Assessment of flow characteristics through a grassed canal

摘要

Vegetation is defined as a kind of surface roughness, which reduces the capacity of the channel and retards the flow by causing loss of energy through turbulence and drag forces of moving water. In water channels, vegetation maybe used to stabilize the water surface and prevent erosion caused by concentrated water flow. In terms of channel lining, grass offers the least expensive option and is the most esthetically pleasing. Most previous studies on this subject were conducted using artificial vegetation in laboratory experiments; research conducted in canals with natural vegetation is lacking. In this study, flow characteristics through a natural grassed canal are studied. In addition, the Manning coefficient and specific energy were determined based on field measurements for both grassed and un-grassed canals. The fieldwork was conducted on Ganabia 9B, southeast of El-Mahalla El Kubra, El-Gharbia, Egypt. The average heights of vegetation in the grassed canal ranged from (28 to 100) cm and the values of flow rate range from (0.0504 to 0.1127) m(3)/s. The number of tests is 42 tests for the un-grassed canal and 205 tests for the grassed canal. This study indicated that the Manning coefficient increased with an increase in energy and momentum coefficients. The energy coefficient increased with a decrease in relative specific energy and an increase in grass height. The un-grassed canal exhibited a smaller relative specific energy than the grassed canal. The momentum coefficients for the non-grassed and grassed canals were 1.015-1.517 and 1.003-1.655, respectively. Therefore, the un-grassed canal showed a higher momentum coefficient than the grassed canal at a ratio of 0.264-1.196%. The relative specific energy was calculated in two ways: (i) using the actual energy coefficient values and (ii) setting it equal to one. As such, the error percentages of the specific energy for the grassed and un-grassed canals were 0.000105-0.1652% and 0.0048-0.2194%, respectively. The results obtained herein were compared with those obtained in previous studies, showing good agreement, but differ in the accuracy of predicting values. Three programs, i.e., gene expression programming, Statistical Package for the Social Sciences, and artificial neural network, were employed to create empirical formulas for modeling the Manning coefficient and relative specific energy for the grassed and un-grassed canals. Gene expression programming was found to be the best modeling program.