Ignition

• Background
• Strategy objectives
• Current knowledge of ignition hazards and modelling capabilities
• Ignition probability estimates & modelling
• Industry practice in assessment of ignition hazards
• Strategy development issues

Background

Ignition causes a release of flammable liquid or gas to become a fire (jet fire, flash fire, pool fire etc.) or explosion.  There are many possible sources of ignition and those that are most likely will depend on the release scenario. Sources of ignition include electrical sparks, static electricity, naked flames, hot surfaces, impact, friction, etc.

Strategy objectives

For gas releases, the timing of any ignition is important in determining the risk and consequences. Early ignition will tend to give rise to a jet fire, whilst delayed ignition can cause an explosion or flash fire, depending on the degree of confinement, congestion, ventilation regime etc. Objectives in this area of fire assessment are:

• To identify areas of uncertainty in the characterisation of ignition sources;
• Identify where the uncertainty is significant in relation to ignition characteristics and offshore risk assessments;
• Initiate research to increase knowledge and understanding in ill-defined areas of ignition characterisation; and
• Promote the use of a consistent methodology in the use of realistic ignition source characterisation.

Current knowledge of ignition hazards and modelling capabilities

Known ignition sources for vapour mixes

• Electric sparks and arcs (from electrical circuits, motors, switches etc.);
• Mechanical sparks (from friction and falling objects);
• Static electrical sparks;
• Lightning;
• Flame (including flaring, boilers, smoking);
• Hot surfaces (including hot work, hot processing equipment, electrical equipment);
• Heat of compression;
• Chemical reactions (eg auto-ignition of oil-soaked lagging on hot piping); and
• High energy radiation, microwaves, RF, etc.

Ignition of non-vapour mixes

There is little data available relating to ignition characteristics of:

Diesel Oil Mists - a dispersion of droplets with diameters <10 microns - such as might be produced when a saturated vapour condenses. It is significant because oil mists may be in a physical form that gives the lowest ignition energies.

Crude oil mists and crude oil-water mixtures - a significant number of UKNS reservoirs now produce high water cut fluids. Ignition of oil-water mixes are not well understood.

Ignition probability estimates & modelling

Ignition sources identification

The generation of ignition probabilities or development of sophisticated ignition probability models is highly dependent on available data.

Friction or the impact energy required to cause ignition - there is renewed interest in friction because of the ATEX Directive, which requires all mechanical equipment for use in flammable atmospheres to be classified in the same way as electrical equipment from 2003.

Ignition of high-pressure releases caused by electrostatic discharges occurring in, or as a result of the release. These events still occur and are very relevant where pipework exists which contains multiphase, high pressure fuels.

Uncertainties in data

Current available ignition data contains uncertainties, specifically:

1. the leak duration, type and dimensions of module, ventilation rate and types of ignition sources;
2. the use of expert judgement in their development to compensate for lack of statistical data; and
3. lack of relevant statistical data for ignition model validation.

Industry practice in assessment of ignition hazards

Safety cases in general do not contain the level of detail for identifying individual ignition sources.

Hazardous area classification is presented as read and some information given about permit to work systems.

Different ignition probability models appear to be in use by different dutyholders.

The Cox, Lees and Ang (1990) model, is used extensively even though Cox, Lees and Ang state that it was speculative only.

Strategy development issues

An updated ignition probability model is being developed based on the latest OIR12 data.

• Disseminate the findings of the UKOOA / OSD ignition probability work;
• Actively support Phase II of the UKOOA project, which proposes a second phase to develop a platform-specific ignition model;
• Raise awareness of ignition sources within the industry since ignition probability is a fundamental input to fire and explosion QRA.