Erosion as an aggravating act causes loss of soil productivity and good yields of agricultural lands. Erosion in water supply-related construction structures causes damage to these structures and designs. Erosion damage is not limited to agricultural-related structures. It can also affect many projects such as beaches, ports, water supply structures, etc. Estimation of erosion deposition volume is of great importance. Machine learning and data mining methods can be used to estimate the deposition process. In this study, a method based on machine learning tools to estimate sedimentation was presented. To do this, after sedimentation data were collected from Minab and Abnama stations, Pre-processing was performed on the data and the data were normalized. Then the normalized data were sent separately to the estimation models based on the artificial neural network. The models used in this study are; Multilayer artificial neural networks and wavelet neural networks. To eva‎luate the proposed method, the multi-piece validation method with the number of pieces equal to 5 was used. The reason for using this validation method was the small number of samples available. The results were eva‎luated based on R2 criteria, correlation, root mean square error and normalized root mean square error. Studies show that the amount of sediment (expected range) covers a wide range from 1 to 500 thousand units. More data is needed to achieve better results. However, the models used in the proposed method with the small amount of data they had, They provided good performance. In general, the wavelet neural network performed better than the normal neural network.

3/1/2021 12:00:00 AM

397

NE88

حليمه يار