Last time I involved in quantitative risk analysis (QRA) project in LNG regasification facility as supporting process engineer. Actually, I did not involve in QRA project itself, but I involved in back pressure study. In order to get to know about this field, I want to share you some basics quantitative risk analysis techniques as mentioned in several literatures, such as BS EN standards and AiChE publications.
Quantitative Risk Analysis is a procedure to accommodate management with event scenarios, measure of likelihood, and magnitute of potential losses. After risk analysis, we want to manage the risk by several means.
- Reduce the frequency of accidents
- Reduce the consequence of accidents
- Provide insurance
Figure below depicts main quantitative risk analysis techniques that normally used.
Let’s take a look one by one.
The first step in quantitative risk analysis techniques is to identify hazards. This can be done by several approaches, such as a Hazard and Operability Study (HAZOP) or a what-if checklist.
HAZOP is simply to challenge if current detail design have sufficient safeguards to respond deviation in operation. However, HAZOP solely is insufficient for QRA. For example in LNG facility, it is better to break the leak possibility in every pipe or vessel based on hole sizes. This is example of leak likelihood in LNG facility based on its hole size.[table "” not found /]
The most important factor is the duration of leak. For a system that linked leak to flammable gas detector and linked to ESD system, isolation may take place in 3-5 min, depending on number and distribution of detectors. For manual response system, isolation may take place in 10 min or more depending on leak location, leak size, and training of the operator.
Cause and Frequency Analysis
Frequency information is required to calculate consequence. Frequency can be estimated by two methods: general historical leak statistics and modeling approach, such as fault tree analysis (FTA) or event tree analysis (ETA).
General historical leak statistics is used for common items, such as pipes, vessels, pumps. On the other hand, modelling approach is used for items with extremely low failure frequencies, such as LNG double wall cryogenic storage tank, etc.
General historical leak statistics can be obtained from hydrocarbon release data base (HCRD).
Because this topic is a big topic, I will share this topic in several parts.
LNG risk based safety – modeling and consequence analysis – Woodward & Pitblado