30 replies to this topic

[Olidin]

Hello,

The required relief rates suggested from API 2000 seems excessive to me. I have a 16ft diameter x 17ft tank and with the required venting, it would be emptied in 25 seconds during fire. This seems so odd to me. (read more below to see how I get those 25 seconds)

I was wondering if someone noticed and/or know the reason why they are being so conservative.

For example, for a cylinder tank 16ft diameter and 17 ft high, on the ground, not elevated, with hexane inside. The wetted area for this would be the tank wall:

16 x pi x 17 = 854 ft^2

According to API 2000 Section 4.3.3.3.3 Table 8, required venting requirement is 493000 scfh (for 900ft^2). So let's put that number in perspective. Tank volume is 8^2 x pi x 17 = 3418 ft^3

So the total time to empty the tank is 3418 / 493000 = 0.0069 hr or 25 seconds.

[Raj Mehta]

Well I feel you have to use the value corresponding to the wetted area i.e. 854 sq. ft. You can do this by interpolation as follows, 

 

How is the tank oriented ? Horizontal or Vertical ? By your calculation of Wetted area, it seems you have taken it to be vertical. 

 

By interpolation: 

                            800 sq. ft -> 462,000 CFH

                            900 sq. ft -> 493,000 CFH

                             -----             -----------

     Difference :      100              31000   CFH

 

31000/100 = x/(854-800)

 

thus, x = 16,740 CFH

 

Total CFH required = (16,740 + 462,000) = 478,740 CFH.

 

Minimal opening nominal pipe size = 8 inches.

 

Thank you. 

Edited by Raj Mehta, 09 April 2013 - 06:45 AM.

[fallah]

Olidin,

 

Your calculation for the total time needed to empty the tank is basically wrong because:

 

- The tank volume you used is in liquid state while the venting capacity is in SCFH of Air

- The venting capacity is going to be reduced v.s. the time as per reduction of wetted area due to liquid level reduction

- ....

Edited by fallah, 08 April 2013 - 01:24 PM.

[proinwv]

Hey guys am I missing something but the venting referred to is out breathing for fire exposure.

[latexman]

Fallah is right. 1 ft³ boiling hexane releases roughly 200 ft³ of sat'd vapor.  You calculated a peak rate, which is ok for the relief, but for an average it'll be half the peak, assuming the rate goes to zero at the end.  So refigure that as 25 seconds x 200 x 2 = 10,000 seconds = 2.8 hours.

 

Sound more reasonable?

[Raj Mehta]

Latexman & fallah, can you elaborate your thinking & also can you check my calculations? Why do you need to calculate the time ? 

 

Thanks.

[fallah]

Raj Mehta,

 

We didn't need to calculate the time. We just checked the calculated time by Olidin.

[Olidin]

Ah, there we go. I completely missed that. What a rookie mistake. Now it makes sense. 

 

Thanks guys. 
 

[Olidin]

Raj,

 

thanks for the note on interpolation. I often don't interpolate on venting requirement since valves are often sized a bit larger than the required size so there was no need to be exact on venting requirement. For venting capacity, I would interpolate (or even sometimes extrapolate with manufacturer's help) if needed.

Thanks.

[Olidin]

Latexman,

 

I think I understand your note on the volume of vapor generated during fire. However, I did not follow your calculation.

Though here is what I thought. Let's use your number. Tank is 3418 ft³, let's say full of hexane. I'm not going to do dynamic analysis, so I have no clue on relief rates. Let's say all vaporize, we get 3418 x 200 = 683600 ft³ vapor. Then we got 683600/493000 = 1.38 hours. Sounds better to me.

Of course this time is only for perspective and do not reflect the true time it takes for relieving. The dynamic behaviors were not considered. Like you and Fallah said, relieving rate would change over time.

[Steve Hall]

The essesntial thing that hasn't been mentioned in this thread, is that the API table is really dumbing down the problem by giving the final answer -- relieving rate -- without the underlying theory.

 

I don't have a copy of API 2000 to refer to, but I do have NFPA-30, the U.S. Flammable and Combustible Liquids Code, that has Annex B which gives background to the numbers.

 

Yes, NFPA-30 has the same table as API 2000, giving CFH vs Wetted Area. Don't neglecct the reduction factors for tanks having qualified drainage, tanks protected with water spray and tanks protected with qualified insulation.

 

But in the Annex, you find that the basis for the table is the heat input, Q, from a fire. Like any heat transfer calculation, you can calculate or measure the heat transfer from the heat source (fire - radiation and convection) to the liquid inside the tank. Somebody did that and the results were correlated with the wetted area of tanks. So the heat input per unit of area changes depending on the area - this is due to curvature of the tank (projected area of the tank to the radiation source changes), and other factors.

