Sediment transport in irrigation canals.

The world population is rapidly increasing and is expected to double to about 10 billion by the year 2050. To support an increasing population in terms of food sufficiency, more and more water will be required. Irrigation is the most critical component of the modern package of inputs to effect high crop production. Irrigation has been the largest recipient of public agricultural investment in the developing world. Hence, continued investment in irrigation along with reforms in institutional arrangements for management of water are very much necessary to ensure adequate supply of food. Simultaneously, water requirements for other purposes, domestic, industrial and hydropower will steadily increase as well. Under this competing situation irrigation will have to become increasingly more efficient in the future.Improved management and operation activities must be implemented to prevent recurring degradation of irrigation projects. Clogging of turnouts and reduction of the conveyance capacity of canals by siltation are problems frequently met in irrigation systems. Annually, high investments are required for rehabilitation of irrigation systems in order to keep them suitable for their purposes. New development of irrigation projects or upgrading of existing schemes will require a better understanding of the sediment transport process under the prevailing flow conditions in irrigation canals. Applicability of the existing sediment transport relationships on irrigation canals has to be better understood. In this way predictions on sediment deposition in irrigation canals will be more reliable.The present study is focused on sediment transport in irrigation canals which may have a serious impact on the operation and maintenance activities. The design of the canal system either should be based on the transport of all the in the water present sediment to the fields or to places in the canal system, where the deposition can be removed with least costs. Sedimentation should be prevented in canals and near structures, as it will hamper and endanger a proper irrigation management. In the design and operation of irrigation canals with sediment-laden water several aspects related to irrigation criteria and sediment transport must be taken into consideration. The need for conveying different discharges at a required water level to meet the irrigation requirement and at the same time to convey the sediment load with a minimum deposition and/or erosion in the canal system should be the main criteria for the canal design. Irrigation canals are generally designed upon the assumption of uniform and steady flow.It is also assumed that there exists an equilibrium situation where the sediments entering into the irrigation canals will be transported without settling or erosion. However, uniform and steady flow are seldom found in reality. In the operation of an irrigation system the flow is predominantly non-uniform. While the sediment transport is highly dependent on the flow conditions it is obvious that the sediment transport capacity of the canals varies as well.Development on sediment transport in open channels have been mainly focused on river engineering. Even though certain similarities between rivers and irrigation canals are present, these concepts are not fully applicable to irrigation canals. A description and analysis of the sediment transport concepts under the specific conditions of irrigation canals will contribute to improve the understanding of these concepts and will help to decide on the applicability of them on simulation of the sediment transport processes for particular conditions of water flow and sediment inputs. A mathematical model which includes sediment transport concepts for the specific conditions of irrigation canals will become an important and timely tool for designers and managers of those systems.The aim of this research is to present a detailed analysis of the sediment transport processes, a physical and mathematical description of the behaviour of sediment transport under flow conditions encountered in irrigation canals and to develop a model to predict sediment transport and the deposition or entrainment rate for various flow conditions and sediment inputs.Sediment transport processesSediment transport and water flow are interrelated and cannot be separated. From a mathematical point of view the interrelation can be described for a one-dimensional phenomenon without changes in the shape of the cross section by the following equations:<UL>governing water flow equations : continuity and dynamic equations;governing sediment equations : resistance to flow, sediment transport equations, continuity equation for sediment mass.</UL>Water flow equations : although one-dimensional flow hardly can be found in nature, water flow in an irrigation canal will be considered to be one-dimensional. Under this assumption, the general equations for one dimensional flow can be described by the Saint Venant equations. The amount of water flowing into irrigation canals during the irrigation season and moreover during the life time of irrigation canals is not constant. For the time depending changes in the bottom of the canal the water flow can be easily schematized as quasi-steady which means that the time depending factors in the Saint Venant equations can be neglected.Resistance to flow : the resistance to flow in open channels is affected by several factors, among which the development of bed forms play an important role. Determination of the friction factor of a movable bed is a complex problem that requires knowledge of an implicit process of flow conditions and bed form development. In order to predict the type of bed forms in irrigation canals the theories developed by Liu, Simons and Richardsons, Bogardi and van Rijn were compared to a selected set of laboratory and field data. Also a comparison of the most widely used methods to predict the resistance to flow with field and flume data has contributed to select an appropriate method for similar situations. The selected methods for predicting the resistance to flow were: White, Bettes and Paris (1979), Brownlie (1983) and van Rijn (1984c).The objective was to find the appropriate theories to describe the bed form and to estimate the resistance to flow (friction factor) in irrigation canals.From the performance of each predictor of bed form type and friction factor method when compared with selected field and laboratory data some conclusions can be drawn:<UL>the theories of van Rijn and Simons and Richardson behave as the best to predict the bed form in irrigation canals;all the bed forms described for the lower regime (ripples, mega-ripples and dunes) can be expected in irrigation canals;the prediction of the friction factor by using the previously described methods takes only into account the bottom friction;the van Rijn method for predicting the friction factor shows the best results when compared with the selected data.