|Summary Distribution power systems experience a lot of faults due to high network utilisation and reduced investment in the infrastructure. Single phase faults are the most common. The standard distribution systems are radial operated. Failure of any one component in the series path would result into service disruption to all customers located downstream. Current methods used to localise faults are manual and lengthy. For permanent faults, it is a common practice to sectionalise a feeder to reduce the fault investigation area. The exact location of the fault is obtained by visual inspection. However, with the introduction of the new marketing policy resulting from deregulation and liberalisation of the power industry, utility companies are required to compensate consumers for non delivery of service. Distribution networks also records high growth rates due to new customer connections arising from increased demand for electric energy. An increase in consumption introduces transmission constraints on the distribution network, demanding for reinforcement on the main circuits. The penetration of distributed production facilities intends to curtail such investment costs and provide continuity of supply to the affected customers while the fault is being resolved. The objective of the thesis was to investigate the impacts caused by different power system configurations and parameters on automatic single phase fault localisation in power systems with distributed production using power frequency quantities. It was also required to reveal the present ground fault protection schemes and recommend requirements on a system with distributed generation.The approach used in the thesis included informal conversation, field research, literature review on similar topics, modelling and simulations using PSCAD/EMTDC software. A field research was conducted in Zambia on the pretext of understanding the problem from a practical perspective. The main focus was placed on power system configuration and layout, current methods used to localise faults and fault statistics. Secondary research on automated fault localisation methods for distribution systems based on fundamental power frequency quantities was done. Much of the literature available focuses on typical radial distribution feeders. In the thesis, two different methods were used in the simulation of distances to the fault location so that the findings could be consolidated. One of the methods is a standard approach used in numerical distance protection relays and the other, was derived, taking into account the presence of a distributed generator.The distribution system can be solidly grounded, ungrounded, grounded through an arc suppression coil or through a resistor. Each type of grounding configuration has its own impact on fault localisation. Different system grounding possibilities were used in the model when investigating the effect of parameter change, on fault localisation. Simulation results revealed that the distances to the fault locations are affected by all parameters selected for investigation. An increase in load with the same load angle was seen to have an impact of reducing the distance to the fault. An increase in the fault resistance had an effect of increasing the distance to the fault location. Moving the distributed generator away from the substation was seen to have had an effect of reducing the distance to the fault location. However, investigations were carried out without compensation for any of the parameters. This could have contributed to inaccuracies in the results obtained. While automatic localisation of ground faults is at the premium, it is required to minimise damages at the fault position and neighbouring equipment so that restoration of supply is prompt. This could be achieved with effective operation of the protection system. Integration of distributed production facilities can affect the dynamic behaviour of the distribution network and change the power flow. This affects operation of the conversional protection schemes deployed on the radial distribution systems. Therefore, protection schemes with a directional feature are recommended in order to improve selectivity.It can be deduced that precise ground fault localisation based on the use of power frequency quantities is a challenge due to many influencing parameters. However, automated fault localisation in power systems with distributed production is the future solution, especially from the operational perspective.