 

Ultimately, the committees decided that Q is related to A in three parts:

from 20 ft2 to 200 ft2 ............. Q = 20000 A

200 ft2 to 1000 ft2 ................. Q = 199300 A^.566

1000 ft2 to 2800 ft2 ............... Q = 963400 A^.338

"For areas beyond 2800 ft2 it has been concluded that complete fire involvement is unlikely and loss of metal strength from overheating will cause failure in the vapor space before development of maximum possible vapor evolution rate. Therefore, additional venting capacity beyond the vapor equivalent of 14090000 Btu/h will not be effective or required."

 

The venting tables are decribed as venting hexane with the free air at 60F and 14.7 psia. But, responding to a different post on this question, the tables assume the heat of vaporization of hexane at its boiling point.

 

Since the table is based on the formula CFH = (70.5 Q) / (L M^.5), Its easily adjusted to other fluids by comparing physical properties of the fluid with those of hexane. For heaxane, L = 144 Btu/lb at boiling point and M = 86.17, L M^.5 = 1337. Take, for example, n-Pentane where L M^.5 = 1300. Adjust the values in the table by multiplying 1337/1300.

 

An additional note in Annex B states, "No consideration is given to possible expansion from the heating of the vapor above the boiling point of the liquid, its specific heat, or the difference in density between the discharge temperature of 60 F, since some of these changes are compensating." They also state, "The use of hexane in deriving the table provides results that are within an acceptable degree of accuracy for the listed liquids." (And here they are referring to a table of common petrochemicals that's included in the Annex).

[ankur2061]

Dear All,

 

Why don't you refer the spreadsheet on emergency venting due to fire from atmospheric storage tanks which is attached with the blog entry at the following link:

 

http://www.cheresour...-storage-tanks/

 

Regards,

Ankur.

[Raj Mehta]

Steve sir, that was really good explaination & Ankur sir, the spreadsheet is worth appreciation.

 

API 2000 speaks about the hexane, and for liquids other than hexane, I was confused if i have to consider the physical properties of the stored liquid at its boiling point or not ? 

 

The other confusion I have is in the units of the stored liquid in case of emergency venting.

 

Say the liquid stored is pentane. Now when I calculate the CFH as mentioned in the spreadsheet (also mentioned in Steve sir's post) i.e. using the corrected venting for pentane (other liquid), the unit would still be SCFH of air or would it be CFH of Pentane ?

 

I am confused with the 1st point of post # 3, I think Fallah wants to highlight about the units or may be something which I am not able to capture. 

 

Eagerly waiting to get this concept clear. 

 

Thanks.

[ankur2061]

Raj,

 

Why the confusion? It is very clear that API STD 2000 addresses the venting rate either for normal venting or emergency venting in terms of Nm3/h or SCFH of air. Table 5 & 6 of API STD 2000 explicitly mention that the emergency venting rate is in Nm3/h and SCFH of air respectively.

 

Regards,

Ankur.

[Raj Mehta]

Sir the confusion in my mind is because, If i have liquid say pentane stored, then during the emergency situation, it will be my pentane vapors going out from the tank and not the air, so the unit / venting capacity should be in CFH of pentane vapors instead of SCFH of air.

 

Am not considering SCFH because standard would mean 15 deg C, but my vapors are leaving at its boiling point which is liquid specific. This thought is very well confusing me and giving me sleepless nights. 

 

Kindly explain why do we always consider / express the unit as SCFH of air? 

 

Thanking you.

[ankur2061]

Raj,

 

An open to atmosphere tank has always a mixture of air plus the hydrocarbon liquid vapor in the vapor space. The vapor is generally a lean mixture (more air and very less of hydrocarbon vapor) at the tank operating conditions. Since API STD 2000 is targeted at liquid hydrocarbons, generally having a molecular weight greater than air, the vapor in the vapor space is predominantly air. 

 

Due to an external fire and rise in pressure it is the air+hydrocarbon lean vapor mixture of lower molecular weight that first gets vented out before the actual hydrocarbon vapors start venting. The venting rate is expressed as an instantaneous rate and this will be the highest (conservative value for design of vent or PVRV) for the lowest molecular weight vapor compared to a vapor of higher molecular weight. This is the reason that air is considered for venting rather than the actual stored liquid of higher molecular weight.

 

The venting rate will actually decrease once the pure hydrocarbon vapors are generated and removed from the vent after all the air+hydrocarbon lean mixture is vented out.

 

Hope this gives you a clear picture of why air is considered for venting as per API STD 2000.

 

Regards,

Ankur.

