Overview of the necessary requirements of the IGC-Code

IGC Code requirements was developed by the International Maritime Organization (IMO) to ensure the safe transportation of dangerous chemicals by sea. It was first adopted in 1971 and has been updated several times since then to reflect new technologies and safety standards.

СодержаниеСвернуть

Introduction
Review of the IGC-Code Requirements
Relation between Volume Flow of Free Air and Heat Flux into the Tank
Origins of Section the 8.5 of the IGC-Code
Other Standards
Evaluating the Criteria
Heat flux q from the fire
Affected tank area A
Fire factor F

The purpose of this Code is to establish an international standard for the safe maritime transport of liquefied gases and the equipment they must carry to minimize the danger to the ship, its crew and the environment.

Introduction

In accordance with the Terms of Reference for the working group outlined in the introduction to this report, detailed determination was made of the methods allowed and used to size pressure relief valve systems on liquefied natural gas carriers. This review included the following.

A review of all working papers related to the requirements for cargo tank pressure relief valves generated between 1971 and 1974 at IMO Subcommittee on Design and Equipment. This included Sessions DE VII, VIII, IX and DE XII plus a number of intercessional meetings of the special working group that was established to develop the Gas Code, which was replaced in 1983 by the IGC Code, (The International Code for the Construction and Equipment of Ships Carrying Liquefied Gases in Bulk). The final paper of DE IX contained the exact text of 8.5 of IGC as it exists today.

A review of the requirements for sizing of pressure relief valves contained in other standards and codes of practice in use today. This includes the CGA-S.1, API 521, API 2000 and NFPA 59-A as well as the requirements of the US Coast Guard.

A detailed review of work carried out by the National Academy of Sciences at the request of the USCG over a period of six (6) years as presented in the report «Pressure-Relieving Systems for Marine Cargo Bulk Liquid Containers» 1973 (ISBN 0-309-02122-7) Ref-02.

Review of the IGC-Code Requirements

Section 8.2.1 of the IGC requires that each cargo tank with a volume exceeding 20 m³ should be fitted with at least two pressure relief valves of approximately equal capacity, suitably designed and constructed for the prescribed service.

Section 8.5.2 IGC provides the following criteria for determining the combined minimum relieving capacity, under fire exposure, with not more than a 20 % rise in A Study on Support Arrangement of a Cargo Tank for Tank Type A LPG Shipscargo tank pressure above the maximum allowable relief valve setting.

Q = F G A^{0, 82} (m^{3} / s) E q . 1

Where Q is the minimum required rate of discharge in cubic meters per second of air at standard conditions 0 °C and 1,013 bar.

A is the external surface area of the tank in m². This is a stronger requirement compared to the other regulations named that require the wetted surface of the tank to be used (comp. also below).

F is the fire exposure factor for different tank types and defined as follows:

F = 1,0 for tanks without insulation located on deck;
F = 0,5 for tanks above the deck when insulation is approved by the Administration. (Approval will be based on the use of an approved fireproofing material, the thermal conductance of the insulation and its stability under fire exposure);
F = 0,5 for uninsulated independent tanks installed in holds
F = 0,2 for insulated independent tanks in holds (or uninsulated tanks in insulated holds);
F = 0,1 for insulated independent tanks in inerted holds (or uninsulated independent tanks in inerted, insulated holds);
F = 0,1 for membrane and semi-membrane tanks.

For independent tanks partly protruding through the open deck, the fire exposure factor should be determined on the basis of the surface areas above and below deck.

Equation (1) is equivalent to the equations which give a heat flux instead of a volume flow of free air (comp. eg NAS report). The relation to these expressions is given by the Gas factor G.

G = \frac{12, 4}{L D} \sqrt{\frac{Z T}{M}} E q . 2

Where:

L is the latent heat in kJ/kg;
Z is the compressibility factor;
M is the molecular weight of the cargo;
and D is a constant based on the relation of the specific heats of the cargo being vaporised.
T is the temperature in degrees K at the relieving conditions.

