NEI-NO--1063 N09905100 Hans Kristian Hoidalen Lightning-induced overvoltages in low-voltage systems REO 2 9 ES3 OST1 universitet Trondheim mm&ng ’Srna*"* NT NT teknisk-naturvitenskapelige Norges LIGHTNING-INDUCED OVERVOLTAGES IN LOW-VOLTAGE SYSTEMS by Hans Kristian Heidalen A dissertation submitted to the Norwegian University of Science and Technology Department of Electrical Power Engineering in partial fulfilment of the requirements for the degree of Doktor Ingenior December 1997 DISCLAIMER Portions of this document may be illegible in electronic image products. Images are produced from the best available original document. Preface 111 PREFACE This thesis is the result of a research project founded by the Norwegian Research Council. All the work was carried out at the Department of Electrical Power Engineering at the Norwegian University of Science and Technology (NTNU) during the years 1994-1997. I would especially express my gratitude to my supervisors Prof. Jarle Sletbak at NTNU and Dr.ing. Thor Henriksen at the Norwegian Electric Power Research Institute (EFI) for being an outstanding source of inspiration (and perspiration) during this work. I would also like to thank all the other institutions and individuals making this thesis possible. In order of appearance: Jostein Huse at EFI, Institute of High Voltage Research at University of Uppsala, University of Florida, the three diploma students Kurt A. Bakke, Abraham T. Gerezgiher and Morten Nordskog, Siemens installasjon, AB Elektro, Det Norske Meteorologiske Institute and Trondheim Energiverk. Finally, I would like to thank all my friends and colleagues at the Department of Electrical Engineering for valuable assistance and encouragement during these three years. Trondheim, December 1997 Hans Kr. Hoidalen Abstract v ABSTRACT Lightning-induced overvoltages (LIOs) from nearby lightning are a main source of failures in low-voltage overhead line systems. Lightning strokes closer than about 1 km can cause harmful overvoltages, which in turn can lead to direct or delayed damages to connected electrical installations or equipment. Insurance companies report an increasing number of damages of electric nature over the years. This increase is probably caused by the introduction of more and more sensitive electrical equipment in an increasing number of installations. This thesis deals primarily with calculations of lightning-induced overvoltages (LIOs) in low- voltage overhead line systems with the objective to enable the design of a proper overvoltage protection. The work is divided in two parts: 1) Development of calculation models 2) Calculations of LIOs in low-voltage systems In the first part models for calculation of LIOs are adapted from the literature or developed based on measurements. An objective when selecting the models is to aim at simple models based on a few measurable quantities, and which show a reasonable accuracy. The models used in this thesis are believed to be fairly accurate for the first few microseconds, which normally is sufficient for prediction of the maximum induced voltage in the system. The lightning channel is modelled by the Modified Transmission Line (MTL) model with the Transmission Line (IL) model as a special case. The coupling between the electrical fields from a lightning channel and an overhead line is modelled by Agrawal ’s model. The attenuation of electrical fields propagating over a lossy ground is modelled by Norton ’s- or the Surface Impedance methods. All these models are well known in the literature and are in this work synthesised to enable calculation of LIOs in practical low-voltage configurations using the electromagnetic transients program, ATP-EMTP. The validity of all the applied models is analysed. In addition measurements have been performed in order to develop models of distribution transformers and low-voltage power installation (LVPI) networks. Simple models of "typical" transformers and LVPIs are developed for calculations when specific data are unavailable. The practical range of values and its influence on the LIOs in a system is investigated. The main frequency range of interest related to LIOs is 10 kHz - 1 MHz in which all the models are accurate. In the second part the adapted or developed models are used to calculate LIOs in low-voltage systems. The influence of various key parameters in the systems is investigated. Of greatest importance are the return stroke amplitude and rise time, the overhead line height and location, the termination of overhead line segments, neutral grounding, and the ground conductivity. • The LIOs in an unprotected system increase proportionally to the return stroke amplitude. Larger rise times of the return stroke result in lower LIOs. • The introduction of lower terminating impedances by connecting e.g. more LVPI networks results in lower LIO. Thus the magnitude of the LIO is likely to be highest in rural areas with few installations connected to an overhead line. A transformer with a Abstract Vi grounded LV neutral can be modelled as a small inductance (4-40 pH) closely related to the transformer ’s rated power and voltage. When the neutral is isolated the model becomes capacitive and the LIOs increases considerably. As a first approximation, LVPI in TN-systems can be modelled as a small inductance (2-20 pH) and in the IT-system as a capacitance (20-200 nF) in series with the inductance found in the TN-system. The influence of type of wiring and apparatus is analysed. The maximum LIO in LVPI networks supplied by an underground cable from an overhead line system is normally little affected by this cable. • An IT-system results in much higher LIO phase-to-ground than a TN-system and this can explain why the number of transients and damages is large in Norway compared to e.g. Sweden. TN-systems result on the other hand in larger phase-to-phase voltages than IT- systems. However, an IT system with a permanent ground fault will experience both high phase-to-phase and phase-to-ground voltages, compare to a TN-system. • The LIOs increase proportionally to the line height when the ground is assumed lossless. The additional contribution from a lossy ground is independent of line height. Lightning strokes near the mid-point of an overhead line results normally in the largest LIOs, but lossy ground effects may modify this. • Lossy ground effects on LIO seem to be very important. Especially in a IT-system the level of calculated LIO increases considerably when a lossy ground is taken into account. The ground losses may reverse the polarity and increase the amplitude of LIOs. However, the effect of a lossy ground is encumbered with uncertainty since a relatively high ground conductivity must be assumed in order to reproduce measurements by calculations. • To protect a low-voltage system completely from LIOs, surge protective devices must be installed at each individual installation. The level of LIOs in a TN-system is much lower and such systems is to some extent self-protected against remote lightning. Even if arresters are installed at the power service entry, large overvoltages can still arise inside the LVPI network. Oscillations due to reflections in the low-voltage system and with frequencies dependent on overhead line segment lengths could excite the natural frequency of connected LVPI circuits, resulting in large internal overvoltages. Such overvoltages can reach amplitudes of several times the protective level of the connected arrester. Contents vii CONTENTS Preface ........................................................................................................... iii Abstract ............................................................................................................... v List of symbols ........................................................................................... xi List of abbreviations ..................................................................................xii Sign conventions ..........................................................................................xii 1. Introduction ...............................................................................................1 1.1 Perspective and motivation ................................................................ 1 1.2 Objectives and contents ....................................................................... 3 2. Background .............................................................................................. 5 2.1 Introduction ......................................................................................... 5 2.2 The lightning discharge ...................................................................... 5 2.2.1 The thunder cloud 5 2.2.2 The charge separation 6 2.2.3 The discharge process 7 2.2.4 Electrical fields from lightning flash 9 2.2.5 Triggered lightning 10 2.2.6 Relative importance of field components 11 2.3 Calculating electrical fields.............................................................. 11 2.4 Lightning flash models .................................................................... 12 2.4.1 Lightning leader model 12 2.4.2 Return stroke model 13 2.5 Lossy ground effects.........................................................................14 2.6 Coupling models ............................................................................... 16 2.7 Calculations versus measurements................................................. 17 2.8 Sources ofLIO ’s..............................................................................