below, with the tube filled 100% with the liquid.  This causes the liquid in the source container to flow upward, and into the lower target container - with no pump, but powered by the fall of the liquid as it flows down the tube under the pull of gravity.  The liquid is forced up the shorter leg and into the longer one by the pressure of the atmosphere.

 

As Bobby Strain has declared, you do not have a siphon taking place.  But you fail to disavow your previous statement and the title of this thread.  For the sake of clear and accurate engineering - as well as for archiving the thread under the correct title, let us all accept the fact that what you have is a simple drain.  You probably are using a positive displacement pump for the sake of measuring the drained flow rate.  Am I correct?  Since both tanks are at atmospheric pressure, the top tank can easily drain all its liquid down 25 meters.  That is not a problem.   What is a problem here is understanding your two queries.

• Since there is no siphoning taking place, how can the pump prevent it?
• Why is it that you expect "siphoning" when you don't have the piping that warrants its existence?

Are you using the correct English word(s) to describe what you have?  I am trying to define the correct, accurate title that should accompany this thread so it can be archived correctly.

A typical siphon is shown on the attached drawing:

 

 Siphon.JPG   28.96KB   6 downloads

[Sherif Morsi]

Hi Art,

 

Sorry for the not-so-accurate sketch. Kindly see attached.

 

The outlet nozzle from the Scav tank is at the bottom then the line goes up around 0.48m to the pump suction then further up on the discharge side until it reaches the pipe rack. It runs for 10-15m on the pipe rack then falls to the ballast water tank 25 m below the Scav tank.

 

I believe this is similar to the explanation and sketch you provided which makes it a siphon. Please correct me if I am wrong.

 

Thanks for your thorough explanation

 

Sherif

Attached Files

•  img-519054635-0001.pdf   507.87KB   50 downloads

[Art Montemayor]

Sherif:

 

Thank you for the explanation and the correct sketch.

[sgkim]

Sherif, 

 

To avoid the siphon effect of your piping, three options are available:  scheme 1) equalizing the top of the line to upstream tank,  scheme 2) equalizing it to downstream tank  or scheme 3) removing the siphon leg.   See the sketch attached shown in cloud marks for three schemes.

 

 img-519054635-0001-R1.pdf   540.21KB   48 downloads

 

~Stefano

Edited by sgkim, 25 May 2017 - 01:32 AM.

[Sherif Morsi]

Thanks Stefano for your input.

 

It has been decided that the pump operation will be manual and manned at all times. The isolation valve upstream the pump will be closed once the pump is shutdown (whether by the operator or by any control switches that can shut it down). The two options considered were an automatic on/off valve upstream the pump or a manned operation.

 

Given that the valve is much more expensive and is a long lead item (at least 30 weeks) and the injection requirement is intermittent at the moment, we decided to go with the manned operation for now.

 

Thank you all for your valuable input. Really appreciate it.

 

Sherif