The dielectric behavior of insulations is a key factor affecting the development of anti-corona materials for generators. Epoxy resin (EP), as the matrix, is blended with inorganic fillers of micron SiC and nano SiO₂ to investigate the effect of micro and nano doping on the conductivity and breakdown mechanism of the composites. Using experimental and simulation analysis, it is found that the effect of nano-SiO₂ doping concentration on the conductivity is related to the dispersion of SiC particles. The lower concentration of SiO₂ could decrease the conductivity of the composites. The conductivity increases with raising the nano-SiO₂ doping concentration to a critical value. Meanwhile, the breakdown field strength of the composites decreases with the rising content of SiC in constant SiO₂ and increases with more SiO₂ when mixed with invariable SiC. When an equivalent electric field is applied to the samples, the electric field at the interface of micron particles is much stronger than the average field of the dielectric, close to the critical electric field of the tunneling effect. The density of the homopolar space charge bound to the surface of the stator bar elevates as the concentration of filled nanoparticles increases, by which a more effective Coulomb potential shield can be built to inhibit the further injection of carriers from the electrode to the interior of the anti-corona layer, thus reducing the space charge accumulation in the anti-corona layer as well as increasing the breakdown field strength of the dielectric.

High-voltage dry-type bushings, serving as the crucial junctions in DC power transmission, represent equipment with the highest failure rate on the DC primary side, underscoring the critical importance of monitoring their condition. Presently, numerical simulation methods are commonly employed to assess the internal state of bushings. However, due to limitations in the efficiency of multi-physics field computations, the guidance provided by numerical simulation results in the field of power equipment condition assessment is relatively weak. This paper focuses on solving the electrical-thermal coupling in high-voltage dry-type bushings. Addressing the most widely used tetrahedral mesh in numerical computations, we propose an efficient solution method based on the concept of "smooth domains." This method involves partitioning the volume centroids of the elements into multiple smooth domains within the computational domain. Electric and thermal conduction matrix calculations occur within these smooth domains, rather than within the grid or element interiors. This approach eliminates the need for traditional element mapping and complex volume integration. To demonstrate the effectiveness of this method, we use high-voltage dry-type bushings as a case study, comparing the performance of our approach with traditional finite element algorithms. We verify the algorithm’s computational efficiency and apply it to the analysis of typical temperature anomalies in bushings, further illustrating its suitability for electrical equipment condition assessment.