</UL>Another important feature related to the resistance of water flow in irrigation canals is the estimation of the friction factor of a irrigation canal with composite hydraulic roughness. The development of bed forms on the bottom, different material on the bottom and side of the canal or vegetated side banks are typical situations for the composite roughness conditions in irrigation canals. The most common cross sections in irrigation canals are the trapezoidal and rectangular cross section with a relatively small value for the bottom width-water depth ratio. In these cross-sections the velocity distribution is strongly affected by the varying water depth on the side slope and the boundary condition imposed to the velocity at the side wall. A method to estimate the effective roughness in a trapezoidal canal with composite roughness along the wetted perimeter which uses the theoretical velocity distribution in the cross section, is proposed.In order to predict the effective roughness in irrigation canals with composite roughness, the existing methods for predicting the effective roughness and the proposed method in this study have been compared with a selected set of laboratory data, which has been collected in the hydraulic laboratory of the Wageningen Agricultural University. The aim of the experiments has been to investigate the friction factor in a trapezoidal canal having varying roughness on side and bottom and to find an appropriate method to estimate the friction factor in a non-wide canal with different roughness along the wetted perimeter. From the comparison the main conclusion can be drawn that the proposed method gives better results than the other methods.For rectangular cross sections with composite roughness the existing methods for estimating the effective roughness can not explicitly be used. Therefor it is proposed to estimate the composite roughness in rectangular cross sections by the same principle as used for the side wall correction. The procedure to estimate the effective roughness in rectangular cross sections has been tested with a selected set of laboratory data used by Krüger. The proposed method predicts more than 95% of the measured values of the composite roughness within a range of error of 15%.Sediment transport equations : sediment transport equations are related to the way in which the sediment is transported: namely in equilibrium and non-equilibrium condition.Sediment transport predictors for equilibrium conditions have been established for different conditions. The use of those equations should be restricted to the conditions for which they were developed. However a comparison of the different equations under similar flow and sediment characteristics, both in irrigation canals and from field and laboratory data will be a useful tool to evaluate the suitability of each equation under these particular flow conditions. In this study, five of the most widely used equations to compute sediment transport have been compared, namely the Ackers and White, Brownlie, Engelund and Hansen, van Rijn and Yang equations. These equations have been compared with field and laboratory data. The objective was to find more reliable predictors of the sediment transport capacity under the flow conditions prevailing in irrigation canals. From that evaluation some remarks can be drawn:<UL>prediction of the sediment transport in irrigation canals within an error factor less than 2 is hardly possible;based on an overall evaluation of all performance criteria for each equation, the Ackers and White and Brownlie equations seem to be the best to predict the sediment transport rate in irrigation canals.</UL>Sediment transport theories have been developed for wide, open channels. Most of the man-made irrigation canals are not considered as wide canals. Recommended values for the ratio of bottom width and water depth (B/h) in those canals are smaller than 8. Existing methods for calculating the total sediment transport capacity for the entire cross section of a non-wide canal do not take into account the velocity distribution over the cross section. A new method to compute the total sediment transport by using a cross section integrated method is proposed, which is based on the assumption of a quasi two-dimensional model. The objective is to consider the effect of the side banks on the distribution of velocities and to adapt the sediment transport predictors for computing the sediment transport for the entire cross section of a non-wide canal. The existing methods and the proposed method to compute the total sediment transport in non-wide canals were compared with a selected set of laboratory data. Based on the overall comparison the proposed method gives better results than the existing methods for computing the sediment transport capacity for the whole cross section.An interesting phenomenon of the non-equilibrium sediment transport in irrigation canals is the adjustment of the actual sediment transport to the sediment transport capacity of the irrigation canal. To simulate the sediment transport under non-equilibrium conditions, the Gallapatti's depth integrated model for adaptation of the suspended load has been used. It has been assumed that the adaptation length for bed load is the same adaptation length for suspended load. Therefore the Gallapatti's depth integrated model can be used to describe the approach of the total sediment concentration to the transport capacity of the irrigation canal.Application of mathematical modelling of sediment transport in irrigation canalsIn order to simulate the sediment transport in irrigation canals, a computer program (SETRIC) has been developed. The computer program can simulate water flow, sediment transport and changes of bottom level in a network composed by a main canal and several laterals with/without tertiary outlets. Also some hydraulics structures are included in the program: overflow and undershot type, submerged culverts and inverted siphons, flumes and drops.The computer program is based on a sub-critical, quasi-steady, uniform or non-uniform flow (gradually varied flow). The water flow can be simulated in open channels, with a rectangular or trapezoidal cross section with single or composite roughness. Only friction losses are considered. No local losses due to changes in the bottom level, cross section or discharges are taken into account. However, changes in the bottom level are included.Sediment characteristics are defined by the