an artificial neural network for prediction of front slope recession in berm breakwater

Berm breakwaters are used as protective structures against the wave attack where larger quarry materials as armor stone is scarce, or large quarry materials are available but using berm breakwater lowers the costs considerably. in addition, wave overtopping in berm breakwaters are significantly lower than the traditional ones for equal crest level because of the wave energy dissipation on the berm.the most important design parameter of berm breakwaters is its seaward berm recession which has to be well estimated. in this paper a method has been developed to estimate the front slope recession of berm breakwaters using artificial neural networks with high accuracy. four different available data-sets from four experimental tests are used to cover wide range of sea states and structural parameters. the network is trained and validated against this database of 1039 data. comparisons is made between the ann model and recent empirical formulae to show the preference of new ann model.