LOSS PREVENTION

Review unit-wide impacts on closed-drain drums

API 521 standard helps decipher the correct operating pressure for this system
R. MUKHOPADHYAY, Consultant, Bangkok, Thailand

losed-drain drums are generally intended to receive hydrocarbondrained liquids from various upstream sources. These drains may be maintenance drains, continuous-process drains and open-hazardous drains from drip pans. The vent line of this vessel type is normally routed to the low-pressure (LP) cold vent headerultimately culminating into a LP cold vent tip. Liquid from the drum is removed periodically. The drum inbreathes during pumping out and out breathes during liquid inflow through the vent line on top. Design pressure. The question is what should the design pressure be without any liquid seal at the vapor outlet to prevent air ingress? To a process designer, a simple design is very straight forward. The process designer will consider gas blowby scenarios from upstream contributory inlet streams and design the vent line of the closed-drain drum, sufficiently large to cater for the largest inflow rate either from non-continuous maintenance drains or the liquid rate from continuous drains coming through restriction orifices (provided for preventing blow-by scenarios). This would prevent the overpressure generation by allowing adequate out breathing facility in case of large inlet flow to the drum. Based on this simplistic concept, the design pressure of an open-to-atmosphere drum system is set at 3.5 barg =50 psig, as per standard practice. However, most process designers may not be able to explain why a 50-psig design pressure was chosen. Why not another number? To find an answer to why 50 psig was chosen, you have to remember that, apart from inlet streams coming from upstream and creating overpressure pos-

sibility in the drum, there may be another source of over-pressure as well. Since these drums handle hydrocarbon liquid/vapor mixture and are operating at pressure slightly above atmosphere, there is a chance of air ingress to the drum in some circumstances (shutdown or other instances of inbreathing/pump-out, etc.), potentially inducing an explosion inside the drum. Vessels need to be strong enough to withstand such overpressure from the internal explosion. Note, in many cases, the LP vent header is not generally purged to prevent air ingress. Reasons are to minimize waste of hydrocarbon/nitrogen and green house gas emissions. Hence, the designer may need to rethink the vessel's design pressure. The explosion overpressure from such a potential internal explosion (deflagration) becomes a key point in correctly assigning the design pressure of such drums. In the past, simulations of explosion scenarios using sophisticated software for limited volume (30-m cloud) had been done (Fig. 1). These calculations used unconfined-explosion models. Maximum over-pressure generated in unconfined explosions is 2.5-3.0 barg or 40 psi. This result is extended to arrive at the extent of over-pressure, suitable for the lower bound of the entire range. Instead of using any arbitrary number in the hydrocarbon processing industry, the number tends to fit the minimum design pressure of such low-pressure atmospheric vessels (non-purged, non-sealed on vapor side) at a minimum 3.5 barg. This is based on the results predicted from mathematical modeling/simulation from an explosion scenario. Following are specific company standard requirements regarding design pressure of closed-drain drums:

TOTAL'S G S - E P - S A F - 2 2 8 (liquid drainage) states: "The design pressure of the closed-drain drum shall be 3.5 barg or more, in line with API RP 521, to provide minimum resistance to an internal explosion. All facilities connected to a closed drain drum shall be designed at no less than the design pressure of the closed-drain drum." This is an open-ended statement, stating 3.5 barg or more. How much more pressure is not concluded. What is the method used to arrive at this pressure? Shell's DEP 8 0 . 4 5 . 1 0 . 1 0 (pressure relief, flare and vent system design), Section - 4.1.3.2 states: "The knockout drum shall be designed as an ASME pressure vessel with a design pressure of at least 3.5 bar (ga) (50 psig). I f no seal vessel is used, the design pressure shall be at least 7 bar (ga) (100 psig). The minimum design pressure of 7 bar (ga) (100 psig) is specified forfiare knockout vessels so that the vessel will safely withstand the overpressures from an internal deflagration (i.e., flashback)." BP RP-44-1 also mentions a minimum design pressure of 7 barg. A P I - 5 2 1 , S e c t i o n 7 . 3 . 2 . 4 , states: "Most knockout drums and seal drums are operating at relatively low pressures. To ensure sound construction, a minimum design gauge pressure of 345 kPa (50 psi) is suggested for knockout drums in subsonic flare or other low-pressure applications. A vessel with a design gauge pressure of 345 kPa (50 psi) should not rupture if a deflagration occurs. Stoichiometric hydrocarbon-air mixtures can produce peak explosion pressures on the order of seven to eight times the absolute operating pressure. Most subsonic-flare seal drums
H Y D R O C A R B O N PROCESSING MAY 2011 107

