Ochi 



problems of slamming and wetness. The test results leave no doubt about the 

 possibility to apply these methods to ship predictions from now on with full con- 

 fidence. 



In his explanation about the basic concept the author distinguished two 

 problems; that is the prediction of the probability of slamming per cycle and 

 the prediction of the number of cycles per unit time. I would like to make a 

 minor remark on both points. 



In Appendix 1 it is mentioned that it is possible to determine the relative 

 motion at the bow using an accelerometer on the model and a wave probe in 

 front of it. This seems to be a method containing some uncertainties. In the 

 first place it will be necessary to know the smooth water level at the station 

 which is considered critical for slamming. 



In Eq. (5) of the paper this is assumed to be equivalent to the draft. This 

 may be true for a vessel like the "Mariner" at a speed of only 10 knots but at 

 higher Froude numbers a significant difference may be found because of the 

 smooth water bow wave system of the ship. The bow wave of a ship usually de- 

 creases the probability of slamming and increases the wetness. The second 

 objection against the use of a wave height transducer in front of the model is 

 that the bow of a pitching and heaving ship creates an oscillating bow wave 

 which will affect the variance of the relative motions. This will be somewhat 

 less important for fine ships than for full ships. It is therefore concluded that 

 the only reliable way to measure the relative motion is to do so at the critical 

 station which is used for the determination of the probability of slamming per 

 cycle. 



The second remark concerns the use of the second moment of the spectrum 

 in order to obtain the expected number of zero upcrossings. Our experience 

 indicates that the quotient of the spectrum area and the first moment gives a 

 better approximation to the number of zero upcrossings. This is only of impor- 

 tance when the spectrum is not narrow because otherwise the two methods yield 

 the same result. The sea spectrum, however, is not always narrow, for instance 

 when it is desired to simulate a Neumann spectrum which has a width of E = 

 0.815. Using the first moment of the spectrum will result in the prediction of 

 less slams per unit time as is shown in the Table 1 where results are shown 

 from some relative bow motion and wave height measurements. 



The width of the spectrum was estimated with the quotient of the number of 

 maxima and the number of zero upcrossings. According to the results in Table 1, 

 Eq. (7) from the paper is more accurate in predicting the number of maxima 

 than in predicting the number of zero upcrossings. 



592 



