Advances in Maintenance of Ports and Waterways: Water Injection Dredging

*Alex Kirichek, Katherine Cronin, Lynyrd de Wit and Thijs van Kessel*

## **Abstract**

The main objective of this chapter is to demonstrate developments in port maintenance techniques that have been intensively tested in major European ports. As regular port maintenance is highly expensive, port authorities are considering alternative strategies. Water Injection Dredging (WID) can be one of the most efficient alternatives. Using this dredging method, density currents near the bed are created by fluidizing fine-grained sediments. The fluidized sediment can leave the port channels and be transported away from the waterways via the natural force of gravity. WID actions can be successfully coupled with the tidal cycle for extra effectiveness. In addition, WID is combined with another strategy to reduce maintenance dredging: the nautical bottom approach, which enables the vessel to navigate through the WID-induced fluid mud layer. The nautical bottom approach uses the density or the yield stress of sediment to indicate the navigability after WID rather than the absolute depth to the sediment bed. Testing WID-based port maintenance requires thorough preparation. Over the years modeling and monitoring tools have been developed in order to test and optimize WID operations. In this chapter, the application of the recently developed tools is discussed.

**Keywords:** fluid mud, dredging, sailing through mud, WID, nautical depth, cohesive sediment

## **1. Introduction**

Navigation in ports, canals and waterways must be safeguarded by maintenance dredging to remove sediments deposited by tide, river flows and currents. In order to keep ports and waterways accessible, this non-contaminated sediment is typically dredged by a trailing suction hopper dredger (TSHD) and reallocated at sea [1].

Maintenance dredging of sediment deposits can be highly expensive and inefficient as it must be done on a regular basis. Therefore, port authorities seek tailormade solutions to reduce the costs and at the same time guarantee safe navigation in ports and waterways. Over the last decades, a number of strategies for port maintenance have been tested by port and governmental authorities. Maintenance dredging can be optimized by techniques to avoid or reduce sedimentation, such as optimization of port design, current deflecting walls, see [2], or by designing a sedimentation trap to focus sediment deposition in order to make reallocation easier and to reduce sediment deposition in other port areas [3].

**Figure 1.**

*Types of port maintenance methods which are based on the dredging methods keeping sediment in water or bringing dredged material on land.*

Once dredging has conducted, typical strategies for dredged sediment management are either based on the concepts of keeping sediment in the water system or bringing sediment on land (see **Figure 1**). The former is generally considered as the most cost-effective strategy. However, the latter can be utilized for beneficial re-use of dredged sediment, thus better embedded into a circular economy.

It is a well-known fact that in major sea ports fine-grained sediment deposits are routinely reallocated from the port area either further away downstream from the dredged area or directly to the sea depending on the return flow of from the reallocation locations. The choice in reallocation area often consists of finding a balance between minimizing sediment return flows back into the harbor and transport distance and costs. Often, the reallocation of dredged sediment is combined with sediment management within a building with nature concept [4]. These reallocation projects are mainly focused on the reallocation of fine-grained sediment for land creation or improvement, wild habitat restoration, shore nourishment and marsh or wetland development [5–7].

In contrast to reallocation of sediment, conditioning is used for port maintenance with the assumption that the sediment stays in the port area. The goal of conditioning the sediment is to create navigable conditions in waterways while keeping the sediment in place. In this case, the nautical bottom concept is often applied for navigation through mud [8–10]. One of the examples for applying sediment conditioning for port maintenance is in the Port of Emden. The sediment first dredged and then conditioned by reducing the strength of dredged sediment in the dredging vessel [8]. The created fluid mud is then pumped back to the port mouth creating a weak navigable fluid mud layer. If the transport of fluid mud towards the river equals the import of suspended mud by exchange flows, a dynamic equilibrium is achieved without residual import, hence dredging.

These techniques do not apply to contaminated dredged sediment which is either stored in confined disposal facilities [1, 11] or processed in sediment treatment facilities [12, 13]. The latter technology uses mechanical treatment to prepare the sediment for further beneficial re-use options. Recently, mechanical treatment is also used for non-contaminated sediment as dredged sediment is being recognized as a resource. The treated material can be used as a constructional component for building and re-enforcement of infrastructure [14, 15].

Water injection dredging (WID) can be used as a tool for both reallocation and conditioning of the deposited sediment. The efficiency of this dredging method has been recognized over the past 30 years. However, the successful application of WID can be only achieved by combining technical approaches with knowledge of the system where WID is to be applied. Particularly, the following key questions have to be answered in order to understand better the impact of WID on reallocation and conditioning of cohesive sediment:


The goal of this chapter is to provide an overview of the developed knowledge and tools that can be used for addressing the abovementioned questions. In addition, recently-developed numerical modeling, field and laboratory experiments can provide the necessary information for optimizing WID and defining the boundary conditions for its application. Finally, the recent findings on navigable conditions in ports and waterways, where WID is used for conditioning the sediment and keeping fluid mud in place, are discussed.

## **2. Working principles behind water injection dredging (WID)**

## **2.1 Fluidization of fine-grained sediment**

The principle of the water injection process is based on fluidization of deposited sediment by a water jet (see **Figure 2**). Water injection is performed by injecting large volumes of water (approx. 12,000 m3 /h) under relatively low pressure (approx. 1-1.5 bar) from water jet nozzles, that are distributed over an equal distance on the jet [16, 17]. The injected water penetrates the cavities between the individual sediment particles weakening the forces between them and destroying the formed structure of the bed. The water-sediment mixture forms a fluid mud layer of about 0.5-3 m thickness right above the bed. Most investigations show that the sediment

## **Figure 2.**

*Phases of WID: I. water injection and fluidization; II. Transition zone, where a density flow is created; III. Transport of the density flow. Adapted from [21, 22].*

material hardly mixes into the upper water volume, and sediment transport of the fluidized mud layer remained predominantly close to the bottom [18, 19].

## **2.2 Transport of fluidized sediment**

A sketch of WID performed in a navigational channel with a bed mainly consisting of fine-grained cohesive sediment is shown in **Figure 3**. The near bed fluidized sediment deposit generates a gravity driven density flow up to few meters high, transporting the sediment in a horizontal direction as a result of the density difference [17, 20–22]. This density flow can be described as a homogeneous suspension layer with a solid concentration of up to 200 g/l. Since the density between the fluid mud layer and the surrounding water body is different, fluid mud sets in motion under the action of natural hydrodynamic processes. Thus, WID is different from agitation dredging in which sediment is deliberately mixed over the full water column and then transported in horizontal direction as a passive plume by the ambient currents resulting in a less environmental-friendly outcomes.

The velocity of fluidized sediment is reported in the range between 0.3 m/s and 1 m/s [16, 21, 22]. Based on the hydrodynamic conditions in a port basin, WIDconditioned sediment can either settle over time in a low-energy area or be transported by means of gravity currents to deeper areas such as sediment traps [3].

Different transport distances from a few hundred meters to a few kilometers are reported for fluidized sediment [19, 21–23]. Natural transport of coarse-grained sandy sediment is substantially shorter. Therefore, the sediment composition of the bottom can be altered by WID operations. Fine-grained sediment can be generally more easily fluidized than coarse-grained material and has better transport properties. Since the fine grain fraction is transported away sooner and further than the coarse grain fraction, over time the particle size distribution of the sediment bed can be segregated as a result of dredging. Therefore, the coarse-grained component increases as a result of WID operations.

#### **Figure 3.**

*Illustration of WID performed in a navigational channel during the ebb tide. a) Initial conditions for WID. b) Fluidization of deposited sediment during WID. c) WID-induced fluid mud layer. d) Final result after WID in case WID is conducted for sediment reallocation purposes.*

## **2.3 Efficiency of WID**

The effectiveness of the WID process can be influenced by various factors. The direction, velocity and achieved transport distances of the fluidized layer depend on the interaction of different physical forces. The important influencing factors are sediment composition and characteristics, WID operation characteristics, resulting density of the fluidized layer, bathymetry and natural currents and bed shear stresses in the WID area. The efficiency of the process is also influenced by the bathymetry of the dredging area and the prevailing natural currents. Productivity is generally increased when WID can be carried out so that fluid mud can flow with a natural gradient from higher to lower-lying bathymetry.

The composition and strength of the sediment are also essential for fluidization process. Although it is reported that WID has been also performed for removing coarse-grained sandy sediment and even consolidated soils [16, 21, 22], the best efficiency of WID has been achieved by fluidizing fine-grained sediment deposits. In [20] WID productions are reported in the order of a few thousand m3 /h for very fine-grained sediments and in the order of a few hundred m3 /h for coarser sediments.

The operational parameters for execution of WID are playing an important role for WID. The determining factors are the nozzles diameter, the flow velocity of the water from the jet, jet penetration, the forward movement of the jet pipe, and the distance between the jet nozzle and the surface of the sediment [24]. A WID operator can find the optimal combination of the aforementioned factors to achieve the maximum production of loosened material. However, not only the mass flux of loosened material should be optimized, but also the initial density, layer height and velocity. A thin but dense layer with little initial momentum will hardly spread, whereas a thick, diluted layer with high velocity will quickly mix with ambient water, with negative consequences for turbidity and focus of sedimentation footprint.

WID is generally considered as a relatively low-cost process [3, 25]. As the fluidized sediment is transported in the form of a density flow on the bed and is not distributed throughout the entire water column, WID is also characterized by a high level of environmental compatibility competing to traditional port maintenance dredging [3, 18]. Recently, it was also shown that WID is more CO2 efficient than the regular TSHD maintenance because WID requires less fuel consumption than TSHD. All these aspects suggest that WID can be more attractive tool for port maintenance.

## **3. Modeling of WID**

In recent years, different tools have been developed for optimizing WID processes and better prediction of sediment plume movement during WID. Numerical modeling tools can be used for estimating sediment dynamics in ports and waterways after WID.

Mid-field modeling is often used for calculating the sediment footprint on the areas up to about 1 km away from WID. The obtained knowledge on sedimentation can help to better design WID operations including real bathymetry of a navigational channel. Existing and hypothetical infrastructure can be included in mid-field modeling allowing for testing of WID in combination with sediment transport steering management solutions such as sediment traps, sills and current-deflection walls.

Far-field modeling evaluates the impact of WID on the scale of the entire port or estuary area. This kind of modeling is used for estimating WID reallocation strategies of sediment from the port basins to the sea and for assessing return flows. Simulations can demonstrate the transport of the WID plume during different phases of the tide and the impact of river and sea conditions. Based on the obtained information, the authorities can decide if conducting WID for reallocation purposes is effective in the port.

## **3.1 Mid-field modeling of WID**

Mid-field modeling is carried out by two distinct models: a Lagrangian 1DV model and a 3D CFD model (TUDflow3D). The Lagrangian 1DV model is a rapid assessment tool which can be used for rather uniform bathymetry and slowly varying flows while neglecting lateral spreading. When these assumptions are not valid the more sophisticated 3D CFD model TUDflow3D can be used which includes lateral spreading and simulates a WID density current in three dimensions. TUDflow3D needs much more simulation time as the Lagrangian 1DV model.

The Lagrangian 1DV approach allows us to follow the development of the fluidized layer flow along a user-defined trajectory using a moving frame of reference. The 1DV model determines the thickness and the density (or the sediment concentration) of the fluidized mud layer and correlates these properties to the hydrodynamics in the water column and the slope of the bed. Additionally, it determines the sedimentation flux on the bed. For an equal initial momentum of the fluidized mud layer, the layer will flow further along a downward slope than along a flat bed. In general, the results of 1DV modeling can be used for a better planning of WID.

