Water Injection Dredging for improving and preserving reservoir storage capacity: modelling and measuring tools

Water Injection Dredging (WID) has been successfully applied for removing sediment deposits in reservoirs, which results in an increase of their storage capacity. This dredging method is based on the fluidization of the top sediment layer by pressurized injection of water by a dredging vessel. The fluidized sediment can be transported towards the dead storage of the reservoir or sluiced out of the reservoir through the bottom outlets of a dam. This flow can either occur by gravity induced flow or especially directed by the dredging strategy of the WID vessel. This dredging technique can increase the water storage capacity of the reservoir and prevent the erosion of the river downstream, hence the sediment blockage. Recent developments in modelling and measuring tools have enabled stakeholders to design, optimize and monitor WID in reservoirs. In this paper, we will demonstrate how modelling and measuring tools can be used to evaluate alternative dredging strategies for reservoir maintenance. In particular, we show how a mid-field and far-field modelling can be applied for designing WID actions and predicting sediment plume dynamics in a given reservoir. Additionally, we will present recently-developed in-situ measuring tools, that are currently used for monitoring turbidity in a water column and sediment properties during and after WID actions. Finally, potential benefit of applying WID in Shihmen Reservoir (Taiwan) is discussed. * Corresponding author: alex.kirichek@deltares.nl © The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 (http://creativecommons.org/licenses/by/4.0/). E3S Web of Conferences 346, 01021 (2022) https://doi.org/10.1051/e3sconf/202234601021 Sharing Water: Multi-Purpose of Reservoirs and Innovations


Introduction
Sediment dynamics and specifically, the siltation of fine sediments in reservoirs is of great interest to those responsible for the maintenance of the dams and water quality of reservoirs. The amount of siltation is dependent on the influx of fine sediments (from runoff, landslides, bank erosion etc.) which gets trapped and settles in the reservoir due to the low flow velocities. Sediment management in reservoirs becomes more challenging due to the aging of existing infrastructure.
Typically, the sediment management strategy for a given reservoir depends on the location, infrastructure and circumstances at each site. Often no action is taken to manage sedimentation in reservoirs as a certain amount of siltation in dead storage is taken account for in the design of dams and reservoirs. Siltation is tackled when it is so much that it influences the functioning or safety of the dam or creates undesired environmental impacts downstream, but measures are very costly. Thus, necessary knowledge is needed for better understanding and mitigating sedimentation challenges. Apart from existing reservoir management strategies, there is a need for innovative effective and cost-efficient solutions.
In this paper, water injection dredging (WID) is proposed as an innovative, efficient and cheaper method to tackle siltation in reservoirs. The paper focusses on the numerical and monitoring tools that can be used to determine the necessary knowledge for applying and optimizing this method.
Density current venting is a technique used in reservoirs to sluice fine suspended sediments that are transported by turbidity currents. Turbidity currents, i.e. density currents driven by suspended-sediment concentration, can be generated naturally by river inflows, or artificially with some method of agitation of muddy deposits. Turbidity currents can transport fine sediments over much longer distances than a normal suspension and can transport the sediments to the dam. As in many situations dam operations do not always allow for sluicing fine sediments from normal suspended loads and turbidity currents. Therefore, the fine sediments arriving at the dam tend to deposit in front of it, forming a bottom-set deposit. After settling and consolidation, a major part of these sediments cannot be remobilized again with normal operations or successive turbidity current venting operations. Low-energy currents in the deep reservoir do not have enough energy to entrain significant amounts of (consolidated) deposits (unless with full water-level draw-down flushing operations). In these situations, it is useful to explore the use of water-injection dredging to artificially remobilize the turbidity currents that move sediments to deeper areas, where they can be vented through outlets.
Water injection dredging (WID) has long been a part of the dredging portfolio for port and waterways maintenance. It has been predominantly applied in ports and waterways with fine grained sediment, however examples of WID applications in coarse sediment are also known within the dredging industry. The principle of the water injection process is based on fluidizing the sediment bed using water jets ( Figure 1). The more water jet nozzles used, the greater the amount of water entering the bed. Since water is injected with relatively low pressure, re-suspension and dispersion of the fine sediment throughout the water column is avoided. Instead, the water injection creates a water-sediment mixture closer to the bed. After WID operation, fluidized mud layers are formed with densities lower than the bed sediment, but higher than the surrounding water density. This WID-made turbidity current spreads itself in the water under the influence of gravity induced hydrodynamic processes. Depending on sediment properties and operational parameters, the thickness of the fluidized mud layer and resulting density current can vary between 0.25 and 3 m. Over the last decades, WID becomes more popular dredging method for port maintenance because economical, operational and ecological aspects of WID seem to be highly competitive in comparison to standard maintenance strategies. In recent years, different tools have been developed for optimizing WID processes and to enable better prediction of the resulting sediment plume transport. Generally, numerical tools that are used in dredging engineering are categorized in 3 groups depending on the scale of the models: near-field, mid-field and far-field models. In this paper, the area around a water jet, where a dynamic plume is formed, is called the near-field. When the dynamic plume is flowing in the WID area, we refer to this area as a mid-field. Finally, the larger area that is beyond the immediate area being dredged, e.g. the reservoir, is referred as a far-field. This paper is structured as follows: first, show few some examples of TUDflow3d model applications for estimating mid-field WID sediment plume dispersion and sedimentation. Outcome of a mid-field model can be used as input to a far field model. Here we present a Delft3D-based modelling tool that is used for far-field prediction of sediment distribution once the turbidity current as a result of WID is dispersed over a wider area in the reservoir. Then, several examples of lab experiments that help to better understand near-field effects during WID is demonstrated. The monitoring tools that can be used for surveying during and after WID are presented. Finally, benefits of applying WID in Shihmen Reservoir (Taiwan) is illustrated.

