Analysis on power oscillation smoothening techniques for 1MWp PV plant

The shift from the conventional sources to the eco-friendly and non-polluting inexhaustible sources of energy such as Solar Photovoltaic (PV) and wind introduces many issues due to their high penetration into the electricity grid where power oscillations is to be considered with highest priority. Thus, the proper selection and usage of smoothening techniques is detrimental in controlling the fluctuated oscillations from the photovoltaic power before being injected to the grid. This study investigates the smoothening quality of photovoltaic power output using Moving Average (MA), Artificial Neural Network (ANN) and Adaptive Neuro-Fuzzy Inference Systems (ANFIS) techniques on a 1MWp PV plant for smoothening the fluctuated oscillations that engenders due to intermittent irradiance and temperature for a real time data profile of a 24-hour day simulated in MATLAB
^®
. The smoothening resulted power output of these techniques were compared; based on the analysis using the performance indicators namely; Smooth Power Variation Rate (SPVR) index, Root Mean Square Error (RMSE) and Mean Absolute Error (MAE). However, an analogy of the three techniques indicated that ANFIS is relatively superior to ANN and MA considering the lowest value of SPVR index, RMSE and MAE that in turn proved to be the best smoothening technique for suppressing the PV power oscillations at the DC side.