Strategic Wind Farm Siting in South Africa: A Resource Feature-Based Clustering Approach

Abstract

South Africa's transition from coal to renewable energy sources is a critical step toward a sustainable future. With an increasing share of variable renewable energy in the energy mix, particularly wind power, there is a pressing need to identify optimal locations for wind farm placement. This report leverages machine learning algorithms, specifically K-means clustering, to analyse time-of-use wind speed data, in order to categorize regions with similar wind patterns. The methodology hinges on data acquisition, feature extraction, and innovative clustering techniques, backed by rigorous validation methods such as the elbow plot and silhouette analysis. An important aspect of this study is the development of a negative map that delineates areas unsuitable for wind farms due to environmental, cultural, or logistical reasons. The research contrasts newly gathered wind speed data against the Wind Atlas for South Africa (WASA) and presents a comparative analysis for a five-year period from 2015 to 2019. The findings illustrate potential sites for wind farm development, excluding restricted areas, for both low and high-demand seasons, contributing to the country's strategic energy planning. This work not only guides the placement of wind farms but also underscores the importance of sustainable development in the renewable energy sector.