1. Figure 4: Pressure contours over the VLS forebody for MOO = 0.73. Simulation uses a four-block grid. Figure also shows block interface lines. except that it is topologically divided into four blocks. Fig. 3 shows these blocks separated solely to facilitate the reader's understanding of the block breakup. Simulations for freestream Mach numbers of 0.5 and 0.73 were performed with the four-block grid. Results for the MOO =0-73 case are shown in Fig. 4 in terms of pressure contours. The Reynolds number for this calculation was 20 million, based on the afterbody diameter. One can observe from Fig. 4 that there is very good continuity of pressure contour lines across block interfaces. Incidentally, the block interfaces with their two overlapped points are also indicated in Fig. 4. The good continuity characteristics of the property contour lines across block interfaces is an indication of the correction and efficiency of the communication between computational blocks.
2. It is interesting to observe that a first analysis of the bottom plot in Fig.11, i.e., the one corresponding to the MOO = 0-94 case, would seem to indicate the existence of discontinuities in the contour lines across the interface between blocks 1 and 3. However, a closer look at this interface reveals that there are no such discontinuities. This can be verified considering the inset in Fig. 11 and the corresponding enlarged plot in Fig.12. A similar behavior is observed for the other contour lines at this interface. The two vertical lines in Fig.12 are exactly the block interface lines. Therefore, this figure is clearly indicating a very smooth crossing ofthe interface bythe pressure contour lines. The reasons for the apparent break in the overall smoothness of the pressure contours for the MOO -0.94 case are not fully understood at the moment. Nevertheless, the most important aspect, from the point of view of the present discussion, is that the communication of information across block boundaries is not causing any error. This is clearly demonstrated by the previous figures, including Fig. 12 for the M^ = 0.94 case.
3. The present work was partially supported by Conselho Nacional de Desenvolvimento Cientifico e Tecnologico, CNPq, under the Integrated Project Research Grants No. 530109/93-0 and 522413/96-0. The availability of computational resources at Centre Nacional de Supercomputagao, CESUP-UFRGS, and at Nucleo deAtendimento em Computac.ao deAlto Desempenho da COPPE/UFRJ, NACAD-COPPE/UFRJ, is alsogratefully acknowledged.