The SET-1 test replicated the LOT SPL-19 test at a 1:10 Froude scale. Damage 

 trend refers to the increasing cumulative damage with increasing wave height. 

 Fitted to the data for both tests are curves of the form 



D- = aN^ (4) 



where D' is the dimensionless damage, Ng the dimensionless wave height stabil- 

 ity number, and a,b the dimensionless regression coefficients. Figure 5 shows 

 that the regression curves fit the data well, particularly at low levels of 

 damage, and provide a convenient method of defining the damage trend. Also 

 shown in Figure 5 is the zero-damage level used in this study (i.e., D' = 2.0). 

 D' = 2.0 is about the lowest level of damage that can be consistently detected 

 in the inherent scatter in the survey data. 



The two tests in Figure 5 had a relative water depth of d/gT^ = 0.0144, 

 where d is the water depth in the tank, T the wave period, and g the 

 acceleration of gravity. In comparing damage trends, both the prototype and 

 small-scale tests were grouped by relative depth (see Table 1) to eliminate 

 the possible influence of wave period effects (Ahrens and McCartney, 1975) . 



Curves of the form of equation (4) were fitted to the model and prototype 

 data, and the following two parameters were chosen to characterize the damage 

 trend (Table 1): the stability number, N^., for D' = 2.0 which characterizes 

 the zero-damage level, and the regression coefficient b (eq. 4) which charac- 

 terizes the rate of increase in damage with increasing wave height. The 

 parameters Nz and b are tabulated and grouped by relative depth in Table 1 

 to facilitate comparison of small-scale and prototype values for similar wave 

 conditions. 



IV, COMPARISON OF MODEL AND PROTOTYPE DATA 

 1. Damage . 



The values of Nz and b from Table 1 are plotted versus d/gT^ in Figure 

 6 to demonstrate the influence of both the scale effects and the wave period 

 effects on N^ and b. The figure shows that the small-scale tests had lower 

 values of N2 and generally lower values of b than the prototype tests with 

 similar wave conditions. This finding indicates that damage is initiated 

 earlier in the small-scale tests than in the prototype tests but proceeds at a 

 slower rate, with respect to increased wave height. The convergence in the 

 damage trends typical of small-scale and prototype tests can be seen in Figure 

 5. The regression curves cross; however, the actual data indicate that while 

 the damage levels in the small-scale tests may approach those of the prototype, 

 they do not surpass them for similar values of the dimensionless wave height. 

 Figure 7 is similar to Figure 5 except it shows all the data for tests where 

 d/gT^ = 0.0144, which includes the data in Figure 5. The small-scale and 

 prototype data fields overlap somewhat, but the crossover suggested by the 

 regression curves in Figure 5 does not occur. For the tests where d/gT = 

 0.0264 and 0.0065, there is more overlap or convergence of small-scale and 

 prototype data fields than shown in Figure 7; for tests where d/gT = 0.0037 

 there is no overlap and little convergence in the damage trends. The reason for 

 the convergence of damage trends is unclear, but it may reflect the influence 

 of breaker characteristics or may be caused by the size of the data set and the 

 inherent scatter in the data. Convergence in the damage trend indicates a 

 reduction in scale effects from the zero-damage level. 



16 



