22 replies to this topic

[ilyes13]

Dear All,

I am currently working on a non-refrigerated LPG storage and I have to review the design done by a colleague.
The LPG is stored in a sphere with PSVs sized for fire case which will discharge to atmosphere.
In addition to the PSVs, my colleague has added a vent to atmosphere (pressure control valve set to 90% of PSV set pressure). The purpose of this vent is to avoid opening the PSV in case of overpressure due to solar heating for example.
I am not very confortable with this system and I would like to know if anyone of you has already experiment this kind of system (combination of a vent + PSV).
Thanks in advance.

[fallah]

Dear All,

I am currently working on a non-refrigerated LPG storage and I have to review the design done by a colleague.
The LPG is stored in a sphere with PSVs sized for fire case which will discharge to atmosphere.
In addition to the PSVs, my colleague has added a vent to atmosphere (pressure control valve set to 90% of PSV set pressure). The purpose of this vent is to avoid opening the PSV in case of overpressure due to solar heating for example.
I am not very confortable with this system and I would like to know if anyone of you has already experiment this kind of system (combination of a vent + PSV).
Thanks in advance.

ilyes13,

We have such configuration but with below differences:

-PV oulet line connected to LPG flare network
-Tank is cylinderical
-PV set is 60% of PSV set

Fallah

[ilyes13]

Thanks Fallah for your valuable answer.
Did you use API 2000 (normally applicable to low pressure storage) to determine the discharge coefficient for the vent?

Ilyes

[fallah]

Did you use API 2000 (normally applicable to low pressure storage) to determine the discharge coefficient for the vent?

ilyes,

I don't think API 2000 would cover determination of discharge coefficient in a pressure vent through a PCV (pressure control valve).

Fallah

Edited by fallah, 26 December 2011 - 07:13 AM.

[ankur2061]

ilyes13,

API 2510 - "Design and Construction of LPG Installations" permits LPG being vented to atmosphere subject to a dispersion analysis.

The venting calculations can be done according to API STD 2000. However, there is an additional aspect which is not covered in detail in API STD 2000 with only a brief description about it. This is the evaporation rate. API STD 2000 mentions that the "outbreathing" rate needs to be increased for evaporation if the product is stored at 40 deg C or it's vapor pressure is greater than 5 kPa. Since LPG vpaor pressures are very high, the venting calculations for the tank "outbreathing" need to add the evaporation rate to the outbreathing due to liquid pump-in. Annex - A, A.3.2.3 mentions that the vapor generation rate may be estimated as 0.5% ot the incoming liquid. An example of gasoline being pumped into an essentially empty tank is cited for the 0.5% evaporation rate in the same clause.

LPG evaporation mechanism is very nicely described at the following link:

http://www.corken.co...ining/cp226.pdf

As Fallah mentioned API STD 2000 does not mention about any discharge coefficients for the vent device although the Annex C - "Types and operating characteristics of venting devices" does provide a detailed description of the types and operation principles of venting devices.

Hope this helps.

Regards,
Ankur.

Edited by ankur2061, 26 December 2011 - 07:31 AM.

[S.AHMAD]

Dear member

Please take note that API 2000 is applicable to storage tank from full vacuum to 15 psig. Non-refrigerated LPG sphere is not covered by API 2000. The storage pressure for non-refrigerated LPG is much higher. In Annex D, the basic equation used is isentropic nozzle flow. However, for LPG storage I believe adiabatic expansion choked flow will be more appropriate one to use, if the vent is short. Otherwise you may check possibility of non-critical flow scenario.

[fallah]

Non-refrigerated LPG sphere is not covered by API 2000.

Ahmad,

Then, which standard covers these tanks instead API 2000 in venting design point of view?

Fallah

[ilyes13]

ilyes13,

API 2510 - "Design and Construction of LPG Installations" permits LPG being vented to atmosphere subject to a dispersion analysis.

The venting calculations can be done according to API STD 2000. However, there is an additional aspect which is not covered in detail in API STD 2000 with only a brief description about it. This is the evaporation rate. API STD 2000 mentions that the "outbreathing" rate needs to be increased for evaporation if the product is stored at 40 deg C or it's vapor pressure is greater than 5 kPa. Since LPG vpaor pressures are very high, the venting calculations for the tank "outbreathing" need to add the evaporation rate to the outbreathing due to liquid pump-in. Annex - A, A.3.2.3 mentions that the vapor generation rate may be estimated as 0.5% ot the incoming liquid. An example of gasoline being pumped into an essentially empty tank is cited for the 0.5% evaporation rate in the same clause.

