In the ongoing energy transition and renewable energy era, maintaining the integrity and resilience of power distribution networks is paramount. It is necessary to minimize power outages and deliver electricity to consumers safely and reliably. Underground power cable failures are one of the leading causes of power disruption. To meet the increasing demand for electricity and adhere to safety and aesthetic considerations, power distribution lines are frequently installed in extended lengths, often in remote and challenging locations. This poses an issue for inspection and maintenance. Despite the development of diagnostics methods such as domain reflectometry and very low frequency methods, coupling standalone diagnostics methods to develop more efficient and robust fault detection models have not really been explored. This thesis work investigates the feasibility of detecting cable anomalies using a focused impedance spectroscopy approach. Such a focused approach aims to propose an efficient and highly fault-sensitive diagnostic technique for anomalies in power cables.

Amongst several aging and damage mechanisms in underground power cables, utility operators tend to pay more attention to thermal anomalies as they can possibly enhance other damage mechanisms. They also bear higher risks such as fire outbreaks in extreme scenarios. This thesis project presents a multi-physics model for thermal fault identification in underground power cables using cable resistive impedance. Resistive impedance spectroscopy is a potential diagnostic technique which utilizes impedance measuring instruments to measure and record real impedance for fault analysis.

Three damage models were investigated with an emphasis on thermal anomalies. A steady and transient model for the temperature distribution of a healthy cable was first developed in python and MATLAB respectively. The thermal model was then coupled to the resistive portion of impedance to assess its behaviour with temperature change. Since cable faults increase cable temperature during operations, the anomaly detection method relied on observing the temperature-induced alteration in cable resistive impedance as a distinctive fault indicator.

An experimental study was further carried out to validate the developed models for a healthy cable and to extend the hypothesis to faulted cables. For this purpose, single core XLPE insulated cables were subjected to increasing temperatures and their impedance recorded using a vector network analyzer. The results from the extracted resistive impedance showed that, although resistive impedance increased with increasing temperatures, it was highly sensitive and fluctuated with temperature changes. For this reason, it stands a chance of being a thermal anomaly detection, provided there is an existing healthy-cables-impedance baseline derived from historical data in electric utility operating systems. The proposed research work stretches to other damage models including mechanical damage and water trees in underground power cables.

Mechanical damages such as abrasion and bending weakens the cable’s protective layers, leaving it susceptible to short circuits. Cable impedance sensitivity to such faults was investigated. The results revealed that, resistive impedance of the line increases in the presence of a notch, but it depends on the depth of the notch. This implies that damages on the outermost layers of the cable may not be perceived as a fault alarm, however, when such damages are left unattended, they tend to gradually escalate, potentially leading to more severe faults over time. According to these findings, resistive impedance spectroscopy may not be a suitable early mechanical anomaly diagnostic method. Resistive impedance of water-trees-induced coaxial cables in the lab showed subtle changes compared to those of thermally and mechanically damaged cables.

The result from this study brings significant potential contribution to the cable health monitoring and fault diagnostic fields. Future work beyond this study includes additional modelling and characterization for simple cable and damage models, and large-scale cable testing of three-phase cables with realistic insulating layers and damage accumulation, and finally field testing and implementation in real-operating power cable networks.