LOSS PREVENTION
operate in the range of gauge pressure from 0 kPa to 34 kPa (0 psi to 50 psi). Seven to eight times the absolute working pressure gives the range of explosion overpressure in such drums as: (0+14.7)* 7 psi to (5+14.7) *7; i.e., 102.9 psi - 137.9 psi or 7 barg - 9.5 barg for most of the subsonic flare seal drums/vent knock-out drums (KODs)unpurged/unsealed atmospheric closed drain drums. Analyzing the API 521 standard, starting from a minimum 50 psi design gauge pressure, it states that the vessel would not rupture under deflagrations. However, based on the drum operating pressure, the explosion overpressure can go up to 137.9 psi (9.5 barg) in commonly known cases. Hence, you can't always stick to a single design pressure value of 50 psig (3.5 barg) for all atmospheric drums. Along with a design pressure arrived at based on purely process considerations, an internal explosion overpressure might become significant if the possibility of such deflagration exists. These higher design pressure range values can be justified by the API 521 interpretation, from a mechanical design point of view of carbon steel, which is the most commonly used material for such vessels. Let's look at API 69 (latest edition) clauses to find a relationship. NFPA 69 has taken it a step further to define two different acceptability criteria with respect to vessel deformation/failure that is: 1) Designpressure for a scenario where explosion causes vessel deformation, but no rupture occurs and 2) Design pressure for a scenario where explosion causes no vessel deformation. Obviously, vessels designed for the second scenario will require higher design pressure (may be higher plate thickness). Some clients prefer design 1, but others prefer design 2. Once the design criteria is established and in line with the NFPA the design matches harmoniously with the API-521 interpretation of two design pressure ranges: minimum 3.5 barg and the higher side up to 9.5 bar. This depends on the drum's observed operating pressure. ASME has a method of calculating deflagration containment design pressure for vessels as per Section V I I I , Division 1, of the B o i l e r a n d Pressure Vessel Code. A S M E B o i l e r a n d Pressure Vessel Code defines two terms to explain the previously mentioned "no rupture" and "no deformation" concepts. It defines the ratio of "ultimate stress" and "enclosure material allowable stress" as a dimensionless entity symbolized by F . It also defines the ratio of "yield stress" and "enclosure material allowable stress" as another dimensionless entity symbolized by F . F corresponds to the "no rupture but deformation" concept, whereas F corresponds to the "no deformation" concept.
u y u y

F = 60,000/20,000 = 3.00; carbon steel plate; SA-516 60 K02100; Temperature < 150F F = 32,000/20,000 = 1.6; carbon steel p; SA-516 60 K02100; Temperature < 150F The following excerpt illustrated in Table 1 from ASME Section I I , Part D , to justify the values mentioned previously. Line 22 values from Table 1 are relevant for this case o f F and F (60,000, 32,000 and 20,000). Using the concepts of NFPA 69's standing and following mentioned clauses, the intended range of deflagration design pressures of 3.5 barg to 9.5 barg can be established and verified. Relevant clauses from NFPA 69 are defined: 13.3.4* Given an initial pressure and dimensionless pressure ratio for the potenu y u y

Select 174 at www.HydrocarbonProcessing.com/RS

108

LOSS PREVENTION
T A B L E 1. Allowable stress (ksi), ASME Section II, Part D, USC Units
Line Nominal Product no. composition form 21~ Carbon steel Carbon steel Plate Plate Plate Spec. No. SA-515 SA-516 SA-283 Alloy Type/design/Units Pgrade No. No. 60 60 D K02401 K02100 K02702 1 1 1 Group No. 1 1 Min. tensile Min. yield External strength, strength, pressure ksi ksi chart No. 60 60 60 32 32 33 CS-2 CS-2 CS-2 Max. use temp. 1,000 1,000 700

