RISK MANAGEMENT FOR LNG IN PORTS

Olga Aneziris

24th September 2020

1

OUTLINE

• Introduction

• Risk Definition

• Basic steps for Quantitative Risk Assessment

• Hazard Identification

• Methods for Hazard Identification

• Accident Sequences

• Methods for Accident Sequence Modelling andQuantification

• Risk Integration

2

LNG CYCLE

3Ref: https://www.fluxys.com/

SHIP TO SHORE LNG BUNKERING

4

Ref: http://www.emsa.europa.eu/

WHAT IS RISK?

5

• ”Organizations of all types and sizes face internal and externalfactors and influences that make it uncertain whether theywill achieve their objectives. The effect this uncertainty has onan organization’s objectives is “risk” (ISO 31000 Riskmanagement – Principles and guidelines)

• Risk refers always to the future and it refers to the multitudeof different possible outcomes each with differentconsequences some of which are undesirable.

WHAT IS RISK?

6

• Combination of probability of occurrence ofharm and the severity of that harm (ISO 18683Guidelines for systems and installations for supplyof LNG as a fuel to ships)

• Risk is now a series of possible consequences andthe associated probabilities

• (𝑝1, 𝑐1), … , (𝑝𝑖 , 𝑐𝑖), … , (𝑝𝑛, 𝑐𝑛)

TYPES OF CONSEQUENCES IN CASE OF ACCIDENTS

7

QUANTITATIVE RISK ASSESSMENT

9

ANSWERS TO THE FOLLOWING QUESTIONS

• WHAT CAN GO WRONG?

• HOW OFTEN THIS CAN HAPPEN?

• WHAT ARE THE POSSIBLE CONSEQUENCES?

• HOW FREQUENT THESE CONSEQUENCES ARE?

BASIC QUESTIONS OF QRA

10

• WHAT CAN GO WRONG?

• HOW OFTEN THIS CANHAPPEN?

• WHAT ARE THE POSSIBLECONSEQUENCES?

• HOW FREQUENT THESECONSEQUENCES ARE?

ACCIDENT SEQUENCES

CONSEQUENCE ASSESSMENT

PROBABILITY OFCONSEQUENCES

ACCIDENT FREQUENCY

MAJOR STEPS FOR QRA

11

HAZARD IDENTIFICATION

12

Plant Familiarization

Hazard source identification

Initiating events

• Description of installation

• Document review, interviews

• Systems, operations

• LNG quantities, storage and transportation

• HAZOP

• FMEA

• Master Logic diagrams

• Accident Databases

13

HAZOP process is a qualitative technique which identifies:

• DEVIATIONS from intended design/ operation

• CAUSES of deviations

• UNDESIRABLE EVENTS

• SAFEGUARDS to prevent causes

Recommends ACTIONS in design to improve safety and operability.

HAZOP is based on use of guide words which question how operating

conditions might not be achieved. It is carried out by a multi-

disciplinary team during a set of meetings.

HAZOP (HAZARD AND OPERABILITY ANALYSIS)

DEVIATION CONSEQUENCE SAFEGUARDS ACTIONS

HAZOP RESULTS

15

HAZOP results are recorded in special TABLES.

KEYWORD DEVIATION CONSEQUENCE SAFEGUARDS ACTIONS

High Level in tank

Overfilling of tank

LT 901ALAH 902

Stop loading

Example: Loading a tank

MASTER LOGIC DIAGRAMS

16

It is a Logic Diagram and starts with a Top Event,considered undesirable.

The top event is “Loss of Containment”.

It continues decomposing into simpler events untilevents challenging safety systems are identified.

These are the initiating events.

MASTER LOGIC DIAGRAM FOR LOSS OF CONTAINMENT

17

LOSS OF

CONTAINMENT

STRUCTURAL

FAILURELOSS OF

BOUNDARY

CONTAINMENT

BYPASS

EXCESS

TEMPERA-

TURE

LOW

LEVEL

LOW

TEMPERA-

TURE

DIRECT

PRESSURE

INCREASE

FROM GAS

COOLING

MALFUN-

CTION

EXCESS

HEATOVRFILLING

INTERNAL EXTERNAL

FIRE

CHEMICAL

INCOMPATIBLE

MATERIAL

RUN AWAY

REACTIONCOMBUSTI-

ON

ROLL OVER PRESSURE

SHOCK IN

HOSE

INTERNAL

PRESSURE

INCREASE

FLOODINGSNOW, ICE SEISMIC HIGH

WINDS

NATURAL

PHENOMEN

A

SUPPORTS

FAIL

EXTRA

LOADS

HIGH

TEMPE-

RATURE

UNDER-

PRESSURE

VIBRATION EXTERNAL

LOADING

OVERPRESSU-

RE

CONTAINMENT

OPENED

CONTAINMENT

OPENS

CORROSION ERROSION

LAYOUT OF LNG STORAGE IN A PORT

18

Decomposition of port in areas where HAZOP orMLD may be applied

Areas of the port

• Hose-loading arm• Ship receiving LNG• Buffer ship• Storage tanks

19

HAZARD IDENTIFICATION IN PORTS HANDLING LNG

20

LNG storage tank

• Boiloff removal malfunction, during unloading or storage

• External fire in storage tank area

• Level rise beyond safety height, or overfilling

• Extra loads

INITIATING EVENTS OF A PORT HANDLING LNG Loading arm section

• Excess external heat in jetty area

• Water hammer in loading arm, due to inadvertent valve closure

• Inadequate cooling of loading arm

• High winds during loading-unloading

• High winds

• Extra loads

Sendout section

• Inadvertent closureof valve in send out

Truck

Buffer Ship

Ship receiving LNG

SAFETY SYSTEMS

22

SAFETY SYSTEMS FOR INITIATING EVENT:Level rise beyond safety limit

Manual stop of Loading - Communication

Emergency Shutdown System (ESD)

Pressure Safety Valves (PSV)

EVENT TREE

23

PSV release

PLANT DAMAGE STATES

24

Definition

• A plant damage state uniquelycharacterizes the installation-dependent conditions of releaseof the hazardous substance.