Hello Ankur,

Thanks for your reply, I agree with you when there is no pressure in the tank but since the tank is pressurized, vaporization will be partially suppressed.
Moreover, a spraying system has been added to promote the condensation of the vapor space during the filling.
My main concern is then the thermal effect during hot days and estimating a realistic flowrate going through the pressure control valve.

Ilyes

[S.AHMAD]

Dear member

1. Please take note that LPG sphere is under it own vapor pressure. That means that the space above the liquid is full of LPG vapor. The operating principle is not the same as atmospheric tank. As we pumping in, pressure rises and the rate of condensation increases, so normally there is no requirement for venting as long as the temperature is kept within the operating range. LPG storage pressure, which is the LPG vapor pressure, is a function of temperature and LPG composition. Similarly, when the LPG is pumping out, pressure drops but rises again due to evaporation such that the equilibrium pressure will be the same as the vapor pressure at the operating temperature.

2. Thus, the main concern is when the temperature rises such that the pressure also rises correspondingly and exceeds the design limit.
3. Higher LPG rundown temperature due to fouling of LPG water cooler is one possible causes besides hotter atmospheric conditions.
4. My personal experience so far, I have never seen the need for venting LPG even though provision for venting is available. In my current place, the vent is sent back to process unit (crude overhead unit) for LPG recovery.

[S.AHMAD]

Dear ilyes

I propose you do energy balance. Take into consideration worst case scenario of LPG cooler fouling and abrupt change of ambient temperature.

Gradual change in ambient temperature is not the main concern.

It is better if you design the storage at the worst temperature conditions and coupled with the highest possible propane content in the LPG.

Relieving capacity should be determined as per API RP 520/521 of various scenario. e,g fire, block discharge etc.

Please take note that the current trend is now to consider relieving liquid due to overfill. In this case the LPG vent and PSV discharge should be sent to flare header.

[ankur2061]

Dear Ahmad,

In case the LPG is stored in a vessel (sphere / bullet) which is located in such an area where a flare header does not exist or it is not practicable to run a flare line from the vessel to the flare system or it is just an isolated storage facility, API 2510 permits atmospheric venting. Of course, atmospheric venting design does have to consider the effects of ground level concentration of flammable LPG by performing an appropriate dispersion analysis.

Your contention that the LPG sphere should be design for the worst ambient conditions (high temperature) is right. However, a certain additional protection can be provided at the discretion of the facilities owner for preventing the pressure in the LPG sphere to reach the set pressure of the PSV which as mentioned by the OP is sized for fire case and not for liquid overfilling case.

As Fallah has mentioned, if you know of any other method for calculating "inbreathing" and "outbreathing" rates other than what API STD 2000 mentions we would be very glad to know about it. Please share if you know some other method.

Regards,
Ankur.

[S.AHMAD]

Dear Ankur & members

1. For venting of vapor, the vent high and location must be high enough for LPG dispersion as per API2510. However,please take note that most dispersion model is for buoyancy gas. LPG is a dense vapor that the currently available model is not very convincing (this is only my opinion after based on manual calculation - however I am no well verse with software such as PHAST). Most of all, dispersion model does not apply to liquid LPG.

2. For liquid overfill, if venting to flare header is not possible. What I could suggest is that to install an INDEPENDENT high and high-high level alarm (Higher SIL) such that the vapor space above the HHL gives enough time for operator to take corrective action. Prevention is better than cure.

3. Please take note that "inbreathing and outbreathing" is not applicable to LPG storage since the LPG is stored under it own vapor pressure.

4. For atmospheric tank, inbreathing is required to prevent formation of vacuum (air or inert gas entering the tank and thus break the vacuum formed). Outbreathing is required to prevent overpressure (air or inert gas is push out).

5. As I have mentioned earlier, for determining the emergency capacity refer to API RP 520/521 for guidance.

6. A piece of advice - better be sorry now (higher capital) than be sorry later (catastrophic failure). LPG may leads to "BLEVE". or Vapor Cloud Explosion (VCE).

[ilyes13]

Dear Ahmad,

Thanks for your replies.

Liquid overfill prevention has been well taken into account.
The pupose of this vent is to avoid opening the PSV in case of overpressure due to thermal effect or potential presence of incondensables that can come from the road tanker during the unloading operation (vapor line connection between the road tanker and the sphere).
Maximum operating pressure corresponds to the maximum ambiant temperature at shade. The unit will be located next to the equator where the solar radiations can be quite high. Sunshield proposition has been rejected by the client but a good white painting should reduce quite efficiently the solar absorbance.
For the vent relief rate, I guess I will consider the volume generated upon vaporization with a realistic heat input coming from the sun radiation.

