Optimal Sizing and Selection of Energy Storage System Considering Load Uncertainty

Abstract

Energy storage can play an important role by storing the surplus energy and discharging it whenever required maintaining the demand supply balance. Deploying energy storage helps in restoring the imbalance in the grid due to the integration of renewable energy sources and aids in proper utilization of these sources by minimizing the wastage of surplus energy generated at times. Energy storage systems have mostly flexible capacity and it is important to utilize an optimal capacity of the storage system for proper functioning, stability of the system and also for enhanced economic benefits by minimization of total cost of the system. In this paper, to obtain the optimal size of the energy storage system, its operation is worked out as an optimization problem with an objective function of minimization of cost of the system considering different operational constraints. The optimization model is formulated as a Nonlinear programming (NLP) model in a 24-hour time frame. It is tested with the hourly generation of a solar farm and the load demand along with other operational parameters of the storage system. It is important to select suitable energy storage system considering different technical aspects and cost of the system as per the service requirements. Two different energy storage systems have been considered in the study. The results obtained with the sizing model with the respective storage system are stated. Load uncertainty is a major uncertainty in power systems, to see the impact of the uncertain load in the sizing of the storage system, probabilistic load for 24 hours is generated using Monte Carlo Simulation (MCS) considering Normal Probability density function and fed to the sizing model as inputs. The optimization problem is solved using GAMS optimization platform. The results obtained with both the normal load and the probabilistic load are compared and studied. Difference in results is observed for both the cases. It is seen that the probabilistic load offers a realistic study on the impact of varying load demand on the sizing model of the storage system.