This study presents a new explicit empirical formula to estimate wave transmission on Cubipod Homogeneous Low-Crested Structures (HLCS) under depth-limited breaking wave conditions. The formula was derived using Artificial Neural Networks (ANN) to identify and quantify the influence of nineteen candidate explanatory variables on the squared wave transmission coefficient, (Formula presented). A total of 210 two-dimensional physical model tests conducted at the Universitat Politècnica de València (Spain) were used to calibrate the formula. The dimensionless crest freeboard using the nominal diameter (Rc/Dn50) and the dimensionless incident wave height at the structure toe using the water depth (Hm0,I/hs) were identified as the most relevant explanatory variables. A new two-variable formulation with 3 fitting-parameters was found to estimate the proportion of transmitted energy, (Formula presented), with a coefficient of determination R2 = 0.89. The proposed formula was also applied to an external dataset of experimental tests on Cubipod HLCS previously reported in the literature. The results demonstrated a significantly better agreement than existing empirical formulas confirming the robustness and applicability of the new formula. The proposed formula is a reliable and easy-to-apply new tool for the preliminary design of emerged and submerged undamaged HLCS in depth-limited breaking wave conditions. The new explicit formula is particularly suited for low-crested structures aimed at combining coastal protection and ecosystem enhancement, such as artificial reefs in coral environments.
