Aiming to precisely predict the workability of dredging operations, Belgian marine contractor Jan De Nul has developed an accurate tool for office studies and onboard decision support. The tool is the culmination of a long-term project, involving careful modelling, verification, validation and large simulation batches.

Bringing engineering insights aboard

The tool is already being used during tenders to estimate the vessel’s standby time using hindcast data from the field of operation. Deployment aboard is being discussed. Ultimately, the goal is to automate all interfaces to the tool, so the crew is continuously provided with the expected workability for the next hours of operation. Different options are being considered to automatically define the sea states to be loaded by the tool, including state-of-the-art wave radars and other weather forecast systems.

In parallel, MARIN has completed the modelling and verification of a second CSD for Jan De Nul. Following the same project approach, the model will be validated by Jan De Nul and then be extended with its own workability tool.

Based on the experience gained on the usage of these tools, Jan De Nul will consider its dissemination to other types of vessels in its fleet or for the modelling of more complex systems, so that the last gap between engineering and operations can be closed. 

One million simulations

A batch of one million simulations was defined and triggered in incremental steps on the cluster, while processing and network capacity were closely monitored by MARIN’s IT personnel. After a couple of days, the one-million batch was successfully completed.

The usage of the tool is very simple. The user must only specify the list of sea states to be analysed and then open the HTML interface, which was developed by MARIN for the specific purposes of this project. The tool automatically interpolates the database and presents the relevant figures. Additionally, the same data analysis methods used to monitor equipment integrity and operational safety aboard the CSD are also used in the tool. This includes, for each sea state, peak distribution plots and a red and green colour scheme that indicates the workability for each predefined criteria and threshold. As a final step, Jan De Nul has extended the tool with a dashboard, which provides simple and accessible usage for its employees.

aNySIM model successfully predicts excessive spud forces

Finally, the model was returned to Jan De Nul for validation based on full-scale data. The same trends that were observed aboard, during control of the spud pile carrier, were observed in the simulations. But most importantly, results demonstrated that the aNySIM model was able to predict - for at least 11 different operations - the significant wave height at which excessive spud forces were measured and halted the CSD operation. Excellent news!

Once the model was validated, the workability tool could be developed. Due to the non-linear characteristics of the modelled operation, efficient frequency-domain calculations could not be used to instantaneously feed the tool. Therefore, a time-domain model was required. The only alternative was to run, a priori, a large batch of time-domain simulations on MARIN’s high-performance cluster and feed the tool with a database.

Stiff constraints imposed by the spud pile carrier and the strong non-linear behaviour of its controller required careful verification and validation. The vessel’s motions and spud loads were verified for different water depths and sea states. Wavelength dependent results were observed, which corroborated the calculated hydrodynamic interaction of the vessel in relatively shallow waters. Viscous damping and time-step sensitivity analysis were also performed to address any model uncertainty.

The project began with the abstraction and modelling of one of Jan De Nul’s cutter suction dredger (CSD) operations, using MARIN’s time-domain software aNySIM. This was not a straightforward task. The exact same control algorithm used aboard the vessel was embedded in aNySIM by the engineers of Jan De Nul, making the implementation and testing of the controller a collaborative task. To better manage several iterations of the model between MARIN and Jan De Nul, a repository for version control was also used.