**Figures 4** and **5** illustrate an example of utility of the 1DV model for water injection dredging. In both figures, the left panel shows the distribution of the sediment concentration and the height of the fluidized mud layer along the slope. The right panel shows the flow velocity of the fluidized mud layer. **Figure 4** shows the simulation of WID for an initial WID plume height of 2 m and **Figure 5** shows the results of WID for an initial WID plume height of 3 m. Both cases start with an initial sediment concentration of 170 kg/m3 and 0.7 m/s flow velocity. It can be seen that a higher fluidized mud layer travels faster and reaches a higher internal velocity.

**Figure 4.** *1DV results for initial WID plume height of 2 m.*

**Figure 5.**

*1DV result for initial WID plume height of 3 m.*

*Advances in Maintenance of Ports and Waterways: Water Injection Dredging DOI: http://dx.doi.org/10.5772/intechopen.98750*

WID density-driven plumes can be also simulated in 3D by the CFD model TUDflow3D [26, 27]. Originally, TUDflow3D has been developed for accurate near field simulations of Trailing Suction Hopper Dredger overflow plumes on real scale. It has also been used for MFE (Mass Flow Excavation) plumes, deep sea mining tailing plumes and salinity driven density flows. TUDflow3D can supplement the 1DV model for complex situations in which the simplifications of the 1DV model make application impossible. TUDflow3D is fully 3D with variable density taken into account in all three dimensions (not just in the vertical), non-hydrostatic pressure and turbulence captured by either the accurate LES (Large Eddy Simulation) approach or by a faster RANS (Reynolds Averaged Navier Stokes) approach.

An instantaneous snapshot of the modeled density current is shown in **Figure 6**. 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 is given in **Figure 7**. Here, different manners of capturing turbulence are compared. In addition to LES with the WALE sub-grid-scale model, the RANS with Realizable K-Epsilon model and Realizable K-Epsilon model with reduced eddy viscosity near the bed are tested. In the latter the eddy viscosity near the bed is adjusted, effectively reduced, to correspond to the correct amount of bed shear stress. The results show that this adjustment improves the Realizable K-Epsilon results for this flow. 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. The Realizable K-Epsilon results with adjusted near bed viscosity are considerably better as the default Realizable K-Epsilon results.

An example a of application of TUDflow3D for WID is given in **Figure 8**. In this CFD run a WID works along a 300 m track which it has done 6 times in a row.

#### **Figure 6.**

*Instantaneous LES snapshot of 3D contour (top) of a turbidity current and SSC at a vertical slice through the center of the turbidity current (bottom).*

**Figure 7.**

*Comparison modeled time averaged velocity and SSC profiles with 3 different turbulence settings (LES; realizable K-epsilon and realizable K-epsilon with reduced near bed viscosity) and measurements from [28].*

**Figure 8.**

*Example of TUDflow3d simulation: Plume distribution from WID action along black dashed line.*

**Figure 9.** *Example of TUDflow3D simulation: Implementing bathymetry in a CFD domain.*

*Advances in Maintenance of Ports and Waterways: Water Injection Dredging DOI: http://dx.doi.org/10.5772/intechopen.98750*

The CFD model uses the real bathymetry of the port. The resulting WID plume is shown in brown and the bathymetry is illustrated as a gray surface. At the moment of this image the WID has just finished the 6th time along the black dashed track of 300 m long. In this example the WID plume flows down the sloping bed in lateral direction under influence of gravity. A top view of the bathymetry is shown in **Figure 9**.

A comparison of TUDflow3D and the Lagrangian 1DV model for WID in a lateral confined situation without bed-slope is shown in **Figure 10**. For this simulation, the following initial conditions were applied: initial WID layer thickens of 2 m, 170 kg/m3 and 0.7 m/s inflow (resulting in an influx of 238 kg/s). The example shows the simulated vertical velocity profiles and density profiles at different distances from the WID. The model also calculates the sedimentation flux out of the WID density current. The results of the 1DV model and full 3D CFD are close to each other for this case. For cases where the assumptions of the Lagrangian 1DV model (neglecting lateral spreading and slowly varying flow conditions) hold it is much faster as the more sophisticated TUDflow3D model and in other cases it is advised to use a 3D near field model like TUDflow3D.

**Figure 10.**

*Comparison of CFD model TUDflow3D and 1Dv simulations for WID in a lateral confined situation. TUDflow3D is compared for two different turbulence settings (LES; realizable K-epsilon).*

## **3.2 Far-field modeling of WID**

Sediment dynamics and specifically, the siltation of mud, in ports is of great interest to those responsible for the maintenance of ports, harbors and access channels around the world. The amount of siltation determines the frequency and volume of maintenance dredging needed to maintain navigable depth. In order to understand sediment dynamics in the system, in particular the processes responsible for suspended mud and fluid mud transport, a range of spatial and temporal scales must be analyzed. A numerical model is an ideal tool with which to investigate both the transport, deposition, and potential resuspension of a WID plume. Such a model was developed, using Delft3D, for the Rhine Meuse Delta in the Netherlands, in order to calculate both background fine sediment dynamics in the Port of Rotterdam and the transportation of a fluid mud layer after a WID operation.

Deltares' open source software Delft3D is a flexible, integrated modeling 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 fluid-mud simulations, density feedback can also occur.

**Figure 11.** *Horizontal near bed plume spreading, WID starts 1 h before HW with a production rate of 500 kg/s.*

#### *Advances in Maintenance of Ports and Waterways: Water Injection Dredging DOI: http://dx.doi.org/10.5772/intechopen.98750*

For this work, a Delft3D model of the entire Rhine Meuse Estuary was setup. Hydrodynamic conditions were simulated for a full month, including wave effects. This hydrodynamic model is then used to force the sediment transport model. Background sediment concentrations are included in the model using three sediment fractions to represent the appropriate range of coarser and finer fractions. Once natural dynamics regarding sediment transport and sediment deposition in the different ports was captured, a range of WID tests could be undertaken. The parameters derived for different WID production rates in the mid-field modeling (described in Section 3.1) are used to define the initial conditions for the WID plume in the far-field model. Numerical experiments could then be performed such as simulating where the WID plume is transported to, the amount of return flow into different parts of the port and the amount of mixing that occurs throughout the water column. Vertical mixing may result in elevated turbidity levels near the surface, which should remain within the environmental limits. The model is also used to investigate the optimum location for sediment traps to capture the WID high density plume.

**Figures 11** and **12** show an example of how the far-field modeling was used to investigate the impact of carrying out WID at different stages of the tidal cycle. WID was carried out in the area of a black rectangle. The colourbar indicates the distribution of suspended sediment concentration (SSC) in the port area. The duration of WID was 8 hours with a production rate of 500 kg/s. During 2 simulations, WID was initiated 1 h before high water (HW) and 1 h before low water (LW). The results of both simulations are shown in **Figures 11** and **12**, respectively.

**Figure 12.** *Horizontal near bed plume spreading, WID starts 1 h before LW with a production rate of 500 kg/s.*

The plume disperses in a distinct way between the simulation starting before high water (HW) compared to a start at low water (LW). **Figure 11** shows that the plume is predominantly dispersed in the seaward direction with the outgoing tide. For WID, this would be the most preferable conditions because in this way the suspended sediment will be relocated from the area where WID is conducted offshore. However, after approximately six hours the flow is reversed, and the plume is pushed in the landward direction.

**Figure 12** show the initial plume dispersion for the simulations in which sediment is released just before LW. The dispersion of the plume in the first 2 hours of the simulations is similar to the experiment with WID release just before HW. However, between four and eight hours a predominant landward plume dispersion is observed. After the flow reversal, it is observed that the plume starts to disperse in the seaward direction. A continuation of the landward spreading is observed in the channel because of the predominant landward flood directed current.

The far-field modeling illustrates the importance of the hydrodynamic conditions during WID. This knowledge can help to choose the most-efficient strategy for WID in ports and waterways with mud layers. The most efficient strategy is not only related to optimizing the sedimentation footprint, but also to minimizing vertical mixing and the contribution of WID to turbidity higher up in the water column. By choosing operational parameters wisely and executing WID operations only during favorable hydrodynamic conditions demands on sedimentation footprint and turbidity are more easily met.

## **4. WID and navigation through mud**

In low-energy regions or in a tidal area of the port, WID-induced sediment can form a fluid mud layer that remains in the port area. The thickness of WID-induced fluid mud layer is often larger than the thickness of original mud layer resulting in a reduced draft for the incoming vessels. In this case, WID is often combined with the nautical bottom approach defined by PIANC for navigation. According to PIANC, 'The nautical bottom is the level where physical characteristics of the bottom reach a critical limit beyond which contact with a ship's keel causes either damage or unacceptable effects on controllability and manoeuvrability' [10, 29]. The nautical bottom allows to use the fluid mud in estimates of under keel clearance (UKC) that the vessels can navigate in the port areas with no unacceptable effects on controllability and maneuvering of the vessels. If accepted by the port authorities, the nautical bottom approach is used for navigation through mud in ports and waterways with fluid mud layers.

Generally, the density of the top sediment layer is used for defining the nautical bottom (see **Figure 13**). The level, where the density of sediment is lower than 1.2 t/m3 , is widely accepted for navigation in ports. Ports in Rotterdam, Zeebrugge, Bordeaux, Saint-Nazaire, Bristol, Bangkok, Tianjin have successfully adapted the density criterium for navigation [29, 30]. However, the Port of Emden relies on the rheological properties rather than density of the sediment for defining the nautical bottom. The yield stress of the top sediment layer gives an indication if the sediment is navigable or not. The sediment with yield stress lower than 100 Pa is considered navigable. The choice of the nautical bottom criterium is related to the conditioning of sediment, that the Port of Emden has been conducting for port maintenance.

The knowledge on in-situ density or rheological properties of the top sediment layer are necessary for implementing the nautical bottom approach. There are in-situ tools that can provide an information about vertical profiles of density and strength in water-mud column. The in-situ devices Rheotune, Graviprobe and DensX have been intensively tested for the nautical bottom approach over last years [3, 29, 31].

*Advances in Maintenance of Ports and Waterways: Water Injection Dredging DOI: http://dx.doi.org/10.5772/intechopen.98750*

**Figure 13.** *Illustration of the nautical bottom concept with the density of 1.2 t/m3*

An example of in-situ measurement of density and yield stress provided by Rheotune is shown in **Figure 14**. The measurements are conducted in a sediment trap that was filled with WID-induced fluid mud during day 1. The development of density and yield stress of WID-induced sediment has been observed for the period of 3 months. The in-situ devices can naturally provide only 1D vertical profiles. However, the thickness of mud layer can be defined from the profiles if the critical value for physical parameters is defined.

*.*

In the example given in **Figure 15**, the critical value for the density is chosen as 1.2 t/m3 providing the density-based nautical bottom shown in red line. In this case, the SILAS software is used for matching the density given by Rheotune (shown by vertical blue line in **Figure 15**) to the seismic data of 38 kHz. The measurements are conducted 7, 21 and 42 days after WID.

The development of WID-induced mud layer be also estimated with the numerical code solving the Gibson Eq. [32]. For instance, settling and consolidation of fluid mud can be predicted by matching the measured data to the model output. **Figure 16** shows the comparison of 1DV model and measured data during consolidation of WID-induced fluid mud layer. The model's output is the density of mud and the water mud interface as a function of time, that can be correlated to measured densities and multibeam data, respectively. The latter can typically provide a reliable water-mud interface for WID operations. For instance, **Figure 17** shows the development of water-mud interface before, during and after WID in the Calandkanaal.

