Predicting failure of dynamic offshore cables by insulation breakdown due to water treeing

Abstract

The offshore wind industry has progressed from traditional fixed bottom platforms for the wind turbines to be installed on, to having floating platforms. This allows access to wind resources further offshore that were previously unavailable due to the greater water depth. Floating wind platforms introduce new challenges for offshore cables, particularly the array cables which will be required to hang from the base of the platform through the water column to the seabed. This arrangement exposes the cable to the dynamics of the marine environment it is installed in, creating concern for the failures of these dynamic cables. One particular failure mechanism of subsea cables is the degradation of the insulation layers by water treeing. Water trees propagate due to mechanical and electrical loadings, and the move to having more dynamic installations for the cables is expected to increase the number of cables failing offshore. A primary concern with water treeing is that with current technologies they remain undetectable until they cause an unexpected failure of a cable, which may have been previously deemed healthy. This work presents a methodology which can model the propagation of a water tree to predict the time taken for it to propagate to a length which would cause failure of the dynamic cable. This ability to predict when a dynamic cable is at risk of failure allows for better planning of maintenance, as opposed to an unexpected failure. To achieve this, both the mechanical and electrical stresses the dynamic cable will be exposed to have been considered. A global model of a floating wind platform has been developed which leads to a local model of the dynamic cable's cross section to deduce the mechanical stresses the environmental force loadings will exert on the insulation. Following this the distortions of the electric field across a dynamic cable's cross section in the presence of a water tree have been modelled to calculate the resulting stresses. Finally these two stresses are combined and a series of fatigue damage and crack propagation methods are employed, leading to the modelling of the microscale breaking of the insulation chemical bonds to predict the propagation of a water tree. This results in the overall prediction of when the modelled dynamic cable is at risk of failure due to water treeing, and a new fatigue life estimate.