Ilyes

[S.AHMAD]

Dear Ilyes

1. In addition to sun radiation, need to include sensible heat from LPG rundown and anticipate the fouling conditions of LPG cooler (if any). Another words, do a complete heat balance that may include heat loss if applicable.

2. From the evaporation rate you can also determine the time taken to reach the design pressure by knowing the initial pressure and volume of space above the liquid level.

3. This means that if the design pressure is chosen correctly, most probably LPG venting could be eliminated except for removing the in-condensibles occasionally.

4. Finally, have you consider the need to have water flooding facility?

Edited by S.AHMAD, 28 December 2011 - 06:59 PM.

[ilyes13]

Dear Ahmad,

I really appreciate your contribution, thanks!

All the points you mentionned have been checked.
I agree with you, I am pretty sure that the vent will be useless on "normal" conditions (i.e. without the presence of incondensables or if the propane content is higher than expected which is actually my assumption for the relief flowrate design).

Are you talking about about fire protection facilities on the sphere? If yes, a water flooding facility has been foreseen.

Ilyes

[kkala]

1.1 I agree with S.AHMAD, since the LPG sphere is a pressure vessel.
1.2 Atmospheric air inbriething into flammable LPG is not safe and has not been observed. The sphere has to resist vacuum due to lowest temperature created during evacuation.
Note: Otherwise vacuum breaker has to insert inert gas into sphere (never seen).
1.3 "Outbreathing" may not be a precise expression for a pressure vessel. No outbreathing occurs during filling, some overhead vapors are merely condensed, increasing temperature and pressure a bit. During first time of filling only, vessel is assumed filled with nitrogen, returning to feeding LPG storage through a (so called) "filling" line. Practical information on other first time filling procedures is welcomed, even though referred "vent" may not have such a purpose (post No 1 by Ilyes13).
2. LPG discharge to atmosphere has been extensively discussed in http://www.cheresour...transfer-system. Sending LPG to flare rather than to atmosphere looks safer, especially in fire (LPG cold gas), even though fire can destroy flare header after some time. And a sphere has much more LPG content than a bullet.
3. Probably a depressurization valve (automatic or remotely operated) can do the task of referred "vent" http://www.cheresour...relieving-rate/. Again discharge to flare (in lieu of safe location with proven effective dispersion) is judged safer.
4. A local LPG station (1997) had several bullets, each one equipped with PSV (no other release observed to air). When pressure was approaching set pressure (due to direct solar heating), a manual spray system (probably the firefighting system) used to temporarily cover bullet walls with water. During 10 years the system got activated once, results were satisfactory. A study for the sphere (considering water temperature, etc) could estimate effectiveness of such an option.

Edited by kkala, 29 December 2011 - 11:43 AM.

[ankur2061]

If an LPG storage is an isolated area with no existing flare system nearby than trying to build a flare system just for a few storage tanks would be totally impractical. Continuously harping on providing a flare system without proper economic justification indicates lack of engineering judgement and practical thinking.

When established codes and practices allow atmospheric venting for LPG vessels (API 2510), I don't see any reason for building a new flare for a few vessels. Presently we are doing a study project for safe routing of PSVs for a middle-east refinery. In this project there are some LPG vessels which are located at quite some distance from the refinery flare system and which have PSVs routed to the atmosphere. The study report has recommended that the LPG spheres PSV tail pipes be elevated for dispersion to atmosphere in order to satisfy the ground level concentration limits of flammable gas based on Shell Fred dispersion software. The study has also concluded that it is impractical from both cost and routing point of view to connect the PSV tail pipes to the flare network which is almost a kilometer away.

As a conclusion I want to reiterate that engineering is not based on personal opinion and judgement which some members think it is. Refer the post I had made regarding personal opinions on engineering

http://www.cheresour...h__1#entry53865

(post # 13 in the link provided above)

[kkala]

