A Trailing Suction Hopper Dredger (TSHD) is a ship that is used to excavate soil at sea, rivers, lakes and ports. The largest ships are used to reclaim land from the sea, such as for the prestigious projects Palm Island and The World in Dubai, whereas the smaller units are maintaining the depth in ports or rivers.
The production process in a TSHD can be divided into three separate phases. At first all the excavated material is stored in the hopper. The second phase starts when the mixture level reaches an overflow height and the excess water is being discharged overboard through the overflow pipe, while the density of the remaining mixture increases. The last loading phase begins after the maximum allowed mass in the hopper has been reached. In order to prevent the ship from sinking a constant-tonnage controller is used. When necessary, the controller lowers the overflow height hence more mixture is disposed through the overflow pipe.
Throughout the production process the dredge operator must decide on the position of the overflow pipe under uncertainty caused by a changing environment and noisy measurements. This results in non-optimal production rate. Therefore, over the past years, extensive research has been conducted to develop an automatic controller of the overflow system. As a result of this research a MPC controller of the overflow system was proposed. This controller automatically adjust the overflow height in order to maximize the production cycle of the TSHD. Furthermore, such a local controller has been integrated into a global MPC controller of the overall ship that aims to improve the production rate, taking into account factors such as: the fuel consumption or environmental restrictions. For the optimal performance of both, global and local controller, certain soil-dependent parameters need to be estimated from the onboard measurements. This is a challenging problem due to the nonlinear characteristics of the processes involved, as well as because the properties of the excavated soil dynamically change during the dredging cycle.
Therefore, the goal of this project is to investigate adaptive filtering algorithms for the sedimentation process. The main objective is to obtain fast and reliable estimates of the soil-dependent parameters which can be further processed in order to compute the set points for the controller of the overflow system. Due to the nature of the process the estimation problem is fitted within a Hybrid Systems framework, which means that the process evolves according to one continuous-time model until it is enforced to proceed according to a different continuous-time dynamics. In the TSHD model these ``jumps'' in the dynamics are caused by the manipulations of the overflow height. The project is in cooperation with IHC Systems for obtaining data and to do experiments.