Vertical density profiles are shown in the right panel of **Figure 16**. The density measurements can be done by different penetrometers [3, 31, 33], in this case the densities are measured by DensX. It can be observed that the measured density profiles show a good resemblance with the results of numerical modeling [31, 33]. Thus, the combination of the model with the in-situ measurements can potentially be used for predicting the development of the nautical bottom in time.

An example of the application of PIANC's nautical bottom approach after WID in the Port of Rotterdam is shown in **Figure 18**. The standard multibeam echosounder survey indicated the bathymetry that corresponds to the water-mud level. However, the WID-induced fluid mud has relatively low densities (<1200 kg/m3 ) and weak strength (<100 Pa). Therefore, the nautical bottom approach can be applied. Adapting either a density-based (1200 kg/m3 ) or yield stress-based (100 Pa) criterium for the nautical bottom results in an additional 1.5 and 2 m of navigable depth, respectively.

20 days after WID, these differences are reduced. However, the yield stress-based nautical bottom still shows an advantage of about 0.5 m of extra navigable depth.

**Figure 14.** *Density and yield stress profiles measured by Rheotune.*

#### **Figure 15.**

*Development of the density-based nautical bottom after WID. Red line shows the level, where the density of sediment is equal to 1.2 t/m3 .*

#### **Figure 16.**

*Estimating consolidation of fluidized mud layer after WID. Left panel shows development of water - fluidized mud interface as well as fluidized mud – Consolidated bed interface. Right panel show model predictions (solid lines) and in-situ measurements (symbols) of densities in water-mud vertical column.*

*Advances in Maintenance of Ports and Waterways: Water Injection Dredging DOI: http://dx.doi.org/10.5772/intechopen.98750*

#### **Figure 17.**

*Multibeam measurements indicating water-mud interface before WID (reference), during WID (day 1) and after WID (day 7 - day 42) in the Calandkanaal.*

#### **Figure 18.**

*An example of applying the nautical bottom approach after WID [3]. The density-based (1200 kg/m3 ) or yield stress-based (100 Pa) criteria brings additional 2 m for nautical depth comparing to the standard multibeambased navigational criterium.*

## **5. Discussion**

Water injection dredging is a widely applicable dredging method. The efficiency of the method for maintaining ports and waterways is generally high. WID operational parameters, knowledge of sediment properties, boundary and hydrodynamic conditions of the maintained area can greatly increase the efficacy of the water injection process. The most important parameters and factors influencing the performance of WID are the following: WID operational parameters (diameter of nozzles, flow velocity from the nozzle, stand-off distance of the jet, trailing speed of the WID vessel), sediment properties (grain size distribution, shear strength, density, oxygen consumption potential and sediment quality), boundary conditions of the maintained area (bathymetry, slope angle, embankments), hydrodynamics conditions (direction and velocity of tidal currents, existing density currents and salinity gradients).

Apart from the operational parameters, other factors and conditions that can increase the performance of WID are site-specific. Currently, the literature on research investigations into WID operational parameters is scarce. Therefore, there is a need for further systematic laboratory investigations for exploring the most-efficient WID operational parameters, which can further maximize the WID production rates in the field.

Sediment properties in the proposed area for WID can be studied before conducting WID. Typically, sediment samples are collected for laboratory analysis. The shear strength and density of sediment are linked to WID operational parameters (such as flow velocity) during the WID fluidization processes. The literature on investigations of sediment properties while testing varying WID operational parameters is very limited. Predominantly, WID is applied in the area with nocontaminated sediment. Therefore, the knowledge of the quality of sediment in the WID area is important.

The geometry of the WID area should be taken into account for planning and execution of WID operations in port and waterways. Bathymetric charts, which, will provide the information about deeper areas in the WID location, which are typically filled in with fluid mud after WID. Furthermore, bathymetric charts will indicate the slopes in the WID area, which can be also used for transporting the fluidized mud more efficiently.

Hydrodynamic conditions in the WID area should be taken into account when determining the final fate of fluid mud generated by WID, whether WID is used for the transport or conditioning of mud. For the transport of mud, the knowledge of the direction of the natural current and current velocities can help to minimize the spread of the WID-induced fluid mud deeper into the port area and maximize the transport of the sediment from the port area. For the conditioning of mud, the hydrodynamic conditions can potentially provide an indication whether fluid mud starts to settle in the allocated area or is transported to other locations of the port. Salinity gradients and local density currents can influence the density currents by damping the velocity of WID-induced fluid mud, thus decreasing production rates in the WID-area.

## **6. Conclusions**

This chapter focusses on presenting an overview of developed knowledge for WID. In particular, new insights gained using a combination of in-situ monitoring and numerical modeling. The research focusea on fluid mud behavior and transport, but also the resulting sediment plume. Both mechanisms are important and depend on the surrounding hydrodynamic conditions. Mid-field modeling was used to investigate the WID plume flow and deposition behavior up to 1 km away from the WID dredger. The WID-induced fluid mud layer thickness and WID production estimates were used as input in to the far-field model. The far-field model was used to determine where the WID-induced plume traveled under different tidal and discharge conditions, how much deposited back in the harbors and how much was flushed out to sea with the ebb tide. The model was also used to test different disposal locations to reduce return flow.

Key factors and parameters influencing the efficiency of WID have been identified from the available literature and discussed further. The modeling tools presented in the chapter can potentially help to analyze the sediment properties, boundary conditions and hydrodynamic conditions in the WID area and in the entire port area. However, more experimental research is needed for defining the *Advances in Maintenance of Ports and Waterways: Water Injection Dredging DOI: http://dx.doi.org/10.5772/intechopen.98750*

most-efficient set of operational parameters. Particularly, the knowledge on linking WID operational parameters with sediment properties for maximizing production rates is very scarce.

By combining measurements from the field, laboratory experiments on fluid mud properties, with a state-of-the-art modeling approach, new insights were gained on the best approach for implementing WID as a maintenance dredging strategy. In addition due to more efficient maintenance, reduction of costs, CO2 emissions and additional environmental impacts is achieved during the application of these techniques.

## **Acknowledgements**

The work in this study is funded by the Port of Rotterdam and by Topconsortium voor Kennis en Innovatie (TKI) Deltatechnologie subsidy. The research is carried out within the framework of the MUDNET academic network https://www.tudelft.nl/mudnet/

## **Conflict of interest**

The authors declare no conflict of interest.

## **Nomenclature**


*Sediment Transport - Recent Advances*

## **Author details**

Alex Kirichek1,2\*, Katherine Cronin2 , Lynyrd de Wit<sup>2</sup> and Thijs van Kessel<sup>2</sup>

1 Faculty of Civil Engineering and Geosciences, Delft University of Technology, The Netherlands

2 Deltares, The Netherlands

\*Address all correspondence to: o.kirichek@tudelft.nl

© 2021 The Author(s). Licensee IntechOpen. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/ by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

*Advances in Maintenance of Ports and Waterways: Water Injection Dredging DOI: http://dx.doi.org/10.5772/intechopen.98750*

## **References**

[1] Kirichek A, Rutgers R, Wensveen M, Van Hassent. Sediment management in the Port of Rotterdam. In: Proceedings of the 10th Rostocker Baggergutseminar; 11-12 September 2005; Rostock

[2] Kirby R. Minimising harbour siltation—findings of PIANC Working Group 43. Ocean Dynamics. 2011; 61: 233-244. https://doi.org/10.1007/ s10236-010-0336-9

[3] Kirichek A, Rutgers R. Monitoring of settling and consolidation of mud after water injection dredging in the Calandkanaal. Terra et Aqua. 2020; 160:16-26

[4] Sittoni L, van Eekelen E, van der Goot F, Nieboer H. The living lab for mud: integrated sediment management based on building with nature concepts. In: Proceedings of the 22nd World Dredging Congress & Exposition (WODCON XXII); 2-26 April 2019; Shanghai

[5] SURICATES - Sediment Uses as Resources In Circular And Territorial EconomieS [Internet]. 2017. Available from: https://www.nweurope.eu/ projects/project-search/suricatessediment-uses-as-resources-in-circularand-territorial-economies/ [Accessed: 2020-12-28]

[6] Baptist JM, Gerkema T, van Prooijen BC, van Maren DS, van Regteren M, Schulz K, Colosimo I, Vroom J, van Kessel T, Grasmeijer B, Willemsen P, Elschot K, de Groot AV, Cleveringa J, van Eekelen EMM, Schuurman F, de Lange HJ, van Puijenbroek MEB, Beneficial use of dredged sediment to enhance salt marsh development by applying a 'Mud Motor', Ecological Engineering. 2019; 127: 312-323.

[7] Marker Wadden - Natuurmonumenten. 2019. Available from: https://www.natuurmonumenten.nl/ projecten/marker-wadden/englishversion [Accessed: 2020-12-28]

[8] Wurpts R, Torn P. 15 Years' Experience with Fluid Mud: Definition of the Nautical Bottom with Rheological Parameters. Terra et Aqua. 2005. 99:22-32

[9] Kirichek A, Chassagne C, Winterwerp H, Vellinga T. How navigable are fluid mud layers? Terra et Aqua. 2018; 151:6-18

[10] PIANC. Harbour Approach Channels - Design Guidelines, Report 121, PIANC, Brussels, 2014.

[11] Slibdepot IJsseloog. 2010. Available from: https://web.archive.org/ web/20100604043705/http://www. rijkswaterstaat.nl/water/plannen\_en\_ projecten/vaarwegen/ketelmeer/ ketelmeer/ketelmeer\_oost/ [Accessed: 2020-12-28]

[12] METHA - MEchanical Treatment and Dewatering of HArbor sediments. 2020. Available from: https://www. hamburg-port-authority.de/en/ themenseiten/metha/ [Accessed: 2020-12-28]

[13] AMORAS - Antwerpse Mechanische Ontwatering, Recyclage en Applicatie van Slib. 2020. Available from: https:// www.maritiemetoegang.be/amoras [Accessed: 2020-12-28]

[14] Kleirijperij. 2019. Available from: https://eemsdollard2050.nl/project/ pilot-kleirijperij/ [Accessed: 2020-12-28]

[15] Brede Groene Dijk. 2019. Available from: https://eemsdollard2050.nl/ project/brede-groene-dijk/ [Accessed: 2020-12-28]

[16] Clausner JE. Water injection dredging demonstrations. U. S. Army Corps of Engineers, Waterways Experiment Station. Dredging Research. 1993. Vol. DRP-93-3

[17] Winterwerp JC, Wang ZB, van Kester JATM, Verweij JF. Far-field impact of water injection dredging in the Crouch River. Proceedings of the Institution of Civil Engineers Water & Maritime Engineering. 2002; 154 (4), 285-296.

[18] Borst WG, Pennekamp JGS, Goossens H, Mullié A, Verpalen P, Arts T, van Deumel PF, Rokosch WD. Monitoring of water injection dredging, dredging polluted sediment. In: Proceedings of the second international conference on dredging and dredged material placement (Dredging ´94); 13-16 November; Walt Disney World, Lake Buena Vista, Florida. New York: ASCE; 1994. Vol. 2: p. 896-905

[19] Meyer-Nehls R, Gönnert G, Christiansen H, Rahlf H. Das Wasserinjektionsverfahren – Ergebnisse einer Literaturstudie sowie von Untersuchungen. In Hamburger Hafen und in der Unterelbe. 2000. ISSN 0177-1191

[20] PIANC - Injection Dredging, Report 120, PIANC, Brussels, 2013

[21] Verhagen HJ. Water injection dredging. In: Proceedings of the 2nd International Conference Port Development and Coastal Environment (PDCE 2000); 5-7 June 2000, Varna, Bulgaria, 2000.

