45 



1 . . 1 . 



, I . . > 1 



40 



■ y\ 



Test Conditions: Significant Wave Height, 16 fl 

 Head Seas 

 Ship Speed, About 6 knots 



35 



- i/ h 



\ 





1 30 





^ 



\ 



V 



--Theocetical Rayleigh Distribution, E = 4.2lf— ^ j 

 /—Experimental Histogta,Ti 



^ 25 





1 



/ 



The ship experienced about 460 variations per hour. 

 \ The test sample consisted of 230 stress variations. 



1 20 



■ 1 











Q- 



N 15 





J 



— 





■ 



10 



-j 







\| 



5 



I 





X^_ " 





L 







J 



, 





stress in Main Deck Amidship, Kips per square inch 

 (1 Kip = 1000 pounds) 



Fig. 10 Distribution of Variation in the Longitudinal Stress, Main Deck, 

 Amidship, Aircraft Carrier 



Atlantic weather stations I and J, see Fig. 12. 

 Darbyshire reported the maximum wave height 

 measured each time observations were made at 

 3-hr intervals, while the ship was at sea. The 

 visual observations made by weather observers, 

 on the other hand, are reported as the significant 

 wave height. It will be of interest to compare 

 the visual observations with the measurements 

 that have been obtained with the wave meter. 

 If the hypothesis is accepted that the short-term 

 distribution of wave height follows the Rayleigh 

 distribution, then the maximum wave height and 

 the significant wave height are related by a con- 

 stant factor. Thus the distributions of maxi- 

 mum and significant wave heights should both be 

 of the same type, log normal in this case, and 

 differ only in their mean values. The U.S. 

 Weather Bureau data indicate that the standard 

 deviation'^ of the Log^ (significant wave height) 

 is 0.62 at Station J and 0.61 at Station I as 

 compared to a value of 0.57 for the Loge (maxi- 

 mum wave height) for the measurements at 

 Stations I and J reported by Darbyshire, Fig. 15. 

 The wave-meter data have been fitted with a 

 log-normal distribution on the assumption that 

 the distribution of the maximum wave heights is 

 log normal. Fig. 15. The experimental data 

 indicate excellent agreement with the fitted 



i given here refer to wave heights measured 



distribution, well within the accuracy of the wave 

 measurements. The latter fact, together with the 

 good agreement between the standard deviations 

 of the distribution of significant and maximum 

 wave heights, support the hypothesis that the 

 distribution of wave heights may be approxi- 

 mated by Rayleigh and log-normal distributions 

 for the short and long term respectively. It is 

 realized that the visual estimates of significant 

 wave heights obtained by the U.S. Weather 

 Bureau may not be accurate estimates of the 

 mean value of the third highest waves. But as 

 long as these estimates are proportional to the 

 actual significant wave height, the form of the 

 cumulative distribution (log-normal type) and 

 its variance remain unchanged; only the mean 

 value would change since log (constant) x is equal 

 to log (constant) plus log x. 



Fig. 14 shows the distribution of the predomi- 

 nant wave periods observed at Ocean Station B. 

 Here again the log-normal distribution appears 

 to provide a reasonably good fit. If these wave 

 periods are converted to wave lengths by means 

 of the formula,'^ L = 5.12 P, where L is the 

 length of the wave in feet and T is its period in 

 seconds, which is applicable to gravity waves 

 in deep water, it may be seen that the wave 

 length will follow the log-normal distribution in- 



" The numerical valu 

 type of distribution. 



of the constant (5.12) does not affect the 



16 



