Ochi 



encouraging that in the case of slamming the number of critical factors has been 

 reduced from three in regular seas to two in irregular seas. This favorable 

 result overcomes otherwise unsurmountable calculation difficulties. 



The formula that Dr. Ochi has developed for the probability of slamming 

 rests on the assumption that the relative motion of an arbitrary ship point is a 

 narrow-band Gaussian process. Although the satisfactory correlation of meas- 

 ured and predicted results which Dr. Ochi shows suggests that this appears to 

 be the case, it must be stressed that one cannot a priori assume that the relative 

 motion will indeed be a narrow-band process because the wave motion is not 

 always a narrow-band process, except perhaps for severe sea conditions. Fur- 

 ther, the sum of two narrow-band processes need not necessarily be a narrow- 

 band process itself. Any absolute ship response, however, such as bow motion 

 for example, can safely be regarded as a narrow-band process since the wave 

 motion is mostly wide-band and the ship-system is strongly resonant. 



The next step in Dr. Ochi's analysis follows the approach employed in other 

 engineering fields in that attention is focused on the envelope of the time function 

 rather than its amplitude. In this connection I would like to point out that Eq. (2) 

 can indeed be regarded as the definition of the envelope and which, stated other- 

 wise, essentially regards |r^(t) | as the instantaneous radius of the image point 

 on the phase plane diagram of Fig. 3(b). Dealing with the envelope rather than 

 with the actual amplitude turns out to be very convenient for we can immediately 

 obtain a closed form expression for the probability of slamming, such as Eq. (5). 

 I cannot precisely follow the steps leading to (5), but I assume that Dr. Ochi 

 multiplies the integrated Rayleigh probability density functions for the relative 

 motion and the relative velocity. This is, of course, permissible since both 

 processes are Gaussian and hence linearly as well as statistically independent. 



The author employs the nomenclature "probability of slamming per cycle of 

 wave encounter." For a narrow- band process one may perhaps speak of cycles 

 in an extended sense and even then the precise meaning of cycle is not very 

 clear. But for a wide-band process, like the wave motion record, is it really 

 possible to identify a cycle of wave encounter ? Also, Fig. 3 seems to indicate 

 that slamming only occurs when r = H and f > r*. Is it not more correct to 

 say that slamming can occur as long as r > H and provided that the relative ve- 

 locity has assumed at least its threshold value? 



The paper deals with the wetness problem in a similar and more simplified 

 way and thus provides prediction methods for the propeller emergence problem 

 also. The author has obtained a fascinating result with regard to the distribu- 

 tion of slamming occurrences. It seems to me that utilization of the exponential 

 distribution of the time intervals between slams together with the expected num- 

 ber of slams per unit time as developed by Tick can be used to provide an an- 

 swer in a statistical sense of the average sustained speed for a given ship. Has 

 the author perhaps examined whether the wetness phenomenon is also a sequence 

 of events which are Poisson distributed? 



In conclusion, I would like to raise one further point which was so strongly 

 mentioned by Professor Weinblum in his paper presented during the First 



594 