[22] Murphy AM. DRP site visit: Water injection dredging. U. S. Army Engineer Waterways Experiment Station, Vicksburg, Miss. Dredging Research. 1993; DRP-93-1

[23] Nasner H. Injektionsbaggerung von Tideriffeln. Hansa. 1992; 129 (2): 195-196.

[24] Estourgie ALP. Theory and practice of water injection dredging. Terra et Aqua. 1988; 38:21-28

[25] Bray RN. Maintenance dredging: where do we go from here? The Dock & Harbour Autority LXX. 1989; 810:57-60 [26] de Wit L, Bliek A.J., van Rhee C. Can surface turbidity plume generation near a Trailing Suction Hopper Dredger be predicted? Terra et Aqua. 2020 September, 6-15

[27] de Wit L. 3D CFD modelling of overflow dredging plumes. [thesis]. Delft: Delft University of Technology; 2015. https://doi.org/10.4233/ uuid:ef743dff-6196-4c7b-8213 fd28684d3a58

[28] Parker G, Garcia M, Fukushima Y, Yu W. Experiments on turbidity currents over an erodible bed. J. of Hydraulic Research. 1987; 25(1):123-147. DOI: 10.1080/00221688709499292

[29] Kirichek A, Chassagne C, Winterwerp H, Vellinga T. How navigable are fluid mud layers? Terra et Aqua. 2018; 151: 6-18.

[30] McAnally WH, Teeter A, Schoellhamer D, Friedrichs C, Hamilton D, Hayter E, Shrestha P, Rodriguez H, Sheremet A, Kirby R. Management of Fluid Mud in Estuaries, Bays, and Lakes, Part 2: Measurement, Modeling, and Management. J. Hydraul. Eng. 2007. 133 (1).

[31] Kirichek A, Shakeel A, Chassagne C. Using in situ density and strength measurements for sediment maintenance in ports and waterways. J. Soils Sediments. 2020; 2546-2552. DOI: 10.1007/s11368-020-02581-8

[32] Merckelbach LM. Consolidation and strength evolution of soft mud layers. [thesis]. Delft: Delft University of Technology; 2000

[33] Kirichek A, Rutgers R. Water injection dredging and fluid mud trapping pilot in the Port of Rotterdam. In: Proceedings of the CEDA Dredging Days 2019, 7-8 November 2019; Rotterdam

## **Chapter 7**

Non-Intrusive Characterization and Monitoring of Fluid Mud: Laboratory Experiments with Seismic Techniques, Distributed Acoustic Sensing (DAS), and Distributed Temperature Sensing (DTS)

*Deyan Draganov, Xu Ma, Menno Buisman,Tjeerd Kiers, Karel Heller and Alex Kirichek*

## **Abstract**

In ports and waterways, the bathymetry is regularly surveyed for updating navigation charts ensuring safe transport. In port areas with fluid-mud layers, most traditional surveying techniques are accurate but are intrusive and provide onedimensional measurements limiting their application. Current non-intrusive surveying techniques are less accurate in detecting and monitoring muddy consolidated or sandy bed below fluid-mud layers. Furthermore, their application is restricted by surveying-vessels availability limiting temporary storm- or dredgingrelated bathymetrical changes capture. In this chapter, we first review existing nonintrusive techniques, with emphasis on sound techniques. Then, we give a short review of several seismic-exploration techniques applicable to non-intrusive fluidmud characterization and monitoring with high spatial and temporal resolution. Based on the latter, we present recent advances in non-intrusive fluid-mud monitoring using ultrasonic transmission and reflection measurements. We show laboratory results for monitoring velocity changes of longitudinal and transverse waves propagating through fluid mud while it is consolidating. We correlate the velocity changes with shear-strength changes while the fluid mud is consolidating and show a positive correlation with the yield stress. We show ultrasonic laboratory results using reflection and transmission techniques for estimating the fluid-mud longitudinal- and transverse-wave velocities. For water/mud interface detection, we also use distributed acoustic sensing (DAS) and distributed temperature sensing (DTS).

**Keywords:** Safe navigation, non-intrusive monitoring of fluid mud, transmission seismic measurements, reflection seismic measurements, yield stress, distributed acoustic sensing (DAS), distributed temperature sensing (DTS)

## **1. Introduction**

Safe navigation through fluid mud is increasingly important because enhancing the navigability with less dredging can help lower transportation costs and benefit biodiversity. The areas with fluid-mud layers need to be routinely surveyed to provide navigation charts used by the vessels. Fluid mud is described as a highly concentrated non-Newtonian suspension of sediment consisting mainly of water, organic matter, silt and clay minerals [1]. Fluid mud is a crucial factor when determining the nautical depth (nautical bottom). It is typically defined by a density value [2]. For example, the Port of Rotterdam uses the density of 1.2 kg/L as a nautical-depth criterium. Other parameters are though also used – for example, the Port of Emden adopts the yield stress of 100 Pa to define the nautical depth [2, 3]. Thus, it is important to have an accurate parameter that description of the fluid mud and could be used in the same way in different ports.

Full-scale and scaled experiments for safe ship navigation in the ports and waterways have been performed already for several decades [3–5]. Traditional ways of characterizing fluid mud involve its sampling, which inevitably disturbs the mud. Other methods, for instance radioactive probes, such as X- and γ-ray tube, can be used to measure the density of the fluid mud, where the density calculation is based on the Lambert–Beer Law [6]. The density profiler based on X-rays – DensX, and the Graviprobe, which measures the cone-penetration resistance and pressures when sinking freely in the water-mud column, can be used to estimate the density and undrained shear strength, respectively [7, 8]. Although these tools can provide a quantitative information about the densities and strength of mud, non-intrusive characterization and monitoring of fluid mud in ports and waterways is preferable. Currently, echo-sounding measurements are used as non-intrusive techniques for assessment of the nautical depth, for which the relationship between the acoustic impedance and densities of the fluid mud are investigated [9]. Multi-beam echosounders are deployed to detect fluid-mud layers. Utilization of signals at a higher frequency (200–215 kHz) and at a lower frequency (15–40 kHz) provides an estimate of the approximate thickness of the fluid-mud layer [7]. The higher-frequency measurements are used to map the lutocline, while the lower-frequency measurements provide an estimate of the sea-floor depth. Schrottke and Becker deployed a high-resolution side-scan sonar with a frequency of 330 kHz and a parametric subbottom profiler with frequencies of about 100 kHz for detecting the fluid mud with high vertical resolution [10]. The velocimeter, especially the acoustic Doppler velocimeter, was developed on the basis of ultrasonic waves to measure turbulency velocities in the fluid-mud sediments [11]. The mentioned techniques, though, rely on longitudinal (P-) waves, which are related to the bulk properties of the materials.

The propagation velocity and amplitude of transverse (S-) waves strongly depend on the geotechnical properties of the sediment, such as fluid mud [12]. Thus, S-waves could be used to characterize the fluid mud more precisely than when using P-waves and thus bulk properties. However, in seismic exploration in marine environments, the sources and receivers are usually deployed in the water column, more often relatively close to the water surface. Thus, the sources, such as airgun arrays, give rise to P-waves, and the receivers, usually towed by a vessel as streamers, record P-waves as well. This limits the utilization of S-waves because extracting the S-wave information is rather more involving and time-consuming [12]. Still, strong P-to-S-converted waves could be generated at the water bottom, and their utilization for characterization of the fluid mud is possible. A technique that could allow direct extraction of the S-wave velocities is seismic interferometry (SI) for retrieval of non-physical reflections. SI is a method that retrieves new recordings from existing recordings most often by cross-correlation [13–15] of the

#### *Non-Intrusive Characterization and Monitoring of Fluid Mud: Laboratory Experiments with… DOI: http://dx.doi.org/10.5772/intechopen.98420*

existing recordings. When the required assumptions for the practical application of SI are not met, non-physical arrivals are also retrieved. Some of the non-physical arrivals arise from internal reflections between layer boundaries [16–18]. SI can thus be applied for targeted retrieval specifically of non-physical (ghost) reflections to estimate the layer-specific velocities for layers in the subsurface [17, 18].

Ultrasonic transmission measurements of marine sediments have been performed, and it was reported that the P-wave attenuation coefficients indicate changes in the sediment composition more distinctly than the velocity of the P-waves [19]. Additionally, relationships between the porosity and P- and S-waves velocities were examined [19]. Leurer [20] carried out pulse-transmission measurements with a center frequency of 50 kHz and reported that in a foraminiferal mud the P-wave velocities range between 1840 m/s and 2462 m/s. Using a center frequency of 100 kHz under different effective pressures, it was also found that the S-wave velocities range between 450 m/s and 975 m/s [20]. Other studies showed that the Swave velocity in mud samples can be as low as 7 m/s when using signals with a center frequency of 200 Hz [21, 22]. These different values for the P- and S-wave velocities show that it is necessary to perform seismic (ultrasonic) measurements for characterization of the fluid mud for each specific location, i.e., for each port or waterway. This would favor utilization of reflection measurements like in seismic exploration as they can be performed more easily. Additionally, seismic reflection experiments can be conducted with the aid of synthetic seismogram analysis to investigate the shearwave velocity structure of the shallow-water sediments [23].

Seismic measurements for characterization and monitoring of the subsurface targets are also performed by means of distributed acoustic sensing (DAS) and with distributed temperature sensing (DTS). DAS has already been successfully used in the field of earthquake seismology [24–28], vertical seismic profiling [29, 30], and ambient noise velocity inversions [31]. DTS measurements have been used for monitoring of subsea structures [32] and of carbon capture, utilization and storage [33]. Thus, DAS and DTS could also be very useful in for characterization of fluid mud.

Utilization of DAS and DTS to measure seismic waves in the water and fluid mud offers advantages over the conventional electrical sensors such as electric isolation, immunity to electromagnetic interference, but also that they are nonconductive and non-corrosive, making them well-suited with regard to safety and durability for utilization in liquid-level sensing [34–37]. Such practical advantages are complemented by economical ones. There has been a rapid development in the optical fibers due to their wide usage by the communication industry. This has led to a substantial decrease in price, as well as an increase in performance. For instance, a single-mode optical fiber that used to cost \$ 20\$/m in 1979 costed just 0.1 \$/m in 2008 [38]. Given that the optical fibers are relatively cheap and require little to no maintenance, they could be very useful, from an economical point of view, as receivers for monitoring the nautical depth in ports and waterways. With the experiments we describe below, we investigate the utilization of optical fibers as receivers for fluid-mud level detection and characterization.

In the following, we use laboratory ultrasonic experiments to investigate how Pand S-wave measurements can be used for fluid-mud characterization. We discuss the latest results of seismic (ultrasonic) measurements of P- and S-waves propagation through fluid mud. In Section 2, we first describe the materials, sample preparation, and the rheological experiments for measuring the yield stress. We then introduce the ultrasonic measurements systems we use for transmission and reflection measurements. Subsequently, we describe the DAS and DTS measurement setups.

In Section 3, we present the results from the transmission measurements for monitoring possible changes of the P- and S-wave velocities when the ultrasonic signals propagate through fluid mud at different stages of consolidation. We link the observed transmission velocity changes to the measured yield stress during the same consolidation stages of the fluid mud. Further, we describe results from the reflection setup for estimating the layer-specific P- and S-wave velocities of the fluid mud. Finally, we validate the utilization of DAS and DTS as seismic and temperature receivers in laboratory experiments for detecting the fluid-mud/water interface.