sediment concentration at the head of the canal and sediment size is characterized by the mean diameter d ₅₀ . The range of values is 0.05 mm≤d ₅₀ ≤0.5 mm. A uniform sediment size distribution has been assumed.The simulation periods take into account the variation of the irrigation water requirement during the growing season. The growing season is divided into four stages depending on the crop development and climate conditions. The program assumes a maximum of four different periods in which the discharges along the system can be varied.Maintenance activities can also be included into the program. Those maintenance activities are referred to the obstruction degree due to weed growth on the banks and by its effect on the roughness condition of the canal. From that point of view three types of maintenance are included in the program: ideal maintenance, well maintained and poor maintained.Some applications of the model to simulate sediment transport in irrigation canals are shown. The results can not be generalized so that they can only be applied for the local flow conditions and sediment characteristics of each application. The applications are meant to show the applicability of the model and to improve the understanding of the sediment transport process for situations usually encountered in irrigation systems. The sediment deposition in an irrigation canal during a certain period will be simulated for each of the different applications. The sediment transport capacity of the irrigation canal is computed according to the Ackers and White's predictor method. The adjustment towards the sediment transport capacity is according to the Gallapatti's depth integrated model. A sediment mass balance in each reach of the canal will give either the net deposition or net entrainment between the two boundaries of a specific canal reach. From the application cases some conclusions are drawn:Changes of discharges : during the simulations for reductions of discharge to 80% of the design value (equilibrium condition), more than 40% of the incoming sediment load was deposited.Changes in the incoming sediment load : the effect of changes in the incoming sediment load on the sediment transport include the effect of variations in the incoming sediment concentration and in the median sediment size during the irrigation season and/or the lifetime of an irrigation canal. For 100% of variation in the incoming sediment concentration about 30% of the incoming sediment load is expected to settle into the canal. A similar behaviour is observed for the case of changes in the design value of the median size of the incoming sediment. For instance a total of about 45% of the incoming sediment during the simulation period is deposited when the sediment size deviates 100% from the equilibrium size.Controlled sediment deposition : two scenarios to concentrate the sediment deposition at the head reach of a canal were simulated. They can be described as: widening (scenario 1) and deepening (scenario 2). No additional considerations for optimizing economical cost and sediment deposition were done. For the specific flow and sediment transport conditions scenario 2 trapped 4 times more sediment than an irrigation canal without control and 1.3 times more than scenario 1.Sediment transport predictor : large differences in the computed sediment deposition were observed among the sediment transport predictors. The hydraulic conditions during the simulation period gave a low sediment transport capacity for the Engelund and Hansen predictor and larger for Brownlie and Ackers and White predictors. By using the Engelund and Hansen's predictor the sediment deposition was 2 and 3 times more that the Brownlie and Ackers and White's predictors respectively.Flow control structures : two types of flow control structures were compared: overflow type and undershot type. The observed total deposition in both cases is rather similar. A larger difference was observed in the distribution of the sediment deposition along the canal. That difference was mainly concentrated in the upstream part of the structure.Maintenance activities : maintenance was related to weed infestation and it was simulated by assuming optimal maintenance and no maintenance at all during the irrigation season. No direct effect of the growth of the weed on the sediment transport is considered. More sediment deposition was observed in the ideally maintained canal than the non-maintained canal. Due to the constant water level at the downstream side of the irrigation canal the flow condition within the canal behaved as: in the ideally maintained canal a gradually varied flow (backwater curve) remained constant during the simulation period. A continuous deposition was observed during all the time along the irrigation canal. In the non-maintained canal the initial flow condition changed in time from a backwater curve to a drawdown curve due to the constant water level at the downstream end and due to the variation of the water level within the canal imposed by the variation of the roughness condition. A sediment deposition period followed by an entrainment period was observed during the irrigation season.Operation activities : for simulating the effect of the operation procedures on the sediment deposition in the main canal four scenarios were investigated. The four scenarios are: scenario 1 (continuous flow); scenario 2 (rotational flow by hour); scenario 3 (rotational flow by day); scenario 4 (rotational flow by week). From the comparison the following conclusions can be drawn:<UL>the largest total sediment deposition was observed in scenario 1. Total sediment deposition in scenarios 2, 3 and 4 was rather similar;large differences were observed in the distribution of the sediment deposition within the reaches of the main canal.</UL>By considering the results of the applications of the mathematical modelling, it can be concluded that model is a useful tool for assessing the sediment deposition within irrigation canals under different flow conditions and sediment characteristics. Nevertheless, the mathematical model's performance can most probably be improved when it is applied in more situations. Monitoring of the sediment deposition in irrigation networks is required to evaluate the model under specific conditions and to investigate the response in time and space of the bottom level to determined water flows and sediment characteristics. Influences of the type and operation of flow control structures, geometrical characteristics of the canals, water flow and incoming sediment characteristics on the deposition, which the mathematical model predicts, will contribute to a better understanding of the sediment transport processes in irrigation canals