IAEA0953 BASACF.

1. NAME OR DESIGNATION OF PROGRAM:  BASACF.

3. DESCRIPTION OF PROGRAM OR FUNCTION

Adjustment of a neutron spectrum based on integral detector measurements and calculation of  an integral dosimetric quantity (integral flux, d.p.a., dose equivalent) and its variance. The program requires measured data (activities and their covariance matrix) and apriori information (spectrum, dosimetry cross sections, integral quantity conversion factor and their covariance matrices). ALL apriori covariance matrices can be read in from a file prepared by some other code or can be generated by means of three different methods (by subroutines included in the program). A subroutine which can normalize the apriori flux to measured data is also included.
The program provides also adjusted domisetry cross sections (with covariance matrix) so that it can be used for an adjustment of cross sections (or response functions of e.g. Bonner balls) by measurements in well-known neutron spectra.

4. METHOD OF SOLUTION

Bayesian theorem on conditional probability applied to linearized relation between activities, dosimetry cross sections and flux. All probability distributions are supposed to be  normal and this supposition leads to minimizing of the same functional as least squares method (STAY'SL). This task is solved by a covariance filter method which avoids any matrix inversion and is  numerically robust and stable.

5. RESTRICTIONS ON THE COMPLEXITY OF THE PROBLEM

This version can use 45 energy groups and 5 detectors and occupies 310 KB of main memory. This restriction can be modified according to available memory.
   The covariance matrix of activities is supposed diagonal.
A solution is produced for any set of input data but in the case of non-consistent data, when measured activities do not match the apriori flux, the solution is not very meaningful.

6. TYPICAL RUNNING TIME

The adjustment of one spectrum with ready covariance matrices takes approximately 90 s. Generation of an apriori covariance matrix of 5 detectors in 45 groups depends on the used method: 9 min for block (STAY'SL) and Gaussian methods and 70 min for AILDV method. This time is relative to the ES-4 computer. Typical time on DEC-VAX 6000 is between 8-24 seconds.

The three test problems included in this package were executed at the NEA DB on a DEC VAX6600 computer in 40 seconds.

7. UNUSUAL FEATURES OF THE PROGRAM

BASACF solves the same problem as  STAY'SL, but on a different theoretical basis, and employs a different numerical algorithm. Compared to STAY'SL it is equipped with additional features: utilization of SAND II type spectrum and cross section libraries, two more methods for generation of covariance matrices, normalisation of the apriori flux and output of the adjusted cross sections and their covariance matrix. These advantages call for larger requirements in memory and computing time in comparison with STAY'SL. The tests do not show any difference of results between these programs.
The main program is a framework of many subroutines and can be easily tailored to a special problem (adjustment of dosimetry cross  sections, special prints etc.). Especially the subroutine which provides all apriori data can be replaced by user written one in case that a different format is used for the apriori data.

8. RELATED AND AUXILIARY PROGRAMS

The program can use SAND II type libraries of spectra and cross sections generated by auxiliary programs CSTAPE and SLTAPE.
Data for STAY'SL auxiliary programs FCOV and XCOV can be used for generation of covariance matrices by the block method.

10. REFERENCES

- Tichy M.
  Bayesian Approach to Neutron Spectrum Adjustment,
  Nucl. Instr. & Meth. in print.

- Milos Tichy:
  The Program BASACF - PC Version, Description and User's Guide
  UDZ 254/88 (March 1990)

11. MACHINE REQUIREMENTS

310 KB of the main memory. Maximum 8 and minimum 3 logical units are used (all files are sequential i.e. magnetic tapes or discs can be employed).

13. OPERATING SYSTEM UNDER WHICH PROGRAM IS EXECUTED:  OS/EC MVT 6.1 M9B; compatible with IBM 360/370 systems.

The tests were performed at the NEA DB on a VAX-6600 computer running under VMS and on a PC/80486 running MS-DOS.

14. OTHER PROGRAMMING OR OPERATING INFORMATION OR RESTRICTIONS

The program can be used without any change on any computer with a Fortran compiler and a sufficient amount or directly addressable memory.

15. NAME AND ESTABLISHMENT OF AUTHORS

          Milos Tichy
          SEVEN
          The Energy Efficiency Centre
          Slezska 7
          CS-120 56 Prague 2
          Czech Republic

File name	File description	Records
IAEA0953_02.001	Information File	102
IAEA0953_02.002	Manual in CHI-writer Format	3929
IAEA0953_02.003	BASACF Control Statements for VMS	79
IAEA0953_02.004	BASADAT Fortran Source	1043
IAEA0953_02.005	BASAUNF Fortran Source	1024
IAEA0953_02.006	TEST.CSI Cross Sections for Sample Problems	37
IAEA0953_02.007	TEST.FLI Flux File for Sample Problems	6
IAEA0953_02.008	TEST1D.DAT BASADAT S.P. 1 Input	24
IAEA0953_02.009	TEST1.BDP BASADAT S.P. 1 Output	54
IAEA0953_02.010	TEST1U.DAT BASAUNF S.P. 1 Input	14
IAEA0953_02.011	TEST1.BUP BASAUNF S.P. 1 Output	233
IAEA0953_02.012	TEST2D.DAT BASADAT S.P. 2 Input	33
IAEA0953_02.013	TEST2.BDP BASADAT S.P. 2 Output	53
IAEA0953_02.014	TEST2U.DAT BASAUNF S.P. 2 Input	14
IAEA0953_02.015	TEST2.BUP BASAUNF S.P. 2 Output	233
IAEA0953_02.016	TEST3D.DAT BASADAT S.P. 3 Input	76
IAEA0953_02.017	TEST3.BDP BASADAT S.P. 3 Output	121
IAEA0953_02.018	TEST3U.DAT BASAUNF S.P. 3 Input	14
IAEA0953_02.019	TEST3.BUP BASAUNF S.P. 3 Output	233
IAEA0953_02.020	Manual in ASCII Format	1288