In Section 4, we discuss the accuracy of our results and their applicability to other ports, while in Section 5 we draw conclusions.

## **2. Characterization and monitoring of fluid mud in a laboratory**

We develop laboratory ultrasonic measurement systems for transmission and reflection seismic measurements for characterization and monitoring of fluid mud while it is consolidating. The transmission seismic-measurements systems are designed for direct, fast, point-to-point measurements in the fluid mud using ultrasonic transducers or DAS as receivers. The reflection seismic-measurements system uses ultrasonic transducers to record waves that have reflected or refracted at different layer boundaries including the bottom of the water layer and the bottom of the fluid-mud layer. The reflection measurements can be used to record common-source gathers, which can subsequently be utilized to characterize velocity changes in the fluid mud during the consolidation using seismic-exploration techniques. We also describe the laboratory setup for rheological measurements of the fluid mud and the setup for DTS measurements.

## **2.1 Fluid-mud sample preparation and handling**

For the transmission and reflection measurements, we use fluid-mud samples extracted from the Calandkanaal (Port of Rotterdam) at the location indicated in **Figure 1a**. Before conducting the measurements, we stir a sample using a mechanical mixer in order to obtain a homogeneous volume of fluid mud with a uniform density. The density of the homogenized sample is 1197 kg/m3. After the homogenization, the fluid-mud sample appears like a mud slurry (**Figure 1b**). The samples are consecutively left to consolidate through a self-weight process. We perform ultrasonic measurements while the fluid mud is consolidating. Synchronously with the ultrasonic measurements, we also perform rheological measurements to investigate the yield stress. We investigate the fluidic yield stress using a recently

#### **Figure 1.**

*(a) Map of the port of Rotterdam illustrating the location of the site from where the fluid-mud samples had been collected (source: Google maps). (b) The process of homogenizing fluid mud with a mechanical mixer.*

## *Non-Intrusive Characterization and Monitoring of Fluid Mud: Laboratory Experiments with… DOI: http://dx.doi.org/10.5772/intechopen.98420*

developed protocol for the fluid mud [39, 40]. We use a HAAKE MARS I rheometer (Thermo Scientific) with two measuring geometries (Couette and vane) and apply stress ramp-up tests to measure the yield stress. The stress ramp-up tests are performed using a stress increase from 0 to 500 Pa at a rate of 1 Pa/s, until the shear rate reaches 300 s<sup>1</sup> , under a stress-control mode.

### **2.2 Transmission seismic measurements with transducer receivers**

The transmission seismic laboratory setup is equipped with two pairs of piezoelectric ultrasonic transducers (**Figure 2b** and **c**). Each pair consists of a source and receiver transducer, with one of the pairs using P-wave transducers and the other pair – S-wave transducers. The direct transmission measurement represents a pointto-point measurement with both transducer pairs placed along the horizontal direction. Because of this source-receiver geometry, the estimated velocities of the P- and S- waves correspond to transmissions along horizontal layers inside the fluid mud, if such layers are developed.

As shown in **Figure 2a**, the laboratory setup includes a fluid-mud tank, a signalcontrol part, and the two pairs of ultrasound transducers. The signal-control part in turn consists of a source-control part and a receiver-control part. In the sourcecontrol part, a function generator produces a desired signal, which signal is subsequently passed to a power amplifier to be finally passed to the source transducer, which sends it through the fluid mud. The fluid-mud tank is a plastic box that has

#### **Figure 2.**

*(a) Sketch of the transmission seismic laboratory setup with the fluid-mud box viewed from above and showing the horizontal arrangement of the two transducer pairs. (b) Side view of the fluid-mud box showing the vertical alignment of the ultrasonic transducers. (c) Photo of the fluid-mud box showing also the two source transducers.*

opening for the installation of the transducer end-caps. The receiver-control part of the setup consists of the receiver transducers, attached to the fluid-mud tank using end-caps, an oscilloscope for digitalization and displaying, and a computer, connected to the oscilloscope, to record the sensed signals. The generated source signal is also visualized on the oscilloscope for quality control.

For the transmission measurements, we use as a source signal a gated sine-wave pulse with a center frequency of 1 MHz. A measurement is performed using a pulsetime delay. To increase the signal-to-noise ratio, especially needed for the S-wave velocity estimations, a measurement at each stage of consolidation consists of 1024 repeated recordings summed together to obtain a final transmission recording. This is done for both the P- and S-wave pair.

For each stage of the measurements, the first step in estimating the propagation velocities is to pick the first arrivals of the P- and S-waves. The second step is to calculate the P- and S-wave velocities by dividing the travel distance of waves, which is the distance from the source to the receiver transducer within each pair (equal for both pairs), by the travel times estimated from the picked first arrivals.

#### **2.3 Reflection seismic measurements with transducer receivers**

Similar to the transmission seismic laboratory setup, the reflection system consists of a signal-control part, a fluid-mud tank, and ultrasound transducers, but further to that also includes a transducer-placement part (**Figure 3**). While the signal-control part is the same as for the transmission measurements (**Figure 3b**), the fluid-mud tank is different and only one pair of ultrasonic transducers is used (**Figure 3c**). The transducer-placement part allows changing the positions of the transducers by moving them along horizontal and vertical bars (**Figure 3a** and **c**). This facilitates recording of reflections at multiple horizontal positions to obtain reflection common-source gathers, if desired with sources and receivers at different depths.

In the measurements we perform, the transducers are placed a certain distance above the top of the fluid-mud layer to better mimic a geometry of a marine seismicexploration survey. While placing sources and receivers during a field measurement campaign directly at the top of the fluid mud would allow direct recording of Swaves, a recording geometry with seismic sources and receivers towed at a certain height above the bed in the navigational channel is more practical – the surface of the sediments is seldomly flat, and hard object protruding from the sediments could damage the sources and/or receivers. On the other hand, towing the sources and receivers at a distance above the top of the fluid-mud layer inevitably brings uncertainty in the estimated seismic velocities caused by the salinity and temperature of the water. It is possible to monitor the changes in the salinity and temperature at specific locations, but the uncertainty still remains when using such point measurements for larger-area surveys due to the dynamics of the marine environments.

In order to eliminate these uncertainties, we apply SI for retrieval of ghost reflections from inside the fluid-mud layer and eliminate the travel-paths of the waves in the water layer. For pressure measurements in water, like in our laboratory setup, a general representation of SI by cross-correlation is [41].

$$p(R2, R1, t) + p(R2, R1, -t) \propto \sum\_{S=S1}^{SN} p(R2, S, t) \bigotimes p(R1, S, t),\tag{1}$$

where *p R*ð Þ 2, *R*1, *t* is the retrieved pressure response at receiver at *R*2 from a virtual source at the position of a receiver at *R*1, *p R*ð Þ 2, *S*, *t* is the pressure response measured at *R*2 from a source at *S*, with *S*1 … *SN* sources distributed evenly over

*Non-Intrusive Characterization and Monitoring of Fluid Mud: Laboratory Experiments with… DOI: http://dx.doi.org/10.5772/intechopen.98420*

#### **Figure 3.**

*Reflection seismic-measurements system. (a) Cartoon of the fluid-mud tank with the transducer-placement part and the signal-control part (identical to the one in the transmission measurements). Red star indicates the source and black probe indicates the receiver. The transducer-placement part allows vertical (blue arrows) and horizontal (white arrows) displacement of the source and receiver. (b) Photo of the signal-control part. (c) Photo of the fluid-mud tank and the transducer-placement part with a source and receiver ultrasonic transducers.*

surface that effectively surrounds the two receivers, �*t* indicates time reversal (acausal time), and ⨂ indicates correlation. As mentioned above, when the assumptions for this simplified representation are not met [14], e.g. as in a seismic reflection survey when the sources are only at the surface and thus do not surround the receivers, ghost reflections are retrieved [17, 18], and we can write

$$p(R2, R1, t) + p(R2, R1, -t) + \text{photons} \propto \sum\_{S=S\text{K1}}^{S\text{K2}} p(R2, S, t) \bigotimes p(R1, S, t),\tag{2}$$

where *ghosts* represents retrieved non-physical arrivals, including ghost reflections, and *SK*1 and *SK*2 now indicate that the summation is only over sources on a limited surface. For practical purposes, in our laboratory setup we choose to have only two source positions and multiple receiver positions. Using source-receiver reciprocity, we can thus rewrite relation (2) as

$$p(\text{S2,S1},t) + p(\text{S2,S1},-t) + \text{photons} \propto \sum\_{R=RK1}^{RK2} p(\text{S2,R},t) \bigotimes p(\text{S1,R},t),\tag{3}$$

where the summation is now over receiver positions and we retrieve a pressure recording at a virtual receiver at the position of source *S*2 from a sourse at *S*1. Thus, to apply SI, we use two common-source gathers (CSGs). The two source positions (labeled Source 1 or S1 and Source 2 or S2 in **Figures 4** and **5**, respectively) at the same height and distanced in the horizontal direction 50 mm from each other. We record the reflected signal at a receiver, labeled Receiver 1 (R1) in **Figure 4**, aligned with the two source positions and distanced 100 mm from S1 (and thus 50 mm from S2). Following the nomenclature in [17], the source and virtual receiver redatumed by SI to the top of the fluid-mud layer during the ghost-reflection retrieval are referred to as ghost source and ghost receiver, respectively. Assuming a favorable geometry, to explain the retrieval of a ghost reflection inside the fluidmud layer, the travel-path of the reflection from the fluid-mud bottom, i.e., the travel-path starting from S1, transmitted at the water/mud interface, reflected by the fluid-mud bottom, transmitted at the mud/water interface, and then arriving at R1 is labeled 1–2–3-4 in **Figure 4**. The travel-path of the reflection from the water/ mud interface, starting from S2 and arriving at R1, is labeled 1<sup>0</sup> -4<sup>0</sup> . Cross-correlation of the recorded reflections at R1 from S1 and S2 will effectively result in removal of the common travel-paths in 1–2–3-4 and 1<sup>0</sup> -4<sup>0</sup> . Thus, the parallel travel-paths 1 and 1<sup>0</sup> and the coinciding travel-paths 4 and 4<sup>0</sup> are eliminated, and only the travel-path 2–3 is left over representing a ghost reflection only inside the fluid-mud layer from a ghost source and ghost receiver placed directly at its top (**Figure 4**). In reality, the exact receiver position ensuring that travel-paths 1 = 1<sup>0</sup> and 4 = 4<sup>0</sup> is unknown. Because of that, recordings at multiple receiver positions from both sources are required, i.e., two CSGs. To obtain such gathers, we displace the receiver from position R1 to the right along the horizontal bar by 5 mm multiple times and record

#### **Figure 4.**

*Illustration of the geometry needed for retrieval of ghost reflections from inside the fluid-mud layer. See text for explanation of the symbols.*

*Non-Intrusive Characterization and Monitoring of Fluid Mud: Laboratory Experiments with… DOI: http://dx.doi.org/10.5772/intechopen.98420*

#### **Figure 5.**

*(a) Illustration of the travel-paths of the expected arrivals from S1 to a receiver in the reflection measurements. (b) Wiggle plot of the recorded CSG from S1. (c) Wiggle plot of the recorded CSG from S2. (d) Sketch of the travel-paths of the primary reflections of the mud top in the CSG from S2. (e) Sketch of the travel-paths of the primary reflection of the mud bottom (PPPP) in the CSG from S1. The ghost reflection is retrieved by summing the individual arrivals highlighted in green in (e) obtained from cross-correlating the primary reflection from the fluid-mud top in the CSG from S2 with the primary reflection PPPP in the CSG from S1.*

for the same source at each receiver position. In this case, we record at 20 positions. That is, the CSGs for S1 and S2 consist of 20 traces each.

