1                                            Frequency Analysis

This Annex contains the details of the Frequency Analysis for the QRA study of the terminal.

1.1                                      Failure Frequencies

A detailed discussion on all the hazard scenarios identified was given in Annex 13A5. It was concluded that internal process hazards, natural hazards and external impacts are included in the generic failure frequencies. These failure frequencies are summarised in this section.

Table 1.1 lists all the failure frequencies adopted for the various release scenarios. Codes are assigned for various source terms. Refer to Annex 13A7 for code definitions.

Table 1.1        LNG Release Event Frequencies

Code

No. of Items

Length of Section (m)

Hole Size (mm)

Initiating Event Frequency

Unit

Reference

L01

1

450

10

3.00E-07

per meter per year

Hawksley [1]

 

 

 

25

3.00E-07

 

 

 

 

50

1.00E-07

 

 

 

 

100

1.00E-07

 

 

 

 

FB

5.00E-08

 

L02

3

20

10

4.05E-03

per year

COVO Study [2]

 

 

 

25

4.05E-03

 

 

 

50

4.05E-03

 

 

 

100

4.05E-04

 

 

 

FB

4.05E-05

L03

2

300

10

1.00E-07

per meter per year

Hawksley

 

 

 

25

1.00E-07

 

 

 

 

50

7.00E-08

 

 

 

 

100

7.00E-08

 

 

 

 

FB

3.00E-08

 

L04

1

30

10

3.00E-07

per meter per year

Hawksley

 

 

 

25

3.00E-07

 

 

 

 

50

1.00E-07

 

 

 

 

100

1.00E-07

 

 

 

 

FB

5.00E-08

 

L05

2

900

10

1.00E-07

per meter per year

Hawksley

 

 

 

25

1.00E-07

 

 

 

 

50

7.00E-08

 

 

 

 

100

7.00E-08

 

 

 

 

FB

3.00E-08

 

L06

1

N/A

10

1.00E-05

per year

Crossthwaite et al [3]

 

 

 

25

5.00E-06

 

 

 

50

5.00E-06

 

 

 

100

1.00E-06

 

 

 

FB

1.00E-06

G07

1

390

10

1.00E-07

per meter per year

Hawksley

 

 

 

25

1.00E-07

 

 

 

 

50

7.00E-08

 

 

 

 

100

7.00E-08

 

 

 

 

FB

3.00E-08

 

G08

1

108

10

1.00E-07

per meter per year

Hawksley

 

 

 

25

1.00E-07

 

 

 

 

50

7.00E-08

 

 

 

 

100

7.00E-08

 

 

 

 

FB

3.00E-08

 

G09

1

450

10

1.00E-07

per meter per year

Hawksley

 

 

 

25

1.00E-07

 

 

 

 

50

7.00E-08

 

 

 

 

100

7.00E-08

 

 

 

 

FB

3.00E-08

 

G10

1

24

10

3.00E-07

per meter per year

Hawksley

 

 

 

25

3.00E-07

 

 

 

 

50

1.00E-07

 

 

 

 

100

1.00E-07

 

 

 

 

FB

5.00E-08

 

G11

1

720

10

3.00E-07

per meter per year

Hawksley

 

 

 

25

3.00E-07

 

 

 

 

50

1.00E-07

 

 

 

 

100

1.00E-07

 

 

 

 

FB

5.00E-08

 

G12

1

300

10

3.00E-07

per meter per year

Hawksley

 

 

 

25

3.00E-07

 

 

 

 

50

1.00E-07

 

 

 

 

100

1.00E-07

 

 

 

 

FB

5.00E-08

 

G13

1

150

10

3.00E-07

per meter per year

Hawksley

 

 

 

25

3.00E-07

 

 

 

 

50

1.00E-07

 

 

 

 

100

1.00E-07

 

 

 

 

FB

5.00E-08

 

G14

1

20

10

4.05E-03

per year

COVO Study

 

 

 

25

4.05E-03

 

 

 

 

50

4.05E-03

 

 

 

 

100

4.05E-04

 

 

 

 

FB

4.05E-05

 

P15

1

1

10

1.00E-04

per year

COVO Study

 

 

 

25

1.00E-04

 

 

 

 

50

1.00E-04

 

 

 

 

100

1.00E-04

 

 

 

 

FB

1.00E-05

 

P16

2

1

10

1.00E-04

per year

COVO Study

 

 

 

25

1.00E-04

 

 

 

 

50

1.00E-04

 

 

 

 

100

1.00E-04

 

 

 

 

FB

1.00E-05

 

P17

5

10

10

3.00E-07

per meter per year

Hawksley

 

 

 

25

3.00E-07

 

 

 

 

50

1.00E-07

 

 

 

 

100

1.00E-07

 

 

 

 

FB

5.00E-08

 

P18

5

10

10

3.00E-07

per meter per year

Hawksley

 

 

 

25

3.00E-07

 

 

 

 

50

1.00E-07

 

 

 

 

100

1.00E-07

 

 

 

 

FB

5.00E-08

 

P19

10

10

10

3.00E-07

per meter per year

Hawksley

 

 

 

25

3.00E-07

 

