HUMIDITY EFFECT ON THE PROPERTIES OF CORONAS PRECEDING BREAKDOWN IN SHORT POSITIVE ROD - PLANE GAPS P.N. Mikropoulos and C.A. Stassinopoulos Department of Electrical & Computer Engineering, Aristotle University of Thessaloniki, Greece ABSTRACT In short rod - plane gaps under positive impulse voltages the properties of phenomena occurring during the early stages of the discharge, namely the first and second coronas, have been investigated. Measurements are reported of the variation of their inception time and voltage as a function of the wavefront duration, the gap length, the shape of the end of the rod and the absolute humidity. For purposes of further elucidation, the gaps were divided into two categories depending on to whether second corona emerged during the front or the tail of the applied impulse, and the properties of the coronas into two groups, on the basis of whether breakdown occurred or not. Special interest was given to the influence of the absolute humidity as the experiments and the analysis that followed showed that humidity acts interactively with the other discharge parameters. As a result, humidity exerts sometimes inhibitory sometimes facilitating action on the inception of the first corona whereas on the inception of the second its action is always inhibitory. INTRODUCTION Following earlier work on longer gaps [1], this study is concerned with 25 and 50 cm positive rod-plane gaps in air. The aim is to obtain a better understanding of the effect of various discharge parameters on the properties of the phenomena preceding breakdown, namely the first and second coronas. Such parameters were the wavefront duration, the gap length, the ge- ometry of the tip of the rod and the absolute humidity. Properties, subjects of investigation, were the first and second corona inception times and voltages. Emphasis was given to the influence of the absolute humidity, since it was argued that more information is required on the effect of humidity on air breakdown of short gaps [2, 3]. The gap consisted of a cylindrical brass rod with a di- ameter of 23 mm hanging over an earthed aluminium 2x2 m plane placed 80 cm above the floor of the labo- ratory. Two gap lengths (25 and 50 cm) and three types of interchangeable rod tips have been used: a conical tip with a 30 o end, a square - cut tip and a hemispherical tip. Impulse voltages were applied with wavefront varying from 2 to 250 μs and wavetail around 2000 μs. The spark gaps of the Marx generator were located at a distance of around 5 m from the rod- plane gap, facing it at an angle. Absolute humidity varied normally achieving values between 4 and 20 gm -3 . EXPERIMENTAL RESULTS General Class 1 (multiple-level) tests [4], were performed in order to obtain breakdown probability curves. During these tests information on the discharge was obtained by the oscillographic monitoring of the electric field at the centre of the plane; this way the inception times for the first and second coronas were measured and from them the inception voltages for the first and second co- ronas were calculated. All the inception voltages were corrected for air density by applying the equation (1) for m = 1 [4]. m t p d U U − ∗ = δ , (1). Influence of the wavefront, gap length and tip of the rod For the full range of absolute humidity, the average values of the first and second corona inception volt- ages U 1 and U 2 respectively, were plotted against the wavefront of the applied impulses t f ; such plots for the 50 cm gaps can be seen in Figures 1 and 2. As t f in- creases, initially a decrease of the values of both U 1 and U 2 is evident, more marked for U 1 , and eventually, for long t f , the rate of decrease declines, tending to zero (Figs. 1 and 2). This pattern also applies for the 25 cm gaps, although in a less pronounced way. 0 50 100 150 200 250 t f , μs 0 55 110 165 220 275 U 1 , kV Hemispherical tip Square-cut tip Conical tip Figure 1 U 1 plotted against t f for all humidities, 50 cm gaps Vertical bars represent 90% confidence limits of U 1