The source signal we use is similar to the one for the transmission measurements but with a center frequency of 100 kHz.

Also with these measurements, to increase the signal-to-noise ratio of the recorded signals, a measurement at each receiver position from each source is repeated 1024 times and the 1024 measurements are summed together to obtain a final trace for that source and receiver positions.

Using the travel-path sketches in **Figure 5a**, we explain several arrivals of interest in the CSGs. **Figure 5b** and **c** present wiggle plots of the recorded CSGs from S1 and S2, respectively. We calculate expected arrival times based on the source/ receiver offsets and the thicknesses of the water and fluid-mud layers, each of which we can directly measure. For propagation through the water layer, we use P-wave velocity of 1500 m/s. For the waves propagating through the fluid mud, we use values estimated from the transmission measurements – 1570 m/s for the P-wave velocity and 958 m/s for the S-wave velocity. The calculated reference times are illustrated by dashed lines superimposed on the CSGs to assist in interpretation of the arrivals. In **Figure 5b** and **c**, the reflection arrivals of interest in this study are the primary reflection from the fluid-mud top (magenta) and the three primary reflections from the fluid-mud bottom that are labeled as PPPP (blue), PPSP (red), and PSSP (orange). The S-waves in the experiment appear as waves converted from P to S at the top or the bottom of the fluid-mud layer. For example, the P-to-S converted wave in PPSP is generated when the P-wave impinging on the fluid-mud bottom is reflected as an S-wave; the P-to-S converted wave in PSSP is generated when the P-wave impinging on the fluid-mud top in transmitted to the fluid mud as an S-wave (**Figure 5a**) and continues to propagate as an S-wave until reaching the fluid-mud top again.

To retrieve ghost reflections, one can use relation (3) and correlate the CSGs. Such an approach could result in other retrieved arrivals interfering with the desired ghost reflections. To avoid that, we follow [17] and correlate only specific arrivals. To retrieve a P-wave ghost reflection from inside the fluid mud, we crosscorrelate the primary reflection from the fluid-mud top in the CSG from S2 (**Figure 5d**) with the primary reflection PPPP in the CSG from S1 (**Figure 5e**). In a similar way, the P-to-S converted ghost reflection and S-wave ghost reflection are retrieved using the reflections PPSP and PSSP in the CSG from S1, respectively.

### **2.4 Transmission seismic measurements using DAS and measurements with DTS**

We use a standard single-mode communication fiber for both the DAS and DTS measurements. This means that we can combine the two methods and compare the difference in their performance With DAS, such fibers can act as seismic receivers that measure the dynamics of a strain field acting on a fiber [42]. With DTS, such fibers can act as strain and temperature sensors (and thus also labeled DT(S)S), which measure the static strain and temperature along the fiber [43].

To verify that these fibers can serve as receivers for fluid-mud level detection and characterization, we conduct seismic and temperature laboratory experiments using commercially available interrogators. These interrogators are the iDAS from Silixa and DITEST STA-R from Omnisens for measuring the acoustic impedance and temperature, respectively. For a more detailed explanation of the iDAS system, the reader is referred to [42].

Our fiber is coiled around a PVC pipe with a diameter of 0.125 m, which allows us to use more fiber and, hence, have more measuring points than when using a

*Non-Intrusive Characterization and Monitoring of Fluid Mud: Laboratory Experiments with… DOI: http://dx.doi.org/10.5772/intechopen.98420*

straight fiber. In addition, the coining increases the vertical resolution by compressing the gauge length of 10 m of the cable (the length over which the backscattered signal is averaged to increase the signal-to-noise ratio of the detected dynamic deformation) only over a few vertical centimeters. Due to the coiling, we also change the directional sensitivity [44], making the cable more sensitive to horizontal waves, with respect to the column. The PVC pipe with the fiber coiled on it is placed inside a transparent column. We first perform experiments with two types of synthetic clay, namely kaolinite and bentonite, and subsequently with two types of fluid mud – one from the Port of Rotterdam, which is the same sample mud as described above, and the other from the Port of Hamburg. For the experiments with the synthetic clays, we fill the lowest part of the column, without coiled optical fiber, with sand. Above the sand, we put one of the clays, and then we fill the remainder with water. For the fluid-mud experiments, we instrument also the lowest part of the column with fiber and start filling the column with one of the fluid muds starting already at the bottom, while we again fill the remainder of the column with water. A schematic overview and pictures of the setup are shown in **Figure 6**. Note that for the measurements with kaolinite and bentonite, we have 0.5 m in depth, which is 123 m in fiber length, acting as sensors. For the measurements in the muds, we added 0.2 m in depth, giving us a total of 171 m of fiber length, acting as sensors. For both setups, we have 10 m of fiber outside of our column to use as a reference.

With DAS, we try to capture the water/mud interface and measure the shear strength build-up. We test various sources for these purposes. Our sources include a small transducer with a center frequency of 500 kHz, a larger transducer with a center frequency of 200 kHz (**Figure 6b** and **c**) and a common duo echo-sounder with a center frequency of 38 kHz and 200 kHz, which is also used by marine vessels to measure depth. We connect these sources to the same source-side signalcontrol part as described above. We use a frequency range from 25 kHz to 45 kHz, since preliminary results indicated that this range should give the best results. The sampling frequency of the DAS system is set at the maximum of the system, which is 100 kHz.

For the DTS measurements, we use two standard heating rods, which we place 5 cm away from the fiber, to heat the column and measure the difference with respect to time along the column. This we only do for the kaolinite sample, since a very similar result is expected for the other clay and two mud samples.

**Figure 6.**

*(a) Schematic overview of the setup for DAS and DTS measurements. A photo of the column with the optical fiber wound around the PVC pipe when using mud from the (b) port of Rotterdam and (c) port of Hamburg.*

## **3. Results**

We describe the results of the ultrasonic transmission measurements with ultrasonic transducers and correlate them to the results from the rheological measurements. We further report the results from the reflection measurements and how they were used to retrieve ghost reflections. We then show the results from the DAS and DTS measurements.

## **3.1 P- and S-waves velocities in the fluid mud from transmission measurements with ultrasonic transducers**

We examine the first arrivals of transmitted P- and S-waves and estimate their velocity variations during the consolidation of the fluid mud. We do not observe a detectable change in the P-wave velocity – the P-wave first arrivals appear to be constant throughout the consolidation process (**Figure 7a**). This finding agrees with previous results reporting that the S-wave velocity is more sensitive to changes in lithology and mechanical properties than the P-wave velocity [45]. The traveltime of the direct arrivals of the P-wave is 0.074 ms (**Figure 7a**), and thus the corresponding velocity is 1570 m/s. By examining the change in arrival time of the

*(a) Transmission recordings of the direct P- and S-wave arrivals as a function of consolidation time. (b) Estimated S-wave velocity as a function of the consolidation time.*

*Non-Intrusive Characterization and Monitoring of Fluid Mud: Laboratory Experiments with… DOI: http://dx.doi.org/10.5772/intechopen.98420*

first S-wave arrival (**Figure 7a**), we find that the S-wave traveltime decreases with consolidation time, indicating that the S-wave velocity increases with the consolidation progress (**Figure 7b**).

We can see from **Figure 7**, that during the first three days the S-wave velocity is nearly stable exhibiting very little fluctuations. Starting from Day 3, the S-wave velocity shows a strong increase from 959 to 995 m/s during the next two days. In the second week, the S-wave velocity only experiences a small increase and eventually reaches 998 m/s. By comparing the velocity variations of the P-wave and Swaves, we can summarize that the relative increase in the S-wave velocity is much stronger than in P-wave velocities, validating the statement that the S-waves are much more sensitive to the consolidation of the fluid mud than the P-waves. This finding agrees with a previous in-situ seismic exploration results using pulsetransmission techniques [45].

By drawing the estimated S-wave velocities from **Figure 7b** as a function of the concurrently estimated fluidic yield stresses (**Figure 8**), we see a positive correlation during the consolidation of the fluid mud. The correlation appears to indicate that the S-wave velocity starts increasing after the fluidic yield stress exceeds some critical value (for each of the Couette and vane geometry). Once the critical value is surpassed, the S-wave velocity increases with the increasing fluidic yield stress caused by the consolidation.

#### **3.2 P- and S-wave velocities inside the fluid-mud layer from ghost reflections**

The recorded primary reflections from the fluid-mud top and bottom are identified and shown in **Figure 9**. We apply SI using the reflection from the mud top in the CSG from S2 and the primary reflections PPPP, PPSP, and PSSP from the mud bottom in the CSG from S1 (**Figure 9**). As explained in Section 2.2, the ghost reflections are retrieved by eliminating the P-wave travel-paths inside the water. The ghost reflections in **Figure 10**, retrieved using the primary reflections PPPP, PPSP, and PSSP, are labeled PP, PS, and SS, respectively. In **Figure 10**, we also show the length of each of the legs of the reflection travel-paths of the retrieved ghost reflections. We use these lengths to estimate the wave velocities using the arrival times of the retrieved ghost reflections.

#### **Figure 8.**

*Relationship between the estimated S-wave velocities (Figure 7b) and the concurrently estimated fluidic yield stress, using Couette and vane geometry, as a function of the consolidation time.*

#### **Figure 9.**

*Identified primary reflections in the common-source gather from (a) source 1 and (b) source 2. We apply seismic interferometry (SI) by correlating (the* ⊗ *symbol) the reflection from the mud top with each of the three identified reflections from the mud bottom followed by summation over the receivers (Eq. 3).*

#### **Figure 10.**

*The travel distances of the travel-paths of the ghost reflections PP, SS, and PS when the fluid-mud thickness is 86 mm, which is the thickness on day 11 of the consolidation.*

As explained, the retrieved result is obtained by stacking the correlated traces. When the receiver array is sufficiently long, the stacking would have resulted in the retrieved ghost reflections only, with the contribution to the retrieved signal coming from summation inside the so-called stationary-phase region [46], i.e., the region where a curve appears nearly horizontal. In **Figures 11a**–**13a**, we indicate the stationary-phase regions with green dashed rectangles. Because our receiver array is of a limited length and is further only on one side of the sources, summation of all traces produces more or less erroneous results (**Figures 11b**–**13b**). Because of this, to retrieve the ghost reflections we use for the summation only traces in the stationary-phase region (**Figures 11c**–**13c**). We then pick from those results the two-way traveltimes to estimate the velocities inside the fluid-mud layer.

*Non-Intrusive Characterization and Monitoring of Fluid Mud: Laboratory Experiments with… DOI: http://dx.doi.org/10.5772/intechopen.98420*

#### **Figure 11.**

*Two-way traveltime pick of the ghost reflection PP. (a) Correlation result of the reflection from the fluid-mud top from Figure 9b with the PPPP reflection from Figure 9a. The stationary-phase region is indicated by the dashed green rectangle. The retrieved ghost reflection PP when summing over (b) all traces in a and (c) the traces inside the stationary-phase region in a.*

#### **Figure 12.** *As in Figure 11 but for ghost reflection PS. The correlation in (a) is with the PPSP reflection.*

Dividing the travel distance of 179.2 mm, which ghost reflection PP has traversed inside the fluid-mud layer (**Figure 10**) by the picked two-way traveltime from **Figure 11c**, we estimate the P-wave velocity to be 1592 m/s. To estimate directly the S-wave velocity inside the fluid-mud layer, we divide the travel distance the ghost reflection SS has traversed inside the fluid-mud layer, again 179.2 mm (**Figure 10**), by the picked two-way traveltime from **Figure 13c**, and obtain 995 m/s. Comparing this value with the estimated value from the transmission measurements on day 11 of 998 m/s (**Figure 7b**), we see that the difference is only 0.3%, which is negligible. Comparison of the estimated P-wave velocity to the value from the transmission measurements of 1570 m/s, we see that the difference is 1.4%, which is a bit higher but still acceptable.

