Design optimization for the insulation of HVDC converter transformers under composite electric stresses

Abstract

This paper proposes a design optimization method for the insulation structure of a High Voltage Direct Current (HVDC) converter transformer, which is exposed to AC, DC, and DC polarity reversal (DCPR) stresses. Since DC and DCPR stresses provide significantly different electric field distributions, conventional transformer insulations designed for AC systems cannot guarantee safety against electrical breakdown. Focusing on main-winding and end-winding insulation including pressboard barriers and oil between windings, this paper presents a design solution based on genetic algorithm (GA). By exploiting modern computing power, the proposed approach can save a considerable amount of effort and time compared to conventional approaches which depend on experimental data and human experience. From the comparison with example reference insulation structures, we verified that the optimized structures maintain the required insulation performance with the reduction of length by 2.8% for the main-winding insulation and by 13.8% for the end-winding insulation.