Characterization of Different Interface Types for HVDC Extruded Cable Applications

Licentiate thesis, 2018

In extruded HVDC cable systems a variety of interface types can be found, and the interfacial properties depend on the application within the system. Such applications can be joints, terminations or the cable itself and they will all have different material combinations and manufacturing methods. To ensure at least 40 years of faultless operation of the cable system and its interfaces, proper design and quality control is essential. This requires detailed knowledge on how physical, measurable quantities of polymer surfaces relate to electrical characteristics such as electric field and space charge accumulation. This work aims to expand our understanding by assessing polymer surfaces created with different, industrializable preparation methods resulting in varying degree of surface roughness. Breakdown tests, charge decay measurements, optical profilometry and other characterization methods have been performed on cable peelings created from full-sized HVDC cables. It was found that, depending on preparation method, very different roughness levels and textures are created on the polymer surfaces. There were also other impacts of the preparation method found, such as altered surface morphology or the introduction of surface states. A clear relation between surface roughness and the localized field distribution was found, which mainly depended on the arithmetical surface height parameter Sa, but also to a minor degree on the shape of the surface texture. Field dependent bulk conductivity was found to reduce the tail in the surface field distribution histogram, but was not able to fully remove the distribution itself. The surface field distribution showed to have a clear impact on space charge injection, and was capable of creating a field threshold effect. This effect could not be accounted for by a fixed field parameter in the Schottky equation. In this work, charge injection was assumed to be the sum of Schottky and Fowler-Nordheim injection, which resulted in a very significant increase in charge injection above the threshold field. This led to bulk-limited charge injection above the threshold field, along with significant homocharge accumulation. Increased roughness was observed to reduce the breakdown field in the tested samples, which for the larger test objects scaled well with the calculated threshold fields. For the cable peelings, an impact of surface roughness was found as well, but at significantly higher field levels. The origin of this scaling effect is further discussed in this thesis. The threshold fields were also found for rough surfaces in the charge decay measurement. By utilizing rough ground electrodes, significantly higher decay currents were measured above the field thresholds. The increase in decay speed is related to opposite charge injection during the poling phase before the measurement. The findings of this work are believed to be applicable to many high voltage applications, especially in cases where solid insulation with low conductivity is used at high electric fields.