UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Incorporating high-impedance-fault detection functionality in distribution system online monitoring : an alternative to existing methods Tabakhpour Langeroudi, Adel


The increasing deployment of informatics and communication systems in distribution systems (DS)s enables many desirable functionalities such as optimal control of distributed energy resources, online fault detection, online topology identification and energy theft detection. In this thesis, we focus on exploiting this emerging monitoring and control infrastructure to address high-impedance faults (HIF)s. HIFs have long been recognized as a serious protection issue in DSs, which can adversely impact the DS safety and protection as well as its operation and profitability. Trees contacting overhead lines and deteriorated insulators (due to cracks, dust, humidity, ice, etc.) are the main triggers of HIF in DSs. Once created, HIF imposes a large impedance between two phases or to the ground (through a third object like a tree or the utility poles). Given its high impedance value, HIF usually draws a relatively small current, less than the nominal load level, making it undetectable for relay protection systems. However, the persistent existence of HIF dissipates significant electric power on the long-term, which is undesirable for DS profitability and operation. More importantly, the existence of such HIFs presents weak points in the DS that can easily deteriorate to arcing faults, raising major safety concerns. Existing literature on HIF detection exploit the nonlinear and transient properties of the HIF current, which depends on the use of measurements with high sampling rates (several kHz to MHz) to enable the extraction of the features corresponding to the HIF by applying appropriate signal processing techniques. As such, these approaches demand high-resolution measurements, which are not necessarily economical or available for many distribution systems. On the other hand, today’s distribution systems are mostly equipped with other types of measurements that are far slower than those required by the existing HIF detection algorithms. These measurements are however transmitted to one central unit and are typically processed together for DS static state estimation. This thesis proposes to take advantage of this and to enable the detection of HIFs using the static DS state estimation. The HIF detection problem is accordingly re-defined in a static framework and the HIF is modeled as a parameter to be estimated from the measurements available to the central state estimator from different sources including SCADA, smart meters and PMU measurements. The distribution system state estimation model is augmented to allow for the proposed HIF detection functionality. The IEEE 13-bus and 123-bus systems are used to evaluate the effectiveness of the proposed approach.

Item Citations and Data


Attribution-NonCommercial-NoDerivatives 4.0 International