step 10 — verify by physical 

 modeling damages and wave transmis- 

 sion of apparent optimum alternative 



102. This step is necessary to assure that all the compounded assump- 

 tions and analytical inaccuracies are within acceptable limits. This is the 

 case with any analytical design procedure for rubble-mound breakwaters since 

 the empirical relations have been shown to all have limited confidence. Each 

 laboratory test of analytical assumptions applied to a specific design will 

 narrow the confidence limits and improve the reliability of future analytical 

 efforts. A simple proof test with monochromatic waves of varying period con- 

 stitutes a minimum effort in this direction, but it is inadequate to test the 

 accuracy of an optimization procedure such as that proposed above. 



103. The damage function %D(H/U.) must be verified by model testing, 

 including simulation of conditions for incipient motion and a range of more 

 severe conditions. The design conditions should be simulated as accurately as 

 possible in order to include the effects of the numerous physical parameters 

 not explicit in the analytical stability formula that was applied. Wave pe- 

 riod, wave groupiness, storm duration, and static stability, among other fac- 

 tors, should be considered. The static friction factor y from the Iribarren 

 formula (Equation 2) should be measured by sliding tests, as proposed by Price 

 (1979) and Graveson, Jensen, and Sorensen (1980). 



104. The fully described incident' wave conditions cannot be simulated 

 with monochromatic waves. Either an average (for example JONSWAP) spectral 

 shape or one adjusted to be similar to measured spectra for extreme storms 

 near the site can be applied in flume tests of the apparent optimum cross sec- 

 tion. Simulation of a gradual rise to peak conditions, then 1,000 waves or 

 more at the peak (stability criterion) condition, followed by a gradual de- 

 crease of wave energy, would be most useful for tests to verify damage func- 

 tions. A test or tests at the design condition should be followed by tests at 

 more extreme conditions related to the extremal distribution of wave heights 

 (and periods) derived in Step 1. Enough %D{H/H^) points must be measured 



to verify or refine the %I){H/U^) analytical function that was applied in 

 Step 8. A minimum of three tests would be useful, including the ^D(H-) 

 point and at least two more severe conditions. More stability tests should 

 be conducted if agreement with the predicted damage function is not good. 

 Techniques to detect gross rocking motion should be applied in identifying 



67 



