QUANTITATIVE RISK ANALYSIS – HOW TO PREVENT INCIDENTS LIKE THE EXPLOSION IN WEST, TEXAS
The incident in Texas prompted us to think about understanding the risks associated with operating a facility handling hazardous materials, and how to prevent these types of incidents. (Current estimates reveal that the facility contained as much as 54,000 pounds of toxic anhydrous ammonia, and reported to the Texas State Health Services Department that it possessed 270 tons of ammonium nitrate.)Since the explosion has also likely made many manufacturers re-evaluate their own facilities, we thought providing a starting point for this evaluation would be helpful. A quantitative risk analysis (QRA) is a good place to start because it attempts to quantify risk from potential incident scenarios so that you can focus on those with the highest risk. It does this by evaluating the consequence of a particular event vs. the probability of that event occurring, to determine the overall level of risk. This type of evaluation is key for low probability, but high consequence events. A QRA allows a company to identify high risk scenarios, and allocate resources towards preventing these.QRA Steps:
Then identify the consequences that result if the hazards are released upon the plant or the surrounding community. There are a number of tools readily available to conduct consequence analysis, such as PHAST .
Once a consequence has been established, the second key part of the QRA is needed: determining the probability of an event. This probability can be determined through various methods, such as fault tree analysis, or layer of protection (LOPA) analysis. Executing a fault tree analysis can be prohibitively expensive, and the results depend on the team recognizing all of the failure modes of complicated systems. For this reason, many organizations are using LOPA to estimate the consequences of a loss of containment of a hazardous material.
A company can then analyze its risk based on the position of various events on a consequence vs. probability plot, which describes the exposed risk. The organization can then make the determination whether the risk of a particular event is too great based on a number of factors, and implement changes or improvements to bring that risk in line with the organization’s goals.
Like any method for estimating probabilities, the limitations of the QRA include such factors as predicting human behavior during stressful conditions (although one can apply a factor for this.) The reliability of any safeguards is also important to consider. This is highly dependent on the safety culture of an organization, because maintenance and testing of safeguards is a key part of making sure these work when called upon.
Organizations such as the Center for Chemical Process Safety (CCPS at www.aiche.org) provide a wealth of reference materials and training courses that can help an organization determine how best to perform a QRA and implement the changes needed to mitigate the risk associated with a facility, specific operating unit, or a process.