WID modelling and monitoring 2.1 Mid-field modelling of WID
Mid-field modelling of WID is carried out by the 3D CFD model TUDflow3d [2]. Originally, TUDflow3d has been developed for accurate mid-field simulations of Trailing Suction Hopper Dredger overflow plumes on real scale. It has also been used for WID density currents in harbour basins, MFE (Mass Flow Excavation) plumes, deep sea mining tailing plumes, sedimentation in a hopper or caisson and salinity driven density flows. TUDflow3D is fully 3D with variable density taken into account in all three dimensions (not just in the vertical), non-hydrostatic pressure and turbulence modelled by either a RANS (Reynolds Averaged Navier Stokes) approach or by the accurate LES (Large Eddy Simulation) approach employing a fine grid. The sediment bed is treated with an immersed boundary technique. In the model, sediment can settle out of the WID density current, deposit on the bed and the WID density current can accelerate itself by eroding sediment from the bed. Hindered settling of the fine sediments near the gelling concentration is taken into account. The feedback of the sediment concentration on the mixture density is captured and this feedback derives the spreading of a WID plume as a density current. An instantaneous snapshot of a modelled density current along a sloping erodible bed in a flume is shown in Figure 2. The individual turbulent eddies and whirls resolved on the grid in LES are clearly visible. Comparison for time averaged velocity and Suspended Sediment Concentration (SSC) profiles with measured ones in an experiment of Parker et al. [3] is given in Figure 3. Here, different manners of capturing turbulence are compared. In addition to LES, the Reynolds averaged Navier Stokes (RANS) and RANS with reduced eddy viscosity near the bed are tested. Both RANS runs use a damping function to take turbulence damping on the upper edge of the turbidity current into account. An advantage of LES is that it does not need such damping function as the influence of a sharp gradient in density is automatically captured in the resolved eddies in LES. Of the two RANS results the ones with reduced near bed viscosity are slightly better. The LES results are most accurate. The vertical SSC profile and layer thickness of the density current is captured very well in the CFD LES model and the velocity profiles are captured reasonably well with a small overprediction of the near bed velocity. Additionally, TUDflow3d is used for an example application of modelling WID in a reservoir. The Shihmen Reservoir (in Taiwan) is used for this example. Fictitious application of WID is modelled for a location in the reservoir with visible silt deposits, see Figure 4. The bathymetry of the reservoir and the domain of CFD model is shown in Figure 5. In this CFD run a WID works along a 300m long track indicated with a black dashed line. The resulting WID density current is shown in Figure 6. It moves down the slope of the bathymetry. This example shows that TUDflow3d can be used as a tool to model mid-field WID density current behaviour in a reservoir. It can be used to design and optimize WID actions and it can provide input for the far field model runs that are described in Section 2.2.

Far-field modelling of WID
In order to assess the optimum dredging criteria and the resulting turbidity plume generated as a result of WID, a combination of mid and far field models is used. The turbidity current behaviour and transport can be assessed, but also the suspended transport, as both mechanisms are important, and will determine the amount of fine sediments that can be flushed. A mid-field model was used to determine the initial behaviour of the turbidity current generated through WID and transport mechanisms as described in the previous section. This model provided input into a larger scale Delft3D model that was set-up to create a tool that could be used to test the driving processes of turbidity current dynamics and to test the impact of different dredging strategies e.g. different locations in the reservoir.
Deltares' open source software Delft3D is a flexible, integrated modelling framework which simulates two and three-dimensional flow, waves, sediment transport and morphology (as well as dredging and dumping) on a time-scale of days to decades. The sediment transport module includes both suspended and bed/total load transport processes for an arbitrary number of cohesive and non-cohesive sediment fractions. It can keep track of the bed composition to build up a stratigraphic record. The suspended load solver is connected to the 2D or 3D advection-diffusion solver of the hydrodynamic module and importantly for fluidmud simulations, density feedback can also occur. For this work, a Delft3D model of the Shihmen Reservoir in Taiwan was set up by Commandeur [4,5], which was originally created to test the efficacy of practical solutions to achieve a higher venting efficiency rate for turbidity currents in that reservoir (see Figure 7). This model can now be taken a step further by also controlling the location and generation of turbidity currents through WID with respect to the location of the bottom outlet in the reservoir. Specific inputs on the density and velocity of the initial plume are provided from the mid-field model. The Delft3D system model can assess the full cycle of siltation and siltation management. The model can be used to pre-determine where siltation is mostly likely to occur in reservoirs by simulating sediment loads into the reservoir. After WID has occurred, the model can be used to track the suspended sediment that exits the reservoir through the outlets and assess subsequent transport or deposition downstream of the dam.

