Stage-Discharge Modelling of Meandering Compound Channels with
Differential Roughness

Abstract

Laboratory experimentations concerning stage-discharge and bed shear stress distribution have been carried out in two sets of meandering compound channels with crossover angles 110_ and 60_ constructed by ‘sine-generated’ curves over a flume of width 4 m. The 110 channel has a very high sinuosity of 4.11, much higher than the available maximum sinuosity of 2.54 in literature. The bed roughness in the floodplain and main channel of the 110 meandering channel is each varied twice in order to obtain a combination of three differential roughness conditions. Variation in bed material is undertaken, not just for the floodplain as is usually provided in literature but also for the meandering main channel.
Velocity profiles across the meander path of the 110_ meandering channel are observed, to analyse the effect of different bed roughness on the profiles. Bed shear stress distribution at the apex section of the 110_ channel is examined for the three different roughness combinations. Shear stress along an entire meander path for the floodplain condition of60_ channel is also analysed. The boundary shear stress study illustrates the position of maximum shear along the apex section as well as across the meander path. These variations are observed for different flow depth. The percentage of shear force shared by the main channel and floodplains is studied with respect to the effect of flow depth and the change in bed roughness which aids in determining apparent shear stress along with the momentum transfer between the main channel and floodplains.
Discharge prediction methodologies have been evaluated in this thesis based on
three different approaches; i.e. by determination of the overall roughness coefficient in meandering compound channel; by evaluation of the percentage of shear force shared by the floodplains, and by balancing the forces acting on the different zones of a compound meandering cross-section. Stage-discharge modelling of a meandering compound channel is complex as compared to a straight compound or simple meandering channels. So an attempt has been made to model the overall roughness coefficient in meandering compound channels. The overall roughness coefficient is observed to be dependent on parameters such as relative flow depth, width ratio, Reynold’s number, Froude number, sinuosity, roughness ratio between the floodplain and the main channel and bed slope. Two methodologies are adopted for the prediction of overall roughness coefficient, i.e. by dimensional analysis and the other by a computational methodology, Multi-Gene Genetic Programming (MGGP). The dimensional analysis model provides a dimensionally homogeneous equation undertaking different parameters while the MGGP model provides an expression for Manning’s n with excellent predictability. These models have been compared with other methods of discharge prediction and even validated with river datasets.
The percentage of shear force carried by the main channel and floodplains for the case of meandering compound channels with percentage of apparent shear force at the various interfaces is studied. A model is developed to further predict the percentage of shear force carried by the floodplain which is employed in an improved channel division method to predict the discharge capacity for the case of meandering compound channels with differential roughness.
The momentum transfer coefficient at the interacting interfaces of a meandering compound channel is analysed. The meandering compound channel is divided into four zones with respect to the interfaces, i.e. the horizontal interface between the lower main channel and the floodplain; and the two vertical interfaces at the meander belt width dividing the outer floodplains from the belt. Force balance in these zones is analysed and, the zonal velocity is obtained, which helps in the evaluation of the individual discharge of the zones.
Momentum transfer coefficient of value, _T =0.01 is proposed for meandering compound channels with respect to the datasets undertaken. Procedure for the calibration of _T is also illustrated for the case where stage-discharge data is known. MGGP technique is used to model an expression for _T , for the undertaken datasets from literature. Within the range of datasets used, this model expression is proposed to predict the momentum transfer coefficient for meandering compound channels for different geometric and hydraulic parameters which in turn has been used to predict the discharge capacity as well as the discharge distribution between the channel subsections.
Popular conveyance prediction models along with the proposed model are applied to the experimental datasets to substantiate predictability of the MGGP model. Validation of the model is done for the new experimental data as well as in natural rivers to justify its predictability in field study.