Notes G13J3 G13J3

< 100F 21.3 21.3 22.0

< 150F 20 20 20.7

< 200F < 250F 19.5 19.5 20.2 19.2 19.2 19.8

22~

2 3 ~ Carbon steel

i
4.0 ^ 3.5 g 3.0

tial deflagration, Pmawp shall be selected based on the following conditions as defined by Eqs. 1 and 2: Permanent deformation, but not rupture, of the enclosure can be accepted. Pmawp > [R(Pi = l 4 . 7 ) - U . 7 ] / ( % F u ) (1) Permanent deformation of the enclosure cannot be accepted. Pmawp > [ R ( P ^ U . 7 ) - U . 7 ] / ( % F y ) (2) where: Pmawp - enclosure design pressure (psig) according to ASME Boiler and Pressure Vessel Code R = dimensionless pressure ratio Pt = maximum initial pressure at which combustible atmosphere exists (psig) Fu = ratio of ultimate stress of the enclosure to the allowable stress of the enclosure according to ASME Boiler and Pressure Vessel Code. Fy = ratio of the yield stress of the enclosure to the allowable stress of the materials of construction of the enclosure according to ASME Boiler and Pressure Vessel Code. 13.3.4.1* The dimensionless ratio, R, is the ratio of the maximum deflagration pressure, in absolute pressure units, to the maximum initial pressure, in consistent absolute pressure units. 13.3.4.2 For use as a practical design basis (since optimum conditions seldom exist in industrial equipment), the value of R is: For most gas and air mixtures, the value of R shall be 9 For St-1 and St-2 dust-air mixtures, the value of R shall be 11 For St-3 dust-air mixtures, the value ofi? shall be 13 13.3.4.3 A value for R other than the values specified in 13.3.4.2 shall be permitted to be used if such value can be substantiated by test data or calculations. 13.3.4.4 For operating temperatures below 25C (77F), the value of R ' shall be calculated for use in Eqs. 1 and 2: R = R [298/ (273+7; )] (3) where: R ' = deflagration ratio adjusted for operating temperature R = maximum deflagration ratio for the mixture measured at 25C (77F) Tj = operating temperature (C)

0.40

s 0.35
;- o.3o | 0.25

2.5

12.0

I- 0.20
! 0.15 \ 0.10 i 0.5 0.00

1 1.5 o
o n. x
m

1.0 0.5 0.0

/ / / /^J J /
1
0.10

50..

* li

- 30-m cloud-CFD Simulation B - Design pressure-time curve No. 3

o n "InnH PFn Cimnla+inn A OU~lll L I U U U U r U OlIIIUIdLIUM M

m

40:

30

few
0.30 0.40 0.50 0.60 Time, seconds 0.70

%
0.80

20 10.

0.00

0.20

0 0 . 9 0 1.00

345-kPa (50-psig) design pressure-time curve and typical model calculations.

13.3.5 The presence of any pressure relief device on the system shall not cause the design pressure calculated by the methods of 13.3.4 to be reduced. 13.3.6* The maximum initial pressure for positive pressure systems shall be as follows: For positive pressure systems that handle gases and liquids, the maximum initial pressure, Pi, shall be the maximum initial pressure at which a combustible atmosphere is able to exist, but a pressure not higher than the setting of the pressure relief device plus its accumulation. For positive pressure systems that handle dusts, the maximum initial pressure shall be the greater of the following two pressure values: Maximum possible discharge pressure of the compressor or blower that is suspending or transporting the material Setting of the pressure-relief device on the vessel being protected plus its accumulation For gravity discharge of dusts, the maximum initial pressure shall be the atmospheric gauge pressure (0.0 bar or 0.0 psi). 13.3.7 For systems operating under vacuum, the maximum initial pressure shall not be less than atmospheric gauge pressure (O.ObarorO.O psi). 13.3.8 Auxiliary equipment such as vent systems, man ways, fittings and other openings into the enclosure, which could also experience deflagration pressures, shall be designed to ensure integrity of the total system and shall be inspected periodically.

Using the previously stated concepts of "permanent deformation acceptable" and "permanent deformation not acceptable" and the concepts of ultimate tensile stress, yield stress and allowable stress of the material of construction (usually carbon steel material) with the proper number crunching, the following calculation support is: Permanent deformation acceptable. R = 9 per NFPA 69 for most gas and air mixtures P i = 0.0 psig (Assuming Atmospheric conditions) Fu = 60,000/20,000 = 3.00 for carbon steel plate SA-516 60 K02100 temperature < 150F P m a a p > [9* (14.7 + 0.0) -14.7]/ [(2/3) * (60,000/20,000)] Pmawp > 58.8 psig for vessel operating @ 0.0 psig (59.25 for 0.1 psig) Matches perfectly with NIOSH's design pressure vs. time curve (Fig. 1). Pmawp >2 8 3 - 8 psig for vessel operating @ 50 psig Pmawp ^ 81.3 psig for vessel operating @ 5 psig. The hydrocarbon processing industry appears to have accepted 50 psig as the design pressure for vessels operating at or near atmospheric pressure. The case valid for permanent deformation accepted but no rupture, in case of internal explosion. " R = 9" given in NFPA for HC. A value of R = 7.8 would give P m a w p = 50 psig. The value of R = 7.8 is in perfect agreement with the statement in API 521 that stoichiometric
H Y D R O C A R B O N PROCESSING MAY 2011