Relieving conditions are 120 % of the pressure at which the pressure relief valve is set. This pressure is the MARVS pressure (MARVS = Maximum Allowable Relief Valve Setting).

Prior to the 1993 Edition of IGC-Code, G was given as follows:

G = 177 / L C (Z T / M) 1 / 2

with the latent heat L given in terms of Kcal/kg and Q given in m³/min rather than m³/s. For the 1993 edition of the IGC-Code the units have been changed to ISO standard units.

The current USCG regulations 46 CFR 54.15.2, are in full agreement with the requirements of 8.5 of IGC-Code. The only difference is in the wording, where the ICG-Code specifies approval by the Administration and the CFR substitutes the word «Commandant», as is customary in US regulations incorporating IMO instruments.

As already indicated, a heat flux is hidden in the term for G. This heat flux

$q$

is an assumed heat flux emanating from the fire, which we can call q.

In the IGC-Code, as in API 521, API-2000, NFPA 59A and CGA S.1, this heat flux is equal to 34 500 BTU/(ft² h), which is equal to 108,83 kw/m². Because the surface area A used by the IGC-Code is not in ft² but in m² and A is set to the power of 0,82 there is an additional factor included that is related to the unit transfer from US to ISO units. This factor leads to heat flux q in the IGC-Code, which is:

q = 0, 652 q

q = 0, 652 \cdot 108, 83 = 70, 96 k W / m^{2}

This unit transfer related fact often leads to the misunderstanding that the IGC-Code uses a heat flux lower than the other regulation, which is not the case. Further explanation is given below and in the Annex.

The same value of q is used for all 32 cargos covered by IGC-Code.

Relation between Volume Flow of Free Air and Heat Flux into the Tank

It can be shown that the volume flow of free air, as given by the equation in IGC 8.5.2, is equivalent to the volume flow of gas that is evaporated by the heat flux of a fire.

The relation between the heat flux into a tank and the volume flow of free air, as given in 8.5 of the IGC Code, is in detail derived in the Article of Heller/Ref-01 in the following articles. He started with the general flow equation (Eq. 5 in/Ref-01, p. 129) and demonstrated that the volume flow of free air, as given in the IGC Code 8.5.2, resulted from the comparison of the flow of air at standard conditions to the flow of any gas through the same orifice of a safety valve.

The flow of the gas, which is LNG for the purpose of this report or any gas from the list of gases in the IGC Code, is calculated from the heat flux into the tank as explained below. With Eq. 14 (Ref-01, p. 132), Heller gets the formula as given in the IGC, including the gas factor G.

Origins of Section the 8.5 of the IGC-Code

A detailed review of the database of IMO documents related to the requirements of 8.5 of IGC, submitted between 1971 and 1974 when subcommittee on Design and Equipment was drafting the Gas Code, as well as the DE reports to the Maritime Safety Committee (MSC), has provided the following insights:

1 The earliest version of the Liquefied natural gas (LNG) – The Ideal GasGas Code criteria for sizing PRV appeared in paper DE 75 dated 7 June 1972. It was para 8.4 at that time.

2 The source of this criteria was not provided in the IMO documents reviewed, but it should be noted that, in accordance with the NAS report, it is exactly identical, including the fire factors, to the criteria that existed in the USCG regs 46, CFR 54-15-5(d) and 46 CFR 38.10-15 in 1968. This includes reference to ASME Section VIII, div 1, Appendix J, for values of ′D′, the constant based on the relationship of the specific heats.

3 This same formula appeared in the ABS Rules and the BV Rules for gas carriers from 1965.

4 DE 75 proposed that F = 0,5 for pressure vessel type tanks in a completely enclosed space below deck, and F = 0,2 for non-pressure vessel type tanks in holds. By definitions 1.4.17 of DE/75, an independent pressure vessel tank is what later became known as a Type C Independent Tank. By this definition a Type B tank, such as a Moss tank, where the stresses in the tank are primarily due to tank weight, product weight or sheer stresses, is a non-pressure vessel type independent tank and a factor of F = 0,2 would apply as long as the tank was in a hold, regardless of insulation.

