Integrated stochastic reserve estimation and MILP energy planning for high renewable penetration: Application to 2050 South African energy system

The energy transition imposes a shift towards renewable energy sources, and the integration of variable ones introduces significant risks to power system stability. Variable renewable energy sources are mostly unpredictable and can provide limited spare capacity to compensate for imbalance in demand and supply. To meet system adequacy and reliability requirements, the power system is operated with different types of reserve margins to ensure the availability of spare capacity at various time scales. However, despite existing guidelines to operate the current system, limited methodologies have been proposed to estimate reserve requirements for future power systems with high penetration of renewables, including their integration into planning tools. In this study, a comprehensive methodology is proposed to estimate the least-cost power system design which include an endogenous stochastic model for estimating reserve requirements interfaced to a Mixed-Integer Linear Programming model. The proposed stochastic reserve estimation model incorporates generator tripping events, renewable energy variability, and ramping characteristics of the residual demand, extending ENTSO-E guidelines to accommodate future scenarios with high penetration of renewable energy sources. Furthermore, a non-linear parametric function is trained to represent the results of the stochastic reserve estimation model and then integrated into an optimization model to plan future power systems, using an iterative approach. The methodology is validated on the current South African power system. The results indicate the model’s effectiveness in optimizing reserve requirements, showing substantial benefits in including storage and other renewable energy technologies to meet future energy demands, while reducing carbon emissions and enhancing grid reliability. • Non-linear stochastic model for reserve estimation in energy planning. • Iterative MILP approach to handle non-linear reserve requirements. • Validation of the methodology on the South African power system. • Exploration of the enabling role of storage technologies in future energy systems.