### **3.3 Detection of the water/mud interface using DAS and DTS**

**Figure 14** shows DAS measurements of the arrivals recorded along the fiber as a function of arrival time when using the fluid mud from the Port of Hamburg and

**Figure 13.** *As in Figure 11 but for ghost reflection SS. The correlation in (a) is with the PSSP reflection.*

#### **Figure 14.**

*DAS recordings using the setup from Figure 6c showing direct arrivals and multiple reflections. The blue line indicates the water/mud interface, which is at 90.7 m along the fiber.*

the large transducer as a source (**Figure 6c**). We perform the measurements after the mud has consolidated for 9 days. To improve the signal-to-noise ratio, we repeat the recordings 10 times and then stack them. Using the first arrivals, i.e., the direct P-wave, we estimate the P-wave velocity in water to be around 1450–1500 m/s, while in the fluid mud to be 1490–1570 m/s. The reason for the uncertainty is likely related to the relatively low rate of time sampling of 100 kHz for the source frequency we use of 25–45 kHz. For this sampling rate, the Nyquist frequency is 50 kHz, which is very close to the source frequencies and, thus, makes the velocity analysis more ambiguous. The small difference in the P-wave velocity of the water and the fluid mud combined with the uncertainties make the detection of the water/ mud interface rather challenging if the first arrival as used.

### *Non-Intrusive Characterization and Monitoring of Fluid Mud: Laboratory Experiments with… DOI: http://dx.doi.org/10.5772/intechopen.98420*

The recordings in **Figure 14** show that a more accurate and robust criterion to detect the water/mud interface is to look at the multiple reflections and their amplitude attenuation. Looking at the figure, we can see that later arrivals appear to faint, i.e., are more attenuated after the water/mud interface, with the latter indicated by the blue line. Taking a closer look at the multiple reflections, we see that these later arrivals have completely fainted after 93.7 m fiber length, with the water/mud interface at 90.7 m fiber length. This difference of 3 m of fiber might be related to the gauge length of the fiber, i.e., the length over which the DAS system averages the observations, which in our case is 10 m. Another reason could be the uncertainty in the exact position of the fiber.

The measurements with the fluid mud from the Port of Rotterdam and the two clays show similar results.

We also look at the signal attenuation due to the consolidation of the mud, and thus the increase of its shear strength. **Figure 15a** and **b** show the DAS recordings in bentonite clay performed on the first and second day of the consolidation, respectively. We see a clear difference in signal penetration through the bentonite clay – on the first day, there is little to no signal penetration, opposed to the second day, when the waves propagate all the way through the column. This difference is purely related to the buildup of shear strength in the bentonite, since bentonite does not settle but builds up shear strength with time.

From the tests we perform with different types of sources (small and large transducer and duo echo sounder) we observe that the small transducer with resonant frequency of 500 kHz does not generate enough energy when we use it for emitting a P-wave at 25 kHz – 45 kHz. For that reason, it is outperformed by the big transducer whose resonant frequency of 200 kHz is closer to our target sourcesignal frequency of 25 kHz – 45 kHz. The duo echo sounder generated by far the strongest signal; however, because it was mounted on the transparent outer column and was situated right above our PVC pipe, a lot of tube waves and refracted waves are generated, which are undesired in our tests. These strong interfering events could potentially be suppressed applying further signal processing, as we suggest above – for example using a frequency-wavenumber filter.

Besides using the optical fiber as a receiver for seismic waves, we also use it as DTS recorder to measure temperature. Due to the difference in the heat capacity and heat conductivity between water and mud, a difference in heating occurs when we start heating up the column using heating rods in the water and kaolinite. This difference can be observed in **Figure 16**, where we show the measured Brillouin

#### **Figure 15.**

*DAS recordings when synthetic clay (bentonite) is used as fluid mud. The recordings were done after (a) half an hour and (b) 24 hours of consolidation of the bentonite.*

#### **Figure 16.**

*Brillouin frequency changes in (a) water and (b) mud after increasing the temperature of the water in the column by 1°C. the red line indicates the water/mud interface. The brown curve represents a reference measurement without heating. The curves with colors other than brown represent measurements after heating up the water several times by 1°C.*

frequency when we heat up the water and kaolinite. The brown curves show a reference measurement before the heating, while the other colored curves show the measurements after increasing the temperature of the water each time by 1°C. Inside the water layer (**Figure 16a**), we observe a linear increase in the Brillouin frequency per <sup>o</sup> C. Inside the mud layer, however, we see a non-linear trend due to the lower heat capacity and lower heat conductivity. This is especially visible along the red curve, which characterizes the first measurement after we start heating up the column: we see that in the lower part, starting at 99 m, the red curve overlaps the brown curve meaning that the heat from the heating rods has not yet reached the fiber at that level and deeper.

The DTS measurements show that interpretation of the water/mud interface can be achieved with a likely accuracy of around 4 cm.

## **4. Discussion**

The direct transmission measurements of the P- and S-wave velocities inside the fluid-mud layer showed that the P-wave velocity is nearly independent of the consolidation process while the S-wave velocity significantly increases during the consolidation. This can be attributed to the property changes of the fluid mud due to the compaction effect of the consolidation and potentially the production of gas in the mud. The S-wave velocity is principally determined by the grain structure and shear modulus of the frame of the solid phase (minerals). The P-wave velocities on the other hand depend on the elastic moduli of the grains, sediment frame, and bulk modulus of the fluid. Thus, for marine sediments with high porosity, such as the fluid mud, the S-wave rather than the P-wave is strongly affected by the consolidation, and, thus, can be potentially used to characterize the consolidation process.

Using SI for retrieval of ghost reflections inside the fluid-mud layer, we removed the kinematic influence of the water layer above the mud. The estimated velocities of the P- and S-waves using the ghost reflections PP and SS, respectively, were very close to the ones estimated from the direct transmission measurements inside the fluid-mud layer. Because we also had the ghost reflection PS (**Figure 12**), we could estimate the S-wave velocity inside the fluid mud also from this arrival. We did this making use of the already estimated P-wave velocity for the propagation along the P-wave path of 91.6 mm in **Figure 10**. The value we then obtained was 991 m/s, which is quite close to the value of the S-wave velocity obtained from the ghost reflection SS, but is of course inheriting errors from the estimation of the P-wave velocity. Nevertheless, all three values can be used as quality control of each other

#### *Non-Intrusive Characterization and Monitoring of Fluid Mud: Laboratory Experiments with… DOI: http://dx.doi.org/10.5772/intechopen.98420*

or as substitutes when one of the three ghost reflections cannot be reliably retrieved due to, for example, interference from other arrivals.

Observing the multiple reflections in the DAS recordings, we estimated an error of 3 m along the coiled fiber in detecting the depth of the water/mud interface. Since we coiled the fiber around a PVC pipe with a diameter of 0.125 m and because the fiber's thickness is 1.6 mm, the 3-meter error of fiber length translates to 1.2 cm of vertical error in the depth of the water/mud interface. With such an error, to the best of our knowledge, our approach is the most accurate non-intrusive method for determining the depth of the water/mud interface. Note that to achieve this accurate result, the only processing we applied was to increase the signal-to-noise ratio by the summation of the 10 separate recordings. More signal processing could further improve the determination of the water/mud interface. We expect that a similar high accuracy is achievable in the field as well since the upper end of the optical fiber is placed at the very bottom of the water layer, which limits errors caused by differences in, for instance, the water temperature.

The direct transmission measurements with DAS, on the other hand, allowed estimation of the P-wave velocity in the fluid mud in the range 1490–1570 m/s. Comparing these values to the value of 1570 m/s from the direct transmission measurements horizontally inside the fluid mud means an uncertainty of about 5.1%, which is not negligible. This confirms the difficulty when using a source in the water and receivers in the fluid mud, and clearly underlines the advantage of using SI with ghost reflections from reflection measurements. Thus, we argue that another very useful application of DAS could be with direct transmission measurements inside the fluid-mud layer, and thus also for transmission tomography between a vertical array of sources inside the mud and a vertical DAS pole with coiled fiber.

For our laboratory measurements, we used fluid-mud samples from the Port of Rotterdam and the Port of Hamburg. Nevertheless, our results and conclusions can be generalized to fluid-mud samples from other ports. Because the estimated P- and S-wave velocities using the ghost reflections do not depend kinematically on the water layer, this technique could easily be applied to any port or waterway. Of course, the P- and S-wave velocities of the fluid mud will differ from place to place, so those will need to be estimated for each place, for example for correlation with the yield stress. The DAS and DTS techniques for estimating the water/mud boundary can likewise be used at any other port or waterway, as they depend only on the strong contrast in the observed parameters between the layer and fluid-mud layer.

### **5. Conclusions**

We presented recent results for non-intrusive characterization and monitoring of fluid mud in ports and waterways using ultrasonic measurements in transmission and reflection geometry, including measurements with Distributed Acoustic Sensing (DAS), and using temperature measurements with Distributed Temperature Sensing (DTS). We performed the measurements in a laboratory on samples from the Port of Rotterdam, Port of Hamburg, and two synthetic clays.

Using ultrasonic transmission measurements with transducers directly inside fluid mud, we investigated the changes in the velocities of longitudinal (P-) and transverse (S-) waves and their possible relation to the yield stress during the consolidation. We observed no detectable change of the P-wave velocities during the consolidation of the fluid mud. We observed that the S-wave velocities exhibited a relatively strong increase after the fluid mud settles for a certain amount of time, in our study after 3 days. Comparing the estimated S-wave velocities to the concurrently estimated fluidic yield stress, we showed a positive correlation between the two. Our findings verify that the S-wave velocities increase with increasing yield stress caused by the fluid-mud consolidation and can thus be potentially used for indirect in-situ assessment of the yield stress.

Using ultrasonic reflection measurements with transducers, we investigated the direct estimation of the P- and S-wave velocities inside the fluid-mud layer. The source and receiver transducers were placed inside the water layer, but we showed that the kinematic influence of the water layer can be completely eliminated by retrieval of non-physical (ghost) reflections inside the fluid mud by application of seismic interferometry. Using the retrieved ghost reflections to estimate the layerspecific P- and S-waves velocities of the mud, we eliminated possible uncertainty due to salinity and temperature gradients of the water, which affect the velocity estimates using the usual seismic-reflection processing techniques. We show that the reflection-estimated velocities differ from the transmission-calculated values only by 1.4% and 0.3% for the P- and S-waves, respectively.

We also showed that DAS and DTS can be very effective in estimating the depth of the water/mud interface. We showed that a standard communication fiber is sufficient to achieve an accuracy in the estimated depth of the water/mud interface of 1.2 cm. This accuracy, to the best of our knowledge, is higher than what is achievable with any the currently used non-intrusive methods. Furthermore, we showed that the strength of the signal recorded with DAS is linked to changes in the shear strength of clays.

## **Acknowledgements**

The research of X.M. is supported by the Division for Earth and Life Sciences (ALW) with financial aid from the Netherlands Organization for Scientific Research (NWO) with grant no. ALWTW.2016.029. The research of M.B. is supported by the Port of Rotterdam, Hamburg Port Authority, Rijkswaterstaat and SmartPort. The project is carried out also within the framework of the MUDNET academic network https://www.tudelft.nl/mudnet/.

