How to do a Quantitative Risk Assessment (QRA)

QRAs are a very valuable tool to determine the risk of using, handling, transporting, and storing dangerous chemical substances. They allow us to understand the risk to which the employees on-site, nearby population or the environment are exposed, allowing the quantified risk values to be used to decide whether such a risk is acceptable.

QRAs have become very useful, especially to help with urban development planning in many densely populated countries. That is because in many countries, there is very little space left for new constructions, and QRAs give a clear indication of which areas are more suitable for it taking into account their proximity to industrial facilities (see blog post on Societal Risk maps.

QRAs also help demonstrate the risk caused by the activity, providing the competent authorities with relevant information to enable decisions on the acceptability of risk-related developments on-site (or around the establishment). They can also help understand the risk to which employees, the environment, and company assets are exposed and make cost-effective decisions to manage the risks at an industrial facility.

Unfortunately, risk assessments and qualitative or semi-quantitative risk analyses are often referred to as QRAs, as these studies include the generation of a risk matrix. However, a true QRA for a process plant is a complex and extensive study that involves consequence modelling, probability data, vulnerability data, local weather, terrain conditions, and potentially population distribution. QRAs have many useful applications, but only when performed correctly, providing traceable and transparent reporting while following a consistent procedure.

In this blog post, I will explain the main steps required to perform a QRA to evaluate all potential loss of containment scenarios involving the release of dangerous chemical substances in process, chemical and petrochemical facilities.

What is a QRA?

A Quantitative Risk Assessment (QRA) is a tool to quantify the risk generated by an activity, industrial site or area compromised by multiple industrial sites. It can focus on “internal” on-site or “external” off-site risks. The latter includes the risk to which the surrounding population is exposed.

After performing a QRA, risk values can be evaluated to determine whether the risk can be considered acceptable. The risk acceptance criterion is usually referred to as ALARP (or “As Low As Reasonably Practicable”), which is a concept that can mean different things for different people. Individual Risk and Societal Risk are very commonly used to quantify risk. However, the degree of acceptability of those risk values usually depends on local regulations and governmental requirements.

If the risk is concluded to be unacceptable, then risk reduction measures need to be implemented. Once the risk reduction measures have been implemented, the risk can be re-evaluated to determine if the new calculated risk can be considered acceptable.

RISKCURVES as a QRA tool

Gexcon’s comprehensive QRA tool, RISKCURVES, can help you perform advanced QRAs to evaluate the storing, handling, and transporting of dangerous chemical substances in process, chemical and petrochemical facilities. This software tool presents calculation results in a range of ways, including (but not limited to):

• Individual Risk or Location-specific Individual Risk (LSIR) (by means of iso-risk contours)
• Consequence Risk (by means of iso-risk contours)
• Societal Risk (by means of an f-N curve and Societal Risk Maps)
• Potential Loss of Life (PLL)
• Risk ranking reports

QRA procedure

The typical procedure to perform a QRA consists of the following steps:

1. Identify relevant activities, units, and processes
2. Define Loss of Containment scenarios (LoCs)
3. Assess the consequences (both effects and damage) for all LoCs
4. Assess the failure frequencies and probabilities of all LoCs
5. Calculate and present risk (e.g., IR contours, SR f-N curve, etc.)
6. Evaluate and analyse risk

The first three steps above correspond to the consequence modelling phase of a QRA (which are the initial steps), whereas the remaining steps correspond to the risk modelling phase. RISKCURVES can be used as a comprehensive QRA tool that can take care of both the consequence and risk modelling steps. However, EFFECTS can naturally be used independently as a consequence modelling tool.

Let me explain all these steps in more detail.

Identify relevant activities, units and processes

First, we need to identify all the relevant activities for the risk assessment.

Examples of relevant activities are:

• Storage of LPG
• Transport of LNG through a long pipe
• (Un)loading of ammonia

Since the total number of installations in an establishment can be very large, and not all installations contribute significantly to the risk, including all installations in a QRA is often not worthwhile. That is why the Purple Book contain information about how to use the sub-selection method to determine which installations substantially contribute to the risk caused by an establishment and should, therefore, be included in a QRA.

The main advantage of applying this sub-selection method is that it works very well as a simplified screening tool to rank risk. However, it may lead to discarding an activity because it does not lead to consequence effects off-site, while it could lead to significant on-site risk. An example of this is scenarios that don’t create external risks but may still lead to substantial costs due to loss of production capacity or unacceptable risks to personnel. That is why this method should be used only for QRAs focused on “external” risk.

