Optimal Substation Placement: A Paradigm for Advancing Electrical Grid Sustainability

The critical importance of optimal substation placement intensifies as the world experiences sustained economic expansion and firmly pursues the decarbonization process. This paper develops an integrative approach to determining the optimal location for a new substation considering the evolving power framework. To this end, a projected 2% national load growth is taken into account, in accordance with the foresight of the Romanian authorities, emphasizing the need to place new substations to enhance the grid’s sustainability. Leveraging the Weibull distribution, a dataset is generated to simulate the anticipated load increase, starting from real power datasets in Romania. Two algorithms are designed for optimal substation positioning: geometric (center-of-gravity-based) and machine learning (K-means clustering). The primary comparison criterion is the minimization of power losses during energy distribution. The results reveal the machine learning approach (i.e., K-means clustering) as the superior alternative, attaining a 60% success rate in minimizing the power losses. However, acknowledging computational constraints, the concurrent utilization of both algorithms is advocated for optimal substation location selection, indicating a potential improvement in outcomes. This study emphasizes the critical need for advanced algorithms, stressing their role in mitigating power losses and optimizing energy utilization in response to evolving load patterns and sustainability goals.