LOSS PREVENTION
hydrocarbon-air mixture can produce peak explosion overpressure on the order of seven to eight times the absolute operating pressure. Permanent deformation not acceptable (preferred by some clients): R = 9 per NFPA 69 for most gas and air mixtures P i = 0.0 psig (assuming atmospheric conditions) F = 32,000/20,000 = 1.6 for carbon steel plate SA-516 60 K02100 temperature
y

< 150F P > [9* (14.7+0.1) - 1 4 . 7 ] / [(2/3) * (32,000/20,000)] Pmawp > 110.25 psig for vessel operating @0.0 psig (111.1 for 0.1 psig) Pmawp > 532.13 psig for vessel operating @ 50 psig Pmawp ^ 153.4 psig (10.6 barg) for vessel operating @ 5.0 psig; R = 9.0 Pmawp ^ 131.4 psig (9.0 barg) for vessel operating @ 5.0 psig; R = 7.8 Conclusion: This is an unusual case, i f
m a a p

permanent deformation cannot be accepted in case of internal explosion. Usually, after an event like an internal explosion, the vessel integrity would be in question and calls for replacement. However, some clients prefer to have design pressure like this. Hence, for the design concept of the "permanent deformation not acceptable" case, the design pressure of closed-drain drum can be in the range of 9-9.5 barg, in line with the API 521 7.3.2.4 (h) clause and the values mentioned in Shell DEP, TOTALs GS-EP, BP-RP and other reputed company standards. Mechanical design v i e w point. A vessel designed for DP =100 psig will show permanent deformation, but no rupture up to 300 psig (simple ratio of design. Stress vs. rupture stress or UTS). A vessel designed for DP =100 psig will not show permanent deformation up to 160 psig (simple ratio of design stress vs. yield stress). A carbon-steel vessel designed for MAOP= 50 psig and a thickness of !4 in. can be as big as 12 ft diameter. So, do not bother till the vessel diameter exceeds 12 ft. I f the diameter exceeds 12 ft, consider increasing the length, keeping the diameter at 12 ft. Design pressure of atmospheric vessels like closed drain drums (not purged), considering internal explosion case: Min. 50 psig (would not rupture as per API 521 Max. = in the range of 131-153.4 psi (9-10.6 barg), considering a variation of R from 7.0-9.0. It is established with calculation supports and references that 50 psig-153.4 psig is the range of design pressure of the closeddrain drum type atmospheric low-pressure vessels, handling hydrocarbon vapor/liquid, with no liquid seal for the LP vent from the vessel; the usual operating pressure of the vessel varies from 0-5 psig. This is on the basis of deflagration (internal explosion) over-pressure, mentioned in many operating company standards and API 521. HP

(/) _
^^^^^

WHO SHOULD I ATTEND:

OPERATIONS
l Process Superintendent

l CURRICULUM: H
.
S a f e t y

I -Global Best Practices

O
^^^^0 ^
C

p e r s o n n e |

I Asset Protection

I -Operations/Unit Engineer M -Plant Manager Wastewater Manager Operations Manager/ Supervisor

I -Decontamination I Trends+Methodologies
P l a n n i

"g

Logistics

O l

^^.EH t H

^ (Ol

^^^^9

-Maintenance

MAINTENANCE RELIABILITY

H H

I.

I Chemistries I Effluent Management

Economjc Benefjts

^^^^ I Manager/Engineer ^^^H -Turn A r o u n d ^ ^ ^ ^ ^ ^ O^l Manager/Engineer ^ ^^^^ Turn Around ^^/^^^ Planner/Coordina

^^^TM
^T B ?^

ENGINEERS

-Tech Service ^^^LKPP^I I Engineer ( / ) ^ | Process Engin ^^MU (0^1 l-Cpnstruction ^^0^\ O^l Manager ^1 U I^F'1'^ Mjanage

^^^U
^^3fl

^ 1

21

ENVIRONMENTAL
& s A F E T

i k 11

PROFESSIONALS

Rajib Mukhopadhyay
ogy

earned

a BTech d e g r e e i n c h e m i c a l t e c h n o l a n d a BSc d e g r e e i n c h e m i s t r y 17 years o f p r o f e s f r o m t h e U n i v e r s i t y o f C a l c u t t a . He has m o r e t h a n sional e x p e r i e n c e i n oil a n d g a s sector, s p e c i a l i z i n g i n t h e process s a f e t y a n d loss p r e v e n t i o n f i e l d . Mr. M u k h o p a d hyay has extensive h a n d s - o n experience in o p e r a t i o n , c o m m i s s i o n i n g a n d t r o u b l e - s h o o t i n g process operations. He has w o r k e d f o r n u m e r o u s e n g i n e e r i n g a n d production organizations.

Select 175 at www.HydrocarbonProcessing.com/RS

110