Photovoltaic array prediction on short-term output power method in centralized power generation system

• Main Authors: Li, Ling Ling, Wen, Shi Yu, Tseng, Ming Lang, Chiu, Anthony S.F.
• Format: text
• Published: Animo Repository 2020
• Subjects: Photovoltaic power generation; Support vector machines; Industrial Engineering; Operations Research, Systems Engineering and Industrial Engineering
• Online Access: https://animorepository.dlsu.edu.ph/faculty_research/3498; https://animorepository.dlsu.edu.ph/context/faculty_research/article/4500/type/native/viewcontent/s10479_018_2879_y.html
• Institution: De La Salle University

The photovoltaic array directly determines the output power system of the entire photovoltaic power generation system. In order to more accurately predict the output power of the photovoltaic power generation system and reduce the impact of photovoltaic power generation on the power system, this study proposes a prediction model based on an improved firefly algorithm optimized support vector machine. The model introduces the linear decreasing inertia weight and the adaptive variable step-size in the original firefly algorithm that effectively improves the convergence speed and optimization ability of the algorithm. The multiple meteorological factors influencing the photovoltaic power generation were studied. Calculate the correlation coefficient of each meteorological influencing factor between the forecasted date and historical date to determine the training sample. The training samples were used to train the prediction model. The photovoltaic array output power in the sunny, cloudy and rainy days was predicted for the three weather styles using the trained prediction model. The results were compared the prediction results on the standard firefly algorithm-based optimizing support vector machine and particle swarm algorithm-based optimizing support vector machines. The proposed method showed that the mean absolute percentage error of the three-weather style prediction result is reduced by 1.66 and 3.30% and the mean square error is reduced by 0.21 and 0.27 compared to other methods. This method is verified to predict the PV array output power more accurately. © 2018, Springer Science+Business Media, LLC, part of Springer Nature.