## **Conflict of interest**

The authors declare no conflict of interest.

*Non-Intrusive Characterization and Monitoring of Fluid Mud: Laboratory Experiments with… DOI: http://dx.doi.org/10.5772/intechopen.98420*

## **Author details**

Deyan Draganov<sup>1</sup> \*, Xu Ma1 , Menno Buisman1,2, Tjeerd Kiers1 , Karel Heller<sup>1</sup> and Alex Kirichek1,3

1 Faculty of Civil Engineering and Geosciences, Delft University of Technology, Delft, The Netherlands

2 Port of Rotterdam, Rotterdam, The Netherlands

3 Deltares, Delft, The Netherlands

\*Address all correspondence to: d.s.draganov@tudelft.nl

© 2021 The Author(s). Licensee IntechOpen. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/ by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

## **References**

[1] McAnally WH, Friedrichs C, Hamilton D, Hayter E, Shrestha P, Rodriguez H, Sheremet A, Teeter A, ASCE task committee on Management of Fluid mud. Management of fluid mud in estuaries, bays, and lakes. I: Present state of understanding on character and behavior. Journal of Hydraulic Engineering. 2007 Jan;133(1):9-22.

[2] Harbour Approach Channels Design Guidelines. In: Report of Marcom Working Group. 2014. p. 49.

[3] Kirichek A, Chassagne C, Winterwerp H, Vellinga T. How navigable are fluid mud layers. Terra et Aqua: International Journal on Public Works, Ports and Waterways Developments. 2018;151.

[4] Delefortrie G, Vantorre M, Eloot K. Modelling navigation in muddy areas through captive model tests. Journal of marine science and technology. 2005 Dec 1;10(4):188-202.

[5] Vantorre M. Ship behaviour and control in muddy areas: state of the art. InProceedings of the 3rd International Conference on Manoeuvring and Control of Marine Craft (MCMC'94), edited by GN Roberts and MMA Pourzanjani, Southampton 1994 Sep (pp. 7-9).

[6] Claeys S, De Schutter J, Vantorre M, Van Hoestberghe T. Rheology as a survey tool: We are not there yet. Hydro International. 2011;15(3):14-19.

[7] Kirichek A, Rutgers R. Monitoring of settling and consolidation of mud after water injection dredging in the Calandkanaal. Terra et Aqua. 2020; 160: 16-26

[8] Kirichek A, Shakeel A, Chassagne C. Using in situ density and strength measurements for sediment maintenance in ports and waterways.

Journal of Soils and Sediments. 2020 Feb 19:1-7.

[9] Hamilton EL, Bachman RT. Sound velocity and related properties of marine sediments. The Journal of the Acoustical Society of America. 1982 Dec;72(6):1891-1904.

[10] Schrottke K, Becker M, Bartholomä A, Flemming BW, Hebbeln D. Fluid mud dynamics in the Weser estuary turbidity zone tracked by high-resolution side-scan sonar and parametric sub-bottom profiler. Geo-Marine Letters. 2006 Sep 1;26(3): 185-198.

[11] Gratiot N, Mory M, Auchere D. An acoustic Doppler velocimeter (ADV) for the characterisation of turbulence in concentrated fluid mud. Continental Shelf Research. 2000 Sep 10;20(12–13): 1551-1567.

[12] Meissner R, Rabbel W, Theilen F. The relevance of shear waves for structural subsurface investigations. InShear waves in marine sediments 1991 (pp. 41-49). Springer, Dordrecht.

[13] Shapiro NM, Campillo M. Emergence of broadband Rayleigh waves from correlations of the ambient seismic noise. Geophysical Research Letters. 2004 Apr 16;31(7).

[14] Wapenaar K, Fokkema J. Green's function representations for seismic interferometry. Geophysics. 2006 Jul;71 (4):SI33-SI46.

[15] Draganov D, Campman X, Thorbecke J, Verdel A, Wapenaar K. Reflection images from ambient seismic noise. Geophysics. 2009 Sep;74(5):A63- A67.

[16] Draganov D, Ghose R, Ruigrok E, Thorbecke J, Wapenaar K. Seismic

*Non-Intrusive Characterization and Monitoring of Fluid Mud: Laboratory Experiments with… DOI: http://dx.doi.org/10.5772/intechopen.98420*

interferometry, intrinsic losses and Qestimation. Geophysical Prospecting. 2010 Mar 26;58(3):361-373.

[17] Draganov D, Heller K, Ghose R. Monitoring CO2 storage using ghost reflections retrieved from seismic interferometry. International Journal of Greenhouse Gas Control. 2012 Nov 1;11: S35-S46.

[18] King S, Curtis A. Suppressing nonphysical reflections in Green's function estimates using source-receiver interferometrySuppressing nonphysical reflections. Geophysics. 2012 Jan 1;77 (1):Q15-Q25.

[19] Breitzke M. Acoustic and elastic characterization of marine sediments by analysis, modeling, and inversion of ultrasonic P wave transmission seismograms. Journal of Geophysical Research: Solid Earth. 2000 Sep 10;105 (B9):21411-21430.

[20] Leurer KC. Compressional-and shear-wave velocities and attenuation in deep-sea sediment during laboratory compaction. The Journal of the Acoustical Society of America. 2004 Oct;116(4):2023-2030.

[21] Ballard MS, Lee KM, Muir TG. Laboratory P-and S-wave measurements of a reconstituted muddy sediment with comparison to cardhouse theory. The Journal of the Acoustical Society of America. 2014 Dec;136(6):2941-2946.

[22] Ballard MS, Lee KM. Examining the effects of microstructure on geoacoustic parameters in fine-grained sediments. The Journal of the Acoustical Society of America. 2016 Sep 8;140(3):1548-1557.

[23] Collins JA, Sutton GH, Ewing JI. Shear-wave velocity structure of shallow-water sediments in the East China Sea. The Journal of the Acoustical Society of America. 1996 Dec;100(6): 3646-3654.

[24] Ajo-Franklin JB, Dou S, Lindsey NJ, Monga I, Tracy C, Robertson M, Rodriguez Tribaldos V, Ulrich C, Freifeld B, Daley T, Li X. Distributed acoustic sensing using dark fiber for near-surface characterization and broadband seismic event detection. Sientific Reports. 2019; 9:1328.

[25] Jousset P, Reinsch T, Ryberg T, Blanck H, Clarke A, Aghayev R, Hersir GP, Henninges J, Weber M, Krawczyk CM. Dynamic strain determination using fibre-optic cables allows imaging of seismological and structural features. Nature Communications. 2018; 9: 2509.

[26] Lindsey NJ, Martin ER, Dreger DS, Freifeld B, Cole S, James RS, Biondi BL, Ajo-Franklin JB. Fiber-optic network observations of earthquake Wavefields. Geophysical Research Letters. 2017 Dec 16; 44(23):11792-11799.

[27] Wang HF, Zeng X, Miller DE, Fratta D, Feigl KL, Thurber CH, Mellors RJ. Ground motion response to an ML 4.3 earthquake using co-located distributed acoustic sensing and seismometer arrays. Geophysical Journal International. 2018 Jun;213(3):220-236.

[28] Yu C, Zhan Z, Lindsey NJ, Ajo-Franklin JB, Robertson M. The potential of DAS in Teleseismic studies: Insights from the goldstone experiment. Geophysical Research Letters. 2019 Feb 16; 46(3):1320-1328.

[29] Daley TM, Miller DE, Dodds K, Cook P, Freifeld BM. Field testing of modular borehole monitoring with simultaneous distributed acoustic sensing and geophone vertical seismic profiles at Citronelle, Alabama. Geophysical Prospecting. 2016; 64(5): 1318-1334.

[30] Mateeva A, Lopez J, Potters H, Mestayer J, Cox B, Kiyashchenko D, Wills P, Grandi S, Hornman K, Kuvshinov B, Berlang W, Yang Z,

Detomo R. Distributed acoustic sensing for reservoir monitoring with vertical seismic profiling. Geophysical Prospecting. 2014; 62(4):679–692.

[31] Zeng X, Lancelle C, Thurber C, Fratta D, Wang H, Lord N, Chalari A, Clarke A. Properties of noise crosscorrelation functions obtained from a distributed acoustic sensing Array at Garner Valley, California. Bulletin of the Seismological Society of America. 2017 Jan 31; 107(2):603-610.

[32] Ravet F, Rochat E, Niklès M. BOTDA-based DTS robustness demonstration for subsea structure monitoring applications. Proc. SPIE 9634, 24th International Conference on Optical Fibre Sensors. 2015 Sep 28; 96345Z.

[33] Stork AL, Chalari A, Durucan S, Korre A, Nikolov S. Fibre-optic monitoring for high-temperature carbon capture, utilization and storage (CCUS) projects at geothermal energy sites. First Break. 2020 Oct; 38(10):61-67.

[34] Shao M, Qiao X, Zhao X, Zhang Y, Fu H. Liquid level sensor using fiber Bragg grating assisted by multimode fiber core. IEEE Sensors Journal. 2016 Jan 7;16(8):2374-2379.

[35] Xu W, Wang J, Zhao J, Zhang C, Shi J, Yang X, Yao J. Reflective liquid level sensor based on parallel connection of cascaded FBG and SNCS structure. IEEE Sensors Journal. 2016 Nov 16;17 (5):1347-1352.

[36] Yang C, Chen S, Yang G. Fiber optical liquid level sensor under cryogenic environment. Sensors and Actuators A: Physical. 2001 Oct 31;94 (1–2):69-75.

[37] Wei M, McGuire JJ, Richardson E. A slow slip event in the south Central Alaska subduction zone and related seismicity anomaly. Geophysical Research Letters. 2012 Aug 16;39(15).

[38] Shizhuo Y, Ruffin PB, Francis TS. Fiber optic sensors. Talor & Francis Group. 2008;479.

[39] Shakeel A, Kirichek A, Chassagne C. Rheological analysis of mud from port of Hamburg, Germany. Journal of Soils and Sediments. 2020; 1-10.

[40] Shakeel A, Kirichek A, Chassagne C. Yield stress measurements of mud sediments using different rheological methods and geometries: An evidence of two-step yielding. Marine Geology. 2020; 106247.

[41] Draganov D, Hunziker J, Heller K, Gutkowski K, Marte F, high-resolution ultrasonic imaging of artworks with seismic interferometry for their conservation and restoration. Studies in Conservation. 2018; 63(5):277-291.

[42] Lindsey NJ, Rademacher H, Ajo-Franklin JB. On the broadband instrument response of fiber-optic DAS arrays. Journal of Geophysical Research: Solid Earth. 2020 Feb;125(2): e2019JB018145.

[43] Li W, Bao X, Li Y, Chen L. Differential pulse-width pair BOTDA for high spatial resolution sensing. Optics express. 2008 Dec 22;16(26):21616-21625.

[44] Kuvshinov B. Interaction of helically wound fibre-optic cables with plane seismic waves. Geophysical Prospecting. 2016; 64(3): 671–688.

[45] Ayres A, Theilen F. Relationship between P- and S-wave velocities and geological properties of near-surface sediments of the continental slope of the Barents Sea. Geophysical prospecting. 2001 Dec 24;47(4):431-441.

[46] Snieder R. Extracting the Green's function from the correlation of coda waves: A derivation based on stationary phase. Physical review E. 2004; 69(4): 046610.

## **Chapter 8**