5 Japan paper DE/82, 20 November 1972, proposed that the ASME code should not be quoted in this regulation and defined pressure vessel type and non-pressure vessel type tanks for fire exposure factor as follows;

a. pressure vessel type – MAWP exceeding 0,7 kg/cm²;
b. non–pressure vessel type – MAWP not exceeding 0,7 kg/cm².

By this definition Japan had proposed that a Moss type B tank is a non-pressure vessel type and the fire factor should be F = 0,2 if it is installed in a cargo hold with no required insulation.

6 Norway’s paper DE 83, 27 November 1972, states that the fire factor F will, to a large extent, depend on how the properties of the insulation is affected during the fire. They proposed that the Table should be completely reconsidered, the formulae for ′c′ should be in the code, the factor ′A′ should be reconsidered for different tank types. ′A′ should be the area being exposed to heat during the fire.

7 UK paper DE/85, 27 November 1972, proposed modification to line 5/6. The maximum relieving pressure should not exceed 20 % above the maximum working pressure and under conditions of external fire should not exceed the maximum test pressure of the tank.

8 DE-XII – 3 February 28 1974, contains an early consolidated version of the Gas Code. It reports about the results of the «Ad Hoc Group» work, which met between 28 of January and 1 of February 1974 under the chairmanship of Robert Lakey (USA). At this time the NAS report (ISBN-0-309-02122-7) dated 1973 can be assumed to be known at least to the US participants. No details are given as to discussion of Chapter 8. Nevertheless, Chapter 8 was nearly in its final shape and the fire factors F given in annex II of the report are partly set into brackets to indicate that they are under discussion.

A closer look to the definitions indicated the following:

a. Reference to insulation is only given to tanks with fire resistant approved insulation on deck. This is in line with the example in the NAS report discussed below.
b. A footnote to F = 0,5 for pressure vessels in hold spaces, regardless if insulated or not, gives the recommendation to look for the possibility for the reduction of F with regard to insulation. This means that the proposed text itself does not look into the influence of tank insulation.
c. No further direct reference is made to any influence of insulation. From the above it can be concluded that the fire factors in DE-XII-3 were defined mainly giving credit to the installation of the tanks in the ship and not to the insulation. This is in line with NAS report, which also sees the insulation as only one part to limit the heat flux into the tank.
d. An indirect relation to a credit in case of fire is the reduction of F from 0,2 to 0,1 for inerted holds. This gives credit to the fact that risk of fire in the hold has been reduced due to the lack of oxygen.
e. The nomenclature of the Section is given in the original US units and only partly in SI units.
f. The fire factors in brackets (or partly in brackets) are those for tanks installed below deck. Only the F = 1,0 and the F = 0,5 for tanks above deck seemed to be agreed on completely. These factors are the same in the IGC Code.

9 Germany submitted DE 110 on 9 May 1974, which presented the final version of the fire factors ′F′ in then 8.4.1, which became 8.5 in Res. A 328 (IX) and is exactly as they exist today in the IGC.

10 Comparing DE XII-3, DE 110 and the IGC Code the following conclusions can be drawn:

a. F = 0,5 was agreed to be valid for independent tanks without insulation and not only for pressure vessels (type C tanks) as proposed by DE-XII-3. In addition the effect of an insulation system was included by DE-110 by proposing it for uninsulated tanks. This appears to give credit to the fact that these tanks are independent from the ship structure which is subjected to the fire.
b. F = 0,2, which was completely in brackets in DE-XII-3, was agreed to cover independent tanks with insulation. This appears to be the final conclusion on giving credit to tank insulation. It appears there was a decision not to limit this to insulation systems that are fireproof since the insulation is shielded by the cargo hold, the access to air is limited and, therefore, the duration of a fire will only affect small parts of the insulation so that a large heat increase is needed and melted parts of the insulation still remains on the tank surface.
c. F = 0,1 for insulated independent tanks in inerted holds and for membrane tanks gives credit to the exclusion of any fire by lack of oxygen as already stated in DE-XII-3.

