RSICC CODE PACKAGE PSR-407

1. NAME AND TITLE

IMPORTANCE: FTA Basic Event & Cut Set Ranking.

2. CONTRIBUTORS

Lawrence Livermore National Laboratory, Livermore, California, through the Energy Science and Technology Software Center, Oak Ridge, Tennessee.

3. CODING LANGUAGE AND COMPUTER

FORTRAN IV; IBM370 (P00407I037000).

4. NATURE OF PROBLEM SOLVED

IMPORTANCE computes various measures of probabilistic importance of basic events and minimal cut sets to a fault tree or reliability network diagram. The minimal cut sets, the failure rates and the fault duration times (i.e., the repair times) of all basic events contained in the minimal cut sets are supplied as input data. The failure and repair distributions are assumed to be exponential. IMPORTANCE, a quantitative evaluation code, then determines the probability of the top event and computes the importance of minimal cut sets and basic events by a numerical ranking. Two measures are computed. The first describes system behavior at one point in time; the second describes sequences of failures that cause the system to fail in time. All measures are computed assuming statistical independence of basic events. In addition, system unavailability and expected number of system failures are computed by the code.

5. METHOD OF SOLUTION

Seven measures of basic event importance and two measures of cut set importance can be computed. Birnbaum's measure of importance (i.e., the partial derivative) and the probability of the top event are computed using the min cut upper bound. If there are no replicated events in the minimal cut sets, then the min cut upper bound is exact. If basic events are replicated in the minimal cut sets, then based on experience the min cut upper bound is accurate if the probability of the top event is less than 0.1. Simpson's rule is used in computing the time-integrated measures of importance. Newton's method for approximating the roots of an equation is employed in the options where the importance measures are computed as a function of the probability of the top event, and a shell sort puts the output in descending order of importance.

6. RESTRICTIONS OR LIMITATIONS

None noted.

7. TYPICAL RUNNING TIME

The test cases executed in about 1 second on the IBM370/195.

8. COMPUTER HARDWARE REQUIREMENTS

Execution of the sample problem required 140K bytes of storage on IBM 370.

9. COMPUTER SOFTWARE REQUIREMENTS

OS/370 and a FORTRAN IV compiler are required.

10. REFERENCES

a) Included in documentation:

H. E. Lambert and B.J. Davis, "The Use of the Computer Code IMPORTANCE with SETS Input," NUREG/CR-1965, SAND81-7068 (March 1981).

b) Background information:

H. E. Lambert, "Fault Trees for Decision Making in Systems Analysis," (Thesis) UCRL-51829 (October 9, 1975).

H. E. Lambert and F. M. Gilman, "The IMPORTANCE Computer Code," Preprint UCRL-79269 (March 14, 1977).

11. CONTENTS OF CODE PACKAGE

Included are the referenced document in (10.a) and one diskette containing a self-extracting compressed DOS file, which contains a readme file, a Fortran file and a sample problem.

12. DATE OF ABSTRACT

May 1999.

KEYWORD: PROBABILITY SAFETY ANALYSIS