Electrothermal Modeling, Simulation and Optimization of Surge Arresters

Abstract

The protection of power transmission systems against voltage surges relies on station class surge arresters. The core of an arrester consists of a stack of metal oxide resistors. The highly nonlinear U-I-characteristic of these resistors allows the arrester to clip voltage surges by conducting current to ground and thereby absorbing large amounts of energy. The most important arrester design objectives are, first, to ensure a balanced distribution of the field and temperature stress along the arrester column, and, second, to guarantee the thermal stability of the arrester after, e.g., a lightning strike. Standard laboratory test procedures for evaluating thermal stability are limited to worst case surrogate models of the full-scale arrester. Hence, numerical finite element simulation is increasingly valuable for analyzing full-scale surge arresters as an alternative to conventional design based on laboratory or field testing approaches. The analysis of arresters, as presented in this thesis, requires transient and coupled finite element simulation of the mutually-dependent electric and thermal fields employing an accurate electrothermal model. This includes detailed knowledge of both, the field- and temperature-dependent metal oxide material characteristics, and the relevant thermal parameters determining the convective and radiative heat transfer properties of the system. The main difficulty for solving this coupled problem, however, lies in the strong nonlinearity of the metal oxide resistor material. This nonlinearity leads to extremely short electrical time scales, whereas the thermal transients are several orders of magnitude longer. Therefore, this thesis adopts a dedicated multirate time integration technique in order to solve the coupled problem efficiently. The proposed numerical approach is applied to the study of graded and ungraded station class arresters in continuous operation as well as under voltages surges. The simulation results are compared to laboratory measurements. An electrothermal simulation-based procedure for the optimization of the arrester’s field grading systems is introduced employing the developed numerical approach. Here, a modeling and optimization approach to avoid the extremely cumbersome solution of the 3-dimensional and transient nonlinear electric field problem is proposed. The optimization procedure uses an equivalent 2-dimensional-axisymmetric arrester model that can reproduce the electric field stress in the resistor column with high accuracy. This is realized by introducing a virtual electrode geometry whose shape and position are determined by a multi-parametric optimization procedure. Subsequently, the grading system of the 3-dimensional station class arrester is optimized efficiently based on transient, electro-quasistatic simulations of the axisymmetric equivalent model. The detailed electrothermal analysis shows that an immense improvement of the field and thermal stress distribution in the resistor column can be obtained. The most serious limitation in the performance of surge arresters is posed by thermal stability. Overvoltages inject electrical energy that heats up the metal oxide resistors. As a result, the point of operation in the U-I-characteristic is shifted towards a higher electrical conductivity, thus, causing a further increase of the power loss. If not sufficiently compensated by an increase of the heat transfer, this process leads to a thermal runaway which is a catastrophic failure of the arrester. To assess thermal stability, the cooling rate is introduced as a key performance indicator. Finite element simulations provide detailed insights in this complex electro-thermally coupled problem. A station class arrester and its commonly used laboratory surrogate are simulated when subjected to the standard overvoltage stress test procedure. Moreover, various thermally stable and unstable scenarios are analyzed to derive a precise and computationally efficient stability criterion. This criterion allows for the identification of the relevant arrester parameters that influence the thermal stability limit. In the thesis, the effect of the electric characteristics of the resistors on the thermal stability of the arrester as well as selected thermal parameters are investigated. Finally, in order to optimize future arrester designs, a prediction function is introduced to estimate the thermal stability limit based on an affordable set of finite element simulations. Throughout this thesis, the focus lies on the practical applicability of the developed methods to address typical surge arrester design problems.