Condition Monitoring of Spud in Cutter Suction Dredger using Physics based Machine Learning

Dredging operation is happening at an increased rate due to the impetus gained towards inland navigation and land reclamation. The spud system is an integral component of a Cutter suction dredger (CSD) which anchors the hull at the dredging location. The spud embedded in the soil maintains the position of the dredger and offers resistance to the motion of the CSD.

In this present study, the spud is modelled as an Euler–Bernoulli beam using finite element analysis. The resistance offered by the soil is evaluated experimentally. Soil stiffness is modelled using two spring elements which restrain the motion of the spud in the transverse and rotational directions. The spud system is subjected to external force due to wave loading. The wave load is determined along the length of the spud using the Morison equation. Heave and pitch are the degrees of freedom of the dredge hull which are restrained by the spud. The ship’s rigid body dynamics are identified using the experimental study. The identified dredge hull is coupled to the spud. The numerical model of the spud system is subjected to random wave loading. A case study is carried out using Monte-Carlo simulation by varying the soil stiffness. The maximum response of the system is evaluated at the top of the spud. A metamodel for the system is developed based on the maximum response at spud and soil stiffness using the Gaussian process emulator (GPE).

The theoretical responses of the system due to theoretical and experimental wave loading conditions show similar characteristics. During the validation study, it is observed that the metamodel is predicting the soil stiffness accurately.

The predictive analysis of the soil-spud system provides a good indication of the embedment of the spud. The analysis indicates that the condition monitoring of the embedment of the spud can be assessed by the metamodel developed using GPE.