The t/d = 0.61 viewport with its crazing cracks is shown in 

 Figure 5. These cracl<s are much smaller than the ones in Figure 4 and the 

 cracl< band is much narrower. The corner did not deform as much as in the 

 previous case. 



In the most conservative design, t/d = 0.92, there were no cracks 

 visible anywhere. In fact, the corner rounding even appeared to be at a 

 minimum. 



Displacements. Figure 7 contains the absolute displacements of the 

 three pairs of experimental viewports which are presented in this case to 

 provide the designer with a good feel for maximum viewport displacements. 

 The pressure and temperature cycling periods are also shown in Figure 7. 

 Table 2 gives the percentage of the total displacement due to creep. 



To simulate a dive, the pressure was increased at a rate of 50 psi/min, 

 and the cooling of the system was initiated. It should be noted that at the 

 operating depth or highest pressure, the temperature is the lowest. The total 

 time period for a typical cycle was 600 hours. As was expected, the least 

 conservative design, t/d = 0.45, had the highest absolute displacement 

 (Figure 7) and the highest percentage of creep (Table 2). 



Analytical 



Analytical results consist of pertinent plots and data extracted from 

 the finite element analysis results. The data are divided into two areas: 

 (1) stresses and (2) displacements. Both results demonstrate the variations 

 in structural response due to changes in the following parameters: t/d ratio, 

 included angle, and boundary condition. 



Stresses. Figures 8 through 13 include both a contour and surface 

 plot of the effective stresses for each of the 24 computer runs; four t/d 

 ratios, three included angles, and two boundary conditions were used. The 

 stress contour plots provide visual quantitative data, whereas, the surface 

 plots provide visual qualitative data. Figures 8, 9, and 10 present the results 

 for the fixed boundary condition, and Figures 11, 12, and 1 3 present the 

 results for the free boundary case. In all of the plots, the stress concentration 

 is apparent at the corner of the low-pressure face. For either the fixed or free 

 boundary condition, the strength of the viewport is seen to increase with an 

 increase in either the t/d ratio and/or the included angle. This agrees with 

 data developed in Reference 3. 



14 



