Original Date: 04/26/1999
Revision Date: 01/18/2007

Probabilistic risk assessment is a method used to calculate an overall system risk by combining high level event probability distributions with probability distributions from each lower level item or process. Although complicated and laborious, this powerful tool provides accurate results for quantifying the degree of risk. In 1996, NASA recognized the need for such a tool to assess overall system and upgrade risks of the space shuttle. Under contract to Marshall Space Flight Center (MSFC), the University of Maryland developed the software for the Quantitative Risk Assessment System (QRAS).

MSFC uses the QRAS to assess the reliability of the space shuttle and its major components; help perform trade-off evaluations; rank space shuttle failure modes; perform sensitivity analysis; assist in other analysis efforts; and evaluate proposed space shuttle upgrades (e.g., propulsion element). The QRAS is particularly useful for large, complex systems because it arranges data in a hierarchical, tree-like structure. The hierarchy format documents the relationship between elements, components, and failure modes or basic event. Next, a functional event sequence diagram (FESD) is developed for each basic event, showing where each process is inspected, tested, or evaluated. Probability parameters are inserted into each process in the FESD, along with the appropriate justification for using the probability selected. A powerful capability of the QRAS is that probabilities can be entered as time based, and various distribution functions can be included. This approach allows users to evaluate risk over a specific time period as well as obtain confidence levels on the resulting system risk.

Concurrent with the above procedures, an event probability distribution is inserted into the model. The event probability distribution defines failure probability during operation and is based on flight test data, probabilistic structural models, and general engineering judgement. The QRAS then calculates the resulting system level risk based on the input data using Monte Carlo simulation techniques. Since the risk is expressed as a probability density function, a confidence level can be associated with the end result. The QRAS also operates as a database which allows users to input the justification and source of each risk probability used in the model. The hierarchical format lets users manipulate the information to examine the lower levels of the system architecture.

The QRAS calculates an overall system risk by combining the probability of occurrence of all system failure modes. This tool’s windows-based, user friendly design makes it advantageous over traditional probabilistic risk assessment methods. MSFC uses the end results to project system life, estimate inspection intervals, and determine flight readiness support. Repetitive iterations of the risk assessment can also be used to perform trade studies by comparing either two design concepts or a new proposed design against a baseline. Although specifically designed for NASA’s Space Shuttle program, the QRAS can be used on almost any program where probabilistic risk assessment is needed.

For more information see the Point of Contact for this survey.