[Raj Mehta]

Thank you sir for your prompt repsponse to my query.  I now have some conceptual clarity on this topic, you mentioned open tank, what if it is a fixed roof tank ? here in this case the vapors would majroly comprise of the liquid stored right sir ? 

 

And in case of nitrogen blanketing in fixed roof tank, it would be nitrogen leaving out and not air. May I am thinking too much, which is leading me to all sorts of possible complexities in this topic. 

 

But i thank you very much sir for helping me out on this topic. 

[CMA010]

Ankur:

 

Expressing venting rates in Nm³/hr or SCFH of air has absolutely nothing to do with a supposudly initial expellement of air. Venting devices are tested/certified/rated using air. Corresponding flow curves are subsequently expressed in an air flow rate (at a set of reference conditions), in order to avoid comparing apples with pears actual venting rates need to be converted to an equivalent air flow rate (with the same reference conditions). This ensures that same size requirement/capacity is required for actual venting rate and equivalent air flow rate.

[ankur2061]

Raj,

 

Open to atmosphere means tanks defined as per API STD 2000 and whose vapor space can have air or nitrogen (in case it is blanketed with Nitrogen). API STD 2000 is also applicable to Internal Floating Roof tanks with a fixed roof above the floating roof. It is however not applicable to external floating roof tanks.

 

Nitrogen and air have nearly the same molecular weight (28 & 28.95 respectively), so a minor correction only needs to be done as per Graham's law of effusion if you need to correct it for blanketing gas rate. You know about the the post I made (post# 14) in the following link:

 

http://www.cheresour...-sixth-edition/

 

To conclude, the lightest vapor (lowest molecular weight) in the vapor space, for a non-blanketed tank will provide the highest vent flow rate which is generally air for atmospheric storage tanks storing liquid hydrocarbons. In case the tank is blanketed, the vent flow rate may be corrected for the blanketing gas using Graham's law of effusion.

 

Regards,

Ankur

[ankur2061]

CMA010,

 

Venting rate calculations are air flow rates at reference conditions of 101.325 kPaa (14.696 psia) and 0 deg C (60 deg F) as per API STD 2000 and not the way other way around as you suggest.

 

Regards,

Ankur

[CMA010]

Ankur: 

 

I'm not suggesting anything of the sort, i'm saying that actual venting rates (substance X at temperature Y etc.) need to be converted to an equivalent air flow rate for reasons as explained. API has conveniently provided methods (formula and/or table) for certain venting circumstances in which this is done in a single step, e.g. formula 11 includes this conversion. For other venting circumstances you will need to do this conversion yourself. It has nothing to do with an initial expellement of air.

[latexman]

Back in CHE 205, Chemical Process Principles, in 1977, my professor was Dr. Richard M. Felder.  Yes, THE Dr. Richard M. Felder himself, co-author of the world famous, best selling "Elementary Principles of Chemical Processes".  When I started CHE 205 that semester, we got the book in manuscriprt form.  The next semester in 1978, we were give the first edition of the book.  One thing he taught I will never forget is that with a gas/vapor volume or volumetric flow you MUST also specify the temperature, pressure, and molecular weight/chemical name.  Otherwise, the volume/volumetric flow rate is useless!  That information is needed to convert to a mass flow rate.

 

API has done this, and they standardized on air.  Why?  To make it more convenient and practical.  Air is the most common and abundant gas there is.  It is relatively safe and cheap.  It is everywhere.  It is what they used to test vent unit flows for years and years.  Simple stated, it is a no-brainer!

Edited by latexman, 12 April 2013 - 07:52 AM.

[proinwv]

As I understand it the current edition of API 2000 was in part driven by research at Protego in Germany. At that time they had a presentation on their work and how it related to the final standard. I seem to remember that they also believed in the "theory" that vents were cheap in relation to tanks so bigger is okay.

 

Anyway, if anyone has a contact there, it might add some thoughts to this discussion.

[CMA010]

Ankur:

 

Graham's law of effusion? Conversion from actual flow rate to equivalent air flow rate (at reference conditions) is based on isentropic nozzle flow, see API STD 2000, appendix D.

[ankur2061]

CMA010:

I think you are stuck with converting from actual flow rate to equivalent air flow rate (at reference conditions), whereas I am talking about converting air flow rate in Nm3/h as calculated in API STD 2000 to blanketing gas rate.

 

The blanketing gas rate is indeed calculated based on Graham's law of effusion:

 

Volume flow rate of blanketing gas in Nm3/h = Volume Flow Rate of Air in Nm3/h *√(MWof Air / MW of blanketing gas)

 

This is a fair approximation of converting air flow rate at given reference conditions to blanketing gas rate at the same reference conditions and has been used in design calculations for blanketing gas flows when the air flow rate is calculated as per API STD 2000.

 

Regards,

Ankur.