SEDIMENT TRANSPORT IN PRISMATIC AND NONPRISMATIC CHANNELS

Abstract

Understanding the hydraulic aspects of prismatic and nonprismatic reaches of a river channel are important for the design of flood control measures and channel improvement works. The floodplain geometry of a river system may alter over its length due to agricultural and development activities. This can result in the formation of a compound channel, which can either be converging or diverging. Obtaining precise and thorough field measurements in natural rivers under flood flow conditions is challenging. Hence, conducting laboratory experiments is crucial for enhancing the understanding of the flow dynamics in compound channels that include prismatic and nonprismatic floodplains. The investigation was carried out on a nonprismatic compound channel that consisted of a prismatic section followed by a converging section. The research examined five distinct relative flow depths (β), including 0.20, 0.30, 0.40, 0.50, and 0.60. Relative flow depth (β) is a dimensionless parameter and it is defined as the ratio of the overbank flow depth to the total flow depth. As the relative flow depth increases, a larger proportion of the water is found above the main channel, which can significantly affect the flow dynamics. The study conducted experiments on the following four different types of nonprismatic compound channels in a masonry flume: (i) a compound channel with smooth converging floodplains, (ii) a compound channel with rough converging floodplains, (iii) a compound channel with sediment in the main channel and smooth converging floodplains, (iv) a compound channel with sediment in the main channel and rough converging floodplains. The study examines different flow characteristics, including the stage-discharge relationship, the water surface profile, the energy slope, the flow resistance, the distribution of average velocity along the channel length, the depth averaged velocity distribution at different sections, the boundary shear stress distribution, the rate of sediment transport, and the longitudinal bed profile. These characteristics are analyzed in both prismatic and nonprismatic sections of compound channels. In addition to the experimental investigation, the flow rate in nonprismatic compound channels was anticipated using the Gene Expression Programming (GEP) soft computing approach, based on geometric and flow variables. vi The experimental findings revealed that the relationship between stage and discharge follows a power law in smooth and rough floodplains, both with and without sediment, respectively. In compound channels with rough floodplains, the water surface profile decreases as compared to smooth floodplains, due to head loss resulting from converging geometry and rough floodplains. The energy slope is nearly uniform in prismatic sections, whereas it rises in nonprismatic sections. The nonprismatic compound channel with rough floodplains experiences lower velocities, decreasing by up to 5.5% and 16% compared to the nonprismatic compound channel with smooth floodplains, without and with sediment, respectively. Shear stress is significantly higher in nonprismatic compound channels with rough floodplains, increasing by up to 70% and 94% compared to smooth floodplains, without and with sediment, respectively. Additionally, sediment transport rates are up to 10% higher in nonprismatic compound channels with smooth floodplains than in those with rough floodplains. The longitudinal bed profile of nonprismatic compound channels falls within the ripples and dunes category. Sediment deposition occurs in the first half of the converging section, while sediment degradation takes place in the latter half. Manning’s roughness coefficient decreases with increasing longitudinal distance in compound channels with smooth and rough floodplains, without sediment, due to flow acceleration caused by the converging channel geometry. Conversely, in sediment-laden compound channels with smooth and rough floodplains, Manning’s roughness coefficient increases along the longitudinal distance due to additional resistance from sediment transport. The roughness coefficient in nonprismatic compound channels with rough floodplains is up to 7% higher in the absence of sediment and up to 20% higher in sediment-laden conditions compared to smooth floodplains. Using GEP, discharge models were developed based on the experimental dataset from this study and validated with data from previous research. Statistical analysis assured that the proposed GEP model (M7), which incorporates factors such as width ratio, relative flow depth, converging angle, relative distance, Froude number, Manning’s roughness coefficient, and floodplain shear force, outperforms other GEP models, theoretical approaches, and existing methodologies for predicting discharge in nonprismatic compound channels. The model demonstrates superior performance across various statistical measures (R² = 0.990, RMSE = 0.094, MAPE = 3.511, SI = 0.043, AIC = - 632.347). Furthermore, the study proposed a unique equation derived through GEP for predicting discharge in nonprismatic compound channels. The effectiveness of this vii equation was validated using real-world data from the River Main in Northern Ireland. Statistical evaluation (R² = 0.959, RMSE = 2.633, MAE = 2.337, MAPE = 10.467) assures the reliability of the proposed GEP-based equation for forecasting discharge in nonprismatic river systems.