Near-field and mid-field laboratory experiments for WID
Over the last decades, laboratory experiments have been intensively used for generating new knowledge on WID. Understanding physical processes, which take place in sediment during WID, helps to determine the most optimal operational parameters that can guarantee efficient WID actions. For instance, in order to be able to control the WID process, speed of a WID vessel, WID cycles, the standoff distance as well as a number and size of nozzles, and pump pressure in relation to sediment properties should be managed carefully. Ill-defined operational parameters can potentially result in an extensive dilution, turbidity generation or uncontrolled re-suspension, thus in inefficient WID processes.
Few examples of laboratory flume experiments conducted for understanding the physical processes behind WID are shown in Figure 8. In these experiments, a consolidated layer of cohesive sediment (mud) is placed on the bottom of a flume. Jet with a single nozzle as well as with multiple nozzles are used for fluidizing consolidated cohesive bed and creating a density current (shown in left/right panels and in middle panel, respectively). Conducting tests for defining the most optimal operational parameters (e.g. standoff distance, pressure speed of WID vessel, etc.) can greatly help to optimize WID operations in order to achieve the most efficient outcome of WID as function of reservoirs specific properties. Furthermore, the lab measured properties serve as input parameters for far-field and mid-field models that are developed for assessing the impact of WID in reservoirs. For instance, Figure 9 shows a time series of a fluidized density current, that is transported along the flat bed by water injection in a flume. Modern measuring technologies allowed us to measure the physical properties of the WID-induced mud layer, such as the velocity of plume, rheology (yield tress and viscosity) of mud (see [6] for measuring rheological properties), the density and the height of the WID-induced mud layer. Most of these physical properties depend on the operational parameters (e.g. pressure and standoff distance), thus connecting the physical properties of mud and operational parameters of WID is the key in maximizing the effect of WID in the field.

Field monitoring tools for WID
There are various monitoring tools that can help to assess the efficiency and environmental impact of WID. Typically, bathymetric surveys are performed from the water surface using an echo-sounder for measuring the depth of water and a GPS tool for measuring geographic location. These measurements provide knowledge about water-mud interface, which can be used for planning WID operations and monitoring the results of WID. Figure 10 shows an example of a bathymetric measurements performed by a multibeam echosounder. These surveys are carried out using high frequency acoustic measurements (about 200 kHz). In addition to the traditional bathymetric surveys, more insightful surveys can be conducted in order to detect the original bottom surface of the reservoir. For these surveys, sub-bottom profiling tools are generally used in combination with multibeam echosounders. These emitted low-frequency (5-40 kHz) can penetrate soft sediment deposits and reflect from a denser layer consisting of soil or rock. Figure 11 shows an example of low-and highfrequency vertical profiles right after WID and 1 month after WID. Fig. 11. Low-and high-frequency vertical profiles indicating the bed level and the water-mud interface, respectively.
Difference between low-and high-frequency levels indicate a layer of WID-fluidized mud layer. An additional knowledge of the physical properties (e.g. shear strength and density) of this layer can be also measured and used for modeling the transport of this layer or for predicting consolidation. The measured vertical density profiles are shown in the right panel of Figure 12. The measurements can be done by different penetrometers [7]. In this case, the densities are measured by DensX (from dotOcean). It can be observed that the measured density profiles show a good resemblance with the 1DV consolidation modelling (more details in Kirichek et al. [8]). The use of Distributed Temperature Sensing (DTS) with optical cables can be used for near and mid field assessment of the sediment density before and after WID and at deposition sites, monitoring the consolidation rate. Figure 13 gives an example based on thermal heating and the difference in heat capacity of solids versus water in sediment.

Conclusions
This paper focusses on insights gained using a combination of monitoring, numerical modelling and laboratory experiments for WID in reservoirs. By combining measurements from the field, laboratory experiments on fluid mud properties, with a state-of the art modelling approach new insights can be gained not only on the best approach for implementing WID as a maintenance strategy for reservoirs with siltation problems but also on the impact of flushing excess sedimentation downstream.
The research focused on cohesive sediment (fluid mud) behavior and transport, but also the resulting sediment plume after WID. Both mechanisms are important and depend on the hydrodynamic conditions. A WID pilot demonstrated that WID is efficient in fluidizing fluid mud deposits. The pilots provided high quality quantitative and qualitative monitoring of fluid mud behavior after WID. This information was used to fine tune a mid-field 3D CFD model of dredge plume behavior. The mid-field WID plume layer thickness and WID production estimates were used as input in to the far-field model. By combining measurements from the field, laboratory experiments on fluid mud properties, with a stateof-the-art modeling approach new insight were gained not only on the best approach for implementing WID in reservoirs but also cost reduction and sustainable management for the reservoir procurator was achieved during the development of these techniques. Finally, the potential of applying WID is illustrated for the case of the Shihmen Reservoir which faces significant losses in reservoir storage by sedimentation of fine sediments.