CCC-0405 SENSIT.

1. NAME OR DESIGNATION OF PROGRAM:  SENSIT.

3. DESCRIPTION OF PROBLEM OR FUNCTION

SENSIT computes the sensitivity and uncertainty of a calculated integral response (such as a dose rate) due to input cross sections and their uncertainties.
Sensitivity profiles are computed for neutron and gamma-ray reaction cross sections (of standard multigroup cross-section sets) and for secondary energy distributions (SED's) of multigroup scattering matrices. In the design sensitivity mode, SENSIT computes changes in an integral response due to design changes and gives the appropriate sensitivity coefficients. Cross-section uncertainty analyses are performed for three types of input data uncertainties:
(a) cross-section covariance matrices for pairs of multigroup reaction cross sections, (b) spectral shape uncertainty parameters for secondary energy distributions (integral SED uncertainties), and (c) covariance matrices for energy-dependent response functions.

4. METHOD OF SOLUTION

For all three types of data uncertainties, SENSIT computes the resulting variance and expected deviation in an  integral response of interest, based on generalized perturbation theory.

5. RESTRICTIONS ON THE COMPLEXITY OF THE PROBLEM

SENSIT uses one-, two-, and three dimensional arrays to manage the large amount of numerical data involved in the execution. Core storage is reserved for a particular dimensional array only during the time the corresponding data are required to be in-core; at other times, the space is made available for the storage of other data. In order to alleviate bookkeeping chores associated with such dynamic storage allocation techniques, Argonne National Laboratory developed a collection of subroutines, called the BPOINTR package, which is incorporated in SENSIT.

6. TYPICAL RUNNING TIME

A typical execution time on the CDC-7600 for a standard cross-section sensitivity analysis, together with an  SED uncertainty analysis is about 4 seconds CPU (central processor units) time. Here a coupled neutron/gamma group structure of 30 x 12 groups, P3 cross-sections, and 137 spatial intervals was used. A typical vector cross-section sensitivity and uncertainty analysis with 30 x 30 covariance matrices requires about 2 seconds of CPU time per case.

7. UNUSUAL FEATURES OF THE PROGRAM:

8. RELATED AND AUXILIARY PROGRAMS

SENSIT uses angular-flux files from one-dimensional discrete-ordinates codes like ONETRAN, ANISN and DTF and reads multigroup cross-section sets in three different formats.  For covariance data input a convenient ENDF-like input format is adopted which is identical to the output format from the cross- section processing code system NJOY.

10. REFERENCES

- S.A.W. Gerstl, R.J. LaBauve, and P.G. Young:
  "A Comprehensive Neutron Cross-Section and Secondary Energy Dis-
   tribution Uncertainty Analysis for a Fusion Reactor",
   Los Alamos Scientific Laboratory report LA-8333-MS (May 1980).

- S.A.W. Gerstl:
  SENSIT: A Cross Section and Design Sensitivity and Uncertainty
  Analysis Code
  LA-8498-MS UC-32 (August 1980).
- BPOINTER + POINTER Description of Subroutines.

11. MACHINE REQUIREMENTS

Due to the variable dimensioning as described above, large flexibility exists to adjust the storage requirements for a specific problem to the available machine core. On a CDC-7600  all of the 8 sample problems and the realistic applications could be executed within a maximum of 24000 words of fast (SCM) core and 80000 words of extended-core (ECM) memory.

13. OPERATING SYSTEM UNDER WHICH PROGRAM IS EXECUTED:  SCOPE 2.1 ,
IBM/OS 370.

14. OTHER PROGRAMMING OR OPERATING INFORMATION OR RESTRICTIONS:

15. NAME AND ESTABLISHMENT OF AUTHOR

S.A.W. Gerstl
Theoretical Division
Los Alamos Scientific Laboratory
Los Alamos, New Mexico 87545, U.S.A.

File name	File description	Records
CCC0405_01.001	SENSIT MAIN PROGRAM (F4)	697
CCC0405_01.002	COMPUTATIONAL SUBROUTINES	2647
CCC0405_01.003	BPOINTR FORTRAN ROUTINES	1580
CCC0405_01.004	BPOINTR ASSEMBLER ROUTINES	314
CCC0405_01.005	ANISN FORWARD SAMPLE PROBLEM	35
CCC0405_01.006	ANISN ADJOINT SAMPLE PROBLEM	37
CCC0405_01.007	ANISN OUTPUT FORWARD PROBLEM	472
CCC0405_01.008	ANISN OUTPUT ADJOINT PROBLEM	473
CCC0405_01.009	SENSIT S.P. 1 INPUT	42
CCC0405_01.010	SENSIT S.P. 2 INPUT	30
CCC0405_01.011	SENSIT S.P. 1 OUTPUT	330
CCC0405_01.012	SENSIT S.P. 2 OUTPUT	252
CCC0405_01.013	TAPE4 CROSS SECTION FOR S.P.	20
CCC0405_01.014	JCL AND INFORMATION	149