Electrical discharges in long air gaps, referred to as ‘leader discharges’, usually propagate with transient thermal plasma channels. The gas temperature evolution of the channels plays a key role in the inception and development of leader discharges. This paper is aimed at investigating the use of quantitative Schlieren techniques in temperature measurement of the leader discharge channels. To do this, a calibrated Toepler’s lens-type Schlieren system was set up to measure the radial temperature profile of leader discharge channels. The measurements were compared with the simulation results from a detailed numerical model which describes the thermal-hydrodynamic properties of the channel. The comparison shows that although a good agreement is obtained when the leader is propagating in a stable manner, the measurements of the temperature evolution at the axis of leader channel deviate significantly from the numerical simulation during the leader initiation stage. Apart from the limited spatial and temporal resolution of the Schlieren system, the deviation is also caused by the non-isobaric gas heating during leader inception. Based on the theoretical analysis and numerical simulation, we proposed several suggestions to the application of quantitative Schlieren techniques to measure the temperature of leader channels.

When concerning the installation and operation costs, it is always desirable to protect an overhead distribution network with a limited quantity of surge arresters (SAs). This paper presents a novel procedure for designing differentiated protection against lightning, by identifying appropriate positions for those SAs on the overhead line. Rather than the traditional flashover rate of a whole line, indices in terms of flashover numbers at individual poles and associated with possible stroke locations are introduced in this paper. Based on these flashover indices, a differentiated configuration of SAs can be identified using a selection strategy. After several rounds of selection, a differentiated SA protection scheme can be obtained, given by a pre-set quantity of SAs. In this procedure, a simplified probability-based approach is proposed to determine the flashover indices. By comparing the traditional Monte Carlo method, this approach can achieve a 10-fold increase in computational efficiency with reasonable accuracy. Two different selection strategies are introduced in this paper. Compared with the selection strategy using the flashover indices at individual poles, the selection strategy based on a group of poles is recommended because it yields a configuration with better lightning performance with the same quantity of SAs. This design procedure is applied to practical distribution networks with a shield wire and nearby structures. The influence of grounding impedance, a shield wire, and nearby structures is addressed. Finally, a generalized step-by-step procedure is provided to assist in a differentiated SA protection design.

The velocity‐current relation of positive leader discharges is of great importance in the research of lightning physics. However, the measured data in a wide range of current up to several hundred amperes are not available so far. To fill the gap, experiments on positive leader discharges in long air gaps under lightning impulses were carried out on the test site at an altitude of 2.1 km. The velocity and current of positive leaders were measured with the synchronized discharge current and high‐speed video frames. An empirical formula for the velocity‐current relation of positive leaders is given with the leader current up to 500 A. The measurements are also compared with the numerical simulations, and it is suggested that the initial tip radius of a propagating leader grows as the leader current increases.

The relaxation process of the discharge channel near the anode in a long air gap was observed using a Schlieren system with a temporal resolution of 5 µ s and a spatial resolution of 70 µ m. The dynamic characteristics of the decay process in the vicinity of the anode are obtained. The discharge channel evolves just as a growing mushroom in nature during the relaxation phase. Two physical quantities, angle θ and velocity v , are defined to describe the process in this paper. The average value of the angle and velocity under lightning impulses are 71.7° and 3.3 m s ⁻¹ respectively, while 7.7 m s ⁻¹ under switching impulses. A simplified model was established to simulate the formation of mushroom-shaped channel. The simulation and experimental results show that the formation and development of the mushroom-shaped channel are due to two factors. One is the convection of the high temperature and high pressure air near the anode produced by the first corona discharge; the other is the ionic migration. These two factors result in the phenomena that the cooling process in the vicinity of the anode is much more efficient than further into the gap, whereas the thermal conductivity of the anode may have little contribution to that.

No observations show evidence of ionization in long air gap discharge during dark periods. However, for the first time, weak channel illuminations are observed during the so‐called dark periods in this paper. The length of these faint bright channels is almost constant during each dark period and is nearly equal to that of the remaining low‐density channels created by the previous discharge. The phenomenon, that is, the dark period is not really dark, is more likely to occur at the time when there is a greater change rate of the applied voltage. These findings are based on the high‐speed direct imaging and Schlieren photography together, in which the former is used to visualize the bright channels and the latter to detect the low‐density area.

Stem is a special short bright zone for the streamer to leader transition in long air gap discharges, whose characteristics play an important role in leader initiation and development. For the purpose of observing the number of stems around the H.V. electrode at the initial corona discharge, experiments were conducted under positive switching and lightning impulses with different crest values, and the schlieren photographs of stems and the discharge current were recorded synchronously. The associated analysis of the experimental results shows that the number of stems around the electrode at the initial corona discharge is always more than one in most cases. In addition, the average number of stems increases with the radius of the electrode and the crest value of applied impulse. If we focus on the characteristic of only one stem for special study, the conditions with a lower crest value under switching impulse and the electrode with a radius as small as possible are preferable.