153 / 2018-08-25 16:43:45

Optimized Design of Contour Shape for Central Insert of Basin-type Insulators based on Genetic Algorithms

Genetic Algorithm (GA),optimized design,central insert,basin-type insulators

Due to its small footprint, compact structure, reliability, safety and flexibility, GIS has been widely used in high voltage even ultra-high voltage substations in recent years. The basin-type insulators in GIS play an important role in supporting conductors, isolating the air chamber and electrical insulation. The basin-type insulator consists of the metal insert and the epoxy resin body, and generally the bond between the central insert and the epoxy is increased by providing projections on the surface of the cylindrical metal insert. Practical experience has shown that the installation of a voltage equalizing ring on the central insert of the basin-type insulator can effectively improve the electric field distribution inside and along the insulator and reduce the partial discharge and the surface charge accumulation. However, in the current vacuum casting process, the thin-walled flanging grading ring forms a semi-closed cavity, which hinders the flow of the epoxy resin, easily generates bubbles and defect inside, and increases the defective rate of the products.
Considering the needs in actual production, the electric field and the stress field of 420kV basin-type insulators are calculated by finite element simulation software COMSOL. Model the convex shape of the central metal insert surface with parameters, including the radius of curvature and the length of extension. Taking the electrical and mechanical properties into account, the maximum electric field intensity inside and along the basin-type insulator and the maximum stress of the basin-type insulator under the hydraulic pressure test are taken as the decision variables. Use MATLAB to call COMSOL's project file, and optimize the returned decision variables based on genetic algorithms. Genetic algorithms are highly parallel, random, adaptive search algorithms developed from the natural selection and evolutionary mechanisms of the biological world. The genetic algorithms have strong robustness and do not depend on the specific problems. It is especially suitable for dealing with the complex nonlinear problems that are difficult to solve by traditional methods. The optimized shielding structure is solved under the reasonable genetic algorithm parameters including the population size, the crossover fraction, and the fitness limit. The optimized design of contour shape for the central insert without shielding ring makes the maximum electric field intensity in the epoxy resin almost the same as that of the original design with shielding ring. In the water pressure test of 1.5 MPa, the maximum stress of the basin insulator is 50.7% higher than that of the original design, but still lower than the damage strength of the epoxy/alumina composite (66 MPa).
The optimization process for the parametric model is simple, efficient and accurate. The work of this paper can reduce the defective rate and the cost of the products, and also the white pollution, which can guide the factory production to some extent.