0.01 













T 













^ 



r 







0.1 

 0.2 



as 





Lower 95% Confidence Limit- 

 Upper 95% Confidence Limits 











































A 































/^/ 













2 



5 



10 

 15 

 20 



30 

 40 

 50 

 60 

 70 

 80 





















k^ 





























'} 



'/ 



























^ 



r 



























/^ 































/ ' 

























































& 





























& 



r 



























A 



f 



























c 



Y 





























/ 







1 









o Bow Moving Down 





/ 





























90 



95 



98 

 99 



























































































































1.5 2 3 4 5 6 78910 15 20 30 40 50 

 Pitch Angle in degrees. Double Amplitude 



•o 1200- 



o 800- 



Bow Moving Down 



6 



Level 



—1 1 1 r- 



4 8 12 16 



Double Amplitude in degrees 

 Pitch Angle 



Fig. 19 (continued) 



ured by means of electric wire strain gages. The 

 measured data are given in Table 9 and the log- 

 normal distribution which was computed for the 

 truncated data is shown in Fig. 18. The fractiles 

 corresponding to the larger, relatively rarely ex- 

 perienced stress variations do not check the fitted 

 distribution too well. However, inasmuch as 

 the sampling period did not cover an entire 

 winter season it can be seen that a single storm 

 can account for a disproportionate number of 

 the relatively rare, high-stress values. 



Aircraft Carrier 



The stress data were obtained during the ship's 

 winter season '55- '56 operations in the North 

 Atlantic Ocean. The strains were measured by 

 means of electric strain gages mounted on the 

 keel amidships. The measured data are tabu- 

 lated in Table 9 and the corresponding log-normal 

 distribution is shown in Fig. 18. The fractiles 

 of the experimental distribution agree unusually 

 well with the computed log-normal distribution, 

 well within the experimental accuracy of the 

 measurements. Thus it may be concluded, for 

 this ship, that the hull-girder stress or bending- 

 moment variations can be approximated by a 

 log-normal distribution. 



USCGC Unimak 



The stress data were obtained during the 

 vessel's operation as a weather ship in the winter 

 season '54-'55. The log-normal distribution, 

 fitted to the truncated experimental data, is shown 

 in Fig. 18. Excellent agreement between the 

 plotted experimental fractiles and the computed 

 log-normal distribution is indicated. 



It is concluded on the basis of this analysis of 

 extensive measurements of hull-girder stresses 

 obtained on seven ships of widely different size, 

 type, and service that the long-term distribution 

 of hull-girder stress may be approximated by a 

 log-normal distribution. 



Ship Motions 



The long-term distributions of ship motions 

 have been studied for the USCGC Casco and the 

 £550 Asheville. 



USCGC Casco. Cumulative distributions of 

 heave, pitch, and roll motions taken from refer- 

 ence (1), are shown in Fig. 19. It is apparent 

 that the logarithmically normal distribution fits 

 the data quite well. 



Esso Asheville. Tables 11 and 12 give the 

 variations of pitching and heaving accelerations, 

 classified according to magnitude and frequency 

 of occurrence for both the ballast and load con- 

 ditions of the ship. The Fisher method for treat- 



31 



