Risk Analysis

From Riskan Wiki
Jump to: navigation, search

Risk analysis is a generic term used to define a broad group of probabilistic analysis in very distinct areas, such as financial market and accountability. Here we will look only at a specific group of these analysis that are used to prevent or estimate the risk of engineering failure in chemical, petrochemical and nuclear plants.

Historical origins

Since the chemical industries and the industrial facilities in general reached a certain level of complexity, it became clear that, in such complex systems, failure is inevitable. Although inevitable, it is a main goal of each plant to reduce the probability of occurrence and the damage caused by that occurrence. We can trace some great events that triggered the development of the Risk Analysis as we know it.

  • 1960 – Space program - First concerns about risk estimate and measurement
  • 1974 - Flixborough disaster - Accident in UK triggers regulation by HSE
  • 1975 – NRC – Reactor Safety Study - First reliability analysis, using fault tree to estimate de expected frequency of a undesirable event
  • 1979 – Three Mile Island - Great nuclear accident that raised the requirements of risk analysis in nuclear plants
  • 1981 – Fault Tree Handbook
  • 1983 – PRA Procedures Guide

Each of this events lead to advancements in the techniques and methods of estimating the risk of a complex facility.

Types of risk analysis

The goal of all types of risk analysis is to minimize the risk in the target plant. All the techniques and methods are divided in two major groups:

  • Qualitative risk analysis - Analysis that aims to minimize the risk, but do not estimate an specific numeric value to the risk, frequency or damage
    • PHA - Preliminary risk analysis
    • HAZOP - Hazard and operability studies
    • FMEA - Failure modes and effects analysis
    • What-If - What if technique
    • HA - Historical analysis
  • Quantitative risk analysis - Analysis that estimates the frequency, the damage or both in order to define an acceptance criteria or to minimize the risk
    • Fault Tree - Estimates the frequency of an undesired event
    • Vulnerability Analysis - Estimates the damage caused by an undesired event
    • QRA - Quantitative risk analysis per se, names the complete analysis that estimates the expected frequency and damage of a list of events, thus estimating the risk of a plant or facility

Quantitative risk analysis - QRA

QRA is the name of the most complete analysis that are required by several regulatory agencies for licensing installations of new plants or to authorize facilities. The basic step in these analysis are:

  • Select the most relevant accident scenarios for the target plant
  • Estimate the expected frequency of occurrence of each of the accident scenarios
  • Estimate the damage caused by each accident scenarios in each possible atmospheric condition and in each possible development
  • Calculates the risk by multiplying the frequency by the damage
  • Compare the resulting risk with an acceptability criteria
  • Proposes measures for risk reduction and calculate the risk reduction for each measure