 

 

 

50

1.00E-07

 

 

 

 

100

1.00E-07

 

 

 

 

FB

5.00E-08

 

P20

1

150

10

3.00E-07

per meter per year

Hawksley

 

 

 

25

3.00E-07

 

 

 

 

50

1.00E-07

 

 

 

 

100

1.00E-07

 

 

 

 

FB

5.00E-08

 

P21

1

300

10

1.00E-07

per meter per year

Hawksley

 

 

 

25

1.00E-07

 

 

 

 

50

7.00E-08

 

 

 

 

100

7.00E-08

 

 

 

 

FB

3.00E-08

 

P22

1

1

10

1.00E-04

per year

COVO Study

 

 

 

25

1.00E-04

 

 

 

 

50

1.00E-04

 

 

 

 

100

1.00E-04

 

 

 

 

FB

1.00E-05

 

T23

3

N/A

Rupture

1.00E-08

per tank per year

“Purple Book” [4]

 

1.2                                      Event Tree Analysis

The frequency of various outcomes following a loss of containment event is estimated using an event tree model. The various outcomes considered include pool fire, jet fire, flash fire and vapour cloud explosions for liquid releases; jet fire and flash fire for continuous gas releases and fireball and flash fire for instantaneous gas releases. Event Tree Analysis is used to describe and analyse how an initiating event may lead to a number of different outcomes, depending upon such factors as the successful implementation of the various emergency response measures and relevant protective safety systems in place.

A generic event tree used for this study is shown in Figure 1.1. The contributing factors taken into account in the event trees are discussed below.

Figure 1.1       Generic Event Tree

Detection & Shutdown Fails

Immediate Ignition

Delayed Ignition (1)

Vapour Cloud Explosion

Delayed Ignition (2)

Event Outcome

 

Yes

 

Yes

 

 

 

 

 

 

Pool fire/ Jet fire

 

 

 

 

 

 

 

 

 

 

 

No

 

No

 

Yes

 

Yes

 

 

Vapour cloud explosion

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

No

 

No

 

 

Flash fire over plant area

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Yes

Flash fire full extent

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

No

Unignited release

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Yes

 

 

 

 

 

 

Pool fire/ Jet fire

 

 

 

 

 

 

 

 

 

 

 

 

 

No

 

Yes

 

Yes

 

 

Vapour cloud explosion

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

No

 

No

 

 

Flash fire over plant area

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Yes

Flash fire full extent

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

No

Unignited release

 

 

 

 

 

 

 

 

 

 

Detection and Shutdown

For loss of containment events from piping and equipment, it has been assumed that detection and shutdown would occur 90% of the time (based on safety integrity level 1 for emergency shutdown systems which has an associated probability of failure on demand of 0.1).

As discussed in Annex 13A7 on the consequence analysis, if detection and shutdown is successful, a 2 minute release is assumed based on the Emergency Shutdown Device (ESD) provisions in the design (Annex 13D). For shutdown failure, a 10-minute release is assumed. The exception to this is the unloading arms for which 2-minute and 30s releases are considered for isolation failure and isolation successful. The release duration does not have a major influence on the hazard distances determined from dispersion modelling, but slightly different ignition probabilities are assumed for these two cases.

Immediate Ignition

Immediate ignition of an LNG release would result in a pool fire, a jet fire or a fireball (for instantaneous gas releases). For a liquid release under pressure, a jet fire is produced. For a non-momentum liquid release, the liquid is assumed to spill onto the ground producing a pool fire. Gas releases are all pressurised releases and ignition would result in a jet fire. For instantaneous gas releases following a rupture failure, a fire ball is assumed to occur.

In the event of non-ignition, a cloud of natural gas would be formed by the gas release or evaporating liquid pool. A flash fire would occur if this cloud were subsequently ignited.

Delayed Ignition

If immediate ignition does not occur, the dispersing cloud of natural gas may subsequently be ignited. Two delayed ignition scenarios are considered. The first, “delayed ignition 1”, takes into account the possibility that ignition could occur within the plant area due to the presence of ignition sources on site. The second, “delayed ignition 2”, assumes ignition occurs after the cloud has dispersed to its full (steady state) extent.

Delayed ignition for an LNG storage tank failure was treated a little differently given the much larger scale of the release. Vaporisation from the liquid pool was observed to be highly transient in nature. The liquid pool expands to its maximum extent after several minutes and then begins to shrink again as the LNG pool “dries up”. The vapour cloud was observed to expand rapidly with the initial pool expansion. Once vaporisation diminishes, however, a sizable cloud of gas within the flammability limits remains and is convected downwind, gradually shrinking as it goes. Delayed ignition 1 was therefore assigned to the cloud at its maximum footprint area, while delayed ignition 2 was applied to the remnants of the cloud at the maximum downwind extent. Different ignition probabilities were also assigned to LNG tank release (Section 1.3).

If delayed ignition does not occur, the vapour cloud disperses with no effect.

Vapour Cloud Explosion

If a delayed ignition occurs within the plant area (delayed ignition 1), the possibility of an explosion occurring within the congested space of the process area is considered.

1.3                                      Ignition Probabilities