11 / 2018-07-27 20:48:13

Modelling charge transport in a HVDC cable using different softwares: from fluid models to macroscopic model

cable, HVDC, model, fluid, chage transport, macroscopic

Modelling charge transport in a solid organic insulation has been studied intensively this last decade, using homemade models or commercial ones for the case of fluid models, or macroscopic models based on the Maxwell-Wagner theory. These models allow calculating the electric field distribution within the insulation, and, for the case of fluid models, having access to more microscopic variables such as the individual contributions from the different types of carriers to the space charge profile and current, the recombination rate, as a function of the material thickness. Until now, no comparison has been performed to understand the advantages and drawbacks of each type of modelling. Moreover, the availability of 2D or 3D space charge measurements makes it now necessary to develop the available models, being macroscopic or fluid, in 2D or 3D. Three different fluid models simulating charge transport in a model cable have been developed [1]: a 1D homemade model developed in Fortran, a 1D axisymmetric model developed using a commercial software, and a real 2D model developed with the same software. A macroscopic model has also been developed [2], based on the variation of the conductivity as a function of the electric field and the temperature. Each fluid model takes into account charge injection, transport, trapping and detrapping using a hopping type mobility, and recombination of charges of opposite polarity, function of the mobility (Langevin). We first compare the simulated results obtained using these models (macroscopic and fluid) for the case of a polyethylene insulated cable, i.e. a cylindrical geometry where the electric field and the temperature are not homogeneous along the insulation. The comparison shows the differences between the macroscopic model and the fluid ones, due to the fact that injection and bulk trapping of charges are not treated in the macroscopic model. A special attention is also paid to the comparison between the different fluid models, as the numerical treatment and the boundary conditions may differ. Comparison of the different simulated results shows the difficulty to handle the commercial software, moreover when a 2D model is at play. Simulations are also performed on a real cable joint, where an interface exists between polyethylene and a rubber material, and for which the macroscopic model has already been validated. Macroscopic models have the advantage to be easy to implement without knowing the physics behind the experimental behavior of a polyethylene used as insulation in a cable geometry. These models are globally able to predict the electric field distribution or the interfacial charge in the case of a multilayer dielectric, as long as charge injection/transport are reasonably equilibrated in the different layers, i.e. no substantial charge accumulation other than that due to the gradient of stresses.

[1] S. Le Roy, G. Teyssedre and C. Laurent, 'Modelling space charge in a cable geometry', IEEE Trans. Dielectr. Electr. Insul., Vol. 23, pp. 2361−2367, 2016
[2] T.T.N. Vu, G. Teyssedre, S. Le Roy and C. Laurent, 'Maxwell-Wagner effect in multi-layered dielectrics: interfacial charge measurement and modelling', Technologies, Vol. 5, pp.1-14, 2017