Graduation of Ben Sloof: Hopper Loading Model and Overflow Losses

Ben Sloof signing his Master of Science degree

Do you remember Ben Sloof? Our young bright graduation student presenting at the Young CEDA Pitch Talks? Last week he graduated (with honours!) on ‘Numerical modelling of overflow losses and flow in Trailing Suction Hopper Dredgers’. You may have noticed that hopper loading was already a topic for previous posts about Ben Sloof and Jordy Boone at the CEDA Dredging Days. It results from the great interest in this part of the dredging process. And it still is a very dynamic field of research indeed, as in a short time frame various new models are presented. Who would ever think that dredging technology is a boring business? We are already dredging for ages, so what innovation could ever contribute to the progress of dredging? Well, I’ll explain what this new kid on the block found out now.

Overview of a hopper loading process with OpenFOAM simulation

Due to time constraints on a graduation project, Ben was limited to work on only a part of the hopper loading process. He worked only on the internal interface between mixture and sediment. Moving free surfaces of a filling hopper or a lowering water table from a reduced overflow height were left out of the scope. Still, it quite accurately describes the observed phenomena in the hopper loading process. We see a negative buoyant jet, we see the jet scour pit, recirculation and mixing in the fluid body and a ‘clear’ top layer which carries away the fines as it is skimmed away through the overflow weir. Deviation from the measurements can be explained by the difference in single fraction particle size.

Introduction to the key components of the proposed new ‘Layer Model’ (1DV)

With the observations from the CFD, Ben Sloof made his new ‘Layer Model’ with specifically a single fraction of particles. Then, he made an effort to correct the ‘Layer Model’ for physics of multiple fractions, with not too much improvement. Later on, he introduced a ‘fiddle factor’, that magically happened to be spot on. Further research and development of the model is currently under way to build a better foundation for this. Who knows, even Einstein had to introduce the ‘Cosmological Constant’ to fit his Theory of General Relativity to reality. Years later he was proved to be right1.

Comparison of van Rhee, Boone and Sloof

The performance of the models by van Rhee and Boone were already discussed2. Now, the validation of the Sloof ‘Layer Model’ is also included. Only after adjustment for the multi fraction in the results, the model nears the performance of the van Rhee model. Difference here is calculation speed. Sloof manages to estimate the hopper performance in seconds. This is why he rightfully received his degree with honours. He improved the speed AND accuracy of the calculation in a method, that would normally take the better part of a PhD study. I am looking forward to his future achievements.

Although this graduation assignment really helps us in refining our design process and gave some insight in the loading and overflow process, it still is an academic exercise. And any scientific model is just as good as the assumptions that were made at the beginning. In reality, the circumstances might be very different from what the model assumed, resulting in quit some different performance.

Heavy weather dredging (Retrieved from YouTube 18/10/2012, unknown source)

References

  1. Cosmological Constant Confirmed
  2. IADC Young Author Award for 1DH Hopper Loading Model of Jordy Boone

See also

Overview of hopper loading models by Ben Sloof
Nice report with an overview of the various hopper loading models

Dredging Engineering (lectures)

Dredging Engineering (papers)

Damen Standard TSHD

Damen MAD

A Simple Soil Sample Show

Demonstration at a workshop ‘Advise a dredging customer’

Recently I had to give a demonstration on one of the aspects of my work: investigate what type of soils the customer will be encountering, what type of dredging equipment is most suitable for that application and what will the production be under those conditions. At this demonstration, the purpose was, to show how I can make something simple very complicated in a short time. The challenge here will be to take you with me on this precipitous path to enlightenment.

Ingredients for a sand sieving demonstration

The example case was to examine two different soil samples. Each was artificially mixed, but representing a Particle Size Distribution of two different borrow areas. The samples are delivered in widely available standardised containers. Yeah, you recognised them: 1.5L soda pop bottles. As the samples were wet, they had to be dried. Normally, this can take two to three hours. Here we applied a magical temporal acceleration by employing a Calvin style cardboard box as an oven. The literal main ‘activity’ for the participants was to manually shake the sieve tower. Rest assured, in our own soil analysis lab we use an industrial automatic sieve tower. This antique specimen comes from our museum. The separated fractions from the sieves, then have to be weighed and finally plotted in a so called particle size diagram. Horizontally you can take a sieve diameter and vertically you can read what mass percentage will pass that mesh diameter.

Resulting Particle Size Distribution and key parameters

Although both samples have the same median diameter d50 (354µm), which is indicative of the general particle size, the distribution is very different. Sample A is almost totally the same diameter, where sample B has a broad distribution. A measure for the distribution is the Uniformity Index (d60/d10). Another important figure to be taken from this graph is the silt fraction. This is the percentage of particles that is smaller than 63µm. More parameters can be taken form this graph, but these are the most important for now.

Influence of particle size distribution on slurry transport

From the equation for the critical velocity, the uniformity index plays an active role. If the uniformity index increases, there is more fine material. Fines tend to increase the density and viscosity of the fluid. Consequently the resulting mixture behaves like a heavier fluid, carrying larger particles. Effectively, the resistance of a non-uniform sand is higher than for a more uniform sand mixture.

