Techno-economic impacts of battery performance models and control strategies on optimal design of a grid-connected PV system
Masoume Shabani, Erik Dahlquist, Fredrik Wallin, Jinyue Yan
Abstract
A battery storage has emerged as the most widely-used storage option, due to its flexible and complementary functionality for renewable energy systems such as solar PV and wind power. In order to ensure the efficient operation of batteries in energy systems, a proper battery model is essential in predicting realistic battery performance under various operating conditions. Accurate knowledge of the state of charge, state of power, and battery efficiency is a necessity for the development of advanced grid management applications. This paper investigates the techno-economic impacts of two battery modelling scenarios on the sizing and optimization of a grid-connected PV-battery system. Scenario 1 is based on a common simple battery model and control strategy which represents the battery status without reflecting dynamic behavior. By contrast, Scenario 2 is based on a complex battery model involving estimation of battery current-voltage characteristics under various operating conditions. A rule-based operational strategy linked to a non-dominated sorting genetic algorithm is further employed for the simulation and multi-objective optimization of a grid-connected hybrid PV-battery system. The battery life cycle cost and the self-sufficiency ratio are analyzed and optimized as objective functions, and battery capacity constitutes as a decision variable. The results show that in order to reach the same self-sufficiency ratio, the optimization of a hybrid energy system based on Scenario 1 leads to solutions with a higher life cycle cost and requiring bigger battery capacity, compared to that of Scenario 2. Moreover, under the same design parameters, the system optimization based on Scenario 2 delivers more power to the end-user, which leads to a higher self-sufficiency ratio compared to when the system is simulated based on Scenario 1. This study proves that an efficient battery model with sufficient accuracy is techno-economically more beneficial, and leads to more accurate battery sizing.