Below is kkala's response to post No 17 by ankur2061
If an LPG storage is an isolated area with no existing flare system nearby then trying to build a flare system just for a few storage tanks would be totally impractical. Continuously harping on providing a flare system without proper economic justification indicates lack of engineering judgement and practical thinking.
You do not know whether the ifs of the above sentence are valid. Wording of post No 11 about API 2510 seems to permit discharge to atmosphere under certain conditions, so discharge to flare is safer. Post No 16 says just that, without proposal (situation not known). Thread referred there has revealed practices not permitting LPG discharge to atmosphere.
When established codes and practices allow atmospheric venting for LPG vessels (API 2510), I don't see any reason for building a new flare for a few vessels. Presently we are doing a study project for safe routing of PSVs for a middle-east refinery. In this project there are some LPG vessels which are located at quite some distance from the refinery flare system and which have PSVs routed to the atmosphere. The study report has recommended that the LPG spheres PSV tail pipes be elevated for dispersion to atmosphere in order to satisfy the ground level concentration limits of flammable gas based on Shell Fred dispersion software. The study has also concluded that it is impractical from both cost and routing point of view to connect the PSV tail pipes to the flare network which is almost a kilometer away.
API 2510 practice permits it conditionally, other practices do not permit it. See http://www.oknation....t.php?id=563728, where reported LPG spheres / bullets are connected to flare, as told in http://www.cheresour...ransfer-system. However our issue is just which practice is safer.
As a conclusion I want to reiterate that engineering is not based on personal opinion and judgement which some members think it is. Refer the post I had made regarding personal opinions on engineering
http://www.cheresour...h__1#entry53865 (post # 13 in the link provided above)
The link directs to the blog entry "The art of responding to queries..." and the issue of personal opinion has started in http://www.cheresou...g-configuration. Argumentation has been extensively developed from both parties, it is no use of repeating it again and again. Understanding the case and interpretation of code requirements is realized through a personal judgement / opinion, as present thread indicates too. Let the readers judge the arguments according to every time case. Note "extensively discussed" from now on will mean the above.
On the other hand there is no personal opinion from kkala, concerning LPG discharge to atmosphere. A try for right answer is better than arguing as above. Any comments on the multiple issues touched in post No 16?

Edited by kkala, 29 December 2011 - 07:14 PM.

[S.AHMAD]

Dear Ankur

In your case, QRA will be the best tool for decision making. If the storage tanks are isolated and far from public, venting to atmosphere at a safe location probably of very low risk and thus acceptable from process safety and business point of view.

[S.AHMAD]

Dear Ilyes

The water flooding facility as mentioned is water injection facility into the sphere. This is proven to be useful when there is a leak at the bottom of the sphere or anywhere at the discharge point. When leak occurs, the sphere can be "flooded" with water such that water will be discharged to atmosphere but LPG is contained since LPG density is lower than that of water.

I have seen such facility in one of my place of work. Which I think is a good idea.

[kkala]

A. Indeed, qualitative risk assessment (QRA) by experts (after a HAZOP study) could assess the risks (on environment and safety) in the specific case and decide accordinly. Previous posts do not consider it unacceptable to discharge gaseous LPG to atmosphere, permitted by API 2510 under certain conditions.
It is noted that high discharge velocities (150 m/s ?) are needed for efficient LPG dispersion into atmosphere (http://www.cheresour...ansfer-system/'> http://www.cheresources.com/invision/topic/13293-lpg-transfer-system/, post No 16). If 150 m/s is realized for the heat absorption specified in API RP 520 for fire case, a weaker fire may result in (say) 10 times less heat absorption, resulting in 15 m/s discharge velocity (or in PSV pop up, rather worse). Steam (mentioned in above web reference) could promote dispersion in this case, at the expense of system simplicity. Such "difficulties" (and similar, as well as probability of liquid discharge) have to be evaluated for the LPG sphere (much bigger than a bullet), with a view to arrive at a scheme safe enough (even not the safest) for today's standards. PSV and "vent" discharge to atmosphere from a safe location of effective dispersion could be acceptable within this procedure.
Β. I believe referred "vent" had better be named "depressurization valve", as explained in post No 16, to open at 90% of design pressure. This, if you wish such a depressurizing function (not personally observed on LPG vessels).
Γ.

.....During first time of filling only, vessel is assumed filled with nitrogen, returning to feeding LPG storage through a (so called) "filling" line. Practical information on other first time filling procedures is welcomed.....

.

.....In my current place, the vent is sent back to process unit (crude overhead unit) for LPG recovery.

Post by S.AHMAD indicates gas return to crude overhead unit. Can you clarify whether sphere is filled with LPG coming from same unit? If so, gas line between sphere and unit is the (so called) filling line and filling procedure is identical in the two above posts.

Edited by kkala, 30 December 2011 - 11:13 AM.

[kkala]

Above from http://www.cheresources.com/invision/topic/16848-prv-tailpipes-in-sandstorm-regions-of-world indicates common understanding now, concerning PSVs discharging gaseous LPG (to atmosphere or flare).
It is also noted that PSV discharge of non flammables (steam, compressed air, etc) are usually directed to atmosphere (safe location, if needed).

Edited by ankur2061, 23 December 2012 - 11:06 AM.

[kkala]

Now that ankur2061 has edited the above post (No 22), it should be clarified that it refers to his post http://www.cheresour...stem#entry67461 '> http://www.cheresour...stem#entry67461 , first and last paragraph.