Define Loss of Containment scenarios (LoCs)

Then we define the loss of containment scenarios. For example:

• Catastrophic rupture of the tank storing LPG
• Full-bore rupture of the pipe transporting LNG
• Release through a hole in the (un)loading hose containing ammonia

The loss of containment scenarios can occur due to the accidental release of a hazardous chemical substance from a piece of equipment (such as stationary tanks, gas cylinders, pipes, pumps, heat exchangers, pressure relief devices, warehouses, etc.) or transport units (e.g., road tankers, rail tank wagons, ships, etc.).

Assess consequences (effects & damage) for all LoCs

The third step of a QRA is to assess the physical effects of the described loss of containment scenarios. Typical physical effects that need to be assessed are heat radiation (for fires), overpressure (for explosions), and toxic concentrations (for toxic clouds). Some examples are:

• Heat radiation levels of the BLEVE that occur upon the rupture of the LPG tank,
• Heat radiation levels of the jet fire that occurs upon the full-bore rupture of the pipe transporting LNG,
• Heat radiation levels of the pool fire that occur upon the generation of a liquid spill when there is a hole in the hose used to (un)load ammonia
• The behaviour of the toxic cloud generated upon the evaporation from the liquid spill generated when there is a hole in the hose used to (un)load ammonia

Additionally, the corresponding translation to damage may be assessed as well.

Assess the frequency and probability of LoCs

The fourth step in a QRA is to assess all the base failure frequencies for all the loss of containment scenarios and all the probabilities of occurrence.

• How many times per year will the catastrophic rupture of the LPG tank occur?
• What is the probability of direct ignition and subsequent BLEVE?
• What is the probability of delayed ignition and subsequent flash fire?
• What is the probability of exposure (wind direction, meteorological statistics, etc.)?
• What is the probability of certain environmental conditions (temperature, solar heat radiation, etc.)?
• What is the probability of the presence of people (determined by population distribution)?

Event trees are beneficial to determine the probabilities of all possible phenomena upon the accidental release of a hazardous chemical substance. Please refer to the blog post “What to expect if there is a leak” for more information about event trees.

Calculate and present risk

Next, we can calculate risk, which is typically the product between the “chance” and the “lethality effects”. The “lethality effects” are the consequences of an undesired outcome (i.e., lethality), which can be calculated as described in the third step of a QRA. The “chance” combines a failure frequency with a probability, which can be calculated as described in the fourth step of a QRA.

RISKCURVES is a very easy-to-use risk modelling tool which quantifies risk as follows.

Individual Risk

Individual Risk is the risk at a location off-site, expressed as a likelihood per year that a person without protection at that location is killed due to an event on-site leading to the release of hazardous chemicals.

The Individual Risk can be calculated by including factors such as:

• The probability of getting killed by exposure is a location-dependent parameter and depends on the lethality footprint of the event of interest.

• The probability of exposure of a human being (due to certain wind direction, being protected, being able to escape, etc.)
• The summation of all the failure frequencies for each scenario
• Probability of an effect (e.g., probability that a mitigating measure fails, probability of direct or delayed ignition, probability of explosion phenomenon due to confinement, etc.)

Individual Risk can be presented as iso-risk contours in RISKCURVES.

Societal Risk

Societal Risk is the cumulative probability per year that at least 10, 100 or 1000 people will be killed as a direct result of their presence within the impact area of an establishment and the occurrence of a chemical accident.

Therefore, Societal Risk is the calculation of the number of victims per event (i.e., per phenomenon, per each possible weather class, per specific wind direction, etc.).

Societal Risk can be presented in RISKCURVES as an f-N curve or as Societal Risk Maps (see previous blog post “Societal Risk Maps as a clear geographical representation of risk”).

Consequence Risk

Consequence Risk is the risk at a location expressed as the likelihood per year that a specific threshold level (i.e., toxic concentration, heat radiation flux, or explosion overpressure level) is being exceeded.

Consequence Risk can also be presented in RISKCURVES as iso-risk contours.

Evaluate and analyse risk

The last step of a QRA process is to evaluate and analyse risk to determine whether risk reduction measures need to be implemented on-site. Examples of possible risk reduction measures that can be concluded from a QRA’s results are:

• Installation of a smaller storage vessel
• Installation of a new storage vessel in a different location as initially planned
• Urban development needs to take place in a different area as initially predicted
• An increase in production cannot take place without posing a threat to the surrounding population

The evaluation and analysis of risk can be done with a risk modelling tool such as RISKCURVES, which will quantify risk by means of Individual Risk iso-risk contours, Consequence Risk iso-risk contours, f-N curve, Societal Risk Maps, etc. RISKCURVES also allows the calculation of risk transects, risk ranking reports, placement of analysis points in the map, etc.

This risk modelling tool gives very quick and clear information on whether the risk acceptance criterion can be accepted to ensure the safety of the surrounding population.

Free viewing demo of RISKCURVES available for download

Do you want to start making the world a safer place?

Get started by downloading the free viewing demo version of RISKCURVES, available for download via the button below.

---