2014 International Conference on Lightning Protection (ICLP), Shanghai, China Sensitivity analysis of leader channel models used in long air gap positive discharge modelling Oscar Diaz, Vernon Cooray Lightning Research Group, Division of Electricity Ångström Laboratory, Uppsala University Uppsala, Sweden oscar.diaz@angstrom.uu.se, vernon.cooray@angstrom.uu.se Liliana Arevalo Power Systems HVDC ABB AB Ludvika, Sweden liliana.arevalo@se.abb.com Abstract — Leader models used in electrical discharge simulation have been proposed in theoretical works by different authors. Their application can be found in the study of lightning upward connecting leaders or long air gap laboratory testing, and can be considered engineering or physical according to their detail level. Based on simplifications and assumptions, these models are capable of predicting the 50% breakdown voltage for certain electrode arrangements, time evolution of physical phenomena like particle densities, temperatures, electric fields, leader and streamer progress, among others. An important parameter in a leader model is the potential distribution along the channel as it propagates. In present work, we compare an engineering and a physical leader model against experimental data recorded while testing a rod-to-plane 10 m gap with switching-like voltage impulse. A sensitivity analysis was done with some basic input parameters of two leader models in order to compare the outcome for different cases. The results showed a strong dependence of the leader channel evolution with the assumed constant average potential gradient used in most of the leader models. Keywords: long air gap discharge; leader model; leader propagation, leader-corona region; high voltage engineering I. INTRODUCTION The high voltage engineering researchers have been long ago interested in modeling the full electrical discharge occurred while testing long laboratory air gaps [1-8]. Important contributions have been done during the last decades, both experimental [1-2] and theoretical [3-7]. The later works are strongly dependant on the former ones, since most of the proposed models depart from experimental observation of measurable physical variables and possible ways to explain its behavior. The experimental work carried by the ‘Les Renardières’ [1- 2] have been used by several authors [5-13]. Throughout their work, measurements of potential, electric field, radiation and current were reported for different electrode arrangements. Furthermore, the physical processes involved in the positive discharge propagation were analyzed and described, like the leader propagation along the air gap due to the leader-corona region (LCR) in front of the leader tip. The leader channel models can be either engineering or physical. The engineering leader models aim to predict mainly the 50% breakdown voltage and its statistical variability [16, 17]. For the physical leader models, a better description of the phenomena involved is obtained by solving equations of the conservation of mass, momentum and energy, continuity of charged, neutral and excited species coupled with basic electromagnetics. The main focus of the present work is to present a sensitivity analysis of two leader models by using a common experimental data input. A more detailed explanation of different existing leader channel models can be found in a future publication [19]. The detailed analysis of the rest of processes involved in the full long laboratory air gap discharge like the streamer inception, the streamer to leader transition, and LCR representation are out of the scope of this work. A detailed explanation of these elements can be found in [3, 7, 16]. II. LONG AIR GAP POSITIVE DISCHARGE A. Generalities The rod-to-plane configuration has been largely used for switching voltage impulse test of long air gaps. The corona inception is the starting point for the positive discharge event. If a threshold critical gradient is reached at the rod electrode surface vicinity, branched streamers initiate from a common point near the rod electrode. After the first streamers have produced a certain amount of charge, mainly required to thermalize the first section of the leader channel at the place where the streamers were originated, the leader inception takes place. At the tip of the newly created leader channel, the LCR sustains the discharge advance by supplying electric charge, eventually thermalizing new sections of the leader channel. This causes the leader to propagate further downwards along the gap if it has an average potential gradient high enough to sustain ionization processes in the LCR, ca. 450 kV/m for atmospheric air. In continuous propagation, the leader tip moves at an almost constant velocity of 1 - 2 cm/μs, with a This work has been funded by the divisions of Power Systems HVDC and Power Products – High Power Breakers at ABB AB, Ludvika, Sweden. 259 978-1-4799-3544-4/14/$31.00 ©2014 IEEE