Based on the IEC operating duty test, the arrester is subjected to two energy impulses representing repeated operating duties. For a detailed description of the simulation
procedure see [12]. The impulses are applied as adiabatic temperature rises in T = 30 K each. The interval between the impulses, tgap, varies from 30 s up to 10 min. Initially,
the arrester is in continuous operation (see Fig. 4). Fig. 6(a) shows the start temperature after the impulse injection. The analysis of the temperature for longer simulation periods
shows that all scenarios are thermally stable. Based on (8), Fig. 6(b) shows the cooling rate η that varies over the arrester segments. One may observe two distinct phenomena. For large gap times (tgap > 5 min) η diminishes, as, the lower the start temperature, the lower is the radiative heat transfer contribution. For short impulse intervals, η also decreases
significantly. In this region, the strongly nonlinear increase in the Joule losses (see Fig. 5) is present. This regime is, by (8), suspected to cause thermal instability for higher impulse intensities. The variation of η over the segments is attributed to the inhomogeneous field stress distribution resulting in reduced losses and temperatures in the lower
segments. However, the lowest cooling rate is observed for the middle segment (seg. no. 3) and short gap times. Compared with the top segment, which features longitudinal heat transfer via the top flange, the cooling rate decreases earlier and moves toward zero. Clearly, the absolute temperature at the top segment is no decisive for thermal stability. Rather, the cooling
rate is strongly influenced by both, heat transfer capability and the distribution of the Joule losses along the arrester, respectively