8 replies to this topic

[kkala]

A. Collected oily rain water of a tank farm is pumped into a crude oil tank (of floating roof) and drawn off from a nozzle near bottom. Attached “siphonbrake.xls” indicates (in simplified form) process requirements, flow rate can vary from 50 to 600 m3/h (depending on 2x50 m3/h pumps or 1x600 m3/h pump in charge, according to rain intensity). Client does not wish at all to have back flow of oil into water (even if the latter is oily) and according to their experience a siphon can be created even if two check valves are installed in series (the latter get out of order “silently”). So they have requested:
(α) Oily water feed from tank roof, through one existing 12” vertical guide-pall (whose holes shall be plugged and instruments shifted). Discharge line (10”) shall be connected to it.
(β) Effective siphon break, through an open pipe / vent of proper height, connected to the highest part of the feeding pipe line (see siphonbrake.xls).
Note: Their practice for siphon break concerning cone roof tanks is to make some holes on the vertical feeding line between cone roof and high high liquid level (not applicable to floating roof tank).

Β. Briefly looking into requirements, following drawbacks have been understood:
(1) Assuming all discharge pipe full of liquid up to the exit, pressure at highest part of line (feeding the tank) is below atmospheric at any flow rate. If an open vent is installed, this would cause air ingression into the guide-pall (dangerous). Referring to siphonbrake.xls diagram, mentioned pressure is calculated to:

For tank full of crude (sg = 0.84): 22.5/10*0.84- 25/10*1.0 = -0.61 kp/cm2 g .
For tank full of water: 22.5/10*1.0 - 25/10*1.0 = - 0.25 kp.cm2 g
For liquid at LLLL : 2/10*1.0 – 25/10*1.0 = - 2.3 kp/cm2 g (impossible)
Note: Frictional ΔP in 12’’ guide-pal is neglected, as insignificant (0.06 kp/cm2 at 600 m3/h). Static pressure cannot balance frictional pressure drop in the flow range 50-600 m3/h. I imagine flow in the guide-pall is not full. But on what criterion is this based on? End of curve data of existing 2x50 m3/h pumps is 2x 90 m3/h at 31 m of liquid, no such data available for the pump of 600 m3/h.
(2) In the (improbable) case of plugged guide pall, pump shall develop its shutoff pressure, so level in the open pipe will be rised by h = 45-27 = 18 m in case of 50 m3/h pumps operating (probably higher if the 600 m3/h pump operates).
Note: Frictional pressure drop of 10” discharge line to its highest point is calculated to 1.3 kp/cm2 for 600 m3/h flow rate. Assuming an open vent of h=0 m at mentioned point, pressure at pump discharge is 27/10*1.0+1.3 = 4.0 kg/cm2 g, that is 40 m of liquid. But if there is no open vent at the highest point, what is the pressure required at the pump discharge? (see siphonbrake.xls, 600 m3/h).

C. After the above considerations, the possibility of elevating the 10” discharge pipe (feeding the tank) 12 m over tank HHLL is investigated, that is 10 m over tank peripheral wall. This could prevent back flow from tank to (oily) water, seeing that siphon cannot be maintained if height difference exceeds 10 m for water or 10/0.84~11.9 m for crude. And flow patern in guide-pall may be forgotten in this case. However this pipe elevation will be probably not acceptable by statuary authorities for permit.

D. We will try to convince Client that 280 m3/h of flow rate covers the max rainfall (which will reduce discharge pipe size to 8”), yet problem will actually remain. Any comments on above or advice on the matter or proposal would be highly appreciated. Probably someone has faced a similar situation, or my inexperience on downward flow complicates the issue. Some siphon breaks (sensing flow direction) have been seen in WWW, but their reliability is not known. Or probably flow arrangement has to be modified (e.g. control valve?) to give a satisfactory solution.

Attached Files

•  siphonbrake.xls   31.5KB   342 downloads

[breizh]

Hi ,
Will it be possible to install an on/off valve at the discharge of the pump to avoid the open pipe at the top of the tank ?
Breizh

[kkala]

Hi , Will it be possible to install an on/off valve at the discharge of the pump to avoid the open pipe at the top of the tank ? Breizh

In the (over) simplified "siphonbrake.xls" no valve is shown at the pump discharge, indeed. Actually valve and check valve shall be placed at the pump discharge, but this would not solve the problem. You see the discharge line is connected to a complex network "communicating" with a lot of tanks of the farm (or pipes receiving also reversed flows towards drain). One cannot be certain that all valves are closed, after all in some cases they should not. I understand Client is afraid that crude of the tank might pass into waste water to treatment.
A tight shut off valve at the highest point of discharge line (e.g. over the tank roof) could solve the problem, yet this is undesirable. Indeed, it is not easy to step up 25 m through the peripheral ladder just to close or open the valve.
Thinking more of the issue, we are looking into feeding the tank via an ascending - descending pipe to a separate nozzle, not through the vertical guide-pall. Of course the problem remains even so, even though you can place easier e.g. a remote operated valve at the external pipeline.

[Art Montemayor]

Kostas:

This may – or may not – be of help, but I believe what you are describing is an application for a backflow prevention device. One such device is called a “vacuum breaker”.

A backflow prevention device is used to protect drinking water supplies from contamination or pollution. In water supply systems, water is normally maintained at a significant pressure to enable water to flow from the usual, conventional water tap (faucet), shower, toilet, etc. Should the water supply pressure fail or be reduced - as may happen if a water main bursts, pipes freeze or there is unexpectedly high demand on the water system - the pressure in the pipe may be reduced to the point where it allows contaminated water from the ground, from storage of from other sources to be drawn into the system. To prevent such an occurrence, many regulatory regimes require there to be an air gap or a mechanical backflow prevention device between the delivery point of mains water and local storage or use. It is in this way that a backflow prevention device protects the public potable water system from contamination hazards which could be severe and deadly.

