Prediction of cross-shore sandbar volumes using neural network approach


Demirci M., Unes F., AKÖZ M. S.

JOURNAL OF MARINE SCIENCE AND TECHNOLOGY, cilt.20, ss.171-179, 2015 (SCI İndekslerine Giren Dergi) identifier identifier

  • Cilt numarası: 20 Konu: 1
  • Basım Tarihi: 2015
  • Doi Numarası: 10.1007/s00773-014-0279-9
  • Dergi Adı: JOURNAL OF MARINE SCIENCE AND TECHNOLOGY
  • Sayfa Sayıları: ss.171-179

Özet

Correct estimation of bar volumes, wave height, wave period and median sediment diameter is crucial for the designing of coastal structures and water quality problem. In this study, bar volumes caused by cross-shore sediment transport were investigated using a physical model and obtained 64 experimental data considering the wave steepness (H (0)/L (0)) and period (T), the bed slope (m) and the sediment diameter (d (50)). Artificial neural network (ANN) and multi-linear regression (MLR) are used for predicting the bar volumes. A multi layer perceptron is used as the ANN structure. The results show that the ANN model estimates are much closer to the experimental data than the MLR model estimates.