11 Norway submitted DE/144 on 2 July 1974 with the following proposals:

a. Where F = 0,5, add «Foam plastics are not regarded as fire resistant materials».
b. Where F = 0,5, delete the wording within the square brackets (Approval will be based on the use of an approved fire proofing material, the thermal conductance of the insulation, and its stability under fire exposure.)
c. Delete all references to fire exposure factors lower than F = 0,5.
d. Delete footnote and insert «In specific cases consideration should be given by the Administration to fire exposure factors less than specified above, based on a review of insulation and surrounding hull structure».
e. Delete present reference to external surface area of the tank (A) and insert A = external surface of the tank (sq metres). Delete the remainder of this sentence.
f. Insert «In specific cases consideration should be given by the Administration to an A – value less than specified above where fire exposure is not liable to affect parts of the tank surface area».

12 Resolution. 328(IX) dated 17 December 1975 was the first version of IGC Code:

a. Included the equations as proposed in DE/75 with the value of F proposed by Germany in DE/110.
b. It would appear that none of the recommendations of DE/144 were adopted.
c. A metric as well as British version was provided.

In general the review indicated that the fire factors were defined in a well balanced engineering judgment process (as also claimed by NAS to be necessary even if more fire testing data is available). The review of papers submitted and reports from the working group meetings reflect that in the selection of the values of F consideration was given to the suitability of various insulation systems to fire exposure as well as the protection provided by the vessel’s structure, including the tank cover.

Other Standards

It has been noted from the review of the other standards mentioned in item 2 that while the CGA standard, like the IGC Code, provides relief valve capacity requirements in terms of the equivalent air flow through the safety valve, the API codes and the NFPA 59A standard provide requirements for determining the total heat flow into the tank as follows:

H = 34, 500 F A^{0, 82} in U S customary units

H = 71, 000 F A^{0, 82} in S I units

H is total heat flux (Btu/hr or watts) into the tank.

It should be noted that 34,500 Btu/hr ft² is numerically equal to 108,78 Kw/m² but, because of the 0,82 exponent on the area when the conversion is made from US customary to SI units, heat flux value becomes 70,93 Kw/m² (comp. above). The following unit conversion calculation can be applied:

Q = 34, 500 B t u / h r f t^{2} \cdot F {(A)}^{0, 82} 1 K w h = 3 412 B T U a n d 1 f t^{2} = 0, 09295 m^{2}

Q = 34, 500 B t u / h r f t^{2} / 3 412 B t u / k w h \cdot F {(A)}^{0, 82}

Q = 10, 1113 k w / f t^{2} \cdot F {(A)}^{0, 82}

Q = 10, 1113 k w / f t^{2} \cdot \frac{{(0, 09295)}^{0, 18} \cdot {(0, 09295)}^{0, 82}}{(0, 09295) f t^{2} / m^{2}} f t^{2} / m^{2} \cdot F {(A)}^{0, 82}

Q = 108, 78 k w / m^{2} \cdot {(0, 09295)}^{0, 18} \cdot F {(A \cdot (0, 09295))}^{0, 82}

Q = 70, 93 k w / m^{2} \cdot F {(A)}^{0, 82}

Evaluating the Criteria

Since there appears to be no question as to the validity of the constant ′D′ in Eq. (2), a review of the criteria for sizing pressure relief Hazards, Risks and Controls available for Safe Containment of Hydrocarbonsvalves on vessels intended to carry liquefied gases in bulk then calls into question the following, which will be considered independently:

the assumed heat flux q emanating from the fire;
the correct area A of the tank to consider;
finally, the selection of a fire factor F assumed.

