Sizing the Battery Energy Storage System on a University Campus With Prediction of Load and Photovoltaic Generation

In this paper, the charge and discharge strategies were conducted for a future battery energy storage system (BESS) at the National Penghu University of Science and Technology, Makung, Taiwan. OpenDSS software was used to establish the power distribution system. A probabilistic neural network model was used to predict the daily load and photovoltaic (PV) generation curve for controlling the charge-discharge of the BESS. This study considered both the actual and predicted values of PV generation systems as well as the daily charge-discharge control of the BESS used to balance the peak and off-peak electricity consumption to shave peak loads under the two-and three-phase electricity-pricing methods. The average monthly electric bill and contract capacity were calculated, and the effects of different BESS capacities from the load curve were observed. The results were used to evaluate and determine the capacity required by the BESS. The prediction errors for the load and PV generation of the year were 6.22% and 7.14%, respectively. The electric bills and contract capacities of the actual and predicted values were compared, and the resulting difference was low; this implies that the proposed prediction method is practicable.