A vacuum breaker is an attachment commonly placed on a faucet valve, toilet, or urinal flush valve, that prevents water from being siphoned backward into the public drinking water system. This prevents contamination should the public drinking water system's pressure drop.

These devices are explained in detail in the Internet. Spirax / Sarco, for example, has a web page explaining how it works. See http://www.spiraxsar...kers.asp#head25

Your flow rates are much, much higher than those in domestic and some industrial applications – but the principle is the same: positively prevent back flow from occurring.

I have employed vacuum breakers such as these in cooling water lines in the past with success.

I hope this helps.

[breizh]

Hi ,
I was thinking at an on/off valve actuated by the start /stop of the pump like I've seen on cooling towers network to avoid back flow of water to basin .

Breizh

[katmar]

The only way that I can see that you will be sure the flow does not entrain air into the tank, while preventing a back syphon, and also not be subject to operator error or equipment failure issues is to make the last vertical leg down into the tank self venting. For 600 m³/h this probably means replacing the 12" guide-pall with a 24" or 26" pipe. You would still need the open vent at the high point to prevent a back syphon, but at least you would be sure not to draw air into the tank. This scheme would also not be sensitive to the tank level.

The lower you can convince your client to go on the pumping rate the smaller the pipe diameter needs to be to ensure that the pipe is self venting.

[kkala]

Sincere thanks to all those having the patient to reply. The responses were really valuable and promoted insight into the issue, even though it has not been yet fully solved. Rain water is considered flammable, that I was not conscious of. Consequently an open vent (even with a flame arrestor at the end) could not be safe enough against static electricity, since fluid velocity in the pipe exceeds 1 m/s.
This complicates the matter. Moreover I would like to share some thoughts on the opinions posted.
1. Art, a "mechanical backflow prevention device" could be a way out, on the condition that it does not cause continual air ingression into the system (air intrusion once to break siphon could be probably acceptable). During siphon operation subatmospheric pressure is actually created around the highest point of the transfer pipe from tank. But to my understanding there will be also (undesirable) continual air intrusion during normal operation, when flow is pumped to tank. Instead of vacuum breaker (as that of Sarco) a siphon braker sensing flow direction might be preferrable (of proven reliability to be accepted by Client).
2. Katmar, a self venting guide-pall would satisfy stated Client's requirements if flow rate were reduced to about 118 m3/h for the existing guide-pall of 12 in (Client would not change its size). Flow in it would be more stable, and air collection in the tank would be avoided, which is important. But the main problem is that flammable oily water and air would be in contact during the flow, from upper part of guide-pall to the tank liquid level.
However the querry "I imagine flow in the guide-pall is not full. But on what criterion is this based on?" has been answered. Info on vertical line flow (Froude number etc) can be found online in "Norsok Standard P-001" (p.11, vertical lines), which was not known to me.
3. Breizh, an automatic on/off valve on the transfer line close to tank could be a way out, especially after the present investigation. The valve could be fail close and motivated by pump motors connected to critical loads; or through orifices sensing low flow from pumps to tank. It might not cover all extreme cases, yet it would have the advantage of simplicity without the probable event of air intrusion. Besides in this case the transfer (oily rainwater) line need not go up (to the tank height) and down, but directly into a new nozzle near tank bottom. Client requirements have to be broadly revised, as I can see. This can be after a discussion in next meeting.
In first week of October we prepared a cost estimate, based on the (strange) concept of transfer line elevation 12 m over tank HHLL. At least from technical viewpoint this seems sound, yet we clarified that the assumption is quite temporary and under review. Total cost is not much affected by this choice. Information on the adopted solution will be given, when it is decided.

Edited by kkala, 21 October 2010 - 03:44 AM.

[kkala]

Having been in central work offices every now and again, I can inform that adopted solution concerns three check valves (not all of the same type) on the discharge line of centrifugal pumps to the tank, versus one usually placed. Mentioned line enters the tank near the bottom now, not through the guide-pal from the top. A motor operated valve shall be also installed on this line, closed by operator whenever the pumps stop, with an intent to prevent back flow from tank. Flow of oily water to the tank shall vary from 50 to 400 m3/h, implemented though a 10" pipeline.
Adopted solution was applied to a similar tank in the past. Thoughts of alarm in case that liquid level of pump basin increases (indicating back flow from tank into basin) have not been realized.
Again sincere thanks to the forum friends, having the patience to deal with this matter and offer advice.

[RicyVarem]

I am looking for a solution for the below system:

We need to pump water from a pond and transfer to the plant at 45080 Mtr distance. The levels and other details are mentioned as below.


Flow to transfer: 400 m3/ hr. Sp Gr. 1.03
E.L. Water Level at pond: 283.3
E.L at top of proposed tower near pond: 301.0
E.L. Existing Booster set 262.0
Collection tank at plant EL. 244.5

Pipe Details: HDPE Pipe DN 280 PN 16. ( ID 230 MM)
Pipe length: New pond to existing booster set: 308 Mtr
Existing booster to Collection tank at plant: 4200 mtr.

We planned to install a submersible pump in the pond with 400 m3/ hr flow rate at 22 mtr head. It wil deliver water to elevation 301. We need to design a syphon breaker at elevation 301.
The booster pump in installed at Elevation 262 inline ( inseries) that will deliver water to the plant. The booster pump is designed for 400 m3/ hr at 46 Mtr head.

Please confirm that the system will work as per above details. Also propose type of Syphon breaker at elevation 301.