Influence of particle size distribution on cutter production

Another aspect of the dredging process is the ability of the CSD to excavate the material from the bottom. Here the uniformity index has the complete opposite influence. A uniform sand distribution will have a lot of voids between the grains. The particles will move easily over each other. When there are lot of smaller particles available, they tend to clog up the voids and bond the bigger particles in a gridlock. These sediments are very hard to excavate. If no geotechnical investigation is available, the PSD can help to estimate a SPT.
With the grains size and the SPT, the audience consulted our lovely looking assistant ‘Sandy’ for a first selection of the required CSD.

Short evaluation of a selected dredge at ‘www.dredgefinder.com’

References

Wilson, Addie & Clift, (1992), “Slurry Transport Using Centrifugal Pumps”, Chapter 5 “Heterogeneous Slurry Flow in Horizontal Pipes”

Youd, (1970), “Densification and Shear of Sand during Vibration”, Journal of the Soil Mechanics and Foundations Division, 1970, Vol. 96, Issue 3, Pg. 863-880

See also

Geotechnical investigation

Dredge Finder

IADC Young Author Award for 1DH Hopper Loading Model of Jordy Boone

Jordy Boone receives the IADC Young Author Award at the CEDA Dredging Days

OK, one last time we will revisit the CEDA Dredging Days. There was so much to see and experience, that there could be some more posts about them. However, daily life already demands more of my focus and there are fun facts to tell about them also. The conclusion of this series is about a well-deserved IADC award to Jordy Boone. Both, because he did write a terrific article and we will read more about it in the Terra et Aqua journal soon enough.

So, why was his article and presentation so special? Hopper loading is one of the key process components in the production cycle of a trailing suction hopper dredge. And therefore there is a lot of interest in this subject. Countless articles and numerous experiments have been performed on this topic, resulting in a lot of different hopper models. Traditionally a literature review starts with Camp (1936). And there are a lot of models that build on that approach. Camp and the derived models are similar to the Lagrangian approach, where they follow the trajectory of a single particle.

Basics and relations of Camp and derived models

A whole different approach is to follow Euler and fix the frame of reference. The contents of the system has to be modelled as a continuum. Ground breaking work has been done by van Rhee. He modelled the hopper in a 2D environment and based a more comprehensive 1DV model on that. Others have followed up and so does Jordy Boone.

Van Rhee and similar ‘Euler’ models

Other approaches don’t take any physics in their modelling, but consider the hopper as a process block in a control system. At the moment they are only useful for monitoring and controlling an existing system, they don’t have a predictive value, yet.

Control system models

In this overview, we can see, that the 1DH Boone model is part of the Euler family. Normally solving multi-dimensional 2D and 3D Euler problems tend to be slow. Van Rhee already pointed this out and part of his PhD. thesis was the presentation of a more comprehensive 1DV model. Basically a column cross-section through the hopper. Mixture would be deposited on the bed and water flows up and out of the system. Where that model does lack the influence of the density current, Boone literally upended this simplification by using horizontal strata in the hopper. Here, the mixture section can incorporate the horizontal density current conditions. This will give correct mass and momentum equations. Vertical exchange processes are than calculated by closure relations.

Comparison van Rhee and Boone

On top of the cake, Boone also verified his approach in laboratory experiments and prototype measurements. As his manuscript is also well written and accessible, he rightfully received the IADC Young Authors Award. Keep up the good work Jordy, we hope to see more interesting work from you.

Jordy Boone presenting his 1DH model at the CEDA Dredging Days

References

Camp T. R. (1936). ”A study of the rational design of settling tanks.” Sewage Works Journal 8(5), pp, 742-758

Miedema, S.A., Vlasblom, W. (1996). ”Theory of Hopper Sedimentation”. 29th Annual Texas A&M Dredging Seminar, New Orleans: WEDA

Van Rhee, C. (2002). ”On the sedimentation process in a Trailing Suction Hopper Dredger”. PhD thesis, TU Delft, the Netherlands

Braaksma, J., Klaassens, J. B., Babuska R., de Keizer, C.(2007). ”A computationally efficient model for predicting overflow mixture density in a hopper dredger”. Terra et Aqua, 106, pp. 16-25

Spearman, J. (2013) ”TASS Software – User Guide for TASS version 4.0”. HR Wallingford

Jensen, J.H., Saremi, S.(2014). ”Overflow concentration and sedimentation in hoppers” J. Waterw., Port, Coast. Ocean Eng., ASCE, 40

Konijn, B.J. (2016). ”Numerical simulation methods for dense-phase dredging flows”. PhD thesis, Universiteit Twente, the Netherlands

J. Boone, J, de Nijs, M.A.J., (2017) “1DH Modeling of Transport and Sedimentation Inside a Hopper of a Trailing Suction Dredger” CEDA Dredging Days 2017
Note: Not available online yet. Depending on publication by CEDA.

See also

Hopper Loading: What Happens Beneath the Surface

Overview of hopper loading models by Ben Sloof
Nice report with an overview of the various hopper loading models