Akindeji, Kayode TimothyMoloi, KatlehoMazibuko, Ntombenhle2025-09-032025-09-032025https://hdl.handle.net/10321/6205Submitted in fulfilment of the requirements of the degree of Doctor of Engineering in Electrical Engineering, Durban University of Technology, Durban, South Africa, 2024.The 2030 Agenda for Sustainable Development, adopted by all United Nations Member States in 2015, presents a unified framework designed to promote peace and prosperity for individuals and the planet, both in the present and for future generations. At the heart of this agenda are the 17 Sustainable Development Goals (SDGs), which call for immediate action from all countries regardless of their development status through a global partnership. Specifically, SDG 7 seeks to guarantee access to affordable, reliable, sustainable, and modern energy for everyone by 2030. South Africa, recognized as one of the most developed nations in Africa, is also the continent's largest energy consumer. The combination of a growing population and an ongoing power crisis has resulted in heightened electricity demand and a need for alternative energy solutions. In recent years, the country has launched various projects focused on renewable energy sources (RESs). However, despite these investments, the contribution of renewables—particularly wind, solar photovoltaic (PV), and concentrated solar power (CSP)—remains limited, accounting for only 13.7% of the total energy mix, which decreases to 7.3% when hydroelectric sources are excluded. Nuclear and diesel energy make up 4.6% and 1.6%, respectively. The ongoing gap between energy supply and demand remains a critical issue. Additionally, transmission line compensation techniques have emerged as a promising method to enhance transmission capacity, minimize losses, and improve stability within power systems. Although these technique increases the energy availability factor, they frequently present technical challenges in the routine functioning of power systems, especially regarding network protection systems. Protection is crucial not only for ensuring system stability but also for the safety of equipment and personnel. Fundamental principles of any protection strategy encompass accuracy, selectivity, and reliability. This research introduces a machine learning-based protection scheme tailored for a compensated transmission line within a renewable energy network. Initially, a simple two-bus network is created, and a series capacitor compensation method is integrated to assess the impact of transmission line compensation on protection systems. Data is gathered and utilized across three different machine learning detection and classification techniques: K-Nearest Neighbours (K-NN), Medium Neural Network (MNN), and Quadratic Support Vector Machine (QSVM). Additionally, the network is expanded to incorporate both solar and wind energy sources to evaluate performance with the inclusion of renewables. The classifiers are then tested and fine-tuned for enhanced performance. Performance metrics, including confusion matrix analysis, precision-recall curves, and ROC curves, are employed to assess the effectiveness of each machine learning approach. The results indicate that the accuracy and reliability of protection are influenced by the application of these techniques, as evidenced by the rate of fault misclassification. Moreover, machine learning methods show promise in enhancing the protection scheme's performance for a network architecture that includes a compensated transmission line and renewable energy sources. Among the classifiers, the SVM has emerged as the most effective machine learning classifier, achieving an average accuracy of 99.2%.210 penRenewable energy sourcesSolar photovoltaicConcentrated solar power (CSPElectric power systems--ProtectionElectric power transmissionRenewable energy sourcesSmart power gridsThesishttps://doi.org/10.51415/10321/6205