By assuming that a polycrystalline aggregate deforms with the formation
of a particular kind of stress-induced martensite, this program calculates the
"modified Taylor factor" for each grain orientation and its average. See refs.
[1] and [2] for the foundation of the calculation. The output of this program
is useful in predicting the mechanical properties of shape memory alloys[2]
and in obtaining the insights of what is going on in polycrystalline shape
memory alloys when they deform pseudoelastically[3,4,5].
This release consists of three programs:
1) TSIM
Main program to do the above. This reads the crystallographic
parameters of the martensite; the habit plane normal, the shear
direction and the magnitude of shear, and outputs the list of modified
Taylor factors and the set of variants formed for orientations evenly
distributed within a unit stereographic triangle. This also outputs
the average of the modified Taylor factors. This program assumes the
cubic symmetry of the parent phase.
2) CTREC
This program reads the output of TSIM and makes distinction of
the yield surface corner types for individual orientations based on
the list of the activated variants.
3) CONTOR
This program also reads the output of TSIM and draw the contour map
of the magnitude of the modified Taylor factor within the unit
stereographic triangle.
These programs are written in FORTRAN77 and, therefore, supposed to be
portable to any machine which executes FORTRAN77 programs except for the
following fact: The programs CTREC and CONTOR make graphic outputs as a way
of summarizing the calculated results. In doing this, the programs call the
routines in the graphic library of the Tohoku University Computer Center,
LIB/DRAFLIBV. Although they are quite elementary and thus ubiquitous, their names and argument listing must be changed to those appropriate in particular sites.
Other than the purpose of the presentation, these graphic outputs are especially useful in checking the validity of the calculated results. That is,
because of certain weakness in the coding in the linear-programming solver
used in the program TSIM, there is a possibility in some special situations
(that is, in some special orientations and with some particular crystallo-
graphic parameters) that the significance of figures is run out during the
repetition and thus invalid outputs are made. The wrong output as such
stands out as singular points in the corner type map made by CTREC and the
contour map made by CONTOR.
Although the coding of TSIM is made in single precision, it must be
compiled with the double precision option. In some cases, the single precision
may suffice; in that case, however, the figures for the virtual zero and the virtual equality must be re-adjusted appropriately.
On the compilation of TSIM, some warning messages are expected but
they are safely ignored. The calculation is set to made on 861 orientations.
This takes CPU-time somewhere around 25 seconds with a mainframe computer,
NEC/ACOS2000 which should be among the fastest in numerical calculation except
for the supercomputers. I suspect, therefore, if it is practical to try to
run this program on PC's or even on common EWS's.
I hope that these programs help you. Although these programs are offered
without any warranty and thus to be used with your responsibility on the
validity of their outputs, any questions on these programs and troubles
encountered in running them are welcome.
REFERENCES
[1] N. Ono and A. Sato: Trans. Japan Inst. Metals, 29(1988), 267.
[2] N. Ono, A. Sato and H. Ohta: Mater. Trans. JIM, 30(1989), 756.
[3] N. Ono: Mater. Trans. JIM, 31(1990), 381.
[4] N. Ono: Mater. Trans. JIM, 31(1990), 855.
[5] N. Ono: ICOMAT-92, Monterey, CA(1992).
--
Noboru Ono * $@>.Ln(J $@M[(J
Dept. of Machine Intelligence and Systems Engineering, * $@5!3#CNG=9)3X2J(J
Faculty of Engineering, Tohoku University. * $@ElKLBg3X9)3XIt(J