In electricity transmission, the inherent qualities of alternative grids fade in front of the difficulty imposed by the reactive power transportation, when overhead lines or, particularly underground or submarine cables reach critical lengths. In the frame of the thought aiming to operate at best renewable energy resources, coming from wind or hydro power stations, the creation of a high voltage direct current grid is assumed to dispatch those energies to the consumption centers. Indeed, from a certain amount of power to transmit, there is a critical length beyond which AC transmission is not viable anymore, and where DC links becomes both more efficient and less costly. The new conversion technologies, namely the voltage source converters, now enables the connection of several terminals and therefore, the creation of a true grid, which can either have a looped, meshed, or radial structure. The advantages of such grid structure in relation to point-to-point links are obvious in terms of reliability of the transmission grid. Nevertheless, some technological locks remain, and the protection against faults of those grids is currently largely debates. This Ph.D work, advised by RTE and Grenoble Electrical Engineering laboratory (G2Elab) is linked with the TWENTIES project, which aims to analyse the potential obstacles to the integration of renewable resources in the European electrical system, and to propose solutions to lift those technological barriers.This Ph.D thesis aims to propose a complete protection plan against insulation faults in cable or busbar, including a main protection and two backups in case of failure of a component, for a submarine DC grid. The target application would be a grid linking offshore wind farms and countries through the North Sea.