• Accident sequences resultinginto the same conditions ofrelease are grouped into groupseach corresponding to aparticular plant damage state.

DAMAGE STATES

25

LNG Storage Tank• Tank rupture owing to overpressure, overfilling

Loading arm- Unloading section• Pipe rupture• Hose rupture

Truck• Truck rupture• BLEVE of tank

Fueled Ship• Tank rupture owing to overpressure, overfilling

Buffer Ship• Buffer ship rupture

BOWTIES AND SAFETY BARRIERS

26

TARGETBarrier

PREVENTION MITIGATION

ACCIDENTInitiating event

Barrier

28

BOWTIE FOR LOSS OF CONTAINMENT (LOC) FROM LOADING ARM WITH LNG

29

LOC OF PIPE WITH LNG- LEFT HAND SIDE

31

Data and Parameter Assessment

Estimate frequencies of the initiating events, component

unavailability, probabilities of human actions

Data Gathering and parameter value assessment (HSE, NFPA,

RIVM, OREDA)

Accident Sequence and Plant damage State Quantification

Calculate frequency of occurrence of accident sequences and

Damage states using the event trees of the logic model.

FREQUENCY ASSESSMENT OF DAMAGE STATES

EVENT TREE QUANTIFICATION

32

10-1/y

10-3

1.55 10-2

10-2

1.55 10-7/y

FAULT TREES

34

“AND” GATE - “OR” GATE

T: NO WATER AT POINT E

FAULT TREES

35

TOP EVENT: Failure of high level control

MINIMAL CUT SETS OF FT “FAILURE OF HIGH LEVEL CONTROL”

38

TERM (#) MINIMAL CUT SETS PROBABILITY (UNAVAILABILITY)

1 ESD FAILURE LEVEL 1.22E-02

2 COMM 1.13E-03

3 LEVEL TRANSMITTER 2.64E-04

4 AIR COMPRESSOR 3.6 E-04

5 PR. VALVE FAILS 6.84E-04

6 LOSS OF AC DIESEL GER 1.8 E-06

TOTAL 1.46E-02

FAULT TREE –EVENT TREE QUANTIFICATION

39

• In case of event tree quantification a big fault tree isconstructed with a top event the “damage state”, an ANDgate and inputs: Initiating event and all other failures in theET sequence.

HIGH LEVEL IN TANK

FAILURE OF MANUAL STOP

FAILURE OF LEVEL CONTROL

FAILURE OFPSV

LNG TANK RUPTURE OWING TO HIGH LEVEL

FREQUENCIES OF PLANT DAMAGE STATES AND RELEASED LNG

40

TASKS OF QUANTITATIVE RISK ASSESSMENT

41

METHODS Event trees Fault trees Bowties

QUANTITATIVE RISK ASSESSMENT (CONSEQUENCE ASSESSMENT)

42

CONSEQUENCE ANALYSIS

43

CONSEQUENCE ASSESSMENT

44

• DOSE ASSESSMENT: The integrated, over time, exposure of anindividual to the extreme phenomenon generated by theflammable material is calculated. This defines the “dose” anindividual receives. Any emergency response plans or othermitigations action are taken into account at this point.

• Consequence Assessment: Appropriate dose/response modelsreceiving as input the dose of heat radiation or overpressurecalculate the probability of fatality or injury of the individualreceiving the dose.

CONDITIONAL ISO RISK CONTOURS

45

Conditional iso-risk contours from plant damage state pressurised LNG tank rupture, and BLEVE (10-2, 10-1, 1)

RISK INTEGRATION AND RISK INDICES

46

• Integration of the results obtained so far, that is combiningfrequencies of various accidents and correspondingconsequences.

• INDIVIDUAL RISK: is defined as the probability of death peryear of exposure to an individual at a certain distance from thehazardous source. It is usually expressed in the form of isoriskcontours.

• GROUP RISK: is a measure of risk for a group of people,expressed as fatalities per year. It is most often expressed interms of the frequency distribution of multiple casualty events(F/N curve).

TOTAL INDIVIDUAL RISK CONTOURS

47

Total Unconditional iso-risk contours (10-7, 10-6, 10-5/yr) for LNG storage plant at port

CONCLUSIONS

51

• QRA estimates the risk level of a port handling LNG (inside andoutside the port).

• Safety of a port is depicted by accident sequences, theirfrequencies of occurrence and the corresponding consequences.

• QRA provides information to decision makers and assists riskreduction strategies in hazardous installations.

• QRA may prevent the occurrence of Major accidents in portshandling LNG and/or limit their consequences.

Olga AnezirisNCSRD DEMOKRITOSAgia Paraskevi Attikis, 15310 GreeceEmail: olga@ipta.demokritos.gr

Thank you for listening. Any Questions?