Heat flux q from the fire

There is considerable discussion in the NAS report regarding the proper heat flux to be used for sizing pressure relief vales for fire exposure. It is recognized in the NAS report that local heat fluxes from methane fire have been observed to be as high as 90 000 Btu/hr ft². However, it is recommended that 34 500 Btu/hr ft² be used as a good approximation for the average heat flux over the entire wetted surface area exposed to fire. This exact same value is used in the CGA, API codes and NFPA codes as well as in the IGC-Code.

Affected tank area A

In the criteria for sizing pressure relief valves the API and NFPA codes use the wetted surface of the tank. The CGA and the IGC-Code use the full surface area. Considering that LNG is almost always transported in fully loaded tanks the difference is negligible.

The NAS report proposed that in sizing pressure relief valves consideration should be given to effects of supporting structures or other specially introduced features that may serve to confine the fire. It states:

«In the calculation of heat transfer from fire to the cargo containment, special consideration should be given to the design features used to limit the portion of the tank surface directly exposed to the fire, such as bulkheads and weather shields».

This precaution would serve to reduce substantially the effective safety device sizing requirements.

The NAS report proposed that a factor E be used, which is defined as the ratio between the tank area under fire A_e exposure and the total tank area A (E = A_e/A; comp NAS Report page 20, Eq. 3). For this reason E is 1,0 as a maximum if the complete tank is subjected to the fire. So the product of E × A represents the fraction of the wetted area A in direct contact with the fire. It is pointed out that, for marine installations, by taking credit for protective structure afforded by the ship’s structure, the effective area will in almost all cases be less that which would be determined by using the tank area A raised to a power of 0,82.

Pass this test online: Crew Evaluation Test online for seamans about Gas Measurement

It should be noted that the IGC-Code formula given can be derived from the NAS formula by applying A_e = A^0,82. This is demonstrated by a calculation provided in the annex.

Notwithstanding the recommendations of the:

NAS report;
the CGA;
API;
NFPA;

codes as well as the IGC Code use the more conservative approach raising the surface area of the tank to the power 0,82, and use this figure as the part of the tank subjected to the fire (A_e).

For a Moss tank, of approximately 36 m (125 ft) in diameter, using the IGC-Code criteria, the effective area of the tank considered to be exposed to the heat of the fire is approximately 15 % of A.

It should be noted that while the API and NFPA codes use the value of A to the 0,82 factor, areas of the tank more than 9 m above the ground are completely excluded from the total area used to size pressure relief valves. It is stated that tests have shown that the effective heat flux from a pool fire at such elevations is negligible.

The API and NFPA codes also exclude areas of the tank that are protected from direct fire impingement, such as skirts and supporting structure on vertical tanks.

There is no reduction in effective area of the tank provided for in the IGC-Code, which consequently means that the IGC-Code is inherently more conservative in this respect for sizing of pressure relief valves.

Fire factor F

In accordance with the NAS report, if the insulation covers more than, say, 70 % of the tank, the fire factor F can be determined by using the methods in Appendix G. The method of Appendix G is applicable for insulation systems that retain effectiveness at expected high temperatures. Under these conditions the F factor is determined as a direct function of the thermal conductivity of the insulation. For tanks that are 100 % insulated Table G-2 shows F factors below 0,05.

The above is consistent with the API codes and NFPA 59A. API 521 would allow a value of F between 0,026 and 0,3 based on a thermal conductivity of 0,58 W/mK.

The insulation system on an LNG carrier has a typical thermal conductivity of 0,038 W/mK. However, for LNG carriers the value of F for insulated tanks is not lower than 0,1 and that is only applicable to insulated tanks inside an inerted cargo hold as per the above discussion.

Regarding the NAS report in Annex G, it should be noted, that the maximum insulation thickness assumed is 1 inch (25,6 mm). The insulation thickness of an LNG tank on an LNG carrier is at least 300 mm.

Footnotes