Plastic-Free Oceans

Abstract

Plastic has many valuable properties and has become an intrinsic part of our everyday life. An approximated 400 million tonnes is produced every year, consequently, vast quantities of plastic waste find their way into the oceans. The question "Why don't we just clean it up?" led to the founding of The Ocean Cleanup by Dutch entrepreneur Boyan Slat in 2013.

As of the writing of this thesis, The Ocean Cleanup has removed 12,989,690 kilograms of trash from oceans and rivers. System 03, consisting of two towing vessels towing a large net, is utilised to extract plastic waste from the oceans. To catch as much plastic at minimum time, effort and money, it is essential to have a comprehensive understanding of the behaviour of this net in water. A 3D Finite Element model is employed to simulate the system. Drawbacks in this model are the high computational demand and inaccuracies. These inaccuracies are caused by missing knowledge regarding drag coefficients at low angles of attack and the extrapolation of net behaviour to large nets.

The main objective of this research is to develop a simple, yet accurate model to simulate the physical behaviour of a high length-to-depth ratio net towed through water. Achieving accuracy in the model requires a comprehensive understanding of the net's behaviour in water, which has been investigated through a series of experiments. The simplicity of the model ensures low computational demand, enabling its use in offshore operations for swift decision-making on board. The developed model is validated using measurements obtained from campaigns conducted by The Ocean Cleanup. The performance of this simplified model is then compared to that of the 3D Finite Element model. Finally, the simplified model is applied to predict the behaviour of potential future systems.

In the experiment, the loads on two different nets were measured as they were towed through a basin at various velocities and angles of attack. The results show a significant decrease in drag at low angles of attack, attributed to the shielding effect. Experiments with two-meter and six-meter nets revealed that the shielding effect becomes insignificant beyond these lengths. Consequently, the drag coefficients calculated at various angles of attack for the six-meter net are utilised in the model.

To predict loads on the system, the model estimates the system's shape. For an accurate simulation of reality, the model makes an initial guess on how the system could look like. After this guess, an iteration loop starts until convergence is achieved. Convergence indicates that the system has reached equilibrium, accounting for external influences, internal forces, deformations, and the shape of the system. The model uses system properties, vessel positions and environmental conditions as input to generate the predicted shape of the system and the predicted loads acting on the system.

The performance of the simplified model is assessed by examining the error between the predicted and measured loads under various conditions. This performance is then compared to that of the 3D model. Both models demonstrate accuracy, with a MAE of just over 20\%. In the wide-span cases, which occur most frequently in the campaigns, the simplified model performs better than the 3D model. In terms of simulation time, the 2D model significantly outperforms the 3D model. While the 3D model requires approximately one week to complete its simulations, the 2D model finishes in 0.14 seconds, making it a flexible and user-friendly option. Altogether the simplified model is a simple, yet accurate model to simulate the physical behaviour of a high length-to-depth ratio net towed through water.

To demonstrate the model's functionality, it is applied to one of the trips, showing the maximum sailing velocities to ensure the system remains within its limits. The same trip is then simulated using potentially new nets and increased sailing velocities, resulting in a higher plastic catch. To improve the model's accuracy, aspects like wave direction, wind speed and wind direction can be implemented. Additionally, conducting a CFD analysis is recommended to gain a deeper understanding of